Principles of terrestrial ecosystem

F. Stuart Chapin IIIPamela A. MatsonHarold A. Mooney

Principles ofTerrestrialEcosystem Ecology

Illustrated by Melissa C. Chapin

With 199 Illustrations

1 3

Library of Congress Cataloging-in-Publication DataChapin, F. Stuart (Francis Stuart), III.

Principles of terrestrial ecosystem ecology / F. Stuart Chapin III, Pamela A.Matson, Harold A. Mooney.

p. cm.Includes bibliographical references (p. )ISBN 0-387-95439-2 (hc :alk. paper)—ISBN 0-387-95443-0 (sc :alk. paper)1. Ecology. 2. Biogeochemical cycles. 3. Biological systems. I. Matson,

P. A. (Pamela A.) II. Mooney, Harold A. III. Title.QH541 .C3595 2002577¢.14—dc21 2002017654

ISBN 0-387-95439-2 (hardcover) Printed on acid-free paper.ISBN 0-387-95443-0 (softcover)

© 2002 Springer-Verlag New York, Inc.All rights reserved. This work may not be translated or copied in whole or in partwithout the written permission of the publisher (Springer-Verlag New York, Inc.,175 Fifth Avenue, New York, NY 10010, USA), except for brief excerpts in con-nection with reviews or scholarly analysis. Use in connection with any form ofinformation storage and retrieval, electronic adaptation, computer software, or bysimilar or dissimilar methodology now known or hereafter developed is forbidden.The use in this publication of trade names, trademarks, service marks, and similarterms, even if they are not identified as such, is not to be taken as an expression ofopinion as to whether or not they are subject to proprietary rights.

Printed in the United States of America.

9 8 7 6 5 4 3 2 1 SPIN 10866301 (hardcover) SPIN 10866759 (softcover)

www.springer-ny.com

Springer-Verlag New York Berlin HeidelbergA member of BertelsmannSpringer Science+Business Media GmbH

F. Stuart Chapin IIIInstitute of Arctic BiologyUniversity of AlaskaFairbanks, AK [email protected]

Harold A. MooneyDepartment of Biological SciencesHerrin Hall, MC 5020Stanford UniversityStanford, CA [email protected]

Pamela A. MatsonDepartment of Geological and

Environmental ScienceSchool of Earth SciencesGreen 355Stanford UniversityStanford, CA [email protected]

Cover illustration: Waterfall and forests on Valean Poas in Costa Rica. Photographby Peter Vitousek.

Human activities are affecting the global environment in myriadways, with numerous direct and indirect effects on ecosystems.The climate and atmospheric composition of Earth are changingrapidly. Humans have directly modified half of the ice-free terres-trial surface and use 40% of terrestrial production. Our actions arecausing the sixth major extinction event in the history of life onEarth and are radically modifying the interactions among forests,fields, streams, and oceans. This book was written to provide a con-ceptual basis for understanding terrestrial ecosystem processes andtheir sensitivity to environmental and biotic changes. We believethat an understanding of how ecosystems operate and change mustunderlie our analysis of both the consequences and the mitigationof human-caused changes.

This book is intended to introduce the science of ecosystemecology to advanced undergraduate students, beginning graduatestudents, and practicing scientists from a wide array of disciplines.We also provide access to some of the rapidly expanding literaturein the many disciplines that contribute to ecosystem understanding.

The first part of the book provides the context for understand-ing ecosystem ecology. We introduce the science of ecosystemecology and place it in the context of other components of theEarth System—the atmosphere, ocean, climate and geologicalsystems. We show how these components affect ecosystemprocesses and contribute to the global variation in terrestrialecosystem structure and processes. In the second part of the book,we consider the mechanisms by which terrestrial ecosystems func-tion and focus on the flow of water and energy and the cycling ofcarbon and nutrients. We then compare and contrast these cyclesbetween terrestrial and aquatic ecosystems. We also consider theimportant role that organisms have on ecosystem processesthrough trophic interactions (feeding relationships), environmen-tal effects, and disturbance. The third part of the book addressestemporal and spatial patterns in ecosystem processes. We finish byconsidering the integrated effects of these processes at the globalscale and their consequences for sustainable use by human soci-

Preface

v

vi Preface

eties. Powerpoint lecture notes developed by one of the authors are available online (www.faculty.uaf.edu/fffsc/) as supplementarymaterial.

Many people have contributed to the development of this book.We particularly thank our families, whose patience has made thebook possible, and our students from whom we have learned manyof the important ideas that are presented. In addition, we thank the following individuals for their constructively critical review ofchapters in this book: Kevin Arrigo, Teri Balser, Perry Barboza,Jason Beringer, Kim Bonine, Rich Boone, Syndonia Bret-Harte,John Bryant, Inde Burke, Zoe Cardon, Oliver Chadwick, ScottChambers, Melissa Chapin, Kathy Cottingham, Joe Craine, Wolf-gang Cramer, Steve Davis, Sandra Diaz, Bill Dietrich, Rob Dunbar,Jim Ehleringer, Howie Epstein, Werner Eugster, Valerie Eviner,Scott Fendorf, Jon Foley, David Foster, Tom Gower, Peter Groff-man, Paul Grogan, Diego Gurvich, Bill Heal, Sarah Hobbie, DaveHooper, Shuijin Hu, Pilar Huante, Bruce Hungate, Jill Johnstone,Jay Jones, Jürg Luterbacher, Frank Kelliher, Jennifer King, DaveKline, Christian Körner, Hans Lambers, Amanda Lynch, MichelleMack, Steve MacLean, Joe McFadden, Dave McGuire, SamMcNaughton, Knute Nadelhoffer, Jason Neff, Mark Oswood, BobPaine, Bill Parton, Natalia Perez, Steward Pickett, Stephen Parder,Mary Power, Jim Randerson, Bill Reeburgh, Peter Reich, JimReynolds, Roger Ruess, Steve Running, Scott Rupp, Dave Schimel,Josh Schimel, Bill Schlesinger, Guthrie Schrengohst, Ted Schuur,Stephen Parder Mark Serreze, Gus Shaver, Nigel Tapper, MonicaTurner, Dave Valentine, Peter Vitousek, Lars Walker, and KateyWalter.We particularly thank Phil Camil,Valerie Eviner, Jon Foley,and Paul Grogan for comments on the entire book; Mark Chapin,Patrick Endres, and Rose Meier for comments on illustrations; PhilCamil for comments on educational approaches; and Jon Foley andNick Olejniczak for providing global maps.

F. Stuart Chapin IIIPamela A. Matson

Harold A. Mooney

Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v

Part I Context

Chapter 1The Ecosystem Concept

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3Overview of Ecosystem Ecology . . . . . . . . . . . . . . . . . . . 3History of Ecosystem Ecology . . . . . . . . . . . . . . . . . . . . . 7Ecosystem Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10Controls over Ecosystem Processes . . . . . . . . . . . . . . . . . 11Human-Caused Changes in Earth’s Ecosystems . . . . . . . 13Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17Additional Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Chapter 2Earth’s Climate System

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18Earth’s Energy Budget . . . . . . . . . . . . . . . . . . . . . . . . . . 18The Atmospheric System . . . . . . . . . . . . . . . . . . . . . . . . . 21

Atmospheric Composition and Chemistry . . . . . . . . . . 21Atmospheric Structure . . . . . . . . . . . . . . . . . . . . . . . . . 22Atmospheric Circulation . . . . . . . . . . . . . . . . . . . . . . . 24

The Oceans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28Ocean Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28Ocean Circulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Landform Effects on Climate . . . . . . . . . . . . . . . . . . . . . 31Vegetation Influences on Climate . . . . . . . . . . . . . . . . . . 32Temporal Variability in Climate . . . . . . . . . . . . . . . . . . . . 34

Long-Term Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . 34Interannual Climate Variability . . . . . . . . . . . . . . . . . . 38Seasonal and Daily Variations . . . . . . . . . . . . . . . . . . . 40

vii

viii Contents

Relationship of Climate to Ecosystem Distribution and Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44Additional Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

Chapter 3Geology and Soils

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46Controls over Soil Formation . . . . . . . . . . . . . . . . . . . . . 46

Parent Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47Climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48Topography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49Potential Biota . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50Human Activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

Controls over Soil Loss . . . . . . . . . . . . . . . . . . . . . . . . . . 50Development of Soil Profiles . . . . . . . . . . . . . . . . . . . . . . 53

Additions to Soils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54Soil Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . 54Soil Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56Losses from Soils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

Soil Horizons and Soil Classification . . . . . . . . . . . . . . . . 58Soil Properties and Ecosystem Functioning . . . . . . . . . . 61Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67Additional Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

Part II Mechanisms

Chapter 4Terrestrial Water and Energy Balance

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71Surface Energy Balance . . . . . . . . . . . . . . . . . . . . . . . . . . 73

Solar Radiation Budget . . . . . . . . . . . . . . . . . . . . . . . . 73Ecosystem Radiation Budget . . . . . . . . . . . . . . . . . . . . 74Energy Partitioning . . . . . . . . . . . . . . . . . . . . . . . . . . . 75Seasonal Energy Exchange . . . . . . . . . . . . . . . . . . . . . 77

Water Inputs to Ecosystems . . . . . . . . . . . . . . . . . . . . . . 77Water Movements Within Ecosystems . . . . . . . . . . . . . . . 78

Basic Principles of Water Movement . . . . . . . . . . . . . . 78Water Movement from the Canopy to the Soil . . . . . . 79Water Movement Within the Soil . . . . . . . . . . . . . . . . . 80Water Movement from Soil to Roots . . . . . . . . . . . . . . 81Water Movement Through Plants . . . . . . . . . . . . . . . . 83

Water Losses from Ecosystems . . . . . . . . . . . . . . . . . . . . 89Evaporation from Wet Canopies . . . . . . . . . . . . . . . . . 89Evapotranspiration from Dry Canopies . . . . . . . . . . . . 90

Contents ix

Changes in Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92Runoff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93


Chapter 5Carbon Input to Terrestrial Ecosystems

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97Photosynthetic Pathways . . . . . . . . . . . . . . . . . . . . . . . . . 98

C3 Photosynthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98C4 Photosynthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102Crassulacean Acid Metabolism Photosynthesis . . . . . . 103

Net Photosynthesis by Individual Leaves . . . . . . . . . . . . 105Basic Principle of Environmental Control . . . . . . . . . . 105Light Limitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105CO2 Limitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109Nitrogen Limitation and Photosynthetic Capacity . . . . 110Water Limitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113Temperature Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . 114Pollutants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

Gross Primary Production . . . . . . . . . . . . . . . . . . . . . . . . 115Canopy Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115Satellite Estimates of GPP . . . . . . . . . . . . . . . . . . . . . . 117Controls over GPP . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119


Chapter 6Terrestrial Production Processes

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123Plant Respiration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

Physiological Basis of Respiration . . . . . . . . . . . . . . . . 125Net Primary Production . . . . . . . . . . . . . . . . . . . . . . . . . . 127

What Is NPP? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127Physiological Controls over NPP . . . . . . . . . . . . . . . . . 128Environmental Controls over NPP . . . . . . . . . . . . . . . 129

Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132Allocation of NPP . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132Allocation Response to Multiple Resources . . . . . . . . 133Diurnal and Seasonal Cycles of Allocation . . . . . . . . . 134

Tissue Turnover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136Global Distribution of Biomass and NPP . . . . . . . . . . . . 137

Biome Differences in Biomass . . . . . . . . . . . . . . . . . . . 137Biome Differences in NPP . . . . . . . . . . . . . . . . . . . . . . 138

Net Ecosystem Production . . . . . . . . . . . . . . . . . . . . . . . 140

x Contents

Ecosystem Carbon Storage . . . . . . . . . . . . . . . . . . . . . 140Leaching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141Lateral Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145Disturbance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145Controls over Net Ecosystem Production . . . . . . . . . . 145Net Ecosystem Exchange . . . . . . . . . . . . . . . . . . . . . . . 146Global Patterns of NEE . . . . . . . . . . . . . . . . . . . . . . . . 147


Chapter 7Terrestrial Decomposition

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151Leaching of Litter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152Litter Fragmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152Chemical Alteration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

Fungi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153Bacteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154Soil Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

Temporal and Spatial Heterogeneity of Decomposition . 157Temporal Pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157Spatial Pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

Factors Controlling Decomposition . . . . . . . . . . . . . . . . . 159The Physical Environment . . . . . . . . . . . . . . . . . . . . . . 159Substrate Quality and Quantity . . . . . . . . . . . . . . . . . . 163Microbial Community Composition and

Enzymatic Capacity . . . . . . . . . . . . . . . . . . . . . . . . . 168Long-Term Storage of Soil Organic Matter . . . . . . . . . . . 169Decomposition at the Ecosystem Scale . . . . . . . . . . . . . . 170

Aerobic Heterotrophic Respiration . . . . . . . . . . . . . . . 170Anaerobic Heterotrophic Respiration . . . . . . . . . . . . . 173


Chapter 8Terrestrial Plant Nutrient Use

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176Nutrient Movement to the Root . . . . . . . . . . . . . . . . . . . 177

Diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178Mass Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178Root Interception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

Nutrient Uptake . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180Nutrient Supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180Development of Root Length . . . . . . . . . . . . . . . . . . . 181Mycorrhizae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182

Contents xi

Root Uptake Properties . . . . . . . . . . . . . . . . . . . . . . . . 184Nutrient Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189Nutrient Loss from Plants . . . . . . . . . . . . . . . . . . . . . . . . 191

Senescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192Leaching Loss from Plants . . . . . . . . . . . . . . . . . . . . . . 193Herbivory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193Other Avenues of Nutrient Loss from Plants . . . . . . . 194


Chapter 9Terrestrial Nutrient Cycling

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197Nitrogen Inputs to Ecosystems . . . . . . . . . . . . . . . . . . . . 198

Biological Nitrogen Fixation . . . . . . . . . . . . . . . . . . . . 198Nitrogen Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

Internal Cycling of Nitrogen . . . . . . . . . . . . . . . . . . . . . . 202Overview of Mineralization . . . . . . . . . . . . . . . . . . . . . 202Production and Fate of Dissolved Organic Nitrogen . . 203Production and Fate of Ammonium . . . . . . . . . . . . . . . 204Production and Fate of Nitrate . . . . . . . . . . . . . . . . . . 207Temporal and Spatial Variability . . . . . . . . . . . . . . . . . 210

Pathways of Nitrogen Loss . . . . . . . . . . . . . . . . . . . . . . . 211Gaseous Losses of Nitrogen . . . . . . . . . . . . . . . . . . . . . 211Ecological Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . 211Solution Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214Erosional Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215

Other Element Cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . 215Phosphorus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215Sulfur . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219Essential Cations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219Nonessential Elements . . . . . . . . . . . . . . . . . . . . . . . . . 220Interactions Among Element Cycles . . . . . . . . . . . . . . 220


Chapter 10Aquatic Carbon and Nutrient Cycling

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224Ecosystem Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224Oceans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228

Carbon and Light Availability . . . . . . . . . . . . . . . . . . . 228Nutrient Availability . . . . . . . . . . . . . . . . . . . . . . . . . . . 231Carbon and Nutrient Cycling . . . . . . . . . . . . . . . . . . . . 233

Lakes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236Controls over NPP . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236

xii Contents

Carbon and Nutrient Cycling . . . . . . . . . . . . . . . . . . . . 238Streams and Rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238

Carbon and Nutrient Cycling . . . . . . . . . . . . . . . . . . . . 240Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242Additional Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

Chapter 11Trophic Dynamics

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244Plant-Based Trophic Systems . . . . . . . . . . . . . . . . . . . . . . 246

Controls over Energy Flow Through Ecosystems . . . . 246Ecological Efficiencies . . . . . . . . . . . . . . . . . . . . . . . . . 250Food Chain Length and Trophic Cascades . . . . . . . . . . 257Seasonal Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258Nutrient Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

Detritus-Based Trophic Systems . . . . . . . . . . . . . . . . . . . 261Integrated Food Webs . . . . . . . . . . . . . . . . . . . . . . . . . . . 261

Mixing of Plant-Based and Detritus-Based Food Chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261

Food Web Complexities . . . . . . . . . . . . . . . . . . . . . . . . 263Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263Additional Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264

Chapter 12Community Effects on Ecosystem Processes

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266Species Effects on Ecosystem Processes . . . . . . . . . . . . . 268

Species Effects on Resources . . . . . . . . . . . . . . . . . . . . 268Species Effects on Climate . . . . . . . . . . . . . . . . . . . . . . 271Species Effects on Disturbance Regime . . . . . . . . . . . 272

Species Interactions and Ecosystem Processes . . . . . . . . 273Diversity Effects on Ecosystem Processes . . . . . . . . . . . . 274Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277Additional Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278

Part III Patterns

Chapter 13Temporal Dynamics

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281Fluctuations in Ecosystem Processes . . . . . . . . . . . . . . . . 281

Interannual Variability . . . . . . . . . . . . . . . . . . . . . . . . . 281Long-Term Change . . . . . . . . . . . . . . . . . . . . . . . . . . . 283

Contents xiii

Disturbance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285Conceptual Framework . . . . . . . . . . . . . . . . . . . . . . . . 285Disturbance Properties . . . . . . . . . . . . . . . . . . . . . . . . . 285

Succession . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288Ecosystem Structure and Composition . . . . . . . . . . . . 288Carbon Balance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292Nutrient Cycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296Trophic Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298Water and Energy Exchange . . . . . . . . . . . . . . . . . . . . 299

Temporal Scaling of Ecological Processes . . . . . . . . . . . . 301Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303Additional Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304

Chapter 14Landscape Heterogeneity and Ecosystem Dynamics

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305Concepts of Landscape Heterogeneity . . . . . . . . . . . . . . 305Causes of Spatial Heterogeneity . . . . . . . . . . . . . . . . . . . 307

State Factors and Interactive Controls . . . . . . . . . . . . . 307Community Processes and Legacies . . . . . . . . . . . . . . . 307Disturbance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309Interactions Among Sources of Heterogeneity . . . . . . 311

Patch Interactions on the Landscape . . . . . . . . . . . . . . . . 314Topographic and Land-Water Interactions . . . . . . . . . 314Atmospheric Transfers . . . . . . . . . . . . . . . . . . . . . . . . . 317Movement of Plants and Animals on the

Landscape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320Disturbance Spread . . . . . . . . . . . . . . . . . . . . . . . . . . . 320

Human Land Use Change and Landscape Heterogeneity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321Extensification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321Intensification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323

Spatial Heterogeneity and Scaling . . . . . . . . . . . . . . . . . . 325Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330Additional Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331

Part IV Integration

Chapter 15Global Biogeochemical Cycles

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335The Global Carbon Cycle . . . . . . . . . . . . . . . . . . . . . . . . 335

Long-Term Change in Atmospheric CO2 . . . . . . . . . . . 337Anthropogenic Changes in the Carbon Cycle . . . . . . . 339Terrestrial Sinks for CO2 . . . . . . . . . . . . . . . . . . . . . . . 340

The Global Methane Budget . . . . . . . . . . . . . . . . . . . . . . 342

xiv Contents

The Global Nitrogen Cycle . . . . . . . . . . . . . . . . . . . . . . . 343Anthropogenic Changes in the Nitrogen Cycle . . . . . . 344

The Global Phosphorus Cycle . . . . . . . . . . . . . . . . . . . . . 347Anthropogenic Changes in the Phosphorus Cycle . . . . 347

The Global Sulfur Cycle . . . . . . . . . . . . . . . . . . . . . . . . . 348The Global Water Cycle . . . . . . . . . . . . . . . . . . . . . . . . . 350

Anthropogenic Changes in the Water Cycle . . . . . . . . 351Consequences of Changes in the Water Cycle . . . . . . . 352


Chapter 16Managing and Sustaining Ecosystems

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356Ecosystem Concepts in Management . . . . . . . . . . . . . . . 357

Natural Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357Resilience and Stability . . . . . . . . . . . . . . . . . . . . . . . . 357State Factors and Interactive Controls . . . . . . . . . . . . . 358

Application of Ecosystem Knowledge in Management . . 359Forest Management . . . . . . . . . . . . . . . . . . . . . . . . . . . 359Fisheries Management . . . . . . . . . . . . . . . . . . . . . . . . . 359Ecosystem Restoration . . . . . . . . . . . . . . . . . . . . . . . . . 360Management for Endangered Species . . . . . . . . . . . . . 360

Integrative Approaches to Ecosystem Management . . . . 362Ecosystem Management . . . . . . . . . . . . . . . . . . . . . . . . 362Integrated Conservation and Development

Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365Valuation of Ecosystem Goods and Services . . . . . . . . 366


Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371

Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423

Introduction

Ecosystem ecology addresses the interactionsbetween organisms and their environment as anintegrated system. The ecosystem approach isfundamental in managing Earth’s resourcesbecause it addresses the interactions that linkbiotic systems, of which humans are an integralpart, with the physical systems on which theydepend. This applies at the scale of Earth as awhole, a continent, or a farmer’s field. Anecosystem approach is critical to resource man-agement, as we grapple with the sustainable useof resources in an era of increasing human population and consumption and large, rapidchanges in the global environment.

Our growing dependence on ecosystem con-cepts can be seen in many areas. The UnitedNations Convention on Biodiversity of 1992,for example, promoted an ecosystem approach,including humans, to conserving biodiversityrather than the more species-based approachesthat predominated previously. There is a grow-ing appreciation of the role that individualspecies, or groups of species, play in the func-tioning of ecosystems and how these functionsprovide services that are vital to humanwelfare. An important, and belated, shift inthinking has occurred about managing ecosys-tems on which we depend for food and fiber.

The supply of fish from the sea is now declin-ing because fisheries management depended onspecies-based approaches that did not ade-quately consider the resources on which com-mercial fish depend. A more holistic view ofmanaged systems can account for the complexinteractions that prevail in even the simplestecosystems. There is also an increasing appreci-ation that a thorough understanding of eco-systems is critical to managing the quality andquantity of our water supplies and in regulatingthe composition of the atmosphere that deter-mines Earth’s climate.

Overview of Ecosystem Ecology

The flow of energy and materials throughorganisms and the physical environment pro-vides a framework for understanding the diver-sity of form and functioning of Earth’s physicaland biological processes. Why do tropicalforests have large trees but accumulate only athin layer of dead leaves on the soil surface,whereas tundra supports small plants but anabundance of soil organic matter? Why doesthe concentration of carbon dioxide in theatmosphere decrease in summer and increasein winter? What happens to that portion of thenitrogen that is added to farmers’ fields but is

1The Ecosystem Concept

Ecosystem ecology studies the links between organisms and their physical environ-ment within an Earth System context. This chapter provides background on the con-ceptual framework and history of ecosystem ecology.

3

4 1. The Ecosystem Concept

not harvested with the crop? Why has the intro-duction of exotic species so strongly affectedthe productivity and fire frequency of grass-lands and forests? Why does the number ofpeople on Earth correlate so strongly with theconcentration of methane in the Antarctic ice cap or with the quantity of nitrogen enter-ing Earth’s oceans? These are representativequestions addressed by ecosystem ecology.Answers to these questions require an under-standing of the interactions between organismsand their physical environments—both theresponse of organisms to environment and the effects of organisms on their environment.Addressing these questions also requires that we think of integrated ecological systemsrather than individual organisms or physicalcomponents.

Ecosystem analysis seeks to understand thefactors that regulate the pools (quantities) andfluxes (flows) of materials and energy throughecological systems. These materials includecarbon, water, nitrogen, rock-derived mineralssuch as phosphorus, and novel chemicals suchas pesticides or radionuclides that people haveadded to the environment. These materials arefound in abiotic (nonbiological) pools such assoils, rocks, water, and the atmosphere and inbiotic pools such as plants, animals, and soilmicroorganisms.

An ecosystem consists of all the organismsand the abiotic pools with which they interact.Ecosystem processes are the transfers of energyand materials from one pool to another. Energyenters an ecosystem when light energy drivesthe reduction of carbon dioxide (CO2) to formsugars during photosynthesis. Organic matterand energy are tightly linked as they movethrough ecosystems. The energy is lost from the ecosystem when organic matter is oxidizedback to CO2 by combustion or by the respira-tion of plants, animals, and microbes. Materialsmove among abiotic components of the systemthrough a variety of processes, including theweathering of rocks, the evaporation of water,and the dissolution of materials in water.Fluxes involving biotic components include theabsorption of minerals by plants, the death ofplants and animals, the decomposition of deadorganic matter by soil microbes, the consump-

tion of plants by herbivores, and the consump-tion of herbivores by predators. Most of thesefluxes are sensitive to environmental factors,such as temperature and moisture, and to bio-logical factors that regulate the populationdynamics and species interactions in communi-ties. The unique contribution of ecosystemecology is its focus on biotic and abiotic factorsas interacting components of a single integratedsystem.

Ecosystem processes can be studied at manyspatial scales. How big is an ecosystem? Theappropriate scale of study depends on the ques-tion being asked (Fig. 1.1). The impact of zoo-plankton on the algae that they eat might bestudied in the laboratory in small bottles. Otherquestions such as the controls over productiv-ity might be studied in relatively homogeneouspatches of a lake, forest, or agricultural field.Still other questions are best addressed at theglobal scale. The concentration of atmosphericCO2, for example, depends on global patternsof biotic exchanges of CO2 and the burning offossil fuels, which are spatially variable acrossthe globe. The rapid mixing of CO2 in theatmosphere averages across this variability,facilitating estimates of long-term changes inthe total global flux of carbon between Earthand the atmosphere.

Some questions require careful measure-ments of lateral transfers of materials. A water-shed is a logical unit in which to study theeffects of forests on the quantity and quality ofthe water that supplies a town reservoir. Awatershed, or catchment, consists of a streamand all the terrestrial surfaces that drain into it. By studying a watershed we can compare thequantities of materials that enter from the air and rocks with the amounts that leave instream water, just as you balance your check-book. Studies of input–output budgets of water-sheds have improved our understanding of theinteractions between rock weathering, whichsupplies nutrients, and plant and microbialgrowth, which retains nutrients in ecosystems(Vitousek and Reiners 1975, Bormann andLikens 1979).

The upper and lower boundaries of anecosystem also depend on the question beingasked and the scale that is appropriate to the

Overview of Ecosystem Ecology 5

question.The atmosphere, for example, extendsfrom the gases between soil particles all the wayto outer space. The exchange of CO2 between aforest and the atmosphere might be measureda few meters above the top of the canopybecause, above this height, variations in CO2

content of the atmosphere are also stronglyinfluenced by other upwind ecosystems. Theregional impact of grasslands on the moisturecontent of the atmosphere might, however, bemeasured at a height of several kilometersabove the ground surface, where the moisturereleased by the ecosystem condenses andreturns as precipitation (see Chapter 2). For

questions that address plant effects on waterand nutrient cycling, the bottom of the ecosys-tem might be the maximum depth to whichroots extend because soil water or nutrientsbelow this depth are inaccessible to the vegeta-tion. Studies of long-term soil development, incontrast, must also consider rocks deep in thesoil, which constitute the long-term reservoir ofmany nutrients that gradually become incorpo-rated into surface soils (see Chapter 3).

Ecosystem dynamics are a product of manytemporal scales. The rates of ecosystem pro-cesses are constantly changing due to fluctua-tions in environment and activities of organisms

c) Forest ecosystem

1 km

How does acid rain influence forest

productivity?

a) Global ecosystem

5,000 kmHow does carbon loss

from plowed soils influence global climate?

b) Watershed

10 kmHow does

deforestation influence the

water supply toneighboring towns?

d) Endolithic ecosystem

1 mm

rock surface

What are the biological controls over rock

weathering?algal zone

lichen zone

Figure 1.1. Examples ofecosystems that range in sizeby 10 orders of magnitude:an endolithic ecosystem in the surface layers of rocks,1 ¥ 10-3 m in height (d); aforest, 1 ¥ 103 m in diameter(c); a watershed, 1 ¥ 105 m inlength (b); and Earth, 4 ¥ 107 min circumference (a). Alsoshown are examples of ques-tions appropriate to each scale.


on time scales ranging from microseconds tomillions of years (see Chapter 13).Light captureduring photosynthesis responds almost instan-taneously to fluctuations in light availability to a leaf.At the opposite extreme, the evolution of photosynthesis 2 billion years ago addedoxygen to the atmosphere over millions ofyears, causing the prevailing geochemistry of Earth’s surface to change from chemicalreduction to chemical oxidation (Schlesinger1997). Microorganisms in the group Archaeaevolved in the early reducing atmosphere ofEarth. These microbes are still the only organ-isms that produce methane. They now functionin anaerobic environments such as wetland soilsand the interiors of soil aggregates or animalintestines. Episodes of mountain building anderosion strongly influence the availability ofminerals to support plant growth. Vegetation isstill migrating in response to the retreat of Pleis-tocene glaciers 10,000 to 20,000 years ago.Afterdisturbances such as fire or tree fall, there aregradual changes in plant, animal, and microbialcommunities over years to centuries. Rates ofcarbon input to an ecosystem through photo-synthesis change over time scales of seconds todecades due to variations in light, temperature,and leaf area.

Many early studies in ecosystem ecologymade the simplifying assumption that someecosystems are in equilibrium with their envi-ronment. In this perspective, relatively undis-turbed ecosystems were thought to haveproperties that reflected (1) largely closedsystems dominated by internal recycling of elements, (2) self-regulation and deterministicdynamics, (3) stable end points or cycles, and(4) absence of disturbance and human influ-ence (Pickett et al. 1994, Turner et al. 2001).One of the most important conceptualadvances in ecosystem ecology has been theincreasing recognition of the importance ofpast events and external forces in shaping the functioning of ecosystems. In this non-equilibrium perspective, we recognize that most ecosystems exhibit inputs and losses, theirdynamics are influenced by both external andinternal factors, they exhibit no single stableequilibrium, disturbance is a natural compo-nent of their dynamics, and human activities

have a pervasive influence. The complicationsassociated with the current nonequilibriumview require a more dynamic and stochasticview of controls over ecosystem processes.

Ecosystems are considered to be at steadystate if the balance between inputs and outputsto the system shows no trend with time(Johnson 1971, Bormann and Likens 1979).Steady state assumptions differ from equilib-rium assumptions because they accept tempo-ral and spatial variation as a normal aspect ofecosystem dynamics. Even at steady state, forexample, plant growth changes from summer towinter and between wet and dry years (seeChapter 6). At a stand scale, some plants maydie from old age or pathogen attack and bereplaced by younger individuals.At a landscapescale, some patches may be altered by fire orother disturbances, and other patches will be in various stages of recovery. These ecosystemsor landscapes are in steady state if there is no long-term directional trend in their pro-perties or in the balance between inputs andoutputs.

Not all ecosystems and landscapes are insteady state. In fact, directional changes inclimate and environment caused by humanactivities are quite likely to cause directionalchanges in ecosystem properties. Nonetheless,it is often easier to understand the relationshipof ecosystem processes to the current environ-ment in situations in which they are not alsorecovering from large recent perturbations.Once we understand the behavior of a systemin the absence of recent disturbances, we canadd the complexities associated with time lagsand rates of ecosystem change.

Ecosystem ecology uses concepts developedat finer levels of resolution to build an under-standing of the mechanisms that govern theentire Earth System. The biologically mediatedmovement of carbon and nitrogen throughecosystems depends on the physiological properties of plants, animals, and soil micro-organisms. The traits of these organisms are the products of their evolutionary histories and the competitive interactions that sortspecies into communities where they success-fully grow, survive, and reproduce (Vrba andGould 1986). Ecosystem fluxes also depend

History of Ecosystem Ecology 7

on the population processes that govern plant, animal, and microbial densities and age structures as well as on communityprocesses, such as competition and predation,that determine which species are present andtheir rates of resource consumption. Ecosystem ecology therefore depends on information and principles developed in physiological, evo-lutionary, population, and community ecology(Fig. 1.2).

The supply of water and minerals from soilsto plants depends not only on the activities ofsoil microorganisms but also on physical andchemical interactions among rocks, soils, andthe atmosphere. The low availability of phos-phorus due to the extensive weathering anderosional loss of nutrients in the ancient soils ofwestern Australia, for example, strongly con-strains plant growth and the quantity and typesof plants and animals that can be supported.Principles of ecosystem ecology must thereforealso incorporate the concepts and understand-ing of disciplines such as geochemistry, hydrol-ogy, and climatology that focus on the physicalenvironment (Fig. 1.2).

Ecosystem ecology provides the mechanisticbasis for understanding processes that occur at global scales. Study of Earth as a physicalsystem relies on information provided by

ecosystem ecologists about the rates at whichthe land or water surface interacts with theatmosphere, rocks, and waters of the planet(Fig. 1.2). Conversely, the global budgets ofmaterials that cycle between the atmosphere,land, and oceans provide a context for under-standing the broader significance of processesstudied in a particular ecosystem. Latitudinaland seasonal patterns of atmospheric CO2 con-centration, for example, help define the loca-tions where carbon is absorbed or releasedfrom the land and oceans (see Chapter 15).

History of Ecosystem Ecology

Many early discoveries of biology were moti-vated by questions about the integrated natureof ecological systems. In the seventeenthcentury, European scientists were still uncertainabout the source of materials found in plants.Plattes, Hooke, and others advanced the novelidea that plants derive nourishment from both air and water (Gorham 1991). Priestleyextended this idea in the eighteenth century byshowing that plants produce a substance that isessential to support the breathing of animals.Atabout the same time MacBride and Priestleyshowed that breakdown of organic mattercaused the production of “fixed air” (carbondioxide), which did not support animal life.De Saussure, Liebig, and others clarified theexplicit roles of carbon dioxide, oxygen,and mineral nutrients in these cycles in thenineteenth century. Much of the biologicalresearch during the nineteenth and twentiethcenturies went on to explore the detailed mechanisms of biochemistry, physiology,behavior, and evolution that explain how lifefunctions. Only in recent decades have wereturned to the question that originally moti-vated this research: How are biogeochemicalprocesses integrated in the functioning ofnatural ecosystems?

Many threads of ecological thought havecontributed to the development of ecosystemecology (Hagen 1992), including ideas relatingto trophic interactions (the feeding relation-ships among organisms) and biogeochemistry(biological influences on the chemical processes

Earth system science

Climatology

Hydrology

Soil science

Geochemistry

Physiologicalecology

Ecosystem ecology

Populationecology

Communityecology

Context

Mechanism

Figure 1.2. Relationships between ecosystemecology and other disciplines. Ecosystem ecologyintegrates the principles of several biological andphysical disciplines and provides the mechanisticbasis for Earth System Science.


in ecosystems). Early research on trophic inter-actions emphasized the transfer of energyamong organisms. Elton (1927), an Englishzoologist interested in natural history,described the role that an animal plays in acommunity (its niche) in terms of what it eatsand is eaten by. He viewed each animal speciesas a link in a food chain, which described themovement of matter from one organism toanother. Elton’s concepts of trophic structureprovide a framework for understanding theflow of materials through ecosystems (seeChapter 11).

Hutchinson, an American limnologist, wasstrongly influenced by the ideas of Elton andthose of Russian geochemist Vernadsky, whodescribed the movement of minerals from soilinto vegetation and back to soil. Hutchinsonsuggested that the resources available in a lakemust limit the productivity of algae and thatalgal productivity, in turn, must limit the abun-dance of animals that eat algae. Meanwhile,Tansley (1935), a British terrestrial plant ecolo-gist, was also concerned that ecologists focusedtheir studies so strongly on organisms that they failed to recognize the importance ofexchange of materials between organisms andtheir abiotic environment. He coined the termecosystem to emphasize the importance ofinterchanges of materials between inorganicand organic components as well as amongorganisms.

Lindeman, another limnologist, was stronglyinfluenced by all these threads of ecologicaltheory. He suggested that energy flow throughan ecosystem could be used as a currency toquantify the roles of organisms in trophicdynamics. Green plants (primary producers)capture energy and transfer it to animals (consumers) and decomposers. At each trans-fer, some energy is lost from the ecosystemthrough respiration.Therefore, the productivityof plants constrains the quantity of consumersthat an ecosystem can support. The energy flow through an ecosystem maps closely tocarbon flow in the processes of photosynthesis,trophic transfers, and respiratory release ofcarbon. Lindeman’s dissertation research onthe trophic-dynamic aspect of ecology was ini-tially rejected for publication. Reviewers felt

that there were insufficient data to draw suchbroad conclusions and that it was inappropriateto use mathematical models to infer generalrelationships based on observations from asingle lake. Hutchinson, Lindeman’s postdoc-toral adviser, finally (after Lindeman’s death)persuaded the editor to publish this paper,which has been the springboard for many of thebasic concepts in ecosystem theory (Lindeman1942).

H. T. Odum, also trained by Hutchinson,and his brother E. P. Odum further developedthe systems approach to studying ecosystems,which emphasizes the general properties ofecosystems without documenting all the under-lying mechanisms and interactions. The Odumbrothers used radioactive tracers to measurethe movement of energy and materials througha coral reef. These studies enabled them to doc-ument the patterns of energy flow and metab-olism of whole ecosystems and to suggestgeneralizations about how ecosystems function(Odum 1969). Ecosystem budgets of energyand materials have since been developed formany fresh-water and terrestrial ecosystems(Lindeman 1942, Ovington 1962, Golley 1993),providing information that is essential for gen-eralizing about global patterns of processessuch as productivity. Some of the questionsaddressed by systems ecology include informa-tion transfer (Margalef 1968), the structure offood webs (Polis 1991), the hierarchical changesin ecosystem controls at different temporal and spatial scales (O’Neill et al. 1986), and theresilience of ecosystem properties after distur-bance (Holling 1986).

We now recognize that element cycles inter-act in important ways and cannot be under-stood in isolation. The availability of water andnitrogen are important determinants of the rateat which carbon cycles through the ecosystem.Conversely, the productivity of vegetationstrongly influences the cycling rates of nitrogenand water.

Recent global changes in the environmenthave made ecologists increasingly aware of thechanges in ecosystem processes that occur inresponse to disturbance or other environmen-tal changes. Succession, the directional changein ecosystem structure and functioning result-

History of Ecosystem Ecology 9

ing from biotically driven changes in resourcesupply, is an important framework for under-standing these transient dynamics of ecosys-tems. Early American ecologists such as Cowlesand Clements were struck by the relatively pre-dictable patterns of vegetation developmentafter exposure of unvegetated land surfaces.Sand dunes on Lake Michigan, for example, areinitially colonized by drought-resistant herba-ceous plants that give way to shrubs, then smalltrees, and eventually forests (Cowles 1899).Clements (1916) advanced a theory of commu-nity development, suggesting that this vegeta-tion succession is a predictable process thateventually leads, in the absence of disturbance,to a stable community type characteristic of aparticular climate (the climatic climax). He sug-gested that a community is like an organismmade of interacting parts (species) and thatsuccessional development toward a climaxcommunity is analogous to the development of an organism to adulthood. This analogybetween an ecological community and anorganism laid the groundwork for concepts ofecosystem physiology (for example, the netecosystem exchange of CO2 and water vaporbetween the ecosystem and the atmosphere).The measurements of net ecosystem exchangeare still an active area of research in ecosystemecology, although they are now motivated bydifferent questions than those posed byClements. His ideas were controversial fromthe outset. Other ecologists, such as Gleason(1926), felt that vegetation change was not aspredictable as Clements had implied. Instead,chance dispersal events explained much of the vegetation patterns on the landscape. Thisdebate led to a century of research on themechanisms responsible for vegetation change(see Chapter 13).

Another general approach to ecosystemecology has emphasized the controls overecosystem processes through comparativestudies of ecosystem components. This interestoriginated in studies by plant geographers andsoil scientists who described general patterns ofvariation with respect to climate and geologicalsubstrate (Schimper 1898). These studiesshowed that many of the global patterns ofplant production and soil development vary

predictably with climate (Jenny 1941, Rodinand Bazilevich 1967, Lieth 1975). The studiesalso showed that, in a given climatic regime, theproperties of vegetation depended strongly onsoils and vice versa (Dokuchaev 1879, Jenny1941, Ellenberg 1978). Process-based studies oforganisms and soils provided insight into manyof the mechanisms underlying the distributionsof organisms and soils along these gradients(Billings and Mooney 1968, Mooney 1972,Larcher 1995, Paul and Clark 1996). Thesestudies also formed the basis for extrapolationof processes across complex landscapes to char-acterize large regions (Matson and Vitousek1987, Turner et al. 2001). These studies oftenrelied on field or laboratory experiments that manipulated some ecosystem property orprocess or on comparative studies across envi-ronmental gradients. This approach was laterexpanded to studies of intact ecosystems, usingwhole-ecosystem manipulations (Likens et al.1977, Schindler 1985, Chapin et al. 1995) andcarefully designed gradient studies (Vitousek etal. 1988).

Ecosystem experiments have provided bothbasic understanding and information that arecritical in management decisions. The clear-cutting of an experimental watershed atHubbard Brook in the northeastern UnitedStates, for example, caused a fourfold increasein streamflow and stream nitrate concentra-tion—to levels exceeding health standards for drinking water (Likens et al. 1977). Thesedramatic results demonstrate the key role ofvegetation in regulating the cycling of waterand nutrients in forests.The results halted plansfor large-scale deforestation that had beenplanned to increase supplies of drinking waterduring a long-term drought. Nutrient addition experiments in the Experimental Lakes Area of southern Canada showed that phosphoruslimits the productivity of many lakes (Schindler1985) and that pollution was responsible for algal blooms and fish kills that werecommon in lakes near densely populated areasin the 1960s. This research provided the basisfor regulations that removed phosphorus fromdetergents.

Changes in the Earth System have led tostudies of the interactions among terrestrial


ecosystems, the atmosphere, and the oceans.The dramatic impact of human activities on theEarth System (Vitousek 1994a) has led to theurgent necessity to understand how terrestrialecosystem processes affect the atmosphere and oceans. The scale at which these ecosystemeffects are occurring is so large that the traditional tools of ecologists are insufficient.Satellite-based remote sensing of ecosystemproperties, global networks of atmosphericsampling sites, and the development of globalmodels are important new tools that addressglobal issues. Information on global patterns ofCO2 and pollutants in the atmosphere, forexample, provide telltale evidence of the majorlocations and causes of global problems (Tanset al. 1990).This gives hints about which ecosys-tems and processes have the greatest impact onthe Earth System and therefore where researchand management should focus efforts to under-stand and solve these problems (Zimov et al.1999).

The intersection of systems approaches,process understanding, and global analysis is anexciting frontier of ecosystem ecology. How dochanges in the global environment alter thecontrols over ecosystem processes? What arethe integrated system consequences of thesechanges? How do these changes in ecosystemproperties influence the Earth System? Therapid changes that are occurring in ecosystemshave blurred any previous distinction betweenbasic and applied research. There is an urgentneed to understand how and why the ecosys-tems of Earth are changing.

Ecosystem Structure

Most ecosystems gain energy from the sun andmaterials from the air or rocks, transfer theseamong components within the ecosystem, thenrelease energy and materials to the environ-ment. The essential biological components ofecosystems are plants, animals, and decom-posers. Plants capture solar energy in theprocess of bringing carbon into the ecosystem.A few ecosystems, such as deep-sea hydro-thermal vents, have no plants but instead have bacteria that derive energy from the

oxidation of hydrogen sulfide (H2S) to produce organic matter. Decomposer microorganisms(microbes) break down dead organic material,releasing CO2 to the atmosphere and nutrientsin forms that are available to other microbesand plants. If there were no decomposition,large accumulations of dead organic matterwould sequester the nutrients required tosupport plant growth. Animals are critical com-ponents of ecosystems because they transferenergy and materials and strongly influence thequantity and activities of plants and soilmicrobes. The essential abiotic components ofan ecosystem are water; the atmosphere, whichsupplies carbon and nitrogen; and soil minerals,which supply other nutrients required byorganisms.

An ecosystem model describes the majorpools and fluxes in an ecosystem and the factorsthat regulate these fluxes. Nutrients, water, andenergy differ from one another in the relativeimportance of ecosystem inputs and outputs vs.internal recycling (see Chapters 4 to 10). Plants,for example, acquire carbon primarily from theatmosphere, and most carbon released by res-piration returns to the atmosphere. Carboncycling through ecosystems is therefore quiteopen, with large inputs to, and losses from,the system. There are, however, relatively largepools of carbon stored in ecosystems, so theactivities of animals and microbes are some-what buffered from variations in carbon up-take by plants. The water cycle of ecosystems is also relatively open, with water entering primarily by precipitation and leaving by evap-oration, transpiration, and drainage to ground-water and streams. In contrast to carbon,most ecosystems have a limited capacity tostore water in plants and soil, so the activity oforganisms is closely linked to water inputs.In contrast to carbon and water, mineral ele-ments such as nitrogen and phosphorus arerecycled rather tightly within ecosystems, withannual inputs and losses that are small relativeto the quantities that annually recycle withinthe ecosystem. These differences in the “open-ness” and “buffering” of the cycles fundamen-tally influence the controls over rates andpatterns of the cycling of materials throughecosystems.

Controls over Ecosystem Processes 11

The pool sizes and rates of cycling differ substantially among ecosystems (see Chapter6). Tropical forests have much larger pools of carbon and nutrients in plants than dodeserts or tundra. Peat bogs, in contrast,have large pools of soil carbon rather than plant carbon. Ecosystems also differ substan-tially in annual fluxes of materials among pools, for reasons that will be explored in later chapters.

Controls over Ecosystem Processes

Ecosystem structure and functioning are gov-erned by at least five independent control variables. These state factors, as Jenny and co-workers called them, are climate, parentmaterial (i.e., the rocks that give rise to soils),topography, potential biota (i.e., the organismspresent in the region that could potentiallyoccupy a site), and time (Fig. 1.3) (Jenny 1941,Amundson and Jenny 1997). Together thesefive factors set the bounds for the characteris-tics of an ecosystem.

On broad geographic scales, climate is thestate factor that most strongly determinesecosystem processes and structure. Global

variations in climate explain the distribution ofbiomes (types of ecosystems) such as wet trop-ical forests, temperate grasslands, and arctictundra (see Chapter 2). Within each biome,parent material strongly influences the types ofsoils that develop and explains much of theregional variation in ecosystem processes (seeChapter 3). Topographic relief influences bothmicroclimate and soil development at a localscale. The potential biota governs the types anddiversity of organisms that actually occupy a site. Island ecosystems, for example, are frequently less diverse than climatically similarmainland ecosystems because new speciesreach islands less frequently and are morelikely to go extinct than in mainland locations(MacArthur and Wilson 1967). Time influencesthe development of soil and the evolution of organisms over long time scales. Time alsoincorporates the influences on ecosystemprocesses of past disturbances and environ-mental changes over a wide range of timescales. These state factors are described in more detail in Chapter 3 in the context of soildevelopment.

Jenny’s state factor approach was a majorconceptual contribution to ecosystem ecology.First, it emphasized the controls over processesrather than simply descriptions of patterns.Second, it suggested an experimental approachto test the importance and mode of action ofeach control. A logical way to study the role ofeach state factor is to compare sites that are assimilar as possible with respect to all but onefactor. For example, a chronosequence is aseries of sites of different ages with similarclimate, parent material, topography, andpotential to be colonized by the same organ-isms (see Chapter 13). In a toposequence,ecosystems differ mainly in their topographicposition (Shaver et al. 1991). Sites that differprimarily with respect to climate or parentmaterial allow us to study the impact of thesestate factors on ecosystem processes (Vitouseket al. 1988, Walker et al. 1998). Finally, a com-parison of ecosystems that differ primarily inpotential biota, such as the mediterraneanshrublands that have developed on west coastsof California, Chile, Portugal, South Africa, andAustralia, illustrates the importance of evolu-

Time

Topo-graphy

Climate

Parentmaterial

Potentialbiota

Ecosystemprocesses

Modulators

Resources Bioticcommunity

Disturbanceregime

Humanactivities

Figure 1.3. The relationship between state factors(outside the circle), interactive controls (inside thecircle), and ecosystem processes. The circle repre-sents the boundary of the ecosystem. (Modified withpermission from American Naturalist, Vol. 148 ©1996 University of Chicago Press, Chapin et al. 1996.)


tionary history in shaping ecosystem processes(Mooney and Dunn 1970).

Ecosystem processes both respond to andcontrol the factors that directly govern theiractivity. For example, plants both respond toand influence their light, temperature, andmoisture environment (Billings 1952). Interac-tive controls are factors that both control andare controlled by ecosystem characteristics (Fig.1.3) (Chapin et al. 1996). Important interactivecontrols include the supply of resources tosupport the growth and maintenance of organ-isms, modulators that influence the rates ofecosystem processes, disturbance regime, thebiotic community, and human activities.

Resources are the energy and materials in theenvironment that are used by organisms tosupport their growth and maintenance (Field et al. 1992). The acquisition of resources byorganisms depletes their abundance in the environment. In terrestrial ecosystems theseresources are spatially separated, being avail-able primarily either aboveground (light andCO2) or belowground (water and nutrients).Resource supply is governed by state factorssuch as climate, parent material, and topogra-phy. It is also sensitive to processes occurringwithin the ecosystem. Light availability, forexample, depends on climatic elements such ascloudiness and on topographic position, but isalso sensitive to the quantity of shading by vegetation. Similarly, soil fertility depends onparent material and climate but is also sensitiveto ecosystem processes such as erosional loss ofsoils after overgrazing and inputs of nitrogenfrom invading nitrogen-fixing species. Soilwater availability strongly influences speciescomposition in dry climates. Soil water avail-ability also depends on other interactive controls, such as disturbance regime (e.g., com-paction by animals) and the types of organismsthat are present (e.g., the presence or absence ofdeep-rooted trees such as mesquite that tap thewater table). In aquatic ecosystems, waterseldom directly limits the activity of organisms,but light and nutrients are just as important as on land. Oxygen is a particularly criticalresource in aquatic ecosystems because of itsslow rate of diffusion through water.

Modulators are physical and chemical prop-erties that affect the activity of organisms but,

unlike resources, are neither consumed nordepleted by organisms (Field et al. 1992). Mod-ulators include temperature, pH, redox state ofthe soil, pollutants, UV radiation, etc. Modula-tors like temperature are constrained byclimate (a state factor) but are sensitive toecosystem processes, such as shading and evap-oration. Soil pH likewise depends on parentmaterial and time but also responds to vegeta-tion composition.

Landscape-scale disturbance by fire, wind,floods, insect outbreaks, and hurricanes is a crit-ical determinant of the natural structure andprocess rates in ecosystems (Pickett and White1985, Sousa 1985). Like other interactive con-trols, disturbance regime depends on both statefactors and ecosystem processes. Climate,for example, directly affects fire probability and spread but also influences the types andquantity of plants present in an ecosystem and therefore the fuel load and flammability of vegetation. Deposition and erosion duringfloods shape river channels and influence the probability of future floods. Change in either the intensity or frequency of disturbance can cause long-term ecosystem change. Woodyplants, for example, often invade grasslandswhen fire suppression reduces fire frequency.

The nature of the biotic community (i.e., thetypes of species present, their relative abun-dances, and the nature of their interactions) can influence ecosystem processes just asstrongly as do large differences in climate orparent material (see Chapter 12). These specieseffects can often be generalized at the level offunctional types, which are groups of speciesthat are similar in their role in community orecosystem processes. Most evergreen trees, forexample, produce leaves that have low rates ofphotosynthesis and a chemical compositionthat deters herbivores. These species make up afunctional type because of their ecological sim-ilarity to one another. A gain or loss of keyfunctional types—for example, through intro-duction or removal of species with importantecosystem effects—can permanently changethe character of an ecosystem through changesin resource supply or disturbance regime.Introduction of nitrogen-fixing trees ontoBritish mine wastes, for example, substantiallyincreases nitrogen supply and productivity

Human-Caused Changes in Earth’s Ecosystems 13

and alters patterns of vegetation development(Bradshaw 1983). Invasion by exotic grassescan alter fire frequency, resource supply, tro-phic interactions, and rates of most ecosystemprocesses (D’Antonio and Vitousek 1992).Elimination of predators by hunting can causean outbreak of deer that overbrowse their foodsupply. The types of species present in anecosystem depend strongly on other interactivecontrols (see Chapter 12), so functional typesrespond to and affect most interactive controlsand ecosystem processes.

Human activities have an increasing impacton virtually all the processes that govern ecosys-tem properties (Vitousek 1994a). Our actionsinfluence interactive controls such as wateravailability, disturbance regime, and bioticdiversity. Humans have been a natural compo-nent of many ecosystems for thousands of years.Since the Industrial Revolution, however, themagnitude of human impact has been so greatand so distinct from that of other organisms thatthe modern effects of human activities warrantparticular attention. The cumulative impact ofhuman activities extend well beyond an individ-ual ecosystem and affect state factors such asclimate, through changes in atmospheric com-position, and potential biota, through the intro-duction and extinction of species. The largemagnitude of these effects blurs the distinctionbetween “independent” state factors and inter-active controls at regional and global scales.Human activities are causing major changes inthe structure and functioning of all ecosystems,resulting in novel conditions that lead to newtypes of ecosystems. The major human effectsare summarized in the next section.

Feedbacks analogous to those in simple phys-ical systems regulate the internal dynamics ofecosystems. A thermostat is an example of asimple physical feedback. It causes a furnace toswitch on when a house gets cold. The housethen warms until the thermostat switches thefurnace off. Natural ecosystems are complexnetworks of interacting feedbacks (DeAngelisand Post 1991). Negative feedbacks occur whentwo components of a system have oppositeeffects on one another. Consumption of prey bya predator, for example, has a positive effect onthe consumer but a negative effect on the prey.The negative effect of predators on prey pre-

vents an uncontrolled growth of a predator’spopulation, thereby stabilizing the populationsizes of both predator and prey. There are alsopositive feedbacks in ecosystems in which bothcomponents of a system have a positive effecton the other, or both have a negative effect onone another. Plants, for example, provide theirmycorrhizal fungi with carbohydrates in returnfor nutrients. This exchange of growth-limitingresources between plants and fungi promotesthe growth of both components of the sym-biosis until they become constrained by otherfactors.

Negative feedbacks are the key to sustainingecosystems because strong negative feedbacksprovide resistance to changes in interactivecontrols and maintain the characteristics ofecosystems in their current state. The acquisi-tion of water, nutrients, and light to supportgrowth of one plant, for example, reduces avail-ability of these resources to other plants,thereby constraining community productivity(Fig. 1.4). Similarly, animal populations cannotsustain exponential population growth indefi-nitely, because declining food supply andincreasing predation reduce the rate of popu-lation increase. If these negative feedbacks are weak or absent (a low predation rate dueto predator control, for example), populationcycles can amplify and lead to extinction of oneor both of the interacting species. Communitydynamics, which operate within a single eco-system patch, primarily involve feedbacksamong soil resources and functional types oforganisms. Landscape dynamics, which governchanges in ecosystems through cycles of dis-turbance and recovery, involve additional feedbacks with microclimate and disturbanceregime (see Chapter 14).

Human-Caused Changes inEarth’s Ecosystems

Human activities transform the land surface,add or remove species, and alter biogeochemi-cal cycles. Some human activities directly affectecosystems through activities such as resourceharvest, land use change, and management;other effects are indirect, as a result of changesin atmospheric chemistry, hydrology, and


climate (Fig. 1.5) (Vitousek et al. 1997c). Atleast some of these anthropogenic (i.e., human-caused) effects influence all ecosystems onEarth.

The most direct and substantial human alter-ation of ecosystems is through the transforma-tion of land for production of food, fiber, andother goods used by people. About 50% ofEarth’s ice-free land surface has been directlyaltered by human activities (Kates et al. 1990).Agricultural fields and urban areas cover 10 to15%, and pastures cover 6 to 8% of the land.Even more land is used for forestry and grazingsystems. All except the most extreme environ-ments of Earth experience some form of directhuman impact.

Human activities have also altered fresh-water and marine ecosystems. We use about

half of the world’s accessible runoff (seeChapter 15), and humans use about 8% of theprimary production of the oceans (Pauly andChristensen 1995). Commercial fishing reducesthe size and abundance of target species andalters the population characteristics of speciesthat are incidentally caught in the fishery. In themid-1990s, about 22% of marine fisheries were overexploited or already depleted, and anadditional 44% were at their limit of exploita-tion (Vitousek et al. 1997c). About 60% of thehuman population resides within 100km of acoast, so the coastal margins of oceans arestrongly influenced by many human activities.Nutrient enrichment of many coastal waters, forexample, has increased algal production andcreated anaerobic conditions that kill fish andother animals, due largely to transport of nutri-ents derived from agricultural fertilizers andfrom human and livestock sewage.

Land use change, and the resulting loss ofhabitat, is the primary driving force causingspecies extinctions and loss of biological diver-sity (Sala et al. 2000a) (see Chapter 12). Thetime lag between ecosystem change and speciesloss makes it likely that species will continue tobe driven to extinction even where rates of landuse change have stabilized. Transport of speciesaround the world is homogenizing Earth’sbiota. The frequency of biological invasions is increasing, due to the globalization of theeconomy and increased international transportof products. Nonindigenous species nowaccount for 20% or more of the plant species in many continental areas and 50% or more ofthe plant species on many islands (Vitousek et al. 1997c). International commerce breaksdown biogeographic barriers, through both purposeful trade in live organisms and inad-vertent introductions. Purposeful introduc-tions deliberately select species that are likely to grow and reproduce effectively in their new environment. Many biological invasionsare irreversible because it is difficult or prohibitively expensive to remove invasivespecies. Some species invasions degrade human health or cause large economic losses.Others alter the structure and functioning ofecosystems, leading to further loss of speciesdiversity.

+-

Predator

Herbivore

Plant A Plant B

Shared resources

Mycorrhizalfungus

A B

CD

E

F

+

+

++

+

-

Resource uptakeCompetitionMutualismHerbivoryPredationPopulation growth

Process Nature of feedback

AA+B

CDEF

--+--+

+

-

-

Figure 1.4. Examples of linked positive and nega-tive feedbacks in ecosystems. The effect of eachorganism (or resource) on other organisms can bepositive (+) or negative (-). Feedbacks are positivewhen the reciprocal effects of each organism (orresource) have the same sign (both positive or bothnegative). Feedbacks are negative when reciprocaleffects differ in sign. Negative feedbacks resist thetendencies for ecosystems to change, whereas posi-tive feedbacks tend to push ecosystems toward a newstate. (Modified with permission from American Nat-uralist, Vol. 148 © 1996 University of Chicago Press,Chapin et al. 1996.)

Human-Caused Changes in Earth’s Ecosystems 15

Human activities have influenced biogeo-chemical cycles in many ways. Use of fossil fuelsand the expansion and intensification of agri-culture have altered the cycles of carbon, nitro-gen, phosphorus, sulfur, and water on a globalscale (see Chapter 15). These changes in bio-geochemical cycles not only alter the ecosys-tems in which they occur but also influenceunmanaged ecosystems through changes inlateral fluxes of nutrients and other materialsthrough the atmosphere and surface waters(see Chapter 14). Land use changes, includingdeforestation and intensive use of fertilizersand irrigation, have increased the concentra-tions of atmospheric gases that influenceclimate (see Chapter 2). Land transformationsalso cause runoff and erosion of sediments andnutrients that lead to substantial changes inlakes, rivers, and coastal oceans.

Human activities introduce novel chemicalsinto the environment. Some apparently harm-

less anthropogenic gases have had drasticeffects on the atmosphere and ecosystems.Chlorofluorocarbons (CFCs), for example,were first produced in the 1950s as refrigerants,propellants, and solvents. They were heraldedfor their nonreactivity in the lower atmosphere.In the upper atmosphere, however, where thereis greater UV radiation, CFCs react with ozone.The resulting ozone destruction, which occursprimarily over the poles, creates a hole in theprotective blanket of ozone that shields Earth’ssurface from UV radiation. This ozone holewas initially observed near the South Pole. Ithas expanded to lower latitudes in the South-ern Hemisphere and now also occurs at highnorthern latitudes. As a result of the MontrealProtocol, the production of many CFCs hasceased. Due to their low reactivity, however,their concentrations in the atmosphere are onlynow beginning to decline, so their ecologicaleffects will persist for decades. Persistent novel

Human populationSize Resource use

Human enterprisesAgriculture Industry Recreation International commerce

Landtransformation

Land clearingIntensification

ForestryGrazing

Biotic additionsand losses

InvasionHuntingFishing

Globalbiochemistry

WaterCarbonNitrogen

Other elementsSynthetic chemicals

Radionuclides

Climate change

Enhancedgreenhouse effect

AerosolsLand cover

Loss of biologicaldiversity

Extinction of speciesand populations

Loss of ecosystems

Figure 1.5. Direct and indirecteffects of human activities on Earth’secosystems. (Redrawn with permis-sion from Science, Vol. 277 © 1997American Association for theAdvancement of Science; Vitouseket al. 1997c.)


chemicals, such as CFCs, often have long-lastingecological effects than cannot be predicted atthe time they are first produced and whichextend far beyond their region and duration ofuse.

Other synthetic organic chemicals includeDDT (an insecticide) and polychlorinatedbiphenyls (PCBs; industrial compounds), whichwere used extensively in the developed worldin the 1960s before their ecological conse-quences were widely recognized. Many of thesecompounds continue to be used in some devel-oping nations. They are mobile and degradeslowly, causing them to persist and to be trans-ported to all ecosystems of the globe. Many of these compounds are fat soluble, so theyaccumulate in organisms and become increas-ingly concentrated as they move through foodchains (see Chapter 11). When these com-pounds reach critical concentrations, they cancause reproductive failure. This occurs most frequently in higher trophic levels and inanimals that feed on fat-rich species. Someprocesses, such as eggshell formation in birds,are particularly sensitive to pesticide accumu-lations, and population declines in predatorybirds like the perigrine falcon have been notedin regions far removed from the locations ofpesticide use.

Atmospheric testing of atomic weapons inthe 1950s and 1960s increased the concentra-tions of radioactive forms of many elements.Explosions and leaks in nuclear reactors usedto generate electricity continue to be regionalor global sources of radioactivity.The explosionof a power-generating plant in 1986 at Chernobyl in Ukraine, for example, releasedsubstantial radioactivity that directly affectedhuman health in the region and increased theatmospheric deposition of radioactive mate-rials over eastern Europe and Scandinavia.Some radioactive isotopes of atoms, such asstrontium (which is chemically similar tocalcium) and cesium (which is chemicallysimilar to potassium) are actively accumulatedand retained by organisms. Lichens, forexample, acquire their minerals primarily fromthe atmosphere rather than from the soil andactively accumulate cesium and strontium.Reindeer, which feed on lichens, further con-

centrate cesium and strontium, as do peoplewho feed on reindeer. For this reason, the inputof radioisotopes into the atmosphere or waterfrom nuclear power plants, submarines, andweapons has had impacts that extend farbeyond the regions where they were used.

The growing scale and extent of human activ-ities suggest that all ecosystems are being influ-enced, directly or indirectly, by our activities.No ecosystem functions in isolation, and all areinfluenced by human activities that take placein adjacent communities and around the world.Human activities are leading to global changesin most major ecosystem controls: climate(global warming), soil and water resources(nitrogen deposition, erosion, diversions), dis-turbance regime (land use change, fire control),and functional types of organisms (speciesintroductions and extinctions). Many of theseglobal changes interact with each other atregional and local scales. Therefore, all eco-systems are experiencing directional changes in ecosystem controls, creating novel condi-tions and, in many cases, positive feedbacks that lead to new types of ecosystems. Thesechanges in interactive controls will inevit-ably change the properties of ecosystems andmay lead to unpredictable losses of ecosys-tem functions on which human communitiesdepend. In the following chapters we point outmany of the ecosystem processes that havebeen affected.

Summary

Ecosystem ecology addresses the interactionsamong organisms and their environment as anintegrated system through study of the factorsthat regulate the pools and fluxes of materialsand energy through ecological systems. Thespatial scale at which we study ecosystems ischosen to facilitate the measurement of impor-tant fluxes into, within, and out of the ecosys-tem. The functioning of ecosystems dependsnot only on their current structure and envi-ronment but also on past events and distur-bances and the rate at which ecosystemsrespond to past events. The study of ecosystemecology is highly interdisciplinary and builds on

Additional Reading 17

many aspects of ecology, hydrology, climatol-ogy, and geology and contributes to currentefforts to understand Earth as an integratedsystem. Many unresolved problems in ecosys-tem ecology require an integration of systemsapproaches, process understanding, and globalanalysis.

Most ecosystems ultimately acquire theirenergy from the sun and their materials fromthe atmosphere and rock minerals. The energyand materials are transferred among compo-nents within the ecosystem and are thenreleased to the environment. The essentialbiotic components of ecosystems includeplants, which bring carbon and energy into theecosystem; decomposers, which break downdead organic matter and release CO2 and nutri-ents; and animals, which transfer energy andmaterials within ecosystems and modulate theactivity of plants and decomposers. The essen-tial abiotic components of ecosystems are theatmosphere, water, and rock minerals. Ecosys-tem processes are controlled by a set of rela-tively independent state factors (climate, parentmaterial, topography, potential biota, and time)and by a group of interactive controls (includ-ing resource supply, modulators, disturbanceregime, functional types of organisms, andhuman activities) that are the immediate con-trols over ecosystem processes. The interactivecontrols both respond to and affect ecosystemprocesses. The stability and resilience of eco-systems depend on the strength of negativefeedbacks that maintain the characteristics ofecosystems in their current state.

Review Questions

1. What is an ecosystem? How does it differfrom a community? What kinds of environ-mental questions can be addressed byecosystem ecology that are not readilyaddressed by population or communityecology?

2. What is the difference between a pool and aflux? Which of the following are pools andwhich are fluxes: plants, plant respiration,

rainfall, soil carbon, consumption of plantsby animals?

3. What are the state factors that control thestructure and rates of processes in ecosys-tems? What are the strengths and limitationsof the state factor approach to answeringthis question.

4. What is the difference between state factorsand interactive controls? If you were askedto write a management plan for a region,why would you treat a state factor and an interactive control differently in yourplan?

5. Using a forest or a lake as an example,explain how climatic warming or the harvestof trees or fish by people might change themajor interactive controls. How might thesechanges in controls alter the structure of orprocesses in these ecosystems?

6. Use examples to show how positive and neg-ative feedbacks might affect the responses ofan ecosystem to climatic change.

Additional Reading

Chapin, F.S. III, M.S. Torn, and M. Tateno. 1996. Prin-ciples of ecosystem sustainability. American Natu-ralist 148:1016–1037.

Golley, F.B. 1993. A History of the EcosystemConcept in Ecology: More Than the Sum of theParts. Yale University Press, New Haven, CT.

Gorham, E. 1991. Biogeochemistry: Its origins anddevelopment. Biogeochemistry 13:199–239.

Hagen, J.B. 1992. An Entangled Bank: The Origins of Ecosystem Ecology. Rutgers University Press,New Brunswick, NJ.

Jenny, H. 1980. The Soil Resources: Origin andBehavior. Springer-Verlag, New York.

Lindeman, R.L. 1942.The trophic-dynamic aspects ofecology. Ecology 23:399–418.

Schlesinger, W.H. 1997. Biogeochemistry: An Ana-lysis of Global Change. Academic Press, SanDiego.

Sousa, W.P. 1985. The role of disturbance in naturalcommunities. Annual Review of Ecology and Systematics 15:353–391.

Tansley,A.G. 1935.The use and abuse of vegetationalconcepts and terms. Ecology 16:284–307.

Vitousek, P.M. 1994. Beyond global warming:Ecology and global change. Ecology 75:1861–1876.

Introduction

Climate exerts a key control over the distri-bution of Earth’s ecosystems. Temperature and water availability determine the rates atwhich many biological and chemical reactionscan occur. These reaction rates control critical ecosystem processes, such as the pro-duction of organic matter by plants and itsdecomposition by microbes. Climate also con-trols the weathering of rocks and the devel-opment of soils, which in turn influenceecosystem processes (see Chapter 3). Under-standing the causes of temporal and spatialvariation in climate is therefore critical tounderstanding the global pattern of ecosystemprocesses.

Climate and climate variability are deter-mined by the amount of incoming solar radia-tion, the chemical composition and dynamics ofthe atmosphere, and the surface characteristicsof Earth.The circulation of the atmosphere andoceans influences the transfer of heat and mois-ture around the planet and thus strongly influ-ences climate patterns and their variability inspace and time. This chapter describes theglobal energy budget and outlines the roles that the atmosphere, oceans, and land surfaceplay in the redistribution of energy to produce

observed patterns of climate and ecosystem distribution.

Earth’s Energy Budget

The balance between incoming and outgoingradiation determines the energy available todrive Earth’s climate system. An understandingof the components of Earth’s energy budgetprovides a basis for determining the causes ofrecent and long-term changes in climate. Thesun is the source of virtually all of Earth’senergy. The temperature of a body determinesthe wavelengths of energy emitted. The hightemperature of the sun (6000K) results in emis-sions of high-energy shortwave radiation withwavelengths of 300 to 3000nm (Fig. 2.1). Theseinclude visible (39% of the total), near-infrared(53%), and ultraviolet (UV) radiation (8%).On average, about 31% of the incoming short-wave radiation is reflected back to space, due tobackscatter (reflection) from clouds (16%); airmolecules, dust, and haze (7%); and Earth’ssurface (8%) (Fig. 2.2). Another 20% of theincoming shortwave radiation is absorbed bythe atmosphere, especially by ozone in theupper atmosphere and by clouds and watervapor in the lower atmosphere. The remaining

2Earth’s Climate System

Climate is the state factor that most strongly governs the global distribution of terrestrial biomes. This chapter provides a general background on the functioningof the climate system and its interactions with atmospheric chemistry, oceans,and land.

18

Earth’s Energy Budget 19

49% reaches Earth’s surface as direct or diffuseradiation and is absorbed.

Over time scales of a year or more, Earth isin a state of radiative equilibrium, meaning thatit releases as much energy as it absorbs. Onaverage, Earth emits 79% of the absorbedenergy as low-energy longwave radiation (3000to 30,000nm), due to its relatively low surfacetemperature (288K). The remaining energy istransferred from Earth’s surface to the atmos-phere by the evaporation of water (latent heatflux) (16% of terrestrial energy loss) or by the

transfer of heat to the air from the warmsurface to the cooler overlying atmosphere(sensible heat flux) (5% of terrestrial energyloss) (Fig. 2.2). Heat absorbed from the surfacewhen water evaporates is subsequentlyreleased to the atmosphere when water vaporcondenses, resulting in formation of clouds andprecipitation.

Although the atmosphere transmits abouthalf of the incoming shortwave radiation toEarth’s surface, it absorbs 90% of the longwave(infrared) radiation emitted by the surface (Fig. 2.2). Water vapor, carbon dioxide (CO2),methane (CH4), nitrous oxide (N2O), andindustrial products like chlorofluorocarbons(CFCs) effectively absorb longwave radiation(Fig. 2.1). The energy absorbed by these radiatively active gases is reradiated in all directions as longwave radiation (Fig. 2.2).The portion that is directed back toward thesurface contributes to the warming of theplanet, a phenomenon know as the greenhouseeffect. Without a longwave-absorbing atmos-phere, the mean temperature at Earth’s sur-face would be about 33°C lower than it is today and would probably not support life.Radiation absorbed by clouds and radiativelyactive gases is also emitted back to space,balancing the incoming shortwave radiation(Fig. 2.2).

Long-term records of atmospheric gases,obtained from atmospheric measurementsmade since the 1950s and from air bubblestrapped in glacial ice, demonstrate large in-creases in the major radiatively active gases(CO2, CH4, N2O, and CFCs) since the beginningof the Industrial Revolution 150 years ago (see Fig. 15.3). Human activities such as fossilfuel burning, industrial activities, animal hus-bandry, and fertilized and irrigated agriculturecontribute to these increases.As concentrationsof these gases rise, more of the longwave radi-ation emitted by Earth is trapped by the atmos-phere, enhancing the greenhouse effect andcausing the surface temperature of Earth toincrease.

The globally averaged energy budget out-lined above gives us a sense of the criticalfactors controlling the global climate system.Regional climates, however, reflect spatial

Ene

rgy

(W m

-2)

Sol

ar in

com

ing

radi

atio

n

Terrestrial outgoing radiation

0 5 10 15 20 25

Visible region

Wavelength (µm)

Abs

orpt

ivity

0

0

0

0

0

0

0

1

1

1

1

1

1

CH4

N2O

O2 and O3

H2O

5 10 15 20 25

CO2

Atmosphere

Figure 2.1. The spectral distribution of solar andterrestrial radiation and the absorption spectra ofthe major radiatively active gases and of the totalatmosphere.These spectra show that the atmosphereabsorbs terrestrial radiation more effectively thansolar radiation, explaining why the atmosphere isheated from below. (Sturman and Tapper 1996, Barryand Chorley 1970.)

20 2. Earth’s Climate System

variability in energy exchange and in heattransport by the atmosphere and oceans. Earthexperiences greater heating at the equator thanat the poles, and it rotates on a tilted axis. Itscontinents are spread unevenly over thesurface, and its atmospheric and oceanic chem-

istry and physics are dynamic and spatially vari-able. A more thorough understanding of the atmosphere and oceans is therefore needed to understand the fate and processing of energy and its consequences for the ecosystems of theplanet.

Atmosphere

Earth

Space

Outgoing radiation

Shortwave

Incomingsolar radiation

Absorption by H2O,dust, O3

Absorptionby clouds

Absorption of direct

solar radiation

Absorptionof diffuse sky

and cloud radiation

Refle

cted

by

surfa

ce

Backscatterby clouds

Backscatterby air

Net absorption by clouds,

CO2,O3

Emission byH2O, CO2, O3

Emissionby clouds

Convectivemixing

Longwave Convection

Lat

ent

hea

t flu

x

Sens

ible

heat

flux

Net longwave reradiation

100

17

3

7

16

23114

102

12

95

723

69

9

48

8

26

31

Figure 2.2. The average annual global energybalance for the Earth–atmosphere system. Thenumbers are percentages of the energy received asincoming solar radiation. At the top of the atmos-phere, the incoming solar radiation (100% or 342Wm-2) is balanced by reflected shortwave radiation(31%) and emitted longwave radiation (69%).Within the atmosphere, the absorbed shortwaveradiation (20%) and absorbed longwave radiation

(102%) and latent plus sensible heat flux (30%) arebalanced by longwave emission to space (57%) andlongwave emission to Earth’s surface (95%). AtEarth’s surface the incoming shortwave radiation(49%) and incoming longwave radiation (95%) arebalanced by outgoing longwave radiation (114%)and latent plus sensible heat flux (30%) (Graedeland Crutzen 1995, Sturman and Tapper 1996, Baedeet al. 2001).

The Atmospheric System 21

The Atmospheric System

Atmospheric Composition and Chemistry

The chemical composition of the atmospheredetermines its role in Earth’s energy budget.Think of the atmosphere as a giant reactionflask, containing thousands of different chemi-cal compounds in gas and particulate forms,undergoing slow and fast reactions, dissolutionsand precipitations. These reactions control the composition of the atmosphere and manyof its physical processes, such as cloud for-mation. These physical processes, in turn,generate dynamical motions crucial for energyredistribution.

More than 99.9% by volume of Earth’satmosphere is composed of nitrogen, oxygen,and argon. Carbon dioxide, the next most abun-dant gas, accounts for only 0.0367% of theatmosphere (Table 2.1). These percentages arequite constant around the world and up to 80km in height above the surface. That homo-geneity reflects the fact that these gases havelong mean residence times (MRTs) in theatmosphere. MRT is calculated as the totalmass divided by the flux into or out of theatmosphere over a given time period. Nitrogenhas an MRT of 13 million years; O2, 10,000years; and CO2, 4 years. In contrast, the MRTfor water vapor is only about 10 days, so its con-centration in the atmosphere is highly variable,depending on regional variations in surfaceevaporation, precipitation, and horizontaltransport of water vapor. Some of the mostimportant radiatively active gases, such as CO2,N2O, CH4, and CFCs, react relatively slowly inthe atmosphere and have residence times ofyears to decades. Other gases are much more

reactive and have residence times of days tomonths. Reactive species occur in traceamounts and make up less than 0.001% of thevolume of the atmosphere. Because of theirgreat reactivity, they are quite variable in timeand place. Some of the consequences of reac-tions among these trace species, such as smog,acid rain, and ozone depletion, threaten the sus-tainability of ecological systems (Graedel andCrutzen 1995).

Some atmospheric gases are critical for life.Photosynthetic organisms use CO2 in the pres-ence of light to produce organic matter thateventually becomes the basic food source for allanimals and microbes (see Chapters 5 to 7).Most organisms also require oxygen for meta-bolic respiration. Dinitrogen (N2) makes up78% of the atmosphere. It is unavailable tomost organisms, but nitrogen-fixing bacteriaconvert it to biologically available nitrogen thatis ultimately used by all organisms in buildingproteins (see Chapter 8). Other gases, such ascarbon monoxide (CO), nitric oxide (NO), N2O,CH4, and volatile organic carbon compoundslike terpenes and isoprene, are the products ofplant and microbial activity. Some, like tropos-pheric ozone (O3), are produced in the atmos-phere as products of chemical reactionsinvolving both biogenic (biologically produced)and anthropogenic gases and can, at high concentrations, damage plants, microbes, andhumans.

The atmosphere also contains aerosols,which are small particles suspended in air.Some aerosol particles arise from volcaniceruptions and from blowing dust and sea salt.Others are produced by reactions with gasesfrom pollution sources and biomass burning.Some aerosols are hydroscopic—that is, theyhave an affinity for water. Aerosols areinvolved in reactions with gases and act ascloud condensation nuclei around which watervapor condenses to form cloud droplets.Together with gases and clouds, aerosols deter-mine the reflectivity (albedo) of the atmos-phere and therefore exert major control overthe energy budget of the atmosphere. The scat-tering (reflection) of incoming shortwave radiation by aerosols reduces the radiationreaching Earth’s surface, which tends to cool

Table 2.1. Major chemical constituents of theatmosphere.

Compound Formula Concentration (%)

Nitrogen N2 78.084Oxygen O2 20.946Argon Ar 0.934Carbon dioxide CO2 0.037

Data from Schlesinger (1997) and Prentice et al. (2001).


the climate. The sulfur released to the atmos-phere by the volcanic eruption of MountPinatubo in the Philippines in 1991, forexample, caused a temporary atmosphericcooling throughout the globe.

Clouds have complex effects on Earth’s radi-ation budget. All clouds have a relatively highalbedo and reflect more incoming shortwaveradiation than does the darker Earth surface.Clouds, however, are composed of water vapor,which is a very efficient absorber of longwaveradiation. All clouds absorb and re-emit muchof the longwave radiation impinging on themfrom Earth’s surface. The first process (reflect-ing shortwave radiation) has a cooling effect byreflecting incoming energy back to space. Thesecond effect (absorbing longwave radiation)has a warming effect, by keeping more energyin the Earth System from escaping to space.Thebalance of these two effects depends on theheight of the cloud. The reflection of shortwaveradiation usually dominates the balance in highclouds, causing cooling; whereas the absorptionand re-emission of longwave radiation gener-ally dominates in low clouds, producing a netwarming effect.

Atmospheric Structure

Atmospheric pressure and density decline withheight above Earth’s surface. The average ver-tical structure of the atmosphere defines fourrelatively distinct layers characterized by theirtemperature profiles. The atmosphere is highlycompressible, and gravity keeps most of themass of the atmosphere close to Earth’s sur-face. Pressure, which is determined by the massof the overlying atmosphere, decreases expo-nentially with height. The vertical decline in airdensity tends to follow closely that of pressure.The relationships between pressure, density,and height can be described in terms of thehydrostatic equation

(2.1)

where P is pressure, h is height, r is density, andg is gravitational acceleration. The hydrostaticequation states that the vertical change in pres-sure is balanced by the product of density andgravitational acceleration (a “constant” that

dPdh

g= -r

varies with latitude). As one moves above thesurface toward lower pressure and density,the vertical pressure gradient also decreases.Furthermore, because warm air is less densethan cold air, pressure falls off with height moreslowly for warm than for cold air.

The troposphere is the lowest atmosphericlayer and contains most of the mass of theatmosphere (Fig. 2.3). The troposphere isheated primarily from the bottom by sensibleand latent heat fluxes and by longwave radia-tion from Earth’s surface. Temperature there-fore decreases with height in the troposphere.

Above the troposphere is the stratosphere,which, unlike the troposphere, is heated fromthe top. Absorption of UV radiation by O3 inthe upper stratosphere warms the air. Ozone isconcentrated in the stratosphere because of abalance between the availability of shortwaveUV necessary to split molecules of molecularoxygen (O2) into atomic oxygen (O) and a highenough density of molecules to bring about therequired collisions between atomic O and mol-

110

100

90

80

70

60

50

40

30

20

10

Thermosphere

Mesopause

Mesosphere

Stratopause

Stratosphere

Tropopause

Troposphere

Heig

ht (k

m)

-90 -30 0 30-60

Temperature (oC)

Mt. Everest

Figure 2.3. Average thermal structure of the atmos-phere showing the vertical gradients in Earth’s majoratmospheric layers. (Redrawn with permission fromAcademic Press; Schlesinger 1997.)


ecular O2 to form O3. The absorption of UVradiation by stratospheric ozone results in anincrease in temperature with height. The ozonelayer also protects the biota at Earth’s sur-face from damaging UV radiation. Biologicalsystems are sensitive to UV radiation becauseit can damage DNA, which contains the infor-mation needed to drive cellular processes. Theconcentration of ozone in the stratosphere hasbeen declining due to the production and emis-sion of CFCs, which destroy stratosphericozone, particularly at the poles. This results in an ozone “hole,” an area where the trans-mission of UV radiation to Earth’s surface isincreased. Slow mixing between the tropos-phere and the stratosphere allows CFCs andother compounds to reach and accumulate inthe ozone-rich stratosphere, where they havelong residence times.

Above the stratosphere is the mesosphere,where temperature again decreases withheight. The uppermost layer of the atmosphere,the thermosphere, begins at approximately 80km and extends into space. The thermos-phere has a small fraction of the atmosphere’stotal mass, composed primarily of O and nitro-gen (N) atoms that can absorb very shortwaveenergy, again causing an increase in heatingwith height (Fig. 2.3).The mesosphere and ther-mosphere have relatively little impact on thebiosphere.

The troposphere is the atmospheric layer inwhich most weather occurs, including thunder-storms, snowstorms, hurricanes, and high andlow pressure systems. The troposphere is thusthe portion of the atmosphere that directlyresponds to and affects ecosystem processes.The tropopause is the boundary between thetroposphere and the stratosphere. It occurs at aheight of about 16km in the tropics, where tropospheric temperatures are highest andhence where pressure falls off most slowly withheight (Eq. 2.1), and at about 9km in polarregions, where tropospheric temperatures arelowest.The height of the tropopause also variesseasonally, being lower in winter than insummer.

The planetary boundary layer (PBL) is thelower portion of the troposphere, which is influ-enced by mixing between the atmosphere andEarth’s surface. Air within the PBL is mixed by

surface heating, which creates convective tur-bulence, and by mechanical turbulence, whichis associated with the friction of air movingacross Earth’s surface. The PBL increases inheight during the day largely due to convectiveturbulence. The PBL mixes more rapidly withthe free troposphere when the atmosphere isdisturbed by storms. The boundary layer overthe Amazon Basin, for example, generallygrows in height until midday, when it is dis-rupted by convective activity (Fig. 2.4). ThePBL becomes shallower at night when there isno solar energy to drive convective mixing. Airin the PBL is relatively isolated from the freetroposphere and therefore functions like achamber over Earth’s surface. The changes inwater vapor, CO2, and other chemical con-stituents in the PBL thus serve as an indicatorof the biological and physiochemical processesoccurring at the surface (Matson and Harriss1988). The PBL in urban regions, for example,often has higher concentrations of pollutantsthan the cleaner, more stable air above. Atnight, gases emitted by the surface, such as CO2

in natural ecosystems or pollutants in urbanenvironments, often reach high concentrationsbecause they are concentrated in a shallowboundary layer.

Previous day's cloud layer New cloud layer

Hei

ght a

bove

can

opy

(km

)

Local time

1.5

1.0

0.5

022o C 30o C

0600 0800 1000 1200

Figure 2.4. Increase in the height of the planetaryboundary layer between 6:00 a.m. and noon in theAmazon Basin on a day without thunderstorms. Theincrease in surface temperature drives evapotranspi-ration and convective mixing, which causes theboundary layer to increase in height until the risingair becomes cool enough that water vapor condensesto form clouds. (Redrawn with permission fromEcology; Matson and Harriss 1988.)


Atmospheric Circulation

The fundamental cause of atmospheric circu-lation is the uneven heating of Earth’s surface.The equator receives more incoming solar radi-ation than the poles because Earth is spherical.At the equator, the sun’s rays are almost per-pendicular to the surface at solar noon. At thelower sun angles experienced at high latitudes,the sun’s rays are spread over a larger surfacearea (Fig. 2.5), resulting in less radiation re-ceived per unit ground area. In addition,the sun’s rays have a longer path through theatmosphere, so more of the incoming solar radiation is absorbed, reflected, or scatteredbefore it reaches the surface. This unequalheating of Earth results in higher tropospherictemperatures in the tropics than at the poles,which in turn drives atmospheric circulation.

Atmospheric circulation has both verticaland horizontal components (Fig. 2.6). Thetransfer of energy from Earth’s surface to theatmosphere by latent and sensible heat fluxesand longwave radiation generates strongheating at the surface. This warming causes the

surface air to expand and become less densethan surrounding air, so it rises. As air rises, thedecrease in atmospheric pressure with heightcauses continued expansion (Eq. 2.1), whichdecreases the average kinetic energy of air mol-ecules, causing the rising air to cool. The dryadiabatic lapse rate is the change in tempera-ture experienced by a parcel of air as it movesvertically in the atmosphere without exchang-ing energy with the surrounding air and is about9.8°Ckm-1. Cooling also causes condensationand precipitation because cool air has a lowercapacity to hold water vapor than warm air.Condensation in turn releases latent heat,which reduces the rate at which rising air coolsby expansion. This release of latent heat cancause the rising air to be warmer than sur-rounding air, so it continues to rise. The result-ing moist adiabatic lapse rate is about 4°Ckm-1

near the surface, rising to 6 or 7°Ckm-1 in themiddle troposphere. The greater the moisturecontent of rising air, the more latent heat isreleased to drive convective uplift, which con-tributes to the intense thunderstorms and deepboundary layer in the wet tropics. The averagelapse rate varies regionally, depending on thestrength of surface heating but averages about6.5°Ckm-1.

Surface air rises most strongly at the equatorbecause of the intense equatorial heating andthe large amount of latent heat released as thismoist air rises and condenses.This air rises untilit reaches the tropopause. The expansion ofequatorial air also creates a horizontal pressuregradient that causes the equatorial air aloft toflow horizontally from the equator along thetropopause toward the poles (Fig. 2.6). Thispoleward-moving air cools due to emission oflongwave radiation to space. In addition, the airconverges into a smaller volume as it movespoleward because Earth’s radius and surfacearea decrease from the equator toward thepoles. Due to the cooling of the air and its convergence into a smaller volume, the densityof air increases, creating a high pressure thatcauses upper air to subside, which forces sur-face air back toward the equator to replace therising equatorial air. Hadley proposed thismodel of atmospheric circulation in 1735, sug-gesting that there should be one large circu-

Atmosphere

Sun'srays

EarthAxis

Figure 2.5. Atmospheric and angle effects on solarinput at different latitudes.The arrows parallel to thesun’s rays show the depth of the atmosphere thatsolar radiation must penetrate. The arrows parallelto Earth’s surface show the surface area over whicha given quantity of solar radiation is distributed.High-latitude ecosystems receive less radiation thanthose at the equator because radiation at high lati-tudes has a longer path through the atmosphere andis spread over a larger ground area.


lation cell in the Northern Hemisphere andanother in the Southern Hemisphere, driven byatmospheric heating and uplift at the equatorand subsidence at the poles. Based on observa-tions, Ferrell proposed in 1865 the conceptualmodel that we still use today, although theactual dynamics are much more complex. Thismodel describes atmospheric circulation as aseries of three circulation cells in each hemi-sphere. (1) The Hadley cell is driven by expan-sion and uplift of equatorial air. (2) The polarcell is driven by subsidence of cold convergingair at the poles. (3) The intermediate Ferrell cellis driven indirectly by dynamical processes (Fig.

2.6). The Ferrell cell is actually the long-termaverage transport caused by weather systems inthe mid-latitudes rather than a stable perma-nent atmospheric feature. The chaotic motionof these mid-latitude weather systems creates anet poleward transport of heat. These threecells subdivide the atmosphere into three dis-tinct circulations: tropical air masses betweenthe equator and 30° N and S, temperate airmasses between 30 and 60° N and S, and polarair masses between 60° N and S and the poles(Fig. 2.6). The latitudinal location of these cells moves seasonally in response to latitudi-nal changes in surface heating by the sun.

ITCZ

Polar cell

Cold subsiding air

Cold subsiding air

Ferre

ll cell

Had

ley

cell

Hadley cell

Cold subsiding air

Warmrising

air

60o

30o

0o

Subtropical high pressure

Warm rising air

Warm rising air

Cold subsiding air

Ferrell cell

Polar cell

Westerlies

NE tradewinds

SE tradewinds

Subtropical high pressure

Westerlies

Polar front

Polar front

Figure 2.6. Earth’s latitudinal atmospheric circula-tions are driven by rising air at the equator and sub-siding air at the poles. These forces and the Coriolisforces produce three major cells of vertical atmos-pheric circulation (Hadley, Ferrell, and polar cells).Air warms and rises at the equator due to intenseheating. After reaching the tropopause, the equator-ial air moves poleward to about 30° N and S lati-tudes, where it descends and either returns to theequator, forming the Hadley cell, or moves pole-ward. Cold dense air at the poles subsides and movestoward the equator until it encounters poleward-moving air at about 60° latitude. There the air rises

and moves either poleward to replace air that hassubsided at the poles (the polar cell) or movestoward the equator to form the Ferrell cell. Alsoshown are the horizontal patterns of atmospheric cir-culation, consisting of the prevailing surface winds(the easterly trade winds in the tropics and the westerlies in the temperate zones). The boundariesbetween these zones are either low-pressure zonesof rising air (the intertropical conversion zone,ITCZ, and the polar front) or high-pressure zones ofsubsiding air (the subtropical high pressure belt andthe poles).


Earth’s rotation causes winds to deflect to theright in the Northern Hemisphere and to theleft in the Southern Hemisphere. Earth and its atmosphere complete one rotation aboutEarth’s axis every day.The direction of rotationis from west to east. Because the atmosphere inequatorial regions is farther from Earth’s axisof rotation than that at higher latitudes, there is a corresponding poleward decrease in thelinear velocity of the atmosphere as it travelsaround Earth. As parcels of air move north or south, they tend to maintain their angularmomentum (Ma), just as your car tends to main-tain its momentum when you try to stop or turnon an icy road.

Ma = mvr (2.2)

where m is the mass, v is the velocity, and r isthe radius of rotation. If the mass of a parcel ofair remains constant, its velocity is inverselyrelated to the radius of rotation. We know, forexample, that skaters can increase their speedof rotation by pulling their arms close to theirbodies, which reduces their effective radius. Airthat moves from the equator toward the polesencounters a smaller radius of rotation aroundEarth’s axis. Therefore, to conserve angularmomentum, it moves more rapidly (i.e., movesfrom west to east), relative to Earth’s surface,as it moves poleward (Fig. 2.6). Conversely,air moving toward the equator encounters anincreasing radius of rotation around Earth’saxis and, to conserve angular momentum,moves more slowly (i.e., moves from east towest), relative to Earth’s surface. This causes theair to be deflected to the right, relative toEarth’s surface, in the Northern Hemisphereand to the left in the Southern Hemisphere(Fig. 2.6), creating clockwise patterns of atmo-spheric circulation in the Northern Hemisphereand counterclockwise patterns in the SouthernHemisphere. This conservation of angularmomentum that causes air to change its direc-tion, relative to Earth’s surface, is known as the Coriolis force. The Coriolis force is apseudo-force that arises only because Earth is rotating, and we view the motion relative to Earth’s surface. Similar Coriolis forces act on ocean currents, creating clockwise ocean

circulation in the Northern Hemisphere andcounterclockwise circulation in the SouthernHemisphere.

The interaction of vertical and horizontalmotions of the atmosphere create Earth’s prevailing winds (i.e., the most frequent winddirections). As air in the Hadley cell movespoleward along the tropopause, it is deflectedby the Coriolis force to a westerly direction—that is, it blows from the west (Fig. 2.6). Thisprevents the poleward-moving air from reach-ing the poles, as it was supposed to do inHadley’s one-cell circulation model. This re-sults in an accumulation of air and a belt of highpressure at about 30° N and S latitude, whichcauses the air to sink. Some of this subsiding air returns toward the equator at the surface,completing the Hadley circulation cell (Fig.2.6). The interaction between motions inducedby the pole-to-equator temperature gradientand the Coriolis force explains why there arethree atmospheric circulation cells in eachhemisphere rather than just one, as Hadley had proposed. At the boundaries between themajor cells of atmospheric circulation, there are relatively sharp gradients of temperatureand pressure that, together with the Coriolisforce, generate strong winds over a broadheight range in the upper troposphere. Theseare the subtropical and polar jet streams.The Coriolis force explains why these windsblow in a westerly direction (i.e., from west toeast).

At the surface, the direction of prevailingwinds depends on whether air is moving towardor away from the equator. In the tropics, surfaceair in the Hadley cell moves from 30° N and Stoward the equator. In the Northern Hemi-sphere, this air is deflected to the right by theCoriolis force and forms the northeast tradewinds (i.e., surface winds that blow from thenortheast). Equatorward flow from 30° S isdeflected to the left and forms the southeasttrade winds. Thus equatorial winds blow pre-dominantly from the east. The region wheresurface air from the Northern and SouthernHemispheres converges is called the inter-tropical convergence zone (ITCZ). Here therising air creates a zone with light winds and high humidity (Fig. 2.6), an area known to


early sailors as the doldrums. Subsiding air at 30° N and S latitudes also produces relatively light winds, an area known as the horse lati-tudes. The surface air that moves polewardfrom 30° to 60° N and S is deflected toward theeast by the Coriolis forces, forming the prevail-ing westerlies, or surface winds that blow fromthe west.

The locations of the ITCZ and of each cir-culation cell shifts seasonally because the zoneof maximum solar radiation input varies from summer to winter due to Earth’s 23.5°tilt with respect to the plane of its orbit aroundthe sun. The seasonal changes in the location of these cells contribute to the seasonality ofclimate.

The uneven distribution of land and oceanson Earth’s surface creates an uneven pattern of heating that modifies the general latitudinaltrends in climate. At 30° N and S, air descendsmore strongly over cool oceans than over therelatively warm land because the air is coolerand more dense over the ocean than over theland. The greater subsidence over the oceanscreates high-pressure zones over the Atlanticand Pacific Oceans (the Bermuda and Pacifichighs, respectively) and over the southernoceans (Fig. 2.7). At 60° N, where air is rising,there are semipermanent low-pressure zonesover Iceland and the Aleutian Islands (the Ice-landic and Aleutian lows, respectively). Theselows are actually time averages of mid-latitudestorm tracks rather than stable features of thecirculation. In the Southern Hemisphere, thereis little land at 60° S; therefore, there is a troughof low pressure instead of distinct centers.Air that subsides in high-pressure centersspirals outward in a clockwise direction in theNorthern Hemisphere and in a counterclock-wise direction in the Southern Hemisphere(Fig. 2.7) due to an interaction between Cori-olis forces and the pressure gradient force produced by the subsiding air. Winds spiralinward toward low-pressure centers in a coun-terclockwise direction in the Northern Hemi-sphere and in a clockwise direction in theSouthern Hemisphere. Air in the low-pressurecenters rises to balance the subsiding air inhigh-pressure centers.The long-term average ofthese vertical and horizontal motions produces

the vertical circulation described by the Ferrellcell (Fig. 2.6) and a horizontal pattern of high-and low-pressure centers commonly observedon weather charts (Fig. 2.7).

These deviations from the expected easterlyor westerly direction of prevailing winds areorganized on a planetary scale and are knownas planetary waves. These waves are influencedby both land–ocean heating contrasts and thelocations of large mountain ranges, such as the Rockies and the Himalayas. These moun-tain barriers force the Northern Hemispherewesterlies vertically upward and to the north.Downwind of the mountains, air descends and moves to the south, forming a trough, muchlike the standing waves in the rapids of a fast-moving river that are governed by the location of rocks in the riverbed. Temperaturesare comparatively low in the troughs, due to thesouthward movement of polar air, and are com-paratively high in the ridges. The trough overeastern North America downwind of the RockyMountains (Fig. 2.7), for example, results inrelatively cool temperatures and a moresoutherly location of the arctic treeline ineastern North America. Although planetarywaves have preferred locations, they are notstatic. Changes in their location or in thenumber of waves alter regional patterns ofclimate. These step changes in the circulationpattern are referred to as climate modes.

Planetary waves and the distribution ofmajor high- and low-pressure centers explainmany details of horizontal motion in the atmo-sphere and therefore the patterns of ecosystemdistribution. The locations of major high- andlow-pressure centers, for example, explain themovement of mild moist air to the west coastsof continents at 60° N and S, where the tem-perate rainforests of the world occur (thenorthwestern United States and southwesternChile, for example) (Fig. 2.7). The subtropicalhigh-pressure centers at 30° N and S cause coolpolar air to move toward the equator on thewest coasts of continents, creating dry medi-terranean climates at 30° N and S (Fig. 2.7). Onthe east coasts of continents subtropical highscause warm moist equatorial air to move north-ward at 30° N and S, creating a moist subtropi-cal climate (Fig. 2.7).


ITCZ

L

L

LH

H

H

H

H

HH

L

H

L

H

90

90 90

90

90

90

180

180

0

0

60

60 60

60

0 0

30

30

30

30

Latit

ude

Longitude

Icelandiclow

BermudahighPacific

high

Siberian high

Aleutian low

January

A

90

90 90

90

90180

180

0

0

60

60 60

60

0 0

30

30

30

30

Latit

ude

Longitude

90

ITCZ

L

H

H

HHH

Bermudahigh

Pacifichigh

H

Icelandiclow

L

Siberianlow

L

Thermallow

L

July

B

Figure 2.7. Average surface wind-flow patterns and the distribution of sea level pressure for January (A)and July (B). (Redrawn from Essentials of Meteorology: An Invitation to the Atmosphere, 2nd edition,by C. Ahrens © 1998, by permission of Brooks/Cole, an imprint of the Wadsworth Group, a division ofThomson Learning; Ahrens 1998.)

The Oceans

Ocean Structure

Oceans maintain rather stable layers withlimited vertical mixing between them. The sun

heats the ocean from the top, whereas theatmosphere is heated from the bottom. Becausewarm water is less dense than cold water, oceansmaintain rather stable layers that do not readilymix. The uppermost warm layer of surfacewater, which interacts directly with the atmos-

The Oceans 29

phere, extends to depths of 75 to 200m, depend-ing on the depth of wind-driven mixing. Mostprimary production, detrital production, anddecomposition take place in the surface waters(see Chapter 10). Another major differencebetween atmospheric and oceanic circulation isthat density of ocean waters is determined byboth temperature and salinity, so, unlike warmair, warm water can sink, if it is salty enough.

There are relatively sharp gradients in tem-perature (thermocline) and salinity (halocline)between warm surface waters of the ocean andcooler more saline waters at intermediatedepths (200 to 1000m) (Fig. 2.8).These two ver-tical gradients cause the surface waters to beless dense than deep water, creating a stablevertical stratification. The deep layer there-fore mixes with the surface waters slowly overhundreds to thousands of years. These deeperlayers nonetheless play critical roles in elementcycling, productivity, and climate because theyare long-term sinks for carbon and the sourcesof nutrients that drive ocean production (seeChapters 10 and 15). Upwelling areas, wheredeep waters move rapidly to the surface,support high levels of primary and secondaryproductivity (marine invertebrates and verte-

brates) and are the locations of many of theworld’s major fisheries.

Ocean Circulation

Ocean circulation plays a critical role in Earth’sclimate system. On average, ocean circulationaccounts for 40% of the latitudinal heat trans-fer from the equator to the poles, with theremaining 60% of heat transfer occurringthrough the atmosphere. The ocean is the dom-inant heat transporter in the tropics, and theatmosphere plays the stronger role at mid-latitudes. The surface currents of the oceans are driven by surface winds and therefore show global patterns (Fig. 2.9) that are gener-ally similar to those of the prevailing surfacewinds (Fig. 2.7). The ocean currents are,however, deflected 20 to 40° relative to the winddirection by Coriolis forces. This deflection and the edges of continents cause ocean currents to be more circular (termed gyres)than the winds that drive them. In equatorialregions, currents flow east to west, driven by theeasterly trade winds, until they reach the conti-nents, where they split and flow poleward alongthe western boundaries of the oceans, carryingwarm tropical water to higher latitudes. Ontheir way poleward, currents are deflected byCoriolis forces. Once the water reaches the highlatitudes, some returns in surface currentstoward the tropics along the eastern edges ofocean basins (Fig. 2.9), and some continuespoleward.

Deep ocean waters show a circulationpattern quite different from the wind-drivensurface circulation. In the polar regions, espe-cially in the winter off southern Greenland andoff Antarctica, cold air cools the surface waters,increasing their density. Formation of sea ice,which excludes salt from ice crystals (brinerejection), increases the salinity of surfacewaters, also increasing their density. The highdensity of these cold saline waters causes themto sink. This downwelling to form the NorthAtlantic deep water off of Greenland, and theAntarctic bottom water off of Antarctica drivesthe global thermohaline circulation in themiddle and deep ocean that ultimately transferswater between the major ocean basins (Fig.2.10). The descent of cold dense water at high

0

250

500

750

100032 33 34 35 36

Salinity (ppt)

Dep

th (

m)

Temperature (oC)

5 10 150

Sur

face

wat

erIn

term

edia

tew

ater

Tem

pera

ture

Salinity

T / H

Figure 2.8. Typical vertical profiles of ocean tem-perature and salinity. The thermocline (T) and halocline (H) are the zones where temperature and salinity, respectively, decline most strongly with depth. These transition zones usually coincideapproximately.


latitudes is balanced by the upwelling of deepwater on the eastern margins of ocean basins at lower latitudes, where along-shore surface currents are deflected offshore by Coriolisforces and easterly trade winds. Net polewardmovement of warm surface waters balances the movement of cold deep water toward theequator. Because thermohaline circulation con-

trols latitudinal heat transport, changes in itsstrength have significant effects on climate. Inaddition, it transfers carbon to depth, where itremains for centuries (see Chapter 15).

Oceans, with their high heat capacity, heat upand cool down much less rapidly than does landand thus have a moderating influence on theclimate of adjacent land. Wintertime tempera-

Gulf Stream

Be

ngue

laC

.

N. Pacific Drift

KuroshioC.

Hum

bold

tC

.

Nor

thAtla

ntic Drift

West Wind Drift

90

90 90

90

180

180

0

0

60

60 60

60

0 0

30

30

30

30

Longitude

90

90

La

titu

de

Figure 2.9. Major surface ocean currents. Warm currents are shown by solid arrows and cold currents bydashed arrows. (Redrawn from Essentials of Meteorology: An Invitation to the Atmosphere, 2nd edition,by C. Ahrens © 1998, by permission of Brooks/Cole, an imprint of the Wadsworth Group, a division ofThomson Learning; Ahrens 1998.)

Warm shallowcurrent

Cold saltydeep current

Figure 2.10. Circulation patterns of deep and surface waters among the major ocean basins.

Landform Effects on Climate 31

tures in Great Britain and western Europe, forexample, are much milder than at similar lati-tudes on the east coast of North America dueto the warm North Atlantic drift (the polewardextension of the Gulf Stream) (Fig. 2.9). Con-versely, cold upwelling currents, or currentsmoving toward the equator from the poles, cooladjacent land masses in summer. The cold California current, for example, which runsnorth to south along the west coast of theUnited States, keeps summer temperatures innorthern California lower than the U.S. eastcoast at similar latitudes. Windward coastal sit-uations are typically more strongly influencedby prevailing onshore winds and thus havemore moderate temperatures than do coastalsituations that are downwind of predominantlycontinental air. New York City, on the easternedge of North America, therefore experiencesrelatively mild winters compared to inlandcities like Minneapolis, but its winter tempera-tures are lower than those of cities on thewestern edge of the continent. These tempera-ture differences play critical roles in determin-ing the kinds of ecosystems that occur overdifferent parts of the globe.

Landform Effects on Climate

The spatial distribution of land, water, andmountains modify the general latitudinal trendsin climate. The greater heat capacity of watercompared to land influences atmospheric cir-culation at local to continental scales. The sea-sonal reversal of winds (monsoon) in easternAsia, for example, is driven largely by the dif-ferential temperature response of the land andthe adjacent seas. During the Northern Hemi-sphere winter, the land is colder than the ocean,giving rise to cold dense air that flows south-ward across India to the ocean (Fig. 2.7). Insummer, however, the land heats relative to theocean. The heating over land forces the air torise, in turn drawing in moist surface air fromthe ocean. Condensation of water vapor in therising moist air produces large amounts of pre-cipitation. Northward migration of the tradewinds in summer enhances onshore flow of air,and the mountainous topography of northern

India enhances vertical motion, increasing theproportion of water vapor that is converted toprecipitation. Together, these seasonal changesin winds give rise to predictable seasonal patterns of temperature and precipitation thatstrongly influence the structure and functioningof ecosystems.

At scales of a few kilometers, the differentialheating between land and ocean produces landand sea breezes. During the day, strong heatingover land causes air to rise, drawing in cool airfrom the ocean. The rising of air over the landincreases the height at which a given pressureoccurs, causing this upper air to move from landtoward the ocean. The resulting increase in themass of atmosphere over the ocean augmentsthe surface pressure, which causes surface air to flow from the ocean toward the land. Theresulting circulation cell is identical in principleto that which occurs in the Hadley cell (Fig. 2.6)or Asian monsoon (Fig. 2.7). At night, when theocean is warmer than the land, air rises over the ocean, and the surface breeze blows fromthe land to the ocean, causing the circulationcell to reverse. The net effect of sea breezes is to reduce temperature extremes and increaseprecipitation on land near oceans or large lakes.

Mountain ranges affect local atmospheric cir-culation and climate through several types oforographic effects—that is, effects due to pres-ence of mountains. As winds carry air up thewindward sides of mountains, the air cools,and water vapor condenses and precipitates.Therefore, the windward side tends to be coldand wet. When the air moves down the leewardside of the mountain, it contains little moisture,creating a rain shadow, or a zone of low precipitation downwind of the mountains.The rain shadow of the Rocky Mountainsextends 1500km to the east, resulting in astrong west-to-east gradient in annual pre-cipitation from Colorado (300mm) to Illinois(1000mm) (see Fig. 14.1) (Burke et al. 1989).Deserts or desert grasslands (steppes) are oftenfound immediately downwind of the majormountain ranges of the world. Mountainsystems can also influence climate by channel-ing winds through valleys. The Santa Annawinds of southern California occur when high


pressure over the interior deserts causes warmdry winds to be funneled through valleystoward the Pacific coast. These winds createextremely dry conditions that promote intensewildfires.

Sloping terrain creates unique patterns ofmicroclimate at scales ranging from ant hills to mountain ranges. Slopes facing the equator(south-facing slopes in the Northern Hemi-sphere and north-facing slopes in the SouthernHemisphere) receive more radiation than op-posing slopes and thus have warmer drier con-ditions. In cold or moist climates, the warmermicroenvironment on slopes facing the equatorprovides conditions that enhance productivity,decomposition, and other ecosystem processes,whereas in dry climates, the greater drought onthese slopes limits production. Microclimaticvariation associated with slope and aspect (thecompass direction in which a slope faces) allowsstands of an ecosystem type to exist hundredsof kilometers beyond its major zone of distri-bution.These outlier populations are importantsources of colonizing individuals during timesof rapid climatic change and are thereforeimportant in understanding species migrationand the long-term dynamics of ecosystems.

Topography also influences climate throughdrainage of cold dense air. When air cools atnight, it becomes more dense and tends to flowdownhill (katabatic winds) into valleys, whereit accumulates. This can produce strong tem-perature inversions (cool air beneath warm air,a vertical temperature profile reversed from thetypical pattern in the troposphere of decreasingtemperature with increasing elevation; Fig. 2.3).Inversions occur primarily at night and inwinter, when there is insufficient heating fromthe sun to promote convective mixing. Cloudsalso tend to inhibit the formation of winter and nighttime inversions because they increaselongwave emission to the surface. Increases insolar heating or windy conditions, such as mightaccompany the passage of frontal systems,break up inversions. Inversions are climaticallyimportant because they increase the seasonaland diurnal temperature extremes experiencedby ecosystems in low-lying areas. In cool cli-mates, inversions greatly reduce the length ofthe frost-free growing season.

Vegetation Influences on Climate

Vegetation influences climate through its effecton the surface energy budget. Climate is quitesensitive to regional variations in the vegeta-tion and moisture content of Earth’s surface.The albedo (the fraction of the incident short-wave radiation reflected from a surface) deter-mines the quantity of solar energy absorbed bythe surface, which is subsequently available fortransfer to the atmosphere as longwave radia-tion and turbulent fluxes. Water generally has alow albedo, so lakes and oceans absorb con-siderable solar energy.At the opposite extreme,snow and ice have a high albedo and henceabsorb little solar radiation, contributing to thecold conditions required for their persistence.Vegetation is intermediate in albedo, andvalues generally decrease from grasslands (withtheir highly reflective standing dead leaves) todeciduous forests to dark conifer forests (seeChapter 4). Recent land use changes have substantially altered regional albedo byincreasing the area of exposed bare soil. Thealbedo of soil depends on soil type and wetnessbut is often higher than that of vegetation, par-ticularly in dry climates. Consequently, over-grazing may increase albedo, reducing energyabsorption and the transfer of energy to theatmosphere. This leads to cooling and subsi-dence, which can reduce precipitation and thecapacity of vegetation to recover from over-grazing (Charney et al. 1977). The large magni-tude of many land-surface feedbacks to climatesuggests that land-surface change can be animportant contributor to regional climaticchange (Chase et al. 2000).

The energy absorbed by a soil or vegetationsurface is transferred to the atmosphere vialongwave radiation and turbulent fluxes oflatent and sensible heat. The partitioning ofenergy among these pathways has importantclimatic consequences (see Chapter 4). Sensibleheat fluxes and longwave radiation directlyheat the atmosphere at the point at which theenergy transfer occurs. Latent heat transferswater vapor to the atmosphere. Water vaporrepresents stored energy, which is released

Vegetation Influences on Climate 33

when the water vapor condenses to form cloudsor precipitation. The energy released to theatmosphere by condensation typically occurssome distance downwind from the point atwhich the water evaporated. Ecosystem struc-ture influences the efficiency with which sen-sible heat and latent heat are transferred to the atmosphere. Wind passing over tall unevencanopies creates mechanical turbulence thatincreases the efficiency of heat transfer fromthe surface to the atmosphere (see Chapter 4).Smooth surfaces, in contrast, tend to heat upbecause they transfer their heat less efficientlyto the atmosphere, only by convection and notby mechanical turbulence.

The effects of vegetation structure on theefficiency of water and energy exchange influ-ence regional climate. Between 25 and 40% ofthe precipitation in the Amazon basin comesfrom water that is recycled from land by evapo-transpiration (Costa and Foley 1999). Simula-tions by climate models suggest that, if theAmazon basin were completely converted fromforest to pasture, South America would have apermanently warmer drier climate (Shukla etal. 1990). The shallow roots of grasses wouldabsorb less water than the deep tree roots,leading to lower transpiration rates (Fig. 2.11).Pastures would therefore release more of theabsorbed solar radiation as sensible heat, whichdirectly warms the atmosphere.The simulationsalso suggests that warming and drying of aircaused by widespread conversion from forest topasture would reduce the transport of moisturefrom the adjoining oceans, causing a permanent

reduction in precipitation—conditions thatfavor persistence of pastures over forests.Thesesimulations do an excellent job of exploring the consequences of such vegetation effects onprocesses that are well understood. There arestill many uncertainties, however. Changes incloudiness, for example, can have either a pos-itive or a negative effect on radiative forcing,depending on the clouds’ properties and height.Because these models do a poor job of simu-lating the processes that produce clouds, thesimulations should be viewed as a way to syn-thesize the net effect of the processes that weunderstand, rather than as predictions of thefuture.

At high latitudes, tree-covered landscapesabsorb more solar radiation before snow melt,due to their low albedo, than does snow-covered tundra. Model simulations suggest thatthe northward movement of the treeline 6000years ago could have reduced the regionalalbedo and increased energy absorption suffi-ciently to explain half of the climate warmingthat occurred at that time (Foley et al. 1994).The warmer regional climate would, in turn,favor tree reproduction and establishment atthe treeline (Payette and Filion 1985), provid-ing a positive feedback to regional warming(see Chapter 12). Predictions about the impactof future climate on vegetation should there-fore also consider ecosystem feedbacks toclimate.

Albedo, energy partitioning between latentand sensible heat fluxes, and surface structurealso influence the amount of longwave radia-

Forest Pasture Forest Pasture Forest PastureEva

potr

ansp

iratio

n (

mm

d-1

)

Sur

face

tem

pera

ture

(o C

)

Pre

cipi

tatio

n (m

m d

-1)

0

3

6

20

24

28

0

4

8

Figure 2.11. Simulations, using a general circulationmodel, of changes in evapotranspiration, surface airtemperature, and precipitation that would occur if

the rain forests of South America were replaced bypasture (Shukla et al. 1990).


tion transferred to the atmosphere (Fig. 2.2).This is due to the dependence of longwave radi-ation on surface temperature. Surface temper-ature tends to be high when the surface absorbslarge amounts of incoming radiation (lowalbedo), has little water to evaporate, and/orhas a smooth surface that is inefficient in trans-ferring turbulent fluxes of sensible and latentheat to the atmosphere (see Chapter 4).Deserts, for example, experience large net long-wave energy losses because their dry smoothsurfaces lead to high surface temperatures, andthere is little moisture to support evaporationthat would otherwise cool the soil.

Temporal Variability in Climate

Long-Term Changes

Long-term climatic change is driven primarilyby changes in solar input and changes in atmos-pheric composition. Earth’s climate is adynamic system that has changed repeatedly,producing frequent, and sometimes abrupt,changes in climate, manifested by a series ofdramatic glacial epochs (Fig. 2.12) and sea levelchanges. Volcanic eruptions and asteroidimpacts contributed to these changes by influ-encing the absorption or reflection of solarenergy. Mountain building and erosion andcontinental drift have modified the patterns ofatmospheric and ocean circulation.The primaryforce responsible for the evolution of Earth’sclimate, however, has been changes in the inputof solar radiation, which has increased overmuch of the past 4 billion years, as the sun has matured (Fig. 2.13) (Schlesinger 1997). Onshorter time scales, solar input has varied pri-marily due to predictable alterations in Earth’sorbit (Fig. 2.14).

Three types of variations in Earth’s orbitinfluence the amount of solar radiation re-ceived at the surface. These variations can bedescribed by three parameters: eccentricity (thedegree of ellipticity of Earth’s orbit around thesun), tilt (the angle between Earth’s axis ofrotation and the plane of its orbit around thesun), and precession (a “wobbling” in Earth’saxis of rotation with respect to the stars, deter-

mining the date during the year when solsticesand equinoxes occur). At present, Earth’s orbitis nearly circular (minimal eccentricity), leadingto relatively small seasonal variation in solarinput related to Earth–sun distance (Sturman

0

Tertiary

Cretaceous

Jurassic

Triassic

Permian

Carboniferous

Devonian

Silurian

Ordovician

Proterozoic

Archaean

Quaternary

ME

SO

ZO

ICPA

LEO

ZO

ICP

RE

CA

MB

RIA

NC

EN

OZ

OIC

2800

4600

570

Cambrian

4700

2300

500

430

395

345

320

280

245

190

136

65

53

37

26

5

2

0.01Holocene

Pleistocene

Pliocene

Miocene

Oligocene

Eocene

Paleocene

ERA PERIODAGE(Ma)

EPOCHMAJOR EVENTS

AGE OFDINOSAURS

AGE OFMAMMALS

AGE OFHUMANS

Extensive coalformation

First land plants

Primitive fish

Formation ofthe Earth

Formation of the oceans

First life

Final assemblyof Pangaea

Breakupof Pangaea

Anthropocene

*

Figure 2.12. Geological time periods in Earth’shistory showing major glacial events (solid bars) andecological events that strongly influenced ecosystemprocesses. Note the changes in time scale (Ma = mil-lions of years). The most recent geologic epoch (theAnthropocene) began about 1750 with the beginningof the Industrial Revolution and is characterized byhuman domination of the biosphere. (Modified withpermission from Sturman and Tapper, 1996.)

Temporal Variability in Climate 35

and Tapper 1996). Tilt determines the strengthof the seasons. Earth’s tilt is presently inter-mediate (23.5°), providing an intermediatedegree of seasonality. Earth’s precession placesthe solstices in December and June, causing the

Northern Hemisphere winters to be mild andSouthern Hemisphere summers to be relativelywarm. At times in the past, the solsticesoccurred at other times of year. The periodici-ties of these orbital parameters (eccentricity,tilt, and precession) are approximately 100,000,41,000, and 23,000 years, respectively. Togetherthey produce Milankovitch cycles of solar inputthat correlate with the glacial and interglacialcycles over the last 800,000 years, as determinedby isotopic analyses of ocean sediments and icecores.

The chemistry of ice and trapped air bubblesprovide a paleorecord of the climate when the ice formed. The Vostok ice core, drilled atVostok Station in Antarctica in the 1980s and1990s, indicates considerable climate variabilityover the past 400,000 years, in large part relatedto the Milankovitch cycles (see Fig. 15.2).Analysis of bubbles in this and other ice coresindicates that past warming events have beenassociated with increases in CO2 and CH4 con-centrations, providing circumstantial evidence

012345

1.4

1.2

1.0

0.6

0.4 012345

0.8

Sol

ar fl

ux

Time (billions of years ago)

Figure 2.13. Solar flux (relative to present solarradiation) received by Earth since the beginning ofthe solar system (Graedel and Crutzen 1995).

Tilt

Precession

Eccentricity

Per

cent

Ear

th-s

un d

ista

nce

in J

une

Less

Mor

e

Deg

rees

24.0

23.5

23.0

22.5

250 200 150 50 0100

Time (thousands of years ago)

Figure 2.14. Long-term variationsin three orbital parameters thatresult in glacial patterns.


for a past role of radiatively active gases inclimate change. The unique feature of therecent anthropogenic increases in these gases is that they are occurring during an interglacialperiod, when Earth’s climate is already rela-tively warm. The Vostok record suggests thatthe CO2 concentration of the atmosphere ishigher now than at any time in the last 400,000years. Fine-scale analysis of ice cores fromGreenland suggests that large changes fromglacial to interglacial climate can occur indecades or less. Such rapid transitions in the climate system to a new state may berelated to sudden changes in the strength of the thermohaline circulation that drivesoceanic heat transport from the equator to thepoles.

Tree ring records, obtained from living anddead trees, provide information about theclimate during the past several thousand years.The width of tree rings gives a record of tem-perature and moisture, and the chemical com-position of wood reflects the characteristics ofthe atmosphere at the time the wood was

formed. Pollen preserved in low-oxygen sedi-ments of lakes provides a history of plant taxaand climate over the past 10,000 years or more(Fig. 2.15). Pollen records from networks of sites can be used to construct maps of species distributions at various times in the past and can provide a history of species migra-tions across continents after climatic changes(COHMAP 1988). Other proxy records pro-vide measures of temperature (species compo-sition of Chironomids), precipitation (lakelevel), pH, and geochemistry.

The combination of paleoclimate proxiesindicates that climate is inherently variableover all time scales. Atmospheric, oceanic,and other environmental changes that areoccurring now due to human activities must beviewed as overlays on the natural climate vari-ability that stems from long-term changes inEarth’s surface characteristics and orbitalgeometry.

Earth’s climate is now warmer than at any time in the last 1000 years (Fig. 2.16) andperhaps much longer.This warming is most pro-

Warmingtrend

Drierclimate

Moisterclimate

Drier,coolerclimate

Pollen abundance (% of total tree pollen)

0 0 0 0 0 0 4040 60

11.0

8.6

3.9

2.7

Birc

h

Pin

e

Oak

Ald

er

Tim

e (t

hous

ands

of y

ears

ago

)

Fir

Elm

Figure 2.15. Pollen profile from a bog innorthwestern Minnesota showing changes in the dominant tree species over the past11,000 years. (McAndrews 1966).


0.0

0.5

1.0

-0.5

-1.0

1000 1200 1400 1600 1800 2000

Tem

pera

ture

ano

mal

y (o C

)

Year

1998

Figure 2.16. Time course of theaverage surface temperature ofthe Northern Hemisphere overthe last 1000 years. The data are presented as a 40-yearrunning average of the differ-ence (anomaly) in temperaturebetween each year and theaverage 1902–1980 temperature.Note that 1998 was the warmestyear of the last millennium.(Redrawn with permission fromGeophysical Research Letters;Mann et al. 1999.)

Figure 2.17. Sources of evidence that Earth’sclimate is warming. Reduction in stratospheric ozonecauses less energy to be absorbed by the stratos-phere; this causes the stratosphere to cool and allows more energy to penetrate to Earth’s surface.

Warming is most pronounced at Earth’s surface andhas caused the surface air and oceans to warm andglaciers and sea ice to melt. (Adapted from IPCCAssessment Report 2001; Folland et al. 2001.)


nounced near Earth’s surface, where its eco-logical effects are greatest (Fig. 2.17). Some of the recent warming reflects an increase insolar input, but most of the warming resultsfrom human activities that increase the con-centrations of radiatively active gases in the atmosphere (Fig. 2.18). Climate models and recent observations suggest that warming will be most pronounced in the interiors of continents, far from the moderating effectsof oceans, and at high latitudes. The high-latitude warming reflects a positive feed-back. As climate warms, the snow and sea ice melt earlier in the year, which replaces the reflective snow or ice cover with a low-albedo land or water surface. These darker surfaces absorb more radiation and trans-mit this energy to the atmosphere, which warms the climate. Those land-surface changes that produce a darker or drier surface also contribute to greater heat transfer to the atmosphere and to the warming of surfaceclimate.

As climate warms, the air has a higher capac-ity to hold water vapor, so there is greaterevaporation from oceans and other moist surfaces. In areas where rising air leads to condensation, there is greater precipita-tion. Continental interiors and rain shadows on the lee sides of mountains are less likely to experience large precipitation increases.Consequently, soil moisture and runoff tostreams and rivers are likely to increase in coastal regions and mountains and todecrease in continental interiors. The com-plex controls and nonlinear feedbacks in theclimate system make detailed climate pro-jections problematic and are active areas ofresearch.

Interannual Climate Variability

Much of the interannual variation in climate is associated with large-scale changes in theatmosphere–ocean system. Superimposed onthe long-term climate variability are inter-annual variations that have been noted byfarmers, fishermen, and naturalists for cen-turies. Some of this variability exhibits repeat-

ing geographic and temporal patterns. One of these phenomena that has received consid-erable attention is the El Niño/southern oscil-lation (ENSO) (Webster and Palmer 1997,Federov and Philander 2000). ENSO events are part of a large-scale, air–sea interaction that couples atmospheric pressure changes (thesouthern oscillation) with changes in oceantemperature (El Niño) over the equatorialPacific Ocean. ENSO events have occurred,on average, every 3 to 7 years over the pastcentury, with considerable irregularity (Trenberth and Haar 1996). No events occurredbetween 1943 and 1951, for example, and three major events occurred between 1988 and 1999.

In most years, the easterly trade winds pushthe warm surface waters of the Pacific west-ward so the layer of warm surface waters isdeeper in the western Pacific than it is in theeast (Figs. 2.9 and 2.19). The resulting warmwaters in the western Pacific are associated with a low pressure center and promote convection and high rainfall in Indonesia. Theoffshore movement of surface waters in theeastern Pacific promotes upwelling of colder,deeper water off the coasts of Ecuador andPeru. These cold, nutrient-rich waters sup-port a productive fishery (see Chapter 10) and promote subsidence of upper air, lead-ing to the development of a high pressure center and low precipitation. At times, how-ever, the Pacific high-pressure and Indone-sian low-pressure centers weaken, and the easterly trades weaken. The warm surfacewaters then move eastward, forming a deeplayer of warm water in the eastern Pacific.This weakens or shuts down the upwelling of cold water, promoting atmospheric convec-tion and rainfall in coastal Ecuador and Peru.The colder waters in the western Pacific,in contrast, inhibit convection, leading todroughts in Indonesia, Australia, and India.This pattern is commonly termed El Niño.Periods in which the “normal” pattern is par-ticularly strong are termed La Niña. The trig-ger for changes in this ocean–atmosphere system is unknown but may involve large-scale ocean waves, known as Kelvin waves,


that travel back and forth across the tropicalPacific.

ENSO events have widespread climatic,ecosystem, and societal consequences. StrongEl Niño phases cause dramatic reductions inanchovy fisheries in Peru and reproductivefailure and mortality in sea birds and marinemammals. Extremes in precipitation linked toENSO cycles are also evident in areas distantfrom the tropical Pacific. El Niño events bringhot, dry weather to the Amazon Basin, poten-tially affecting tree growth, soil carbon storage,and fire probability. Northward extension ofwarm tropical waters to the northern Pacificbrings rains to coastal California and highwinter temperatures to Alaska. An importantlesson from ENSO studies is that strong cli-matic events in one portion of the globe haveclimatic consequences throughout the globedue to the dynamic interactions (termed tele-

connections) associated with atmospheric cir-culation and ocean currents.

The Pacific North America (PNA) pattern isanother large-scale pattern of climate variabil-ity. The positive mode of the PNA is character-ized by above-average atmospheric pressurewith warm dry weather in western NorthAmerica and below-average pressure and lowtemperatures in the east. Variability in the PNA (or PNA-like) pattern is loosely linked to ENSO phases. Another large-scale climatepattern is the North Atlantic oscillation (NAO).Positive phases of the NAO are associated with a strengthening of the pressure gradientbetween the Icelandic low-pressure and theBermuda high-pressure systems (Fig. 2.7). Thisincreases heat transport to high latitudes bywind and ocean currents, leading to a warmingof Scandinavia and western North America and a cooling of eastern Canada. Although the

x

x

x

xx

x

xx

x

x

xx

xx

x

x

x

CFCs

N2OCH4

CO2Tropospheric

ozone

Aerosols

Fossilfuel

burning(black

carbon)

Fossilfuel

burning(organiccarbon)

Sulfate

Biomassburning

Mineraldust

Rad

iativ

e fo

rcin

g (W

m-2

)

Coo

ling

War

min

g

Troposphericaerosolindirecteffect

Aviation-induced

Contrails Cirrus

Stratosphericozone Land-

use(albedo)

Level of scientific certainty of the impact

High Medium Medium Low Verylow

Verylow

Verylow

Verylow

Verylow

Verylow

Verylow

Verylow

0

1

2

3

-1

-2

Solar

Figure 2.18. Major changes in the climate systemthat have caused Earth’s climate to warm between1750 and 2000. Some changes in the climate systemlead to net warming; others lead to net cooling. Thelargest single cause of climate warming is probably

the increased concentration of atmospheric CO2, pri-marily as a result of burning fossil fuels. (Adaptedfrom IPCC Assessment Report 2001; Ramaswamy etal. 2001.)


factors that initiate these large-scale climatefeatures are poorly understood, the patternsthemselves and their ecosystem consequencesare becoming more predictable. Future climaticchanges will likely be associated with changesin the frequencies of certain phases of theselarge-scale climate patterns rather than anysimple linear trend in climate. Climatewarming, for example, might increase the fre-

quency of ENSO events and positive phases ofthe NAO.

Seasonal and Daily Variations

Seasonal and daily variations in solar input have profound but predictable effects onclimate and ecosystems. Perhaps the mostobvious variations in the climate system are the

Westerly aloft

Easterly trades

Lowpressure

Highpressure

ColdupwellingThermocline

Moderatehigh

pressure

Warmupwelling

Weak windsWeak winds

Thermocline

Moderatehigh

pressureLow

pressureLow

pressure

"Normal " conditions

El Nino event~

(La Nina event)~

Figure 2.19. Walker circulation of the ocean andatmosphere in the tropical Pacific between SouthAmerica and Indonesia during “normal” years andduring El Niño years. In normal years, strong east-erly trade winds push surface ocean waters to thewest, producing deep, warm waters and high pre-cipitation off the coast of Southeast Asia and cold,

upwelling waters and low precipitation off the coastof South America. In El Niño years, however, weakeasterly winds allow the surface waters to move fromwest to east across the Pacific Ocean, leading tocooler surface waters and less precipitation in South-east Asia and warmer surface waters and more pre-cipitation off South America.

Relationship of Climate to Ecosystem Distribution and Structure 41

patterns of seasonal and diurnal change. Theaxis of Earth’s rotation is fixed at 23.5° relativeto its orbital plane about the sun. This tilt inEarth’s axis results in strong seasonal variationsin day length and the solar irradiance—that is,the quantity of solar energy received at Earth’ssurface per unit time. During the spring andautumn equinoxes, the entire Earth’s surfacereceives approximately 12h of daylight (Fig.2.20).At the Northern Hemisphere summer sol-stice, the sun’s rays strike Earth most directly inthat hemisphere, and day length is maximized.At the Northern Hemisphere winter solstice,the sun’s rays strike Earth most obliquely in thathemisphere, and day length is minimized. Thesummer and winter solstices in the SouthernHemisphere are 6 months out of phase fromthose in the Northern Hemisphere. Variations in incident radiation become increasingly pro-nounced as latitude increases. Thus tropicalenvironments experience relatively small sea-sonal differences in solar irradiance and daylength, whereas they are maximized in theArctic and Antarctic. Above the Arctic andAntarctic Circles, there are 24h of daylight atthe summer solstice, and the sun never rises atthe winter solstice. The relative homogeneity oftemperature and light throughout the year inthe tropics contributes to their high productiv-ity and diversity. At higher latitudes, the lengthof the warm season strongly influences the lifeforms and productivity of ecosystems.

Variations in light and temperature also playan important role in determining the types of plants that grow in a given climate and therates at which biological processes occur. Many

biological processes are temperature depen-dent, and slower rates occur at lower tem-peratures. Diurnal variations in day length(photoperiod) provide important cues thatallow organisms to prepare for seasonal varia-tions in climate.

Relationship of Climate toEcosystem Distribution and Structure

Climate is the major determinant of the globaldistribution of biomes. The major types ofecosystems (Plate 1) and their productivityshow predictable relationships to climatic variables such as temperature and moisture(Holdridge 1947, Whittaker 1975) (Fig. 2.21;Plate 2).An understanding of the causes of geo-graphic patterns of climate, as presented in this chapter, therefore allows us to predictthe distribution of Earth’s major biomes withtheir characteristic patterns of productivity anddiversity (Plates 3 and 4).

Tropical wet forests occur from 12° N to 3° Sand correspond to the ITCZ. Day length and

Sun

September

JulyJanuary

March

Figure 2.20. Earth’s orbit around the sun, showingthat the zone of greatest heating (the ITCZ) is southof the equator in January, north of the equator inJuly, and at the equator in March and September.

Tundra

Borealforest

Temperateforest

Woodland

or grassland

SavannaTropical

dry forest

Air

tem

pera

ture

(o C

)

-15

-10

-5

0

5

10

15

20

25

3050 100 150 200 250 300 350 400 450

Precipitation (cm yr-1)

Temperate

wet forest

Tropicalwet forest

Desert

Figure 2.21. Distribution of major biomes in rela-tion to mean annual temperature and precipitation.(Redrawn from Communities and Ecosystems, 2ndedition, by R.H. Whittaker © 1975, by permission ofPearson Education, Inc., Upper Saddle River, NJ;Whittaker 1975.)


solar angle show little seasonal change, leadingto constant high temperatures. High solar radi-ation and convergence of the easterly tradewinds at the ITCZ promote strong convectiveuplift, leading to high precipitation (175 to 400cm annually). Periods of relatively low pre-cipitation seldom last more than 1 to 2 months.Tropical dry forests occur north and south oftropical wet forests. Tropical dry forests havemore pronounced wet and dry seasons becauseof seasonal movement of ITCZ over (wetseason) and away from these forests (dryseason). Tropical savannas occur between thetropical dry forests and the deserts. Thesesavannas are warm and have low precipitationthat is highly seasonal. Subtropical deserts at 25to 30° N and S have a warm dry climate becauseof the subsidence of air in the descending limbof the Hadley cell.

Mid-latitude deserts, grasslands, and shrub-lands occur in the interiors of continents, par-ticularly in the rain shadow of mountain ranges.They have low and unpredictable precipitation,low winter temperatures, and greater tempera-ture extremes than tropical deserts. As precipi-tation increases, there is a gradual transitionfrom desert to grassland to shrubland. Temper-ate wet forests occur on the west coasts of con-tinents at 40 to 65° N and S, where westerliesblowing across a relatively warm ocean providean abundant moisture source and migrating lowpressure centers associated with the polar frontpromote high precipitation. Winters are mild,and summers are cool. Temperate forests occurin the mid-latitudes, where there is sufficientprecipitation to support trees. The polar frontmigrates north and south of these forests fromsummer to winter, producing a strongly sea-sonal climate. Mediterranean shrublands aresituated on the west coasts of continents. Insummer, subtropical oceanic high pressurecenters and cold upwelling coastal currentsproduce a warm dry climate. In winter, as windand pressure systems move toward the equator,storms produced by polar fronts provide unpre-dictable precipitation.

The boreal forest (taiga) occurs in continen-tal interiors at 50 to 70° N. The winter climateis dominated by polar air masses and thesummer climate by temperate air masses, pro-

ducing cold winters and mild summers. The dis-tance from oceanic moisture sources results in low precipitation. The subzero mean annualtemperatures lead to permafrost (permanentlyfrozen ground) that restricts drainage andcreates poorly drained soils and peatlands inlow-lying areas. Arctic tundra is a zone north ofthe polar front in both summer and winter,resulting in a climate that is too cold to supportgrowth of trees. Short cool summers restrictbiological activity and limit the range of lifeforms that can survive.

Vegetation structure varies with climate bothbetween and within biomes. Each biome type is dominated by predictable growth forms ofplants. Tropical wet forests, for example, aredominated by broad-leaved evergreen trees,whereas areas that are periodically too cold ordry for growth are dominated by deciduousforests or, under more extreme conditions, bytundra or desert, respectively. Biomes are notdiscrete units with sharp boundaries but varycontinuously in structure along climatic gradi-ents. Along a moisture gradient in the tropics,for example, vegetation changes from ever-green tall trees in the wettest sites to a mix of evergreen and deciduous trees in areas thatsee the beginnings of a seasonal drought(Ellenberg 1979) (Fig. 2.22). As the climatebecomes still drier, the stature of the trees andshrubs is reduced because there is less lightcompetition and more competition for water.This leads to a shrubless desert with herbaceousperennial herbs in dry habitats. With extremedrought, the dominant life form becomesannuals and bulbs (herbaceous perennials inwhich aboveground parts die during the dryseason). A similar gradient of growth forms,leaf types and life forms occurs along moisturegradients at other latitudes.

The diversity of growth forms within someecosystems can be nearly as great as the diver-sity of dominant growth forms across biomes.In tropical wet forests, for example, continuousseasonal growth in a warm moist climate pro-duces large trees with dense canopies that inter-cept and compete for a large fraction of theincoming radiation. Light then becomes themain driver of diversity within the ecosystem.Plants that can reach the canopy and have

Relationship of Climate to Ecosystem Distribution and Structure 43

access to light compete effectively with talltrees. These growth forms include vines, whichparasitize trees for support without investingcarbon in strong stems. In this way vines cangrow quickly into the canopy. Epiphytes arealso common in the canopies of tropical wetforests, where they receive abundant light but,because their roots are restricted to the canopy,their growth is often water limited. Epiphyteshave therefore evolved various specializationsto trap water and nutrients. There is a widerange of subcanopy trees, shrubs, and herbs thatare adapted to grow slowly under the low-lightconditions beneath the canopy (Fig. 2.22). Lightis therefore probably the general driver ofstructural diversity in the dense forests of wettropical regions.

What determines structural diversity wheremoisture, rather than light, is limiting? Deserts,particularly warm deserts, have a great diversityof plant forms, including evergreen and decid-uous small trees and shrubs, succulents, herba-ceous perennials, and annuals. These growthforms do not show a well-defined vertical partitioning but show consistent horizontal pat-terns related to moisture availability. Competi-

tion for water results in diverse strategies forgaining, storing, and using the limited watersupply. This leads to a wide range of rootingstrategies and capacities to evade or enduredrought.

Species diversity declines from the tropics to high latitudes and in many cases from low to high elevation. Species-rich tropical areassupport more than 5000 species in a 10,000km2

area, whereas in the high arctic there are lessthan 200 species in the same area (Plate 4).Many animal groups show similar latitudinalpatterns of diversity, in part because of theirdependence on the underlying plant diversity.Climate, the evolutionary time available forspecies radiation, productivity, disturbance frequency, competitive interactions, land areaavailable, and other factors have all beenhypothesized to contribute to global patterns ofdiversity (Heywood and Watson 1995). Globalpatterns of diversity correlate most clearly withsome dimension of climate and related eco-system processes. Models that include onlyclimate, acting as a filter on the plant functionaltypes that can occur in a region, can reproducethe general global patterns of structural and

Figure 2.22. The change in life form dominance along a tropical gradient along which precipitation changesbut temperature is relatively constant. (Redrawn with permission from Journal of Ecology; Ellenberg 1979.)


species diversity (Kleiden and Mooney 2000).The actual causes for geographic patterns ofspecies diversity are undoubtedly more com-plex, but these models and other analysessuggest that human-induced changes in climate,land use, and invasions of exotic species mayalter future patterns of diversity.

Summary

The balance between incoming and outgoingradiation determines Earth’s energy budget.The atmosphere transmits about half of theincoming shortwave solar radiation to Earth’ssurface but absorbs 90% of the outgoing long-wave radiation emitted by Earth. This causesthe atmosphere to be heated primarily from thebottom and generates convective motion in the atmosphere. Large-scale patterns of atmos-pheric circulation occur because the tropicsreceive more energy from the sun than theyemit to space, whereas the poles lose moreenergy to space than they receive from the sun.The resulting circulation cells transport heatfrom the equator to the poles to balance theseinequalities. In the process, they create threerelatively distinct air masses in each hemi-sphere, a tropical air mass (0 to 30° N and S), atemperate air mass (30 to 60° N and S) and apolar air mass (60 to 90° N and S). There arefour major areas of high pressure (the two polesand 30° N and S), where air descends and pre-cipitation is low. The subtropical high pressurebelts are the zones of the world’s major deserts.There are three major zones of low pressure(the equator and 60° N and S), where air risesand precipitation is high. These areas supportthe tropical rain forests at the equator and the temperate rain forests of western North and southern South America. Ocean currentsaccount for about 40% of the latitudinal heat transport from the equator to the poles.These currents are driven by surface winds andby the downwelling of cold saline waters at high latitudes, balanced by upwelling at lowerlatitudes.

Regional and local patterns of climate reflectheterogeneity in Earth’s surface. Unevenheating between the land and the oceans mod-

ifies the general latitudinal patterns of climateby generating zones of prevailing high and lowpressure. These pressure centers and the loca-tion of major mountain ranges guide stormtracks that strongly influence regional patternsof climate. Oceans and large lakes also moder-ate climate on adjacent lands because their high heat capacity causes them to heat or coolmore slowly than land. These heating contrastsproduce predictable seasonal winds (mon-soon) and daily winds (land–sea breezes) thatinfluence the adjacent land. Mountains alsocreate heterogeneity in precipitation and in thequantity of solar radiation intercepted.

Vegetation influences climate through itseffects on surface albedo, which determines thequantity of incoming radiation absorbed by thesurface and energy released to the atmospherevia longwave radiation and turbulent fluxes oflatent and sensible heat. Sensible heat fluxesand longwave radiation directly heat the atmos-phere, and latent heat transfers water vapor tothe atmosphere, influencing local temperatureand moisture sources for precipitation.

Climate is variable over all time scales. Long-term variations in climate are driven largely by changes in solar input and atmospheric composition. Superimposed on these long-termtrends are predictable daily and seasonal pat-terns of climate, as well as repeating patternssuch as those associated with El Niño/southern oscillation. These oscillations causewidespread changes in the geographic patternsof climate on time scales of years to decades.Future changes in climate may reflect changesin the frequencies of these large-scale climatemodes.

Review Questions

1. Describe the energy budget of Earth’ssurface and the atmosphere. What are the major pathways by which energy isabsorbed by Earth’s surface? By the atmos-phere? What are the roles of clouds andradiatively active gases in determining therelative importance of these pathways?

2. Why is the troposphere warmest at thebottom but the stratosphere is warmest at


the top? How does each of these atmos-pheric layers influence the environment ofecosystems?

3. Explain how unequal heating of Earth bythe sun and the resulting atmospheric circulation produce the major latitudinalclimate zones, such as those characterizedby tropical forests, subtropical deserts, tem-perate forests, and arctic tundra.

4. How does the rotation of Earth (and theresulting Coriolis forces) and the separa-tion of Earth’s surface into oceans and continents influence the global patterns ofclimate?

5. How does the chemical composition ofEarth’s atmosphere influence climate?

6. What causes the global pattern in surfaceocean currents? Why are the deep waterocean currents different from those at thesurface? What is the nature of the con-nection between deep and surface oceancurrents?

7. How does ocean circulation influenceclimate at global, continental, and localscales?

8. How does topography affect climate at con-tinental and local scales?

9. What are the major causes of long-termchanges in climate? How would you expectfuture climate to differ from that of todayin 100 years? 10,000 years? 2 billion years?Explain your answers.

10. Explain how the interannual variations inclimate of Indonesia, Peru, and Californiaare interconnected. Would these patternsinfluence eastern North America orEurope? How?

11. Explain the climatic basis for the global dis-tribution of each major biome type. Usemaps of global winds and ocean currents toexplain these distributions.

12. Describe the climate of your birthplace.Using your understanding of the globalclimate system, explain why that locationhas its particular climate.

Additional Reading

Ahrens, C.D. 1998. Essentials of Meteorology: AnInvitation to the Atmosphere. Wadsworth,Belmont, CA.

Bradshaw, M., and R. Weaver. 1993. Physical Geog-raphy. Mosby, St. Louis.

Graedel, T.E., and P.J. Crutzen. 1995. Atmosphere,Climate, and Change. Scientific American Library,New York.

Oke, T.R. 1987. Boundary Layer Climates. 2nd ed.Methuen, London.

Skinner, B.J., S.C. Porter, and D.B. Botkin. The BluePlanet: An Introduction to Earth System Science.2nd ed. Wiley, New York.

Sturman, A.P., and N.J. Tapper. 1996. The Weatherand Climate of Australia and New Zealand. OxfordUniversity Press, Oxford, UK.

Introduction

Soils form a thin film over Earth’s surface inwhich geological and biological processes inter-sect. A unique feature of terrestrial ecosystemsis that vegetation acquires its resources fromtwo quite different environments—the air andthe soil. The soil is a multiphasic system con-sisting of solids, liquids, and gases, with solidstypically occupying about half the soil volume,and liquids and gases each occupying 15 to 35%of the volume (Ugolini and Spaltenstein 1992).The physical soil matrix provides a source ofwater and nutrients to plants and microbes and is the physical support system in which terrestrial vegetation is rooted. It provides themedium in which most decomposer organismsand many animals live. For these reasons,the physical and chemical characteristics of soils strongly influence all aspects of ecosystemfunctioning, which, in turn, feed back to influ-ence the physical, structural and chemical properties of soils (see Fig. 1.3). Soils play such an integral role in ecosystem processesthat it is difficult to separate the study of soilsfrom that of ecosystem processes. In open-water (pelagic) ecosystems, phytoplanktoncannot directly tap resources from sediments,so soil processes provide nutrient resources to

plants only indirectly through mixing of thewater column.

Soils are also a critical component of the totalEarth System. They play a key role in the giantglobal reduction–oxidation cycles of carbon,nitrogen, and sulfur. Soils mediate many of thekey reactions in these cycles and provide essen-tial resources to biological processes that drivethese cycles.

Soils represent the intersection of the bio,geo, and chemistry in biogeochemistry. Many of the later chapters in this book address theshort-term dynamics of soil processes, particu-larly those processes that occur on time scalesof hours to centuries. This chapter emphasizessoil processes that occur over longer time scalesor that are strongly influenced by physical andchemical interactions with the environment.This is essential background for understandingthe dynamics of ecosystems.

Controls over Soil Formation

The soil properties of an ecosystem result fromthe dynamic balance of two opposing forces:soil formation and soil loss. State factors differin their effects on these opposing processes and therefore on soil and ecosystem properties

3Geology and Soils

Within a given climatic regime, soil properties are the major control over ecosys-tem processes. This chapter provides background on the factors regulating the soilproperties that most strongly influence ecosystem processes.

46

Controls over Soil Formation 47

(Dokuchaev 1879, Jenny 1941, Amundson andJenny 1997).

Parent Material

The physical and chemical properties of rocksand the rates at which they are uplifted andweathered strongly influence soil properties.The dynamics of the rock cycle, operating overbillions of years, govern the variation and dis-tribution of geological materials on Earth’s sur-face. The rock cycle describes the cyclic processby which rocks are formed and weathered—that is, the chemical and physical alteration ofrocks and minerals near Earth’s surface (Fig.3.1). The rock cycle produces minerals thatbuffer the biological acidity that accounts for

much of rock weathering but also providesmany of the nutrients that allow biology toproduce this acidity. The compounds producedby weathering move via rivers to the oceanswhere they are deposited to form sediments,which are then buried to form sedimentaryrocks. Igneous rocks form when magma fromthe molten core of Earth moves upward towardthe surface in cracks or volcanoes. Either sedimentary or igneous rocks can be modifiedunder heat or pressure to form metamorphicrocks. With additional heat and pressure, meta-morphic rocks melt and become magma. Anyof these rock types can be raised to the surfacevia uplift during mountain-building episodes,after which the material is again subjected toweathering and erosion (Fig. 3.1). The timing

Precipitation in oceans

Uplift

Uplift

UpliftIgneousrocks

Heat and pressure

Cooling

Melting

Metamorphicrocks

Heat and pressure

Sediments

Burial andlithification

Weathering and erosion

Sedimentaryrocks

Magma

Figure 3.1. The rock cycles as proposed by Huttonin 1785. Rocks are weathered to form sediment,which is then buried. After deep burial, the rocksundergo metamorphosis, melting, or both. Later,they are deformed and uplifted into mountain

chains, only to be weathered again and recycled.(Redrawn with permission from Earth by FrankPress and Raymond Siever © 1974, 1978, 1982, and1986 by W.H. Freeman and Company; Press andSiever 1986.)

48 3. Geology and Soils

and locations of uplift and the kinds of rockuplifted ultimately determine the distributionof different kinds of bedrock across Earth’ssurface.

Plate tectonics are the driving forces behindrock formation. The lithosphere, or crust—thestrong outermost shell of Earth that rides onpartially molten material beneath—is brokeninto large rigid plates, each of which movesindependently. Where the plates converge andcollide, portions of the lithosphere buckledownward and are subducted, leading to theformation of ocean trenches, and the overridingplate is uplifted, causing the formation ofmountain ranges and volcanoes (Fig. 3.2).Regions of plate collision and active mountainbuilding coincide with Earth’s major earth-quake belts. The Himalayan Mountains, forexample, are still rising due to the collision of the Indian subcontinent with Asia. If platesconverge in one place, they must diverge or separate elsewhere. Eurasia, Africa, and theAmericas were once the single supercontinentof Pangaea, 200 million years ago. The mid-Atlantic and mid-Pacific ridges are zones ofactive divergence of today’s ocean plates.

Climate

Temperature and moisture influence rates ofchemical reactions that in turn govern the rate and products of weathering and therefore the development of soils from rocks. Tempera-

ture and moisture also influence biologicalprocesses, such as the production of organicmatter by plants and its decomposition bymicroorganisms and therefore the amount andquality of organic matter in the soil (see Chap-ters 5 to 7). Soil carbon, for example, increasesalong elevational gradients as temperaturedecreases (Vitousek 1994b) and decreases inrain shadows downwind of mountain ranges(Burke et al. 1989). Precipitation is a majorpathway by which many materials enter ecosys-tems. Oligotrophic (nutrient-poor) bogs are isolated from mineral soils and depend entirelyon precipitation to supply new minerals. Themovement of water is also crucial in determin-ing whether the products of weathering accu-mulate or are lost from a soil. In summary,climate affects virtually all soil properties atscales ranging from local to global.

Topography

Topography influences soils through its effecton climate and differential transport of fine soilparticles. The attributes of topography that areimportant for ecosystem processes include thesite’s topographic position on a catena or hills-lope complex, the aspect of the slope, and therelationship between the site and hydrologicpathways (Amundson and Jenny 1997).Characteristics such as soil depth, texture, andmineral content vary with hillslope position.Erosional processes preferentially move fine-

Mountain rangesand volcanoes

Granite

Crumpled sedimentary and metamorphic rocks

Trench

Lithosphere

Figure 3.2. Cross-section of a zone of plate collisionin which the oceanic plate is subducted beneath acontinental plate, forming an ocean trench in thezone of subduction and mountains and volcanoes in

the zone of uplift. (Redrawn with permission fromEarth by Frank Press and Raymond Siever © 1974, 1978, 1982, and 1986 by W.H. Freeman andCompany; Press and Siever 1986.)

Controls over Soil Formation 49

grained materials downslope and deposit themat lower locations. Depositional areas at thebase of slopes and in valley bottoms thereforetend to have deep fine-textured soils with a highsoil organic content (Fig. 3.3) and high water-holding capacity. Depositional areas supplymore soil resources to plant roots and microbesand provide greater physical stability than dohigher slope positions. For these reasons valleybottoms typically exhibit higher rates of mostecosystem processes than do ridges or shoul-ders of slopes. Soils in lower slope positions in sagebrush ecosystems, for example, havegreater soil moisture, higher soil organic mattercontent, and higher rates of nitrogen mineral-ization and gaseous losses than do upslope soils(Burke et al. 1990, Matson et al. 1991).

Slope position also determines patterns ofsnow redistribution in cold climates, withdeepest accumulations beneath ridges and inthe protected lower slopes. These differentialaccumulations alter effective precipitation andlength of growing season sufficiently to influ-ence plant and microbial processes well into thesummer.

Finally, the aspect of a slope influences solarinput (see Chapter 2) and therefore soil tem-perature, rates of evapotranspiration, and soilmoisture. At high latitudes and in wet climates,these differences in soil environment reducerates of decomposition and mineralization onpoleward-facing slopes (Van Cleve et al. 1991).At low latitudes and in dry climates, however,the greater retention of soil moisture on

poleward-facing slopes allows a longer growingseason and supports forests, whereas slopesfacing the equator are more likely to supportdesert or shrub vegetation (Whittaker andNiering 1965).

Time

Many soil-forming processes occur slowly, sothe time over which soils develop influencestheir properties. Rocks and minerals are weathered over time, and important nutrientelements are transferred among soil layers ortransported out of the ecosystem. Hillslopeserode, and valley bottoms accumulate materi-als, and biological processes add organic matterand critical nutrient elements like carbon and nitrogen. Phosphorus availability is highearly in soil development and becomes pro-gressively less available over time due to itslosses from the system and to its fixation inmineral forms that are unavailable to plants(Fig. 3.4) (Walker and Syers 1976). This processrequired millions of years of soil developmentin Hawaii, despite a warm moist climate (Crewset al. 1995) and resulted in a change from nitro-gen limitation of plant growth on young soils tophosphorus limitation on older soils (Vitouseket al. 1993).

Some changes in soil properties happen relatively quickly. Retreating glaciers and riverfloodplains often deposit phosphorus-rich till. Ifseed sources are available, these soils are colo-nized by plants with symbiotic nitrogen-fixingmicrobes, allowing such ecosystems to accumu-late their maximum pool sizes of carbon andnitrogen within 50 to 100 years (Crocker andMajor 1955, Van Cleve et al. 1991). Other soil-forming processes occur slowly. Young marineterraces in coastal California have relativelyhigh phosphorus availability but low carbonand nitrogen content. Over hundreds of thou-sand of years, these terraces accumulate organicmatter and nitrogen, causing a change fromcoastal grassland to productive redwood forest(Jenny et al. 1969). Over millions of years, sili-cates are leached out, leaving behind a hardpanof iron and aluminum oxides with low fertilityand seasonally anaerobic soils. The pygmycypress forests that develop on these old ter-

<1

3-4>4

Organiccarbon (%)

2-3

Erosionlikely

Depositionlikely

1-2

Figure 3.3. Relationship between hillslope position,likelihood of erosion or deposition, and soil organiccarbon concentration. (Redrawn with permissionfrom Oxford University Press; Birkeland 1999.)


races have very low productivity. The phenoliccompounds produced by these trees as defensesagainst herbivores also retard decomposition,further reducing soil fertility (Northup et al.1995) (see Chapter 13).

Potential Biota

The past and present organisms at a sitestrongly influence soil chemical and physicalproperties. Most soil development occurs in the presence of live organisms (Ugolini andSpaltenstein 1992). There are often clear asso-ciations between vegetation and soils. Theorganic acids in the litter of many coniferousspecies, for example, acidify the soil. This, incombination with the characteristically lowquality of conifer litter, leads to slower decom-position in conifer than in deciduous forests(Van Cleve et al. 1991) (see Chapter 7). It is frequently difficult, however, to separate the

chicken from the egg. Did the vegetation deter-mine soil properties or vice versa?

One approach to determining vegetationeffects on soils has been to plant monoculturesor species mixes into initially homogeneoussites. Rapidly growing grasses in a nitrogen-poor perennial grassland enhanced the nitro-gen mineralization of soils within 3 years(Wedin and Tilman 1990) (see Fig. 12.5), as did deep-rooted forbs in an annual grassland(Hooper and Vitousek 1998). Another ap-proach has been to examine the consequencesof species invasions or extinctions on soilprocesses. The invasion of a non-native nitro-gen fixer into Hawaiian rain forests, forexample, increased nitrogen inputs to thesystem more than fivefold, altering the charac-teristics of soils and the colonization and com-petitive balance among native plant species(Vitousek et al. 1987) (see Fig. 12.3).

Animals also influence soil properties. Earth-worms, termites, and invertebrate shreddersstrongly influence decomposition rates (seeChapter 7) and therefore soil properties thatare influenced by soil organic content.

Human Activities

Since the 1950s, the tripling of the human population and associated agricultural andindustrial activities have strongly influencedsoil development worldwide. Human activitiesinfluence soils directly through changes innutrient inputs, irrigation, alteration of soilmicroenvironment through land use change(see Chapter 14), and increased erosional lossof soils. Human activities indirectly affect soilsthrough changes in atmospheric compositionand through the deletions and additions of spe-cies. Today and in the future, human activitieswill affect ecosystem properties both directlyand through their effects on other interactivecontrols (see Chapters 14 and 16).

Controls over Soil Loss

Soil formation depends on the balance betweendeposition, erosion, and soil development. Soilthickness varies with hillslope position, with

Primary P

Organic P

Secondary P

Total phosphorus

Pho

spho

rus

cont

ent

Time

OccludedP

Figure 3.4. The generalized effects of long-termweathering and soil development on the distributionand availability of phosphorus (P). Newly exposedgeologic substrate is relatively rich in weatherableminerals, which release phosphorus. This releaseleads to accumulation of both organic and readilysoluble forms (secondary phosphorus, such ascalcium phosphate). As primary minerals disappearand secondary minerals capable of sorbing phos-phorus accumulate, an increasing proportion of thephosphorus remaining in the system is held inunavailable (occluded) forms. Availability of phos-phorus to plants peaks relatively early in thissequence and declines thereafter. (Redrawn withpermission from Geoderma, Vol. 15 © 1976 ElsevierScience; Walker and Syers 1976.)

Controls over Soil Loss 51

erosion dominating on steep hillslopes, deposi-tion in valley bottoms, and soil development on gentle slopes and terraces where the lateraltransport of materials is minimal (Fig. 3.3).Much of Earth’s surface is in hilly or moun-tainous terrain where erosion and depositionare important processes. In these landscapepositions, soils may have developed for only afew thousand years, even though soils on flatterraces of the same geomorphic age may havedeveloped for millions of years. Erosion is animportant process, because it removes the products of weathering and biological activity.In young soils, these erosional losses reduce soilfertility by removing clays and organic matterthat store water and nutrients. On highlyweathered landscapes, however, erosion renewssoil fertility by removing the highly weatheredremnants (sands and iron oxides) that con-tribute little to soil fertility and by exposing lessweathered materials that provide a new sourceof essential nutrients.

Average regional erosion rates vary by twoto three orders of magnitude among areas withdifferent topography and climate (Table 3.1)(Milliman and Meade 1983). Erosion rates tendto approach rates of tectonic uplift, so regionswith active tectonic uplift and steep slopes gen-erally have higher erosion rates than flat weath-ered terrain. In steep terrain, soil creep andother processes unrelated to rainfall can be thedominant erosional processes. Climatic zones

with high rainfall also tend to have high erosionrates. Some semiarid zones and polar montaneregions with active uplift also have occasionalepisodes of active erosion during intense rains,because there is little vegetation to protect soilsfrom erosion. Land use changes that reducevegetation cover can increase erosion rates byseveral orders of magnitude, causing meters ofsoil to be lost in a few years. This compares tobackground erosion rates of 0.1 to 1.0mm per1000 years in many areas. High erosion ratescaused by land use change are only temporaryin a geologic sense. Once the soil mantle isgone, erosion rates decline to values deter-mined by climate, bedrock type, and slope.Much of the erosion on natural landscapesprobably occurs during rare periods of unusu-ally high rainfall or low vegetation cover ratherthan during average conditions.

The dominant erosional processes depend on topography, surface material properties, andthe pathways by which water leaves the land-scape. Mass wasting is a major erosionalprocess in most regions. This is the downslopemovement of soil or rock material under theinfluence of gravity without the direct aid ofother media such as water, air, or ice. The rateof mass wasting depends on the hillslope gradi-ent, length, and curvature. Some mass wastingoccurs rapidly, for example, in rockfalls, land-slides, or debris flows. Landslides can transportsediment volumes ranging from a few cubic

Erosion ratea

Climate zone Relief (mm per 1000 yr)

Glacial Gentle (ice sheets) 50–200Steep (valleys) 1000–5000

Polar montane Steep 10–1000Temperate maritime Mostly gentle 5–100Temperate continental Gentle 10–100

Steep 100–200+Mediterranean — 10–?Semiarid Gentle 100–1000Arid — 10–?Wet subtropics — 10–1000?Wet tropics Gentle 10–100

Steep 100–1000

a Erosion rates are estimated from average sediment yields of rivers in different cli-matic and topographic regimes.Data From Selby (1993).

Table 3.1. Climatic and topo-graphic effects on long-termerosion rates.


meters to cubic kilometers. The probability of amass-wasting event depends on the balancebetween the driving forces for downslopemovement and the forces that resist this move-ment. Gravity is the major driving force formass wasting.The gravitational force (or stress)can be divided into two components: one par-allel to the slope, which drives mass wasting,and one perpendicular to the slope, whichincreases the friction between the material andthe bedrock (Fig. 3.5).The steeper the slope, thegreater the downhill component of the forceand therefore the greater the probability ofmass wasting.

Many factors influence the strength of a soilmass (i.e., the amount of force required to initiate slope failure) (Selby 1993). Theseinclude the sliding friction between the material and some well-defined plane and the internal friction caused by the frictionamong individual grains within the soil matrix.In some cases, there is a well-defined planealong which materials can slide, such as themovement of soils over a frozen soil layer, butcommonly it is the internal friction that largelydetermines the resistance to mass wasting.Cohesion among soil particles and water mole-cules enhances the internal friction that resists mass wasting. A small amount of waterenhances cohesion among particles, explainingwhy sand castles are easier to make with moistthan with dry sand. A high water content,however, exerts pressure on the grains, makingthem more buoyant and reducing the frictionalstrength. They become unstable, leading to liq-uefaction of the soil mass, which can flowdownslope. Fine-particle soils have lower slopethresholds of instability and are more likely tolead to slope failure than are coarse-texturedsoils. Roots also increase the resistance of soilsto downslope movement, so deforestation andother land use changes that reduce rootbiomass increase the probability of landslides.

Mass wasting on soil-mantled, well-vegetatedgentle slopes occurs slowly through soil creep.Displacement of surface soil particles byfreeze–thaw events or the movement of soilsbrought to the surface by burrowing animals,for example, is likely to cause a net down-slope movement of soil. These small-scale pro-

cesses contribute to erosion rates of 0.1mmyr-1 or less.

The pathways by which water leaves thelandscape strongly influence erosion. Watertypically leaves a landscape by one of severalpathways: groundwater flow, shallow subsur-face flow, or overland flow (when precipitationrate exceeds infiltration rate) (see Fig. 14.6).The relative importance of these pathways isstrongly influenced by topography, vegetation,and material properties such as the hydraulic

Fn

FtFp

Fn

Fp

Ft

Figure 3.5. Effect of slope on the partitioning of thetotal gravitational force (Ft) into a component thatis normal to the slope (Fn)—and therefore con-tributes to friction that resists erosion—and a com-ponent that is parallel to the slope (Fp)—andtherefore promotes erosion. Steep slopes have alarger Fp value and therefore a greater tendency toerode.

Development of Soil Profiles 53

conductivity of soils. Groundwater and shallowsubsurface flow dissolve and remove ions andsmall particles. At the opposite extreme, over-land flow causes erosion primarily by surfacesheet wash, rills, and rain splash. This typicallyoccurs in arid and semiarid soil-mantled land-scapes or on disturbed ground. Overland flowrates of 0.15 to 3cms-1 are sufficient to suspendclay and silt particles and move them downhill(Selby 1993). As water collects into gullies,its velocity, and therefore erosion potential,increases. Vegetation and litter layers greatlyincrease infiltration into the soil by reducingthe velocity with which raindrops hit the soil, thereby preventing surface compaction bythe raindrops. Vegetated soils are also lesscompact because roots and soil animals createchannels in the soil. In these ways vegetationand a litter layer substantially increase infiltra-tion and therefore groundwater and subsurfaceflow.

Wind is an important agent of erosion inareas where wind speeds are high at the soilsurface, for example, where vegetation removalexposes the soil surface to strong winds. Some

agricultural areas in China have lost meters ofsoil to wind erosion and have become an impor-tant source of iron to the phytoplankton in thePacific Ocean (see Chapter 10).

Glaciers are an important erosional pathwayin cold, moist climates. Glacial rivers often carry large sediment loads and produce abraided river valley with meandering streamchannels that are important locations ofprimary succession.

Erosion in one location must be balanced by deposition elsewhere. Deposition can rangefrom slow rates of dust or loess input to siltation events during floods to massivemoraines or debris accumulations at the base ofslopes.

Development of Soil Profiles

Soils develop through the addition of materialsto the system, transformation of those materi-als in the system, transfer down and up in thesoil profile, and loss of materials from thesystem (Fig. 3.6).

ADDITIONS

Precipitation (including ions and solid particles);

organic matter

Ground surface

TRANSFORMATIONSOrganic matter humus

Primary minerals

hydrous oxidesclaysions, H4SiO4

LOSSES

H4SiO4

Ions,

TRANSFERS

Ions, H4SiO4

Soil

TRANSFERSHumus

compounds,clays,

ions, H4SiO4

Figure 3.6. Processes leading to addi-tions, transformations, transfers, andlosses of materials from soils. Silica isH4SiO4. (Redrawn with permission fromSoils and Geomorphology by Peter W.Birkeland, © 1999 Oxford UniversityPress, Inc.; Birkeland 1999.)


Additions to Soils

Additions to the soil system can come fromoutside or inside the ecosystem. Inputs fromoutside the ecosystem include precipitation andwind, which deposit ions and dust particles,and floods and tidal exchange, which depositsediments and solutes (see Chapter 9). Thesource of these materials determines their size distribution and chemistry, leading to thedevelopment of soils with specific textural andchemical characteristics. Organisms within theecosystem add organic matter and nitrogen tothe soil, including the aboveground and below-ground portions of plants, animals, and soilmicrobes.

Soil Transformations

Within the soil, materials are transformedthrough an interaction of physical, chemical,and biological processes. Freshly depositeddead organic matter is transformed in the soilby decomposition to soil organic matter,releasing carbon dioxide and nutrients such as nitrogen and phosphorus (see Chapter 7). Morerecalcitrant organic compounds undergophysicochemical interactions with soil mineralsthat contribute to the long-term storage of soilorganic matter.

Weathering is the change of parent rocks and minerals to produce more stable forms.This occurs when rocks and minerals becomeexposed to physical and chemical conditionsdifferent from those under which they formed(Ugolini and Spaltenstein 1992). Weatheringinvolves both physical and chemical processesand is influenced by characteristics of theparent material and by temperature, moisture,and the activities of organisms. Physical weath-ering is the fragmentation of parent materialwithout chemical change. This can occur when rocks are fractured by earthquakes orwhen stresses are relieved due to erosional lossof the weight of overlying rock and soil. In addition, soil particles and rock fragments are abraded by wind or are ground against one another by glaciers, landslides, or floods.Rocks also fragment when they expand andcontract during freeze–thaw, heating–cooling,or wetting–drying cycles or when roots grow

into rock fissures. Fire is a potent force of physical weathering because it rapidly heats the rock surface to a high temperature whileleaving the deeper layers cool. Physical weath-ering is especially important in extreme andhighly seasonal climates. Wherever it occurs, itopens channels in rocks for water and air topenetrate, increasing the surface area for chem-ical weathering reactions.

Chemical weathering occurs when parentrock materials react with acidic or oxidizingsubstances, usually in the presence of water.During chemical weathering, primary minerals(minerals present in the rock or unconsolidatedparent material before chemical changes haveoccurred) are dissolved and altered chemicallyto produce more stable forms, ions are released,and secondary minerals (products that areformed through the reaction of materialsreleased during weathering) are formed. Someprimary minerals can be hydrolyzed by water,producing new minerals plus ions in solution.Hydrolysis reactions, however, typically includeboth water (H2O) and an acid. Carbonic acid(H2CO3) is the most important acid involved inchemical weathering. The CO2 concentration inmost soils is 10- to 30-fold higher than in air,due to the low diffusivity of gases in soil and therespiration of plants, soil animals, and microor-ganisms. Weathering rates are particularly highadjacent to roots because of the high rates of biological activity and CO2 production in the rhizosphere. Carbon dioxide dissolves andreacts with water to form carbonic acid, whichthen ionizes to produce a hydrogen ion (H+)and a bicarbonate ion (HCO3

-). Carbonic acid, for example, attacks potassium feldspar(KAlSi3O8), which is converted into a sec-ondary mineral, kaolinite (Al2Si2O5(OH)4), bythe removal of soluble silica (SiO2) and potas-sium ion (K+) (Eq. 3.1). Kaolinite can, under theright conditions, undergo another dissolution to form another secondary mineral gibbsite(Al(OH)3).

(3.1)

Plant roots and microbes secrete manyorganic acids into the soil, which influence

2KAlSi O 2 H HCO H O

Al Si O OH 4SiO 2K 2HCO3 8 3 2

2 2 5 4 2 3

+ +( ) + Æ( ) + + +

+ -

+ -


chemical weathering through their contributionto soil acidity and their capacity to chelate ions.In the chelation process, organic acids combinewith metallic ions, such as ferric iron (Fe3+) andaluminum (Al3+), making them soluble andmobile. Chelation lowers the concentration ofinorganic ions at the mineral surface, so dis-solved and primary mineral forms are no longerin equilibrium with one another. This acceler-ates the rate of weathering.

The physical and chemical properties of rock minerals determine their susceptibility to weathering and the chemical products thatresult. Sedimentary rocks like shale that formby chemical precipitation, for example, havemore basic cations like calcium (Ca2+), sodium(Na+) and potassium (K+) than does igneousrock and tends to produce soils with a relativelyhigh pH and a high capacity to supply mineralcations to plants. Igneous rocks weather in thereverse order in which they crystallize duringformation (Birkeland 1999). Olivine, forexample, is one of the first minerals to crystal-lize as magma cools. It has a high energy offormation and weathers easily. Feldspar formsand weathers more slowly than olivine, andquartz is one of the last minerals to form(explaining why it forms crystals) and is highlyresistant to weathering (Table 3.2). Secondaryminerals such as the silicate clay minerals andiron and aluminum oxides are among the mostresistant minerals to weathering. Textural dif-ferences in parent material also influence therate of chemical breakdown, with fine-grainedrocks weathering more slowly than coarse-grained rocks.

Warm climates promote chemical weatheringbecause temperature speeds chemical reactions

by increasing the kinetic energy of reactants.The activities of plants and microorganisms are also more rapid under warm conditions.Wet conditions promote weathering throughtheir direct effects on weathering reactions and their effects on biological processes. Not sur-prisingly, the hot wet conditions of humid tropical climates yield the highest rates ofchemical weathering.

The secondary minerals formed in weather-ing reactions play critical roles in soils andecosystem processes. In temperate soils, weath-ering products include layered silicate clay min-erals. These small particles (less than 0.002mm)are hydrated silicates of aluminum, iron, andmagnesium arranged in layers to form a crys-talline structure. Two types of sheets make upthese minerals: A tetrahedral sheet consists ofunits composed of one silicon atom surroundedby four atomic oxygen (O-) groups (Fig.3.7A,B). An octahedral sheet consists of unitshaving six oxygen (O-) or hydroxide (OH-) ionssurrounding an Al3+, magnesium ion (Mg2+), orFe3+ ion (Fig. 3.7C,D). Various combinations of these sheets give rise to a wide variety of clay minerals. Montmorillonite and illite, forexample, have 2 :1 ratios of silica to aluminum-dominated layers. Kaolinite, a more strongly

Table 3.2. Stability of common minerals underweathering conditions at Earth’s surface.

Most stable Fe3+ oxides Secondary mineralAl3+ oxides Secondary mineralQuartz Primary mineralClay minerals Secondary mineralK+ feldspar Primary mineralNa+ feldspar Primary mineralCa2+ feldspar Primary mineral

Least stable Olivine Primary mineral

Data from Press and Siever (1986).

= Hydroxyland

= Aluminum, magnesium, etc.

(C) (D)

(A) (B)

= Oxygen

= Silicon

and

and

Figure 3.7. The molecular structure of a simple claylayer. A, A tetrahedral unit. B, A tetrahedral sheet.C, An octahedral unit. D, An octahedral sheet.(Redrawn with permission from Clay Mineralologyby R.E. Grim, © 1968 McGraw-Hill Companies;Grim 1968.)


weathered clay mineral, has a 1 :1 ratio of thetwo. The structure and concentration of theseclay minerals influences the cation exchangecapacity, water-holding capacity, and othercharacteristics of soils.

Secondary minerals that form in soils can beeither crystalline, with highly regular arrange-ments of atoms, as in silicate clay minerals, oramorphous, with no regular arrangement ofatoms. In many volcanic soils forming on ashdeposits, allophane (Al2O3 ·2SiO2 ·nH2O) is anamorphous mineral that is produced relativelyearly in weathering but is then transformed to crystalline aluminum oxide minerals likeAl(OH)3 with time. In many tropical soils,weathering has occurred in place for millions ofyears in a humid climate with high leachingrates. Here the relatively mobile ions of silicon(Si) and Mg2+ as well as Ca2+, K+, and Na+ arepreferentially leached, leaving behind the lessmobile ions Al3+ and Fe3+. Silicate clay mineralsare therefore no longer present, and the clay-size particles are composed of iron and alu-minum oxides. These oxides form when Fe2+ orAl2+ are released in the weathering of iron- or aluminum-bearing minerals and are then oxidized in solution and react with anions toform a precipitate. Hematite and gibbsite areexamples of the oxides produced.

Soil Transfers

Vertical transfers of materials through soilsgenerate distinctive soil profiles—that is, thevertical layering of soils. These transfers typi-cally occur by leaching (the downward move-ment of dissolved materials) and particulatetransport in water. Soluble ions that are addedin precipitation or released by weathering inupper layers of the soil profile can move down-ward in solution until a change in chemicalenvironment causes them to become reactantsin chemical processes or until dehydrationcauses them to precipitate out of solution. Theamounts of silica and base cations in secondaryminerals therefore frequently increase withdepth. These cations are leached from upperlayers (termed horizons) and form new miner-als under the new conditions of pH and ioniccontent encountered at depth. Chelated com-

plexes of organic compounds and iron or aluminum ions are also water soluble and canmove in water to deeper layers of the soilprofile. Slight changes in ionic content and themicrobial breakdown of the organic matter areamong the processes that can cause the metalions to precipitate as oxides. Clay-size particleslike silicates and iron and aluminum oxides can also be transported downward in solution,sometimes forming deep horizons with highclay content in wet climates. Soil texture affectsthe rate and depth of leaching (Fig. 3.8) andthus the translocation and accumulation ofmaterials in soil profiles. Constituents releasedduring weathering of coarse-textured glacialtill, for example, may be leached from the soilbefore they have a chance to react chemicallyto form secondary minerals.

Soils of arid and semiarid environments alsoaccumulate materials in specific horizons.Thesesystems often have hard calcium (or magne-sium) carbonate-rich calcic horizon or caliche.Downward-moving soil water carries dissolvedCa2+ and bicarbonate (HCO3

-). Precipitation ascalcium carbonate (CaCO3) occurs under con-ditions of increasing pH, which drives reaction3.2 to the left. Precipitation can also occurunder saturating concentrations of carbonateand with evaporation of soil water.

CaCO3 + H2CO3 ´ Ca2+ + 2HCO3- (3.2)

Figure 3.8. Hypothetical depth of leaching relatedto the texture of the original parent material.(Redrawn with permission from Soils and Geomor-phology by Peter W. Birkeland, © 1999 Oxford University Press, Inc.; Birkeland 1999.)


Although most of the transfers in soils occurthrough the downward movement of water,materials can also move upward in water. Thecapillary rise of water from a shallow watertable, for example, transfers water and ionsfrom lower to upper soil layers (see Chapter 4).Because capillary water movement depends onadhesive properties of soil particles, the poten-tial distance for capillary rise is greater in claysoils, which have small pore sizes, than in sandysoils (Birkeland 1999). Soluble ions or com-pounds may accumulate in layers at the top ofthe capillary fringe. Salt pans, for example, format the soil surface in low-lying areas of deserts.Minerals that are added to soils in irrigationwater in dry regions can also accumulate at thesoil surface as the water evaporates. Thisprocess of salinization has led to widespreadabandonment of farmland in dry regions of theworld. In western Australia, for example, wide-spread removal of vegetation for croplandscaused the soils to become so saline that manyof these croplands have been abandoned.Salinization occurs naturally in deserts, whenstreams drain into salt flats, where the waterevaporates rather than runs off.

Some minerals accumulate at the top of awater table that forms when downward perco-lation of water is impeded by an impermeablesoil layer. Poor drainage often leads to lowoxygen availability because oxygen diffuses10,000 times more slowly in water than in airand is readily depleted in water-logged soils byroot and microbial respiration. Low oxygenconcentration creates reducing conditions thatconvert ions with multiple oxidation states totheir reduced form. Iron and manganese, forexample, are more soluble in their reduced(Fe2+ and Mn2+, respectively) than in their oxi-dized state (Fe3+ and Mn4+, respectively). Fe2+

and Mn2+ diffuse through waterlogged soils tothe surface of the water table, where there issufficient oxygen to convert them to their oxidized forms. There they precipitate out ofsolution to form a distinct layer that is rich iniron and manganese. This layering of iron andmanganese is particularly pronounced in lakesediments, where there is a strong gradient inoxygen concentration from the sedimentsurface. The conversion from ferric (Fe3+) to

ferrous (Fe2+) iron gives rise to the characteris-tic gray and bluish colors of waterlogged gleysoils.

Soils that are subjected to repeated wettingand drying and saturation during some seasonscan also develop characteristic accumulations.Plinthite layers (sometimes called laterite) intropical soils, for example, are layers of iron-rich minerals that have hardened irreversiblyon exposure to repeated saturation and dryingcycles. Depending on their location within theprofile, these layers can impede water drainageand root growth.

The actions of plant roots and soil animalstransfer materials up and down the soil profile(Paton et al. 1995). Organic matter inputs to soiloccur primarily at the surface and in upper soil horizons. When plants die, the mineralsacquired by deep roots are also deposited on ornear the soil surface. This contributes, forexample, to the base-rich soils and uniqueground flora beneath deep-rooted oak trees insouthern Sweden (Andersson 1991).Tree wind-throw, which occurs when large trees aretoppled by strong winds, also redistributes rootsand associated soil upward. Finally, animalssuch as gophers transfer materials up and downin the soil profile as they tunnel and feed onplant roots. Earthworms in temperate soils andtermites in tropical soils are particularly impor-tant in transferring surface organic matter deepinto the soil profile and, at the same time, bring-ing mineral soil from depth to the surface (seeChapter 12). These processes play critical rolesin the redistribution of nutrients and in thecontrol of net primary productivity.

Losses from Soils

Materials are lost from soil profiles primarily assolutions and gases. The quantity of mineralsleached from an ecosystem depends on boththe amount of water flowing through the soilprofile and its solute concentration. Manyfactors influence these concentrations, includ-ing plant demand, microbial mineralizationrate, cation- or anion-exchange capacity, andprevious losses via leaching or gas fluxes. Aswater moves through the soil, exchange reac-tions with mineral and organic surfaces replace


loosely bound ions on the exchange complexwith ions that bind more tightly. In this waymonovalent cations such as Na+, ammonium ion(NH4

+), and K+ and anions such as chloride ion(Cl-) and nitrate ion (NO3

-) are readilyreleased from the exchange complex into thesoil solution and are particularly prone toleaching loss. The maintenance of chargebalance of soil solutions requires that the leach-ing of negatively charged anions be accompa-nied by an equal charge of positive ions(cations). Inputs of sulfuric acid (H2SO4) in acidrain therefore increase leaching losses ofreadily exchangeable base cations like Na+,NH4

+, and K+, which leach downward withsulfate ion (SO4

2-).Materials can also be lost from soils as gases.

Gas emissions depend on the rate of produc-tion of the gas by microbes, the diffusionalpaths through soils, and the exchange at thesoil–air interface (Livingston and Hutchinson1995). The controls over these losses are dis-cussed in Chapter 9.

Soil Horizons and Soil Classification

Ecosystem differences in additions, trans-formations, transfers, accumulations, and lossesgive rise to distinct soils and soil profiles. Soilsinclude organic, mineral, gaseous, and aqueousconstituents arranged in a relatively predictable

vertical structure. The number and depth ofhorizons (layers) and the characteristics of eachlayer in a soil profile vary widely among soils.Nonetheless, a series of typical horizons can bedescribed for many soils (Fig. 3.9). The organichorizon, or O horizon, consists of organic mate-rial that accumulates above the mineral soil.This layer of organic material is derived fromthe litter of dead plants and animals. The Ohorizon can be subdivided based on the degreeof decomposition that the majority of materialhas undergone, with the lower portion of the Ohorizon being more decomposed. The Ahorizon is the uppermost mineral soil horizon.Being adjacent to the organic layers, it typicallycontains substantial quantities of organicmatter and is therefore dark in color.The O andA horizons are the zones of most active plantand microbial processes and therefore havehighest nutrient supply rates (see Chapter 9).Many soils in wet climates have an E horizonbeneath the A horizon that is strongly leached.Most clay minerals and iron and aluminumoxides have been leached from the horizon,leaving behind resistant minerals like quartz, inaddition to sand and silt-size particles. The Bhorizon beneath the A and E horizons is thezone of maximum accumulation of iron andaluminum oxides and clays. Salts and precipi-tates sometimes also accumulate here, espe-cially in arid and semiarid environments. The Chorizon lies beneath the A and B horizons.Although it may accumulate some of the

Oi Organic, slightly decomposedOe Organic, moderately decomposedOa Organic, highly decomposed

A Mineral, mixed with humus, dark colored

E

B

C

R

Soi

l

Bedrock

Horizon of maximum leaching of silicateclays, Fe, Al oxides, etc.

Zone of Fe and Al accumulation

Zone of least weathering and accumulation;contains unweathered parent material

O

Figure 3.9. A generic soil profile showingthe major horizons that are formed duringsoil development. Density of dots reflectsconcentration of soil organic matter.

Soil Horizons and Soil Classification 59

leached material from above, it is relativelyunaffected by soil-forming processes. C hori-zons typically include a significant portion ofunweathered parent material. Finally, at somedepth, there is an unweathered bedrock layer,the regolith (R).

Despite the large variation among theworld’s soils, they can be classed into majorgroups that share many of the same propertiesbecause they have formed in response tosimilar soil-forming factors and processes. Soilclassification systems rely on the diagnosticcharacteristics of specific horizons and onorganic matter content, base saturation, andproperties that indicate wetness or dryness.TheU.S. Soil Taxonomy recognizes 12 major soilgroupings, called soil orders (Table 3.3) (Bradyand Weil 1999). Most agronomic and ecosystemstudies classify soils to the level of a soil series,a group of soil profiles with similar profilecharacteristics such as type, thickness, and prop-erties of the soil horizons. Soil series can befurther subdivided into types, based on thetexture of the A horizon, and into phases, basedon information such as landscape position,

stoniness, and salinity.A comparison of soil pro-files from the major soil orders illustrates theeffect of different climatic regimes on soildevelopment (Figs. 3.10 and 3.11).

Entisols are soils with minimal soil develop-ment. They occur either because the soils arerecent or processes that disrupt soil develop-ment dominate over processes that form soils.This is the most widespread soil type in theworld, making up 16% of the ice-free surface.Inceptisols, in which the soil profile has onlybegun to develop, occupy an additional 10% ofthe ice-free surface. Rock and shifting sandaccount for another 14% of the ice-free surface.Thus about 40% of the ice-free surface of Earthshows minimal soil development.

Histosols are highly organic soils thatdevelop in any climate zone under conditionsin which poor drainage restricts oxygen diffu-sion into the soil, leading to slow rates ofdecomposition and accumulation of organicmatter. There is a well-developed O horizon of undecomposed organic material where mostplants are rooted.The high water table preventsthe vertical leaching required for soil devel-

Table 3.3. Names of the soil orders in the United States soil taxonomy and their characteristics and typicallocations.

Area (% ofSoil Order ice-free land) Major Characteristics Typical Occurrence

Entisols 16.3 no well-developed horizons sand deposits, plowed fieldsInceptisols 9.9 weakly developed soils young or eroded soilsHistosols 1.2 highly organic; low oxygen peatland, bogGelisols 8.6 presence of permafrost tundra, boreal forestAndisols 0.7 from volcanic ejecta; moderately developed recent volcanic areas

horizonsAridisols 12.1 dry soils with little leaching arid areasMollisols 6.9 deep dark-colored A horizon with >50% base grasslands, some deciduous forests

saturationVertisols 2.4 high content (>30%) of swelling clays; crack grassland with distinct wet and dry

deeply when dry seasonsAlfisols 9.7 sufficient precipitation to leach clays into a B humid forests; shrublands

horizon; >50% base saturationSpodosols 2.6 sandy leached (E) horizon; acidic B horizon; cold wet climates, usually beneath

surface organic accumulation conifer forestsUltisols 8.5 clay-rich B horizon, low base saturation wet tropical or subtropical climate;

forest or savannaOxisols 7.6 highly leached horizon with low clay; highly hot humid tropics beneath forests

weathered on old landformsRock and sand 14.1

Data from Miller and Donahue (1990) and Brady and Weil (2001).

Figure 3.10. Relationships among the major soilorders, showing the conditions under which theyform, the relative time required for formation, and

Gelisol(Tundra, bog)

Aridisol (Desert)

Spodosol(Acidic conifer

forests)

B

A

C

75

0

100

25

50

Cal

cic

A

B

A

75

0

100

25

50

O

C

E

A

B

Mollisol(Grassland,

deciduous forest)

O

A

C

75

0

100

25

50

Per

maf

rost

75

0

100

25

50

75

0

100

25

50

Oxisol(Tropical wet forest)

Soi

l dep

th (

cm)

Soi

l dep

th (

cm)

Figure 3.11. Typical profilesof five contrasting soil ordersshowing differences in thetypes and depths of horizons.Symbols same as in Figure 3.9.

the types of ecosystems with which they are com-monly associated (Birkeland 1999).

Soil Properties and Ecosystem Functioning 61

opment, so these soils have weak developmentof mineral soil horizons. Gelisols are organicsoils that develop in climates with a meanannual temperature below 0°C and are under-lain by a layer of permanently frozen soil (permafrost).

Andisols are young soils that occur on volcanic substrates and tend to produce amorphous clays.

Aridisols, as the name implies, develop inarid climates. The low rainfall minimizes leach-ing and weathering and may allow accumula-tion of soluble salts. There is no surface Ohorizon. The shallow A horizon has littleorganic matter due to low productivity andrapid decomposition. Low precipitation resultsin a poorly differentiated B horizon. Many ofthese soils develop a calcic layer of calcium andmagnesium carbonates that precipitate at depthbecause there is insufficient water to leach themout of the system. Desert calcic layers cangreatly reduce root penetration, restricting theroots of many desert plants to surface soils.Aridisols are a common world soil type,accounting for 12% of the terrestrial surface(Miller and Donahue 1990).

Vertisols are characterized by swelling andshrinking clays. They tend to occur in warmregions with a moist to dry climate.

Mollisols are fertile soils that developbeneath grasslands and some deciduous forests.They have a deep organic-rich A horizon witha high nutrient content that grades into a Bhorizon. Due to their high fertility, mollisolshave been extensively cultivated and supportthe major grain-growing regions of the world.They account for 25% of soils in the UnitedStates and 7% of soils worldwide (Miller andDonahue 1990).

Spodosols (or podzols by the older termi-nology) are highly leached soils that develop in cold wet climates, usually beneath coniferstands. Beneath the A horizon is a highlyleached, almost white, E horizon and a darkbrown or black B horizon, where leaching pro-ducts accumulate. These soils are acidic with ahigh sand and low clay content. Alfisols developbeneath temperate and subtropical forests,especially in deciduous forests that receive lessprecipitation. They are less strongly leached

than spodosols and therefore have a higher claycontent and lower acidity.

Ultisols develop in warm, humid areas wherethere is substantial leaching. These soils oftenhave a high clay content, a low base saturation,and low fertility. Oxisols are the most highlyweathered leached soils. They occur on oldlandforms in the wet tropics. The A horizon isso highly weathered that it contains iron oxidesbut very little clay and has extremely low fer-tility. This horizon often extends several metersin depth.

Soil Properties and Ecosystem Functioning

Spatial and temporal variations in soil develop-ment result in large variations in soil properties.These soil properties, in turn, modulate theavailability of water and nutrient resources for plant growth and therefore the cycling of water and nutrients through ecosystems. Inthis section, we discuss a few of the importantproperties.

Soil texture is defined by the relative pro-portions of soil particles of different sizes, rang-ing from clay-size particles (less than 0.002mm)to silt (0.002 to 0.05mm), and sand (0.05 to 2.0mm) (Fig. 3.12). Loam soils contain sub-stantial proportions of at least two size classesof particles. Rocks and gravel are large (greaterthan 2mm), unweathered primary minerals that can occupy a substantial proportion of the volume of many soils. Sand- and silt-sizeparticles consist of unweathered primary min-erals and some secondary minerals released by weathering. Clay-size particles consist of alarger proportion of secondary minerals,including layered silicate clays and other smallcrystalline and amorphous minerals. Parentmaterial has a large effect on soil texturebecause rocks differ in their rates of physicaland chemical weathering.

Soil texture also depends on the balancebetween soil development that occurs in place,deposition from wind or water, and erosionalloss of materials. When soils weather in place,the conversion of primary minerals to clay-sizesecondary minerals increases the proportion of


fine soil particles. For this reason, high-latitudesoils, with their slow rates of chemical weather-ing, frequently have a lower clay content thando temperate or tropical soils of similarbedrock and age. Fine particles are, however,more susceptible to erosion by wind or waterthan are large particles. Water erosion trans-ports clays from hilltops to valley bottoms, pro-ducing fine-textured soils in river valleys. If theriver valleys are poorly vegetated, as in braidedrivers that drain glaciated landscapes, wind canthen move fine particles back to hillslopes toform loess soils with a high silt content. Overmillions of years, minerals dissolve and are lostfrom the soil.

Texture is important primarily because itdetermines the total surface area in a volumeof soil. Fine-grained particles in the soil matrixhave greater surface-to-volume ratio. Soils withfine textures and large surface areas hold morewater by adsorption of water films to soil par-ticles.

The packing of small particles in the soilmatrix results in greater water-holding porespace between particles. Under intermediatelevels of rainfall, ecosystems growing on sandy

soils tend to be more xeric (characterized byplants that are tolerant of dry conditions) thanthose occurring on finer-textured soils. In thecase of soils with silicate clay particles, theincreased surface area also represents greatersurface charge and thus greater cationexchange capacity, as described later. Manyother soil properties such as bulk density, nutri-ent content, water-holding capacity, and redoxpotential are related to soil texture, so texturecan be a good predictor of many ecosystemproperties (Parton et al. 1987).

Soil structure reflects the aggregation of soilparticles into larger units. Aggregates formwhen soil particles become cemented togetherand then crack into larger units as soils dry orfreeze. There are many types of glue that bindsoil particles together to form aggregates.Theseinclude organic matter, iron oxides, polyvalentcations, clays, and silica. Soil aggregates rangein size from less than 1mm to greater than 10cm in diameter.

Soil texture strongly affects the formation ofsoil aggregates. Sandy soils form few aggre-gates, whereas loam and clay soils form a rangeof aggregate sizes. Polysaccharides secreted

10

20

30

40

50

60

70

80

90

100

203040 105060708090

100

10

20

30

40

50

60

70

80

9010

0

Sandyclay

Clay

Clay loam Silty clay loam

Silt loam

Sand

Loam

Sandyclay loam

Percent sand

Perc

ent c

lay Percent silt

Loamy sand Silt

Sandy loam

Siltyclay

Figure 3.12. Percentages of sand,silt, and clay in the major soil textural classes. (Redrawn withpermission from Soils and Geomorphology by Peter W.Birkeland, © 1999 Oxford University Press, Inc.; Birkeland1999.)


by roots and bacteria are important sources of organic matter that binds soil particlestogether. Fungal hyphae contribute strongly toaggregation in many soils. For these reasons, theloss of soil organic matter and its associatedmicrobes can lead to a loss of soil structure,which contributes to further soil degradation.Earthworms and other soil invertebrates con-tribute to aggregate formation by ingesting soiland producing feces that retain a coherentstructure. Plant species and their microbialassociates differ in the capacity of their exu-dates to form aggregates, so soil texture, organicmatter content, and species composition allinfluence soil structure.

The cracks and channels between aggregatesare important pathways for water infiltration,gaseous diffusion, and root growth, thus affect-ing water availability, soil aeration, oxida-tion–reduction processes, and plant growth.Thefine-scale heterogeneity created by soil aggre-gates is critical to the functioning of soils. Slowgas diffusion through the partially cementedpores within aggregates creates anaerobic con-ditions immediately adjacent to aerobic sur-faces of soil pores.This allows the occurrence inwell-drained soils of anaerobic processes suchas denitrification, which requires the productsof aerobic processes (nitrification, in this case)(see Chapter 9).

Compaction by animals and machinery fillsthe cracks and pores between aggregates.Plowing reduces aggregation through mech-anical disruption of aggregates and through lossof soil organic matter and associated cementingactivity of microbial exudates and fungalhyphae (Fisher and Binkley 2000). The loss ofsoil structure through compaction preventsrapid infiltration of rainwater and leads toincreased overland flow and erosion.

Bulk density is the mass of dry soil per unitvolume, usually expressed in grams per cubiccentimeter (gcm-3; equivalent to megagramsper cubic meter, or Mgm-3). Bulk density varieswith soil texture and soil organic mattercontent. Bulk densities of mineral soils (1.0 to2.0gcm-3) are typically higher than those oforganic soil horizons (0.05 to 0.4gcm-3). Fine-textured soils have higher internal surface areaand more pore space than coarser-textured

soils, and thus their bulk densities tend to belower. If they are compacted, however, bulkdensities of clay soils can be higher than thoseof coarse-textured soils. Bulk density stronglyinfluences the nutrient and water charac-teristics of a site. Organic soils, for example,frequently have highest concentrations (percentof dry mass) of carbon in surface horizons withlow bulk density but the greatest quantities(grams per cubic centimeter) of carbon atdepth, where bulk density is greater. The quan-tity of nutrient per unit volume is calculated bymultiplying the percentage concentration of thenutrient times soil bulk density. Volumetricnutrient content is generally more relevantthan nutrient concentration in describing thequantity of nutrients directly available to plantsand microbes.

Water is a critical resource for most eco-system processes. In soils, water is held in porespaces as films of water adsorbed to soil parti-cles. The soil is water-saturated when all porespaces are filled with water. Under these con-ditions water drains under the influence ofgravity (saturated flow) until, often afterseveral days, the adhesive forces that holdwater in films on soil particles equals thegravitational pressure. At this point, called fieldcapacity, water no longer freely drains.

At water contents below field capacity, watermoves through the soil by unsaturated flow inresponse to gradients of water potential—thatis, the potential energy of water relative to purewater (see Chapter 4).When plant roots absorbwater from the soil to replace water that is lostin transpiration, there is a reduction in thethickness of water films adjacent to roots, whichcauses the remaining water to adhere moretightly to soil particles. The net effect is toreduce the soil water potential at the rootsurface. Water moves along water films throughthe soil pores toward the root in response tothis gradient in water potential. Plants continueto transpire, and water continues to movetoward the root until some minimal waterpotential is reached, when roots can no longerremove water from the particle surfaces. Thispoint is called the permanent wilting point.Water-holding capacity is the difference inwater content between field capacity and per-


manent wilting point (see Fig. 4.5). Water-holding capacity is substantially enhanced bypresence of clay and soil organic matterbecause of the large surface area of these mate-rials. The water-holding capacity of an organicsoil might, for example, be 300% (300g H2O per100g dry soil), while that of a clay soil may be30% and that of a sandy soil could be less than20%. On a volumetric basis, water-holdingcapacity is normally highest in loam soils. Oneconsequence of this difference is that, for agiven amount of rainfall, sandy soils are wettedmore deeply than clay soils but retain less waterin soil horizons that are accessible to plants.The water-holding characteristics of soils helpdetermine the amount of water available forplant uptake and growth and for microbialprocesses, including decomposition and nutri-ent cycling and loss.

Oxidation–reduction reactions involve thetransfer of electrons from one reactant toanother, yielding chemical energy that can beused by organisms (Lindsay 1979). In thesereactions, the energy source gives up one ormore electrons (oxidation). These electrons aretransferred to electron acceptors (reduction).Ahandy mnemonic is: “LEO the lion says GER,”where LEO stands for loss of electrons—oxi-dation, and GER stands for gain of electrons—reduction. Redox potential is the electricalpotential of a system due to the tendency ofsubstances in it to lose or accept electrons(Schlesinger 1997, Fisher and Binkley 2000).There is a wide range of redox potentialsamong soils due to their ionic and chemicalcompositions. One important set of redox reac-tions, which occurs inside the mitochondria oflive eukaryotic cells, is the transfer of electronsfrom carbohydrates through a series of reac-tions to oxygen.This series of reactions releasesthe energy needed to support cellular growthand maintenance. Many other redox reactionsoccur in the cells of soil organisms, when elec-trons are transferred from electron donors toacceptors (Table 3.4). The greatest amount ofenergy can be harvested by organisms by trans-ferring electrons to oxygen. However, underanaerobic conditions, which commonly occur inflooded soils with high organic matter contentsor in aquatic sediments, electrons must be

transferred to other electron acceptors; thusprogressively less energy is released with thetransfer to each of the following electron accep-tors:

O2 > NO3- > Mn4+ > Fe3+ >

SO42- > CO2 > H+ (3.3)

As soil redox potential declines, the pre-ferred electron carriers are gradually consumed(Table 3.4). As oxygen becomes depleted, forexample, the redox reaction that generates themost energy is denitrification (transfer of elec-trons to nitrate), followed by reduction of Mn4+

to Mn2+, then reduction of Fe3+ to Fe2+, thenreduction of SO4

2- to hydrogen sulfide (H2S),then reduction of CO2 to methane (CH4). Thuspoorly aerated soils with high sulfate concen-trations (e.g., salt marshes) are less likely toreduce CO2 to CH4 than are similar soils withlower SO4

2- concentrations.Many soil organisms carry out only one or a

few redox reactions, although certain bacteriacan couple the reduction of Mn4+ and Fe3+

directly to the oxidation of simple organic sub-strates (Schlesinger 1997).Temporal and spatialvariations in soil redox potential alter the typesof redox reactions that occur primarily by alter-ing the competitive balance among theseorganisms. Organisms that derive more energyfrom their redox reactions (e.g., denitrifierscompared to methane producers) will be com-petitively superior, when they have an adequatesupply of electron acceptors.

Soil organic matter content is a critical com-ponent of soils, affecting rates of weatheringand soil development, soil water-holding capa-city, soil structure, and nutrient retention. Inaddition, soil organic matter provides theenergy and carbon base for heterotrophic soilorganisms (see Chapter 7) and is an importantreservoir of essential nutrients required forplant growth (see Chapter 8). Soil organicmatter originates from dead plant, animal, andmicrobial tissues, but includes a range of mate-rials from new, undecomposed plant tissues toresynthesized humic substances that are thou-sands of years old, whose origins are chemicallyand physically unrecognizable (see Chapter 7).Because soil organic matter is important to somany soil properties, loss of soil organic matter


as a consequence of inappropriate land man-agement is a major cause of land degradationand loss of biological productivity.

The negative log of the hydrogen ion (H+)activity (effective concentration) in solution isreferred to as pH, which is a measure of theactive acidity of the system. The pH stronglyaffects nutrient availability through its effectson cation exchange (see next paragraph) and the solubility of phosphate compounds andmicronutrients such as iron, zinc, copper, andmanganese.

Cation exchange capacity (CEC) reflects thecapacity of a soil to hold exchangeable cationson negatively charged sites on the surfaces ofsoil minerals and organic matter. Cationexchange occurs when a loosely held cation ona negatively charged site exchanges with acation in solution. Values for CEC vary morethan 100-fold among clay minerals. Crystallineclay minerals typically have a negative orneutral charge under ambient soil pH. Thenegative charge originates from unsatisfied neg-ative charges along the interlayer surfaces of thesilicate clay lattices, especially in the 2 :1 clays,and from hydroxide (—OH) groups that are

exposed on the edges of 1 : 1 clay particles. Soilorganic matter has very high CEC that origi-nates from the presence of —OH and carboxyl(—COOH) groups at the surfaces of organiccompounds and within particles of humic materials. Soil organic matter contributes sub-stantially to the total CEC of some soils. Forexample, organic matter accounts for most CEC in tropical soils that consist primarily of iron and aluminum oxides and 1 :1 silicate clay minerals, because there is little CEC in themineral matrix. High-latitude soils also derive alarge proportion of their CEC from organicmatter due to their high organic content and lowclay content. The pool of exchangeable cationsin the soil is much larger than the soil solutionpool and represents the major short-term storeof cations for plant and microbial uptake.

Exchangeable cations are attracted to thenegatively charged surfaces. Base saturation isthe percentage of the total exchangeable cationpool that is accounted for by base cations (thenonhydrogen, nonaluminum cations).The iden-tity of the cations on the exchange sitesdepends on the concentrations of cations in thesoil solution and on the strength with which dif-

Redox potential Energy releaseb

Reactiona (mV) (Kcal mol-1 per e-)

Reduction of O2 812 29.9O2 + 4H+ + 4e- Æ 2H2O

Reduction of NO3- 747 28.4

NO3- + 2H+ + 2e- Æ NO2

- + H2OReduction of Mn4+ to Mn2+ 526 23.3

MnO2 + 4H+ + 2e- Æ Mn2+ + 2H2OReduction of Fe3+ to Fe2+ -47 10.1

Fe(OH)3 + 3H+ + e- Æ Fe2+ + 3H2OReduction of SO4

2- to H2S -221 5.9SO4

2- + 10H+ + 8e- Æ H2S + 4H2OReduction of CO2 to CH4 -244 5.6

CO2 + 8H+ + 8e- Æ CH4 + 2H2O

a The reactions at the top of the table occur in soils with high redox potential andrelease more energy (and are therefore favored) when the electron acceptors areavailable. The reactions at the bottom of the table release less energy and thereforeoccur only if other electron acceptors are absent or have already been consumedby redox reactions. Abbreviations include electrons (e-), nitrite ion (NO2

-), man-ganese dioxide (MnO2), Ferric hydroxide (Fe(OH)3), organic matter (CH2O), uni-versal gas constant (R), temperature (T), and equilibrium constant (K).b Assumes coupling to the oxidation reaction: CH2O + H2O Æ CO2 + 4H+ + 4e- andthat the energy released - RT ln(K).Data from Schlesinger 1997.

Table 3.4. Sequence of H+-consuming redox reactionsthat occur with progressivedeclines in redox potential.


ferent cations are held to the exchangecomplex. In general, cations occupy exchangesites and displace other ions in the sequence

H(Al3+) > H+ > Ca2+ > Mg2+ > K+

ª NH4+ > Na+ (3.4)

so leached soils tend to lose Na+ and NH4+ but

retain Al3+ and H+. This displacement series is aconsequence of differences among ions incharge and hydrated radius. Ions with morepositive charges bind more tightly to theexchange complex than do ions with a singlecharge. Ions with a smaller hydrated radiushave their charge concentrated in a smallervolume and tend to bind tightly to the exchangecomplex.

Minerals like the iron and aluminum oxidesfound in many tropical soils have surfacecharges that vary between positive and nega-tive, depending on pH. At the low pH condi-tions typical of these soils, the net charge issometimes positive (Uehara and Gillman1981), so they attract anions, creating an anionexchange capacity. As with cations, anionabsorption depends on the concentration ofanions and their relative capacities to be heldor to displace other anions. Anions generallyoccupy exchange sites and displace other ionsin the sequence

PO43- > SO4

3- > Cl- > NO3- (3.5)

so leached soils tend to lose NO3- and Cl- but

retain phosphate (PO43-) and sulfate (SO4

3-).This retention reflects both anion exchange andthe formation of covalent bonds that are notreadily broken.

The high CEC and base saturation found inmany soils, especially in many temperate soils,provide buffering capacity that keeps the soilsfrom becoming acid. When additional H+ isadded to the system in solution (e.g., in acidrain), it exchanges with cations that were heldon cation exchange sites on clay minerals andsoil organic matter. Buffering capacity allowsthe pH in forest soils to remain relatively constant for long periods despite chronic expo-sure to acid rain. When the buffering capacity is exceeded, the soil pH begins to drop, whichcan solubilize aluminum hydroxides (Al(OH)x),Al3+, and other cations, with potentially toxic

effects on both terrestrial and downstreamaquatic ecosystems (Schulze 1989, Aber et al.1998). In many tropical soils, the relatively lowCEC does not function as efficiently to buffersoil solution chemistry. Additions of acids tothese already acidic unbuffered systemsreleases aluminum in solution more readily,making these soils potentially toxic to manyplants and microbes.

Summary

Five state factors control the formation andcharacteristics of soils. Parent material is gene-rated by the rock cycle, in which rocks areformed, uplifted, and weathered to produce thematerials from which soil is derived. Climate isthe factor that most strongly determines therates of soil-forming processes and thereforerates of soil development. Topography modifiesthese rates at a local scale through its effects onmicroclimate and the balance between soildevelopment and erosion. Organisms alsostrongly influence soil development throughtheir effects on the physical and chemical envi-ronment. Time integrates the impact of all statefactors in determining the long-term trajectoryof soil development. In recent decades, humanactivities have modified the relative importanceof these state factors and substantially alteredEarth’s soils.

The development of soil profiles representsthe balance between profile development, soilmixing, erosion, and deposition. Profile devel-opment occurs through the input, transforma-tion, vertical transfer, and loss of materials fromsoils. Inputs to soils come from both outside theecosystem (e.g., dust or precipitation inputs)and inside the ecosystem (e.g., litter inputs).The organic matter inputs are decomposed toproduce CO2 and nutrients or are transformedinto recalcitrant organic compounds. The car-bonic acid derived from CO2 and the organicacids produced during decomposition convertprimary minerals into clay-size secondary minerals, which have greater surface area andcation exchange capacity. Water moves thesesecondary minerals and the soluble weatheringproducts down through the soil profile until


new chemical conditions cause them to becomereactants or to precipitate out of solution.Leaching of materials into groundwater orerosion and gaseous losses to the atmosphereare the major avenues of loss of materials fromsoils.The net effect of these processes is to formsoil horizons that vary with climate, parentmaterial, biota, and soil age. These horizonshave distinctive physical, chemical, and biologi-cal properties.

Review Questions

1. What processes are responsible for thecycling of rock material in Earth’s crust?

2. Over a broad geographic range, what are thestate factors that control soil formation?How might interactive controls modify theeffects of these state factors?

3. What processes determine erosion rate?Which of these processes are most stronglyinfluenced by human activities?

4. What processes cause soil profiles todevelop? Explain how differences inclimate, drainage, and biota might affectprofile development.

5. What are the processes involved in physicaland chemical weathering? Give examples ofeach. How do plants and plant products con-tribute to each?

6. How is soil texture defined? How does itaffect other soil properties? Why does itinfluence ecosystem processes so strongly?

7. What is cation exchange capacity (CEC),and what determines its magnitude in tem-perate soils? How would you expect thedeterminants of CEC to differ among his-tosol, alfisol, and oxisol soils?

8. In a warm climate, how will soil processesand properties differ between sites withextremely high and extremely low precipita-tion? In a moist climate, how will soilprocesses and properties differ between siteswith extremely high and extremely low soiltemperature?

9. If global warming caused only an increase intemperature, how would you expect this toaffect soil properties after 100 years? After1 million years?

Additional Reading

Amundson, R., and H. Jenny. 1997. On a state factormodel of ecosystems. BioScience 47:536–543.

Birkeland, P.W. 1999. Soils and Geomorphology. 3rded. Oxford University Press, New York.

Brady, N.C., and R.R. Weil. 1999. The Nature andProperties of Soils. 12th ed. Prentice-Hall, UpperSaddle River, NJ.

Jenny, H. 1941. Factors of Soil Formation. McGraw-Hill, New York.

Selby, M.J. 1993. Hillslope Materials and Processes.Oxford University Press, Oxford, UK.

Ugolini, F.C., and H. Spaltenstein. 1992. Pedos-phere. Pages 123–153 in S.S. Butcher, R.J.Charlson, G.H. Orians, and G.V. Wolfe, editors.Global Biogeochemical Cycles. Academic Press,London.

Introduction

Water and solar energy are essential for thefunctioning of the Earth System. Since neitheris distributed evenly around the globe, themechanisms by which they are redistributed(the global hydrologic cycle and energy budget)are important (see Chapter 2). These processesare so tightly intertwined that they cannot be treated separately (Box 4.1). Solar energydrives the hydrologic cycle through the verticaltransfer of water from Earth to the atmospherevia evapotranspiration, the sum of evaporationfrom surfaces and transpiration, which is thewater loss from plants. Conversely, evapotran-spiration accounts for 75% of the turbulentenergy transfer from Earth to the atmosphereand is therefore a key process in Earth’s energybudget (see Fig. 2.2). The hydrologic cycle also controls Earth’s biogeochemical cycles by influencing all biotic processes, dissolving nutrients, and transferring them within andamong ecosystems. These nutrients provide theresources that support growth of organisms.The movement of materials that are dissolvedand suspended in water links ecosystems withina landscape.

The importance of the hydrologic cycle raisesconcerns about the extent to which it has beenmodified by human activities. Humans cur-

rently use half of Earth’s readily available freshwater, which is about half of the annual meanrunoff in regions accessible to people. Wateruse is projected to increase to 70% by 2050(Postel et al. 1996). This human use of freshwater affects land and water management;the movement of pollutants among ecosystems;and, indirectly, ecosystem processes in un-managed ecosystems. Land use changes havealtered terrestrial water and energy budgetssufficiently to change regional and globalclimate (Chase et al. 2000). Finally, humanactivities alter the capacity of the atmosphereto hold water vapor. Water vapor is the majorgreenhouse gas. It is transparent to solar radia-tion but absorbs longwave radiation from Earth(see Fig. 2.1) and thus provides an insulativethermal blanket. Climate warming caused byemissions of other greenhouse gases increasesthe quantity of water vapor in the atmosphereand therefore the efficiency with which theatmosphere traps longwave radiation. Thiswater vapor feedback explains why climateresponds so sensitively to emissions of othergreenhouse gases (see Chapter 2). Warmingaccelerates the hydrologic cycle, increasingevaporation and rainfall at the global scale (seeChapter 15). Warming also causes the sea levelto rise, due primarily to the thermal expansionof the oceans and secondarily to melting of

4Terrestrial Water and Energy Balance

The hydrologic cycle is the matrix in which all other biogeochemical cycles func-tion. This chapter describes the controls over the hydrologic cycle and ecosystemenergy balance, which drives the hydrologic cycle.

71

72 4. Terrestrial Water and Energy Balance

The unique properties of water are critical tounderstanding the linkages between energyand water budgets. Evapotranspiration isone of the largest terms in both the waterand the energy budgets of ecosystems, sofactors governing the magnitude of evapo-transpiration determine the strength of thelinkage between the water and energy cycles.How much energy is required to melt ice, towarm water, and to evaporate water? Whatdetermines the quantity of water vapor thatthe atmosphere can hold before precipita-tion occurs?

Due to its high specific heat—the energyrequired to warm 1g of a substance by 1°C—water changes temperature relatively slowlyfor a given energy input (Table 4.1). Conse-quently, the summer temperature near largewater bodies fluctuates less and is generallycooler than in inland areas. A wet surfacealso heats more slowly but evaporates morewater than a dry surface.

Considerable amounts of energy areabsorbed or released when water changesstate. The energy required to change 1g ofice to liquid water, the heat of fusion, is 0.33MJkg-1. More than seven times thatenergy (2.45MJkg-1) is required to change 1g of liquid water to water vapor at 20°C, theheat of vaporization. Changes betweenliquid and vapor therefore generally havegreater effects on ecosystem energy budgetsthan do changes between liquid and solid.

A consequence of these properties ofwater is that energy is released from a

Table 4.1. Specific heat of various materials.a

Substance (kJ kg-1 °C-1)

Ice 2.1Water 4.2Steam 2.0Dry sand 0.8Peat soils 1.8Air 1.0

a It takes four times the energy to raise the temperatureof an equal mass of water to that of air.

12

10

8

6

4

2

00 10 20 30 40 50V

apor

pre

ssur

e (k

Pa)

Temperature (oC)

100% RH 50% RH

20

40

60

80

100

Vap

or d

ensi

ty (

g m

-3)

0

Figure 4.1. Relationship between temperatureand the water-holding capacity of the atmosphereat 50% and 100% relative humidity (RH). Notethat relative humidity is not a good predictor ofevaporation rate. The same relative humidity canoccur at a series of different vapor pressures.

Box 4.1. Properties of Water that Link Water and Energy Budgets

transpiring leaf as the water changes from aliquid to vapor, causing the leaf to cool. Con-versely, the atmosphere generally becomeswarmer when water condenses to form clouddroplets because energy is released whenwater changes phase from a vapor to a liquidstate. This energy provides the additionalbuoyancy that forms tall thunderhead clouds(see Chapter 2).

The water vapor density (or absolutehumidity) is a measure of the mass of waterper volume of dry air. The amount of watervapor that can be held in air without itbecoming saturated increases greatly as tem-perature rises (Fig. 4.1). Consequently, asclimate warms, the water-holding capacity ofthe atmosphere increases in a greater-than-linear fashion. The relative humidity (RH) isthe ratio of the actual amount of water heldin the atmosphere compared to themaximum amount that could be held at thattemperature. Since the maximum potentialvapor density is quite sensitive to tempera-ture, the same relative humidity can occur atvery different vapor densities (Fig. 4.1). Rel-ative humidity alone is therefore not a goodindicator of the absolute water vaporcontent of the air.

Surface Energy Balance 73

day in a nonpolluted atmosphere, direct radia-tion accounts for 90% of incoming shortwaveradiation to an ecosystem, and diffuse radiationbecomes proportionately greater on cloudydays or near dawn or dusk when sun angles arelower. For Earth as a whole, direct and diffuseradiation each account for about half of incom-ing shortwave radiation (see Fig. 2.2).

The proportion of the incoming shortwaveradiation that is absorbed depends on thealbedo (a) or shortwave reflectance of theecosystem surface. Ecosystem albedos vary atleast 10-fold, ranging from highly reflective sur-faces such as fresh snow to dark surfaces suchas wet soils (Table 4.2). Conifer canopies, for

Table 4.2. Typical values of albedo of major surfacetypes on earth.

Surface type Albedo

Oceans and lakes 0.03–0.10a

Sea ice 0.30–0.45Snow

Fresh 0.75–0.95Old 0.40–0.70

Arctic tundra 0.15–0.20Conifer forest 0.09–0.15Broadleaf forest 0.15–0.20Agricultural crops 0.18–0.25Grassland 0.16–0.26Savanna 0.18–0.23Desert 0.20–0.45Bare soil

Wet, dark 0.05Dry, dark 0.13Dry, light 0.40

a Albedo of water increases greatly (from 0.1 to 1.0) atsolar angles <30°.Data from Oke (1987), Sturman and Tapper (1996), andEugster et al. (2000).

Vapor pressure is the partial pressureexerted by water molecules in the air. Thedriving force for evaporation is the differ-ence in vapor pressure between the airimmediately adjacent to an evaporatingsurface and that of the air with which itmixes. The air immediately adjacent to anevaporating surface is approximately satu-rated at the temperature of the surface.The vapor pressure deficit (VPD) is the

difference between actual vapor pressureand the vapor pressure of air at the sametemperature and pressure that is saturatedwith water vapor. This term is loosely usedto describe the difference in vapor pressurebetween air immediately adjacent to anevaporating surface and the bulk atmos-phere, although strictly speaking, the airmasses are at different temperatures.

glaciers and ice caps. A rising sea level endan-gers the coastal zone, where most of the majorcities of the world are located. Given the keyrole of water and energy in ecosystem andglobal processes, it is critical that we understandthe controls over water and energy exchangeand the extent to which they have been modified by human actions.

Surface Energy Balance

Solar Radiation Budget

The energy absorbed by a surface is the balancebetween incoming and outgoing radiation.Here we focus on ecosystem-scale radiationbudgets, although the general principles applyat any scale, ranging from the surface of a leafto the surface of the globe (see Fig. 2.2). Thetwo major components of the radiation budgetare shortwave radiation (K), the high-energyradiation emitted by the sun, and longwaveradiation (L), the thermal energy emitted by allbodies (see Chapter 2). Net radiation (Rnet) isthe balance between the inputs and outputs ofshortwave and longwave radiation, measuredas watts per meter squared (Wm-2).

(4.1)

On a clear day, direct radiation from the sunaccounts for most of the shortwave input to anecosystem (see Fig. 2.2). Additional input ofshortwave radiation comes as diffuse radiation,which is scattered by particles and gases in the atmosphere, or as reflected radiation fromclouds and surrounding landscape units such aslakes, dunes, or snowfields. At noon on a clear

R K K L Lnet in out in out= -( ) + -( )


example, have a lower albedo than deciduousforests, and grasslands with large amounts of standing dead leaves have relatively highalbedo. The albedo of a complex canopy is lessthan that of individual leaves, because much ofthe light reflected or transmitted by one leaf isabsorbed by other leaves and stems. For thisreason, deep, uneven canopies of conifer forests have a low albedo. Changes in eco-system albedo explain in part why high-latituderegions are projected to warm more rapidlythan low latitudes. As climate warms, snow andsea ice will melt earlier in the spring, replacinga reflective snow-covered surface with a darkabsorptive surface. This process, together withthe resulting change in temperature, is referredto as the snow (or ice) albedo feedback. Overlonger time scales, the northward movement oftrees into tundra causes an additional reductionin regional albedo in winter because the darkforest canopy masks the underlying snow-covered surface.With the low sun angles typicalof high latitudes, this effect is significant evenwith sparse canopies. As the treeline movesnorth, the land surface absorbs more energy,which is then transferred to the atmosphere,causing a positive feedback to regionalwarming (Foley et al. 1994). Albedo also variesdiurnally; it is about twice as high in earlymorning and evening as at midday. The diurnalchanges in absorbed radiation are thereforegreater than one would expect from diurnalvariations in incoming radiation.

Ecosystem Radiation Budget

The amount of longwave radiation emitted by an object depends on its temperature and its emissivity, a coefficient that describes thecapacity of a body to emit radiation. Mostabsorbed radiation is emitted (emissivity about 0.98 in vegetated ecosystems), so differ-ences among ecosystems in longwave radiationbalance depend primarily on the temperatureof the sky, which determines Lin, and the surfacetemperature of the ecosystem, which deter-mines Lout.

(4.2)R K K L L

K T Tnet in out in out

in sky sky surf surf

= -( ) + -( )= -( ) + -( )1

4 4a s e e

where a is the surface albedo, s is the Stefan-Boltzman constant (5.67 ¥ 10-8 Wm-2 K-4), T isabsolute temperature (K), and e is emissivity.Clouds are warmer than space and effectivelytrap longwave emissions from the surface, soecosystems receive more longwave radiationunder cloudy than under clear conditions. Thisexplains why cloudy nights are warmer thanclear ones and why deserts are generally coldat night, despite the high inputs of solar energyduring the day.

Longwave radiation emitted by the eco-system depends on surface temperature, which,in turn, depends on the quantity of radiationreceived by the surface and the efficiency withwhich this energy is transmitted into the air andsoil. Surfaces that absorb a large amount ofradiation, due to high solar inputs and/or lowalbedo tend to be warmer and therefore emitmore longwave radiation. Dry surfaces andleaves with low transpiration rates also tend tobe warm because they are not cooled by theevaporation of water. Desert sands, recent burnscars, and city pavements, for example, are generally hot. Similarly, a well-watered lawn ismuch cooler than an ecosystem that is dry or isdominated by plants with low transpirationrates. Because Lout is a function of temperatureraised to the fourth power (Eq. 4.2), surfacetemperature has a powerful multiplying effecton Lout.

Canopy structure also influences surfacetemperature and surface energy exchangethrough its effect on the efficiency of energydissipation. The irregular surface of vegetationslows down airflow unevenly, creating mechan-ical turbulence. Tall uneven canopies such asconifer forests are aerodynamically more roughthan are short smooth canopies. The mechani-cal turbulence generated by airflow across thevegetation surface creates eddies, which sweepdown into the canopy, transporting bulk airinward and canopy air out. These eddies trans-fer energy away from the surface and mix itwith the atmosphere (Jarvis and McNaughton1986). Air flowing across short, smoothcanopies such as grasslands or crops tends to beless turbulent, so these canopies are less effi-cient in shedding the energy that they absorb—that is, they are less tightly coupled to the bulk

Surface Energy Balance 75

atmosphere. Smooth canopies therefore tend tohave higher surface temperatures during theday and greater longwave emissions than doforest canopies. In general, albedo and surfacetemperature have the greatest impact on radi-ation balance and therefore net radiation (Eq.4.2). Shortwave radiation varies much morestrongly from day to night than does longwaveradiation (Fig. 4.2).

Energy Partitioning

The net energy absorbed by an ecosystem (netradiation) is approximately balanced by energythat is released to the atmosphere or conductedinto the soil. There are two major ways in whichenergy can be stored (S) in an ecosystem, anincrease in the temperature of biomass and theconversion of light to chemical energy throughphotosynthesis. These two forms of energystorage are generally less than 10% of net radi-ation on a daily basis.Thus, although the energytrapped by photosynthesis is the major ener-getic engine that drives the carbon cycle ofecosystems, it is only a small part of the total energy budget of the ecosystem. Because

ecosystem energy storage is usually small,energy input at the surface approximatelyequals energy loss over a day.

Net radiation is partitioned primarily amongthree major avenues of energy exchangebetween the ecosystem and the atmosphere–soil: ground heat flux (G), latent heat flux (orevapotranspiration, LE), and sensible heat flux(convective heating, H). The latent heat ofvaporization (L; 2.45MJkg-1 at 20°C) is a con-stant describing the quantity of energy requiredto evaporate a given mass of water, and evapo-transpiration rate (E) is the rate of water transfer from a land or water surface to theatmosphere. Net radiation is positive whendirected toward the surface; H, LE, G, and DS are positive when directed away from thesurface.

Rnet = H + LE + G + DS (4.3)

Available energy is that portion of Rnet thatis neither stored nor conducted into the ground;it is the energy available for turbulent exchangewith the atmosphere as H and LE.

(4.4)

Ground heat flux, the heat transferred fromthe surface into and out of the soil, is negligibleover a day in most temperate and tropicalecosystems, because the heat conducted intothe soil during the day is balanced by heat conducted back to the surface at night. Themagnitude of ground heat flux depends on thethermal gradient between the soil surface anddeep soils and the thermal conductivity of soils,which is greatest in soils that are wet and havea high bulk density. In contrast to temperatesoils, permafrost regions of the arctic andboreal forest have substantial ground heat flux(10 to 20% of net radiation), due to the strongthermal gradient between the soil surface andthe permafrost. Clear lakes and the ocean,which transmit substantial radiation to depth,also exchange substantial energy as “ground”heat flux.

Latent heat flux is the energy transferred tothe atmosphere when water is transpired byplants or evaporates from leaf or soil surfaces.Latent heat flux, which is measured in energy

Available energy = netR G SH LE

- +( )= +

D

Figure 4.2. Radiation budget of a Douglas fir forestduring the summer. (Redrawn with permission fromMethuen; Oke 1987.)


units, is identical to evapotranspiration, whichis measured in water units. This heat is trans-ported from the surface into the atmosphere byconvection and is released to the atmospherewhen water vapor condenses to form clouddroplets, often at considerable distances from the point at which evaporation occurred.Dewfall represents a small latent heat flux fromthe atmosphere to the ecosystem at night underconditions of high relative humidity and coldleaf or soil surfaces. Note that latent heat fluxis the process that transfers water from theecosystem to the atmosphere. Equation 4.3therefore also relates ecosystem water loss tothe energy budget that drives evapotran-spiration. It is because of the conservation of energy and mass that the energy and watercycles intersect.

Sensible heat flux is the heat that is initiallytransferred to the near-surface atmosphere byconduction and to the bulk atmosphere by con-vection; it is controlled in part by the tempera-ture differential between the surface and theoverlying air. Air close to the surface becomeswarmer and more buoyant than the air imme-diately above it, causing this parcel of air to rise,the process of convective turbulence. Mech-anical turbulence is caused by winds blowingacross a rough surface; it generates eddies thattransport warm moist air away from the surfaceand bring cooler drier air from the bulk atmos-phere back toward the surface. Surface turbu-lence is the major process that transfers latentand sensible heat between the surface and theatmosphere (see Chapter 2).

There are important interactions betweenlatent and sensible heat fluxes from ecosystems.The consumption of heat by the evaporation ofwater cools the surface, thereby reducing thetemperature differential between the surfaceand the air that drives sensible heat flux. Con-versely, the warming of surface air by sensibleheat flux increases the quantity of water vaporthat the air can hold and causes convectivemovement of moist air away from the evapo-rating surfaces. Both of these processes increasethe vapor pressure gradient that drives evapo-ration. Because of these interdependencies,surface moisture has a strong effect on theBowen ratio—that is, the ratio of sensible tolatent heat flux.

Bowen ratios range from less than 0.1 fortropical oceans to great than 10 for deserts,indicating that either latent heat flux or sensi-ble heat flux can dominate the turbulent energytransfer from ecosystems to the atmosphere,depending on the nature of the ecosystem andthe climate. In general, ecosystems with abun-dant moisture have higher rates of evapotran-spiration and therefore lower Bowen ratiosthan do dry ecosystems. Similarly, ecosystemsdominated by rapidly growing plants, whichhave high transpiration rates (see Chapter 5),have proportionately lower sensible heat fluxesand low Bowen ratios (Table 4.3). Strong windsand/or rough canopies, which generate surfaceturbulence, tend to prevent a temperaturebuildup at the surface and therefore reducesensible heat flux and Bowen ratio. For thesereasons, energy partitioning varies substantiallyboth seasonally and among ecosystems. TheBowen ratio determines the strength of the linkage between the energy budget and thehydrologic cycle, because it is inversely relatedto the proportion of net radiation that driveswater loss from ecosystems: the lower theBowen ratio, the tighter the linkage betweenthe energy budget and the hydrologic cycle.

The spatial patterning of ecosystems influ-ences energy partitioning (see Chapter 14).Heating contrasts between adjacent ecosystemscreate convective turbulence; this turbulence is therefore much greater at boundariesbetween ecosystems with contrasting energybudgets than in the centers of ecosystems. Mostevaporation from large lakes, for example,

Table 4.3. Representative Bowen ratios for differ-ent vegetation types.

Surface type Bowen ratio

Desert >10Semiarid landscape 2–6Arctic tundra 0.3–2.0Temperate forest and grassland 0.4–0.8Boreal forest 0.5–1.5Forest, wet canopy -0.7–0.4Water-stressed crops 1.0–1.6Irrigated crops -0.5–0.5Tropical rain forest 0.1–0.3Tropical ocean <0.1

Data from Jarvis (1976), Oke (1987), and Eugster et al.(2000).

Water Inputs to Ecosystems 77

occurs near their edges, rather than in thecenter, where the overlying air is so stable that it saturates rapidly and supports a rela-tively low evaporation rate. A swamp withinterspersed patches of vegetation would have greater convective turbulence and overall evaporation than homogeneous water or moist vegetation. When ecosystem patches that differstrongly in energy partitioning are larger indiameter than the depth of the planetaryboundary layer (greater than about 10km),they can modify mesoscale atmospheric circu-lations and cloud and precipitation patterns(Pielke and Avisar 1990, Weaver and Avissar2001).

Seasonal Energy Exchange

Over sufficiently long time scales, energyoutputs are tightly coupled to inputs becausethere is little energy storage at Earth’s surface.Most ecosystems have a limited capacity forlong-term energy storage in vegetation andsurface soils. Consequently, energy losses to theatmosphere closely track solar inputs on both adaily and a seasonal basis, although the form inwhich this energy is lost (sensible vs. latent heatflux) varies, depending on moisture availability.Ground heat fluxes are usually negligible whenaveraged over 24h (Oke 1987). Importantexceptions to this generalization are waterbodies such as lakes and oceans, in which thesolar inputs often penetrate to depth, result-ing in substantial warming of the water, andsome high-latitude regions. Water bodies oftenabsorb substantial energy in spring and earlysummer, when the solar angle is greatest,causing the water to warm. There is a netrelease of energy in the autumn, moderatingthe temperatures of adjacent terrestrial sur-faces (see Chapter 2). This seasonal pattern ofenergy exchange drives the annual or semian-nual turnover of lakes (see Chapter 10). Per-mafrost contributes to a seasonal imbalance inenergy absorption and release in cold climates.In the arctic, for example, 10 to 20% of theenergy absorbed during summer is consumedby thawing of frozen soil. This energy isreleased back to the atmosphere the nextwinter, when the soil refreezes (Chapin et al.2000a).

Snow-covered surfaces experience thresh-old changes in energy exchange at the time of snowmelt (Liston 1999). The high albedo of snow-covered surfaces minimizes energyabsorption until snowmelt occurs, at which time there is a dramatic increase in the energyabsorbed by the surface and transferred to the atmosphere. This may result in abruptincreases in regional air temperature aftersnowmelt. Leaf out also alters energy exchangeby both changing albedo and increasing evapo-transpiration at the expense of sensible heatflux.

Water Inputs to Ecosystems

Precipitation is the major water input to mostterrestrial ecosystems. Global and regional con-trols over precipitation therefore determine thequantity and seasonality of water inputs to mostecosystems (see Chapter 2). In ecosystems thatreceive some precipitation as snow, however,the water contained in the snowpack does notenter the soil until snow melt, often monthsafter the precipitation occurs. This causes theseasonality of water input to soils to differ fromthat of precipitation.

Vegetation in some ecosystems, particularlyin riparian zones, accesses additional ground-water that flows laterally through the eco-system. Desert communities of phreatophytes(deep-rooted plants that tap groundwater),for example, may absorb sufficient ground-water that the ecosystem loses more water in transpiration than it receives in precipi-tation. Lakes and streams also receive most of their water inputs from groundwater orrunoff that drains from adjacent terrestrialecosystems. Water inputs to freshwater eco-systems are therefore only indirectly linked toprecipitation.

In ecosystems with frequent fog, canopyinterception of fog increases the water inputs toecosystems, when cloud droplets that might nototherwise precipitate are deposited on leaf sur-faces and are absorbed by leaves or drip fromthe canopy to the soil. The coastal redwoodtrees of California, for example, depend on fog-derived water inputs during summer, whenprecipitation is low, but fog occurs frequently.


Similarly, in areas that are climatically marginalfor Australian rain forests, the capture of fogand mist by trees can augment rainfall by 40% (Hutley et al. 1997). In most ecosystems,however, most of the precipitation that is inter-cepted by the canopy evaporates before itreaches the soil, so canopy interception gener-ally reduces the effectiveness of precipitationthat enters the ecosystem rather than increas-ing water inputs.

Water Movements Within Ecosystems

Basic Principles of Water Movement

The soil behaves like a bucket that is filled byprecipitation and emptied by evapotranspira-tion and runoff. The soil is the major waterstorage reservoir of ecosystems. When waterinputs from precipitation exceed the capacity of the soil to hold water, the excess drains to groundwater or runs off over the groundsurface (Fig. 4.3), just as water added to a fullbucket spills over the edges instead of beingretained in the bucket. The water losses fromthe ecosystem move laterally to other ecosys-

tems such as streams and lakes. Evaporationfrom the soil surface and transpiration byplants are the other major avenues of water lossfrom the soil reservoir. These processes con-tinue only as long as the soil contains water thatplants can tap, just as evaporation from abucket continues only as long as the bucketcontains water.

Water is stored in soil primarily in poresbetween soil particles, so the water-holdingcapacity of a soil depends on its total porevolume. Pore volume, in turn, depends on soildepth and the proportion of the soil volumeoccupied by pores, the spaces between soil particles. Shallow soils on ridgetops hold lesswater than deep valley-bottom soils; rocky orsandy soils, in which soil solids occupy much of the soil volume, hold less water than fine-textured soils; and soils compacted by intensivegrazing or farm machinery hold less water than noncompacted soils (see Chapter 3).As described later, water is held most effec-tively by small soil pores that have a largesurface-to-volume ratio.

Water moves along a gradient from high tolow potential energy. The energy status of waterdepends on its concentration and various pres-sures. The pressures in natural systems can be

Infiltration Absorption

Groundwater

Soil water

Throughfall

InterceptionPrecipitation

Transpiration

Evaporation

Stemflow

Surface runoffSnow

Percolation

Base flow

Evaporation

Figure 4.3. Water balance ofan ecosystem (Waring andRunning 1998).

Water Movements Within Ecosystems 79

described in terms of either hydrostatic pres-sures or matric forces (Passioura 1988). Themajor hydrostatic pressures in natural systemsare (1) gravitational pressure, which dependson height and (2) pressures that are generatedby physiological processes in organisms. Thegravitational pressure is higher at the top of a tree than in the roots, so plants must movewater to leaves against this gravitational force.Plants can generate either positive pressures,(e.g., the turgor pressure that maintains therigidity of plant cells) or negative pressures(which move water from the roots to leaves).Matric forces result from the adsorption ofwater to the surfaces of cells or soil particles.The thinner the water film, the more tightly thewater molecules are held to surfaces by matricforces.

In most cases water moves in response tosome combination of these forces. We can con-sider all these forces simultaneously by express-ing them in units of water potential—that is, thepotential energy of water relative to pure waterat the soil surface.The total water potential (yt)is the sum of the individual potentials.

Yt = Yp + Yo + Ym (4.5)

The pressure potential (yp) is generated bygravitational forces and physiological processesof organisms; the osmotic potential (yo) reflectspresence of substances dissolved in water; andthe matric potential (ym) is caused by adsorp-tion of water to surfaces. In some treatments,matric potential is considered a component of pressure potential (Passioura 1988). By convention, the water potential of pure waterunder no pressure at the soil surface is given avalue of zero. Water potentials are positive ifthey have a higher potential energy than thisreference and negative if they have a lowerpotential energy.

Water Movement from the Canopy to the Soil

In closed-canopy forests, the canopy interceptsa substantial proportion of incoming precipita-tion (Fig. 4.3). Intercepted precipitation can beevaporated directly back to the atmosphere,absorbed by the leaves, drip to the ground

(throughfall), or run down stems to the ground(stem flow). Interception is the fraction of pre-cipitation that does not reach the ground. Itcommonly ranges from 10 to 50% for closed-canopy ecosystems (Waring and Running1998). After light rain or snowfalls, a substan-tial proportion of the precipitation may eva-porate and return directly to the atmospherewithout entering the soil. For this reason,canopy heterogeneity generates heterogeneityin water input to soil and therefore soil mois-ture availability. Throughfall is the process thatdelivers most of the water from the canopy tothe soil.

The capacity of the canopy to intercept and store water differs among ecosystems. Itdepends primarily on canopy surface area, par-ticularly the surface area of leaves (Fig. 4.4).Forests, for example, frequently store 0.8,0.3, and 0.25mm of precipitation on leaves,branches, and stems, respectively. Conifer for-ests typically store about 15% of precipitation,whereas deciduous forests store 5 to 10%(Waring and Running 1998). Epiphytes, whichare rooted in the canopy, depend completely oncanopy interception for their water supply andincrease canopy interception. Factors such asstand age and epiphyte load influence canopyinterception through their effects on canopysurface area.

E. p

auci

flora

Pinus radiata

E. maculata

Inte

rcep

tion

stor

age

(kg)

Leaf area (m2)

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

01 2 3 4 5 6 7 8 9 10 11

Figure 4.4. Interception storage capacity of Euca-lyptus spp. and Pinus radiata with different leaf areas. (Redrawn with permission from Journal ofHydrology, Vol. 42 © 1979 Elsevier Science; Aston1979.)


The bark texture and architecture of stemsand trunks influence the amount and directionof stem flow.Trees and shrubs with smooth barkhave greater stem flow (about 12% of precipi-tation) than do rough-barked plants such asconifers (about 2% of precipitation) (Waringand Running 1998). In the Eucalyptus mallee insouthwestern Australia as much as 25% of theincoming precipitation runs down stems, due tothe parachute architecture of these shrubs. Thestem flow then penetrates to depth in the soilprofile through channels at the soil–root inter-face (Nulsen et al. 1986).

Water Movement Within the Soil

Pressure gradients associated with gravity andmatric forces control most water movementthrough soils. The rate of water flow throughthe soil (Js) depends on the driving force (thegradient in water potential) and the resistanceto water movement. This resistance, in turn,depends on the hydraulic conductivity (Ls) ofthe soil, and the path length (l) of the columnthrough which the water travels.

(4.6)

This simple relationship describes most of the patterns of water movement through soils,including the infiltration of rainwater or snowmelt into the soil and the movement of waterfrom the soil to plant roots. Soils differ strik-ingly in hydraulic conductivity due to differ-ences in soil texture and aggregate structure(see Chapter 3). For this reason water movesmuch more readily through sandy soils thanthrough clay soils or compacted soils. The rateof water flow in saturated soils, for example,differs by three orders of magnitude betweenfine and coarse soils (less than 0.25 to greaterthan 250mmh-1).

Infiltration of rainwater into the soil dependsnot only on hydraulic conductivity but also onpreferential flow through macropores createdby cracks in the soil or channels produced byplant roots and soil animals (Dingman 2001).Variation in flow paths in the surface few millimeters of soil can have large effects on

J Lls s

t=DY

infiltration. Impaction by raindrops on anunprotected mineral soil, for example, canreduce hydraulic conductivity dramatically. Forthis reason the presence of a surface moss or litter layer, which prevents impaction by raindrops, is one of the most important factorsdetermining whether water enters the soil or flows over the surface. Any time that precipitation rate exceeds the infiltration rate,water accumulates on the surface and overlandflow may occur, leading to erosional loss of soil.

Some soils have horizons of low hydraulicconductivity that prevent water percolation todepth. For example, calcic (caliche) layers indeserts, permafrost in arctic and boreal ecosys-tems, and hardpans in highly weathered soilsare horizons with such low hydraulic conduc-tivity that the water table remains close to the surface, rather than moving into a deepgroundwater pool (see Chapter 3).

Once water enters the soil, it moves down-ward under the force of gravity until the matricforces, which account for the adsorption ofwater to soil particles, exceed the gravitationalpotential. Water that is not retained by matricforces drains through the soil to groundwater.The field capacity of a soil is the quantity ofwater retained by a saturated soil after gra-vitational water has drained. The large surfacearea per unit soil volume in fine-textured soilsexplains their high field capacity.A clay soil, forexample, with its high proportion of small particles (Table 4.4), holds four times morewater than a sandy soil. Organic matter alsoenhances the field capacity of soils, because ofits hydrophilic nature and its effects on soilstructure. For this reason the soils beneathshrubs in deserts, which have higher organic

Table 4.4. Typical pore size distribution of differentsoil types.

Pore space (% of soil volume)

Particle size (mm) Sand Loam Clay

>30 75 18 60.2–30 22 48 40<0.2 3 34 53


content, retain more water than do soils thatreceive less litter input.

At field capacity, the water potential of a soilis about -0.03MPa—that is, close to the waterpotential of pure water (0.00MPa). As a soildries, the films of soil water become thinner,and the remaining water is held more tightly toparticle surfaces. The permanent wilting pointis the soil water potential (about -1.5MPa) at which most mesic plants wilt because theycannot obtain water from soils. Many drought-adapted plants, however, can obtain water from soils at water potentials as low as -3.0 to-8.0MPa (Larcher 1995). A second conse-quence of thin water films in dry soils is thatwater cannot move directly across water-filledsoil pores but must move around the edges ofthe air-filled pores along a much longer, moretortuous path. For this reason, the hydraulicconductivity of soil declines dramatically as thesoil dries. The difference in the water contentbetween field capacity and permanent wiltingpoint (water-holding capacity) provides an esti-mate of the plant-available water, althoughsome of this water is held in such small poresthat it moves slowly to roots (Fig. 4.5). Vegeta-tion often extracts 65 to 75% of the plant-avail-able water before there are signs of water stress(Waring and Running 1998). The total quantityof water available to vegetation is the availablewater content per unit soil volume times thevolume exploited by roots.

Water Movement from Soil to Roots

Water moves from the soil to the roots of transpiring plants by flowing from high to lowwater potential. Water moves from the soil intothe root whenever the root has a lower waterpotential than the surrounding soil. Movementof water into the root along a water poten-tial gradient causes the water film on adjacentsoil particles to become thinner. The localizedreduction in water potential near the rootcauses water to move along soil films towardthe root. In this way a root can access mostavailable water within a radius of about 6mm. As the soil dries, hydraulic conductivitydeclines, and the root accesses water lessrapidly. In saline soils, the osmotic potential of the soil solution reduces total soil waterpotential, so roots with a given water potentialcan absorb less water than from nonsaline soils.

A continuous pathway for water move-ment from the soil to the root is provided byroot hairs and mycorrhizal hyphae that extend into the soil and by carbohydrates secreted bythe root that maximize contact between the root and the soil. When this root–soilcontact is interrupted by the shrinking of dryingsoil or the consumption of root cortical cells by soil animals, the root can no longer absorbwater.

Rooting depth reflects a compromisebetween water and nutrient availability. Mostplant roots are in the upper soil horizons wherenutrient inputs are greatest and where nutrientsare generally most available (see Chapter 8). Ina given ecosystem, short-lived herbs are gener-ally more shallow rooted than long-lived shrubsand trees and depend more on surface moisture(Fig. 4.6). In arid ecosystems surface evapora-tion and transpiration dry out the surface soils.For this reason, deserts, arid shrublands, andtropical savannas have many species with deeproots (Fig. 4.7). Phreatophytes are an extremeexample of deep-rooted plants. These desertplants produce roots that extend to the watertable, often a depth of tens of meters. Theseplants have no physiological adaptations todrought and have high transpiration rates. Evenwet ecosystems such as tropical rain forests

40

30

20

10

0Sand Sandy

loamLoam Silt

loamClayloam

Clay

Unavailable water

Permanentwilting percentage

Soi

l wat

er (

% o

f soi

l vol

ume)

Total waterIn situfield

capacity Available water

Figure 4.5. Plant-available water as a function of soil texture. (Redrawn with permission from Academic Press; Kramer and Boyer 1995.)


have dry seasons. This may explain the occur-rence in such forests of deep-rooted trees that tap water from depths of more than 8m(Nepstad et al. 1994). Deep-rooted plants maybe more common than generally appreciated.Deep roots often extend into cracks in bedrock,where they tap water as it drains through rockchannels to groundwater.

Rooting depth has important ecosystem consequences because it determines the soilvolume that can be exploited by vegetation (seeChapter 12). California grassland soils below a1-m depth, for example, remain moist even atthe end of the summer drought, whereas anadjacent chaparral shrub community uses waterto a depth of 2m (Fig. 4.8). This greater rootingdepth contributes to the longer growing seasonand greater productivity of the chaparral. Evenin the chaparral, species differences in rootingdepth (Box 4.2) lead to differences in moisturesupply and drought stress.

0

0 0.2 0.4 0.6 0.8 1

50

100

150

200

Grasses

Shrubs

Trees

Soi

l dep

th (c

m)

Cumulative root biomass (fraction of total)

Figure 4.7. Maximum rooting depths of selectedspecies in the major biome types of the worldshowing that species in each biome differ widely.

Figure 4.6. The cumulative fraction of roots foundat different soil depths for three plant growth formsaveraged over all biomes. (Redrawn with permissionfrom Oecologia; Jackson et al. 1996.)

Woody species in dry environments are often deeplyrooted. (Redrawn with permission from Oecologia;Canadell et al. 1996.)


Water Movement Through Plants

The vapor-pressure gradient from the leafsurface to the atmosphere is the driving forcefor water movement through plants. Watertransport from the soil through the plant to the atmosphere takes place in a soil-plant-atmosphere continuum that is interconnectedby a continuous film of liquid water. Watermoves from the soil through the plant to theatmosphere along a gradient in water potential.The low partial pressure of water vapor in airrelative to that inside the leaves is the major

driving force for water loss from leaves, whichin turn drives water transport along a pressuregradient from the roots to the leaves, which inturn drives water movement from the soil intothe plant. The rate of water movement throughthe plant (Jp) (Eq. 4.7) is determined by thewater-potential gradient (the driving force;DYt) and the resistance to water movement, justas described for water movement through soils(Eq. 4.6). As in soils, the resistance to watermovement through the plant depends onhydraulic conductivity (Lp) and path length (l).

(4.7)

The movement of water into and through theplant is driven entirely by the physical processof evaporation from the leaf surface and in-volves no expenditure of metabolic energy bythe plant. This contrasts with the acquisition ofcarbon and nutrients for which the plant mustexpend considerable metabolic energy (seeChapters 5 and 8).

Roots

Water moves through roots along a water-potential gradient from moist soils to theatmosphere during the day and sometimes todry surface soils at night. In moist soils, the cellmembranes, which are composed of hydropho-bic lipids, provide the greatest resistance towater movement through roots (see Fig. 8.3).This membrane resistance to water flow isgreatest under conditions of low root tempera-ture or flooding, so plants that are not adaptedto these conditions experience substantialwater stress in cold or saturated soils. In drysoils the contact between the root and the soil accounts for the greatest resistance to water flux through the root. Plants overcomethis resistance primarily by increasing allo-cation to the production of new roots (seeChapter 6).

In dry environments, there is a strong verti-cal gradient in soil water potential due to thelow soil water potential in dry surface soils.However, water moves slowly through the soilbecause of the low hydraulic conductivity of drysoils. During the day, when plants lose water

J Llp p

t=DY

ψplant

ψplant

Soil water potential , ψsoil (MPa)

Quercus -0.2

Adenostoma -1.4Heteromeles -2.1

Rhamnus -2.8

Chaparral

Sitanion < -4.0

Grassland0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

-5.0 -4.0 -3.0 -2..0 -1.0 0

0

Dep

th (m

)

Figure 4.8. Soil water profiles in adjacent shrub andgrassland communities at the end of the summerdrought period. Predawn water potentials are a goodindex of the soil moisture and the degree of droughtstress experienced by the plant. (Redrawn with per-mission from Oecologia; Davis and Mooney 1986.)


Ocean Continent

Snow

Snow

Evaporation

-25 Rain

-40 Vapor -80 Vapor -110 Vapor

-45 Rain

Figure 4.9. The effect of evaporation and subsequent condensation during rainfall on the ratio ofhydrogen isotopes. (Redrawn with permission from Academic Press; Dawson 1993.)

The source of water used by plants can bedetermined from the isotopic composition ofplant water.The ratio of the concentration ofdeuterium (D) to hydrogen (H) provides auseful signature of different water sources.Evaporation discriminates against theheavier isotope (deuterium), causing the iso-topic ratio of D :H to decline (become morenegative) relative to the water source thatgave rise to evaporation (Fig. 4.9). Conden-sation, on the other hand, raises the D:H ratio, causing rainfall to have a less negative

hydrogen isotopic ratio than its parent airmass. The D:H ratio of water vapor remain-ing in the atmosphere therefore declineswith sequential rainfall events. There is alsoa linear relationship between air tempera-ture at the time of precipitation and thehydrogen isotope ratio, so summer precipi-tation has a higher D:H ratio than winterprecipitation. These phenomena generate acharacteristic signature that identifies differ-ent sources of water that plants use (Fig.4.10). The isotopic ratios of xylem water, for

Box 4.2. Isotopic Signatures of Water Sources

by transpiration, plant water potential is lowerthan soil water potential, so water moves fromthe soil into the plant, particularly from deepsoils where water is most available (highest soilwater potential) (Fig. 4.11). At night, whentranspiration ceases, plant water potential equi-librates with the water potential of deep soils. When surface soils are drier than those atdepth, the water potential gradient is from deepsoils through roots to shallow soils. Becauseroots have much higher hydraulic conductivitythan soils, this gradient in water potential driveshydraulic lift, the vertical movement of waterfrom deep to shallow soils through roots alonga water potential gradient (Caldwell et al.1998). Hydraulic lift occurs in most arid ecosys-tems and in many moist forests. Sugar maple

trees, for example, acquire all their moisturefrom deep roots, but 3 to 60% of the water usedby shallow-rooted herbs in these forests comesfrom water that has been hydraulically lifted bythe maple trees (Dawson 1993). In the GreatBasin deserts of western North America, 20 to 50% of the water used by shallow-rootedgrasses comes from water that is hydraulicallylifted by deep-rooted sagebrush shrubs. Thewater provided by hydraulic lift stimulatesdecomposition and mineralization in dryshallow soils, augmenting the supplies of bothwater and nutrients to shallow-rooted species.Because deep-rooted plants both provide water to and remove water and nutrients from shallow soils, hydraulic lift complicates theinterpretation of species interactions in many


Soil water andrunoff

Xylem water

Pre

cipi

tatio

n

Sum

mer

Win

ter

Fossilwater

?

Fog

Groundwater

Figure 4.10. Isotopic signa-ture of water from varioussources. By sampling thewater in the xylem of plants,one can determine the main water supply used by a plant. (Redrawn with per-mission from AcademicPress; Dawson 1993.)

Day ψday

ψnight

-30Air-90

-4.2 Leaves -1.2

-4.0 -4.0Surface

soil

-4.1 Roots -1.2

-1.0Deepsoil -1.0

Night

Hyd

raul

ic li

ft

Transpiration

Figure 4.11. Patterns of soil water potential andwater movement in arid environments during theday and at night. During the day, water moves fromsoils (especially deep soils) to the atmosphere in

example, show that some plants derive mostof their water from fog, whereas others usesoil water or ground water. These measure-ments also differentiate water derived fromwinter vs. summer precipitation (Dawson1993). Similarly, D:H ratios of stream water

identify the relative contributions of soilwater from recent precipitation events vs.ground water. Oxygen isotope ratios inwater show patterns of variation similar tothose of hydrogen.

response to the strong water potential gradient fromthe plant to the atmosphere. At night, water movesfrom wet soils at depth to dry surface soils throughthe root system, the process of hydraulic lift.


ecosystems. When surface soils are wetter thandeep soils after rain, roots provide an avenue torecharge deep soils, perhaps explaining howdeep-rooted desert plants grow through drysoils to the water table (Burgess et al. 1998).Thus roots provide an avenue for rapid watertransport from soil of high to low water poten-tial, regardless of the vertical direction of thewater-potential gradient.

Stems

Water moves through stems to replace waterlost by transpiring leaves. The water-conductingtissues in the xylem are narrow capillaries ofdead cells that extend from the roots to theleaves. Water is “sucked up” through these cap-illary tubes in response to the water-potentialgradient created by transpirational water loss. The cohesion of water molecules to oneanother and their adhesion to the walls of the narrow capillary tubes allow these watercolumns to be raised under tension (a negativewater potential) as much as 100m in tall trees.

There is a tradeoff between hydraulic con-ductivity of stems and their risk of cavitation—that is, the breakage of water columns undertension. Hydraulic conductivity of stems varieswith the fourth power of capillary diameter, soa small increase in vessel diameter greatlyincreases hydraulic conductivity. For example,vines, which have relatively small stems andrely on other plants for their physical support,have large-diameter xylem vessels. This allowsrapid water transport through narrow stems butincreases the risk of cavitation and may explainwhy vines are most common in moist environ-ments such as tropical forests. The stems oftropical vines, for example have hydraulic con-ductivities and velocities of sap flow that are 50-to 100-fold higher than those of conifers(Larcher 1995). Broad-leaved deciduous treesare intermediate. Many plants in moist envi-ronments, particularly herbaceous plants,function close to the water potential where cavitation occurs, suggesting that they investjust enough in water transport tissues to allowwater transport for the growing season (Sperry1995). Plants from dry environments produce

stems with a larger safety factor—that is, stemsthat resist cavitation at much lower waterpotential than the plants commonly experience(Fig. 4.12). Small roots are generally more vul-nerable to cavitation than are stems and mayfunction as a “hydraulic fuse” that localizesfailure in relatively cheap and replaceable parts of the plant (Jackson et al. 2000).

Plants in cold environments suffer cavitationfrom freezing. Trees adapted to these cold environments typically produce abundantsmall-diameter vessels that can, in somespecies, refill after cavitation (diffuse-porousspecies). In contrast, many trees in warm envi-ronments produce both small- and large-diam-eter vessels that cannot be refilled aftercavitation and therefore function for only asingle growing season (ring-porous species).

The water transported by a stem depends onboth the hydraulic conductivity of individualconducting elements and the total quantity of

0

0

-2

-4

-6

-8

-10-18 -15 -12 -9 -6 -3

1:1

Min

imum

nat

ive

ψpl

ant(

MP

a)

ψplant at 100% conductance loss

Figure 4.12. The relationship between the waterpotential at which a plant loses all xylem conductiv-ity due to embolism and the minimum water poten-tial observed in nature. Each datum point representsa different species. The 1 : 1 line is the line expectedif there were no safety factor—that is, if each specieslost all conductivity at the lowest water potentialobserved in nature. Species that naturally experiencelow water potentials exhibit a greater margin ofsafety (i.e., a greater departure from the 1 : 1 line).(Redrawn with permission from Academic Press;Sperry 1995.)


conducting tissue (the sapwood). There is astrong linear relationship between the cross-sectional area of sapwood and the leaf area sup-ported by a tree (Fig. 4.13). However, the slopeof this relationship varies strikingly amongspecies and environments. Drought-resistantspecies generally have less leaf area per unit ofsapwood than do drought-sensitive speciesbecause of the small vessel diameter ofdrought-resistant species. The ratio of leaf areato sapwood area, for example, is generally more

than twice as great in trees from mesic envi-ronments as in trees from dry environments,due to the smaller-diameter conducting ele-ments in dry environments (Margolis et al.1995). Any factor that enhances the productiv-ity of a tree increases its ratio of leaf area tosapwood area.This ratio increases, for example,with improvements in nutrient or moisturestatus and is greater in dominant than subdom-inant individuals of a stand.

Water storage in stems buffers the plant fromimbalances in water supply and demand. Thewater content of trunks of trees generallydecreases during the day, causing water uptakeby roots to lag behind transpirational water lossby about 2h (Fig. 4.14). The quantity of waterstored in sapwood is substantial, equivalent toas much as 5 to 10 days of transpiration. Thissapwood water, however, exchanges relativelyslowly, so stores of water in sapwood seldomaccount for more than 10% of transpiration.In dry tropical forests, where trees lose theirleaves during the dry season, this stored wateris critical to support flowering during the dryseason. Trees in these forests that have low-density wood and large stem water storage canflower during the dry season, whereas treeswith high-density wood and low stem waterstorage can flower only during the wet season(Borchert 1994). Water storage in desert succu-lents may allow transpiration to continue for

500

1000 2000 3000

400

300

200

100

00

Sapwood area (cm2)

Leaf

are

a (m

2 ) Douglas fir

Ponderosa pine

Western white pine

Figure 4.13. Leaf area versus sapwood cross-sectional area for three forest trees. Ponderosa pine,which typically occupies dry sites, has smaller vesselsand therefore supports less leaf area per unitsapwood than does Douglas fir from moist sites.(Redrawn with permission from Canadian Journal ofForest Research; Monserud and Marshall 1999.)

14 16 18 20 22

Time (h)

0

30

20

10

06 10 12 24

0

5

10

Tran

spira

tion,

(m

L m

-2 h

-1) Transpiration

Wat

er u

ptak

e (L

h-1

)

2 4 8

Water uptake

Figure 4.14. Diurnal time course of water uptakeand water loss by Siberian larch. During morning,transpiration is supported by water loss from stems,creating a lower water potential in stems and roots,

which generates the water potential gradient toabsorb water from the soil.The water stored in stemsis replenished at night. (Redrawn with permissionfrom BioScience; Schulze et al. 1987.)


several weeks after water uptake from the soilhas ceased.

Leaves

Water loss from leaves is controlled by theevaporative potential of the air, the watersupply from the soil, and the regulation ofwater loss by leaves. Soil water supply and theevaporative potential of the air are the majorenvironmental controls over water loss fromleaves (Fig. 4.15). Stomata are pores in the leafsurface that can be opened or closed to regu-late the entry of CO2 into leaves and the loss ofwater from leaves. Stomata are the valves thatdetermine the resistance to water movementbetween the soil and the air. The low hydraulicconductivity of dry soils minimizes the amountof water that can move directly from dry soil tothe air by surface evaporation. The extensiveroot systems of plants and the high hydraulicconductivity of plant xylem make plants aneffective conduit for moving water from the soilto the atmosphere. Plants adjust the size ofstomatal openings to regulate the loss of waterfrom leaves. Because stomatal conductance

also determines the rate of CO2 entry intoleaves, there is an inevitable tradeoff betweencarbon gain and water loss by leaves (seeChapter 5).

Diurnal and climatic differences in air tem-perature and humidity determine the drivingforce for transpiration. Air inside the leaf isalways saturated with water vapor because it isadjacent to moist cell surfaces. On a sunny day,air temperature rises to a maximum shortlyafter midday, allowing the air to hold morewater. This rise in air temperature and the radiation absorbed by the leaf increases the temperature of the leaf and therefore the water vapor concentration inside the leaf. Thewater vapor concentration of the external air generally increases less than that inside the leaf. The resulting increase in the gradientin water vapor concentration between theinside and the outside of the leaf increases thetranspirational water loss from the leaf. Inevening the temperature decreases, causing adecline in the water vapor concentration insidethe leaf and a decline in transpiration. Varia-tions in weather or climate that cause anincrease in air temperature and/or a decrease

STATE FACTORS

Interactivecontrols

Directcontrols

LONG-TERMCONTROLS

SHORT-TERMCONTROLS

BIOTA

PARENTMATERIAL

CLIMATE

Plantfunctional

types

Soilresources

EVAPO-TRANSPIRATION

Indirectcontrols

TIME

Net radiation/ VPD

Water-holdingcapacity

Boundary-layerconductance

Precipitation

Surfaceroughness

Photosyntheticcapacity

Wateravailability

Stomatalconductance

Figure 4.15. The major factors controlling evapo-transpiration from a plant canopy. The major short-term controls over evapotranspiration include wateravailability, net radiation, and stomatal conductance

(mainly under dry conditions) and boundary layerconductance (mainly under moist conditions), whichin turn depend on climate, parent material, soilresources and disturbance.

Water Losses from Ecosystems 89

in atmospheric moisture content also enhancethe driving force for transpirational water loss. The evaporative potential of desert air istherefore extremely high because it is both hot and dry. Cloud forests generally have lowevaporative potential because the air is satu-rated and because clouds reduce radiationinput. Cold climates have low evaporativepotential because cold air holds relatively littlewater vapor.

Stomatal conductance is the major controlthat plants exert over water loss from a leaf.Some plants reduce stomatal conductancewhen leaves are exposed to warm dry air thatwould otherwise cause high transpirationalwater loss. This response of stomatal con-ductance to dry air makes water loss fromleaves less sensitive to dry atmospheric con-ditions than we might expect. Species differconsiderably in their sensitivity of stomatalconductance to the evaporative potential of theair. Both the mechanism and the ecological patterns in the sensitivity of stomatal conduc-tance to atmospheric humidity are poorlyunderstood.

Stomatal conductance declines in response todrought because plants sense the soil moisturecontent of their root systems. Roots exposed tolow soil moisture produce abscisic acid (ABA),a hormone that is transported to leaves andcauses a reduction in stomatal conductance.Plants from mesic habitats are particularly sensitive to low soil water potential, closing stomates in response to soil drying before theyexperience large changes in plant water poten-tial. Plants from dry environments show lessresponse of stomatal conductance to soil dryingand therefore continue to absorb and losewater, as the soil dries. Drought-adapted plantsin arid ecosystems therefore maintain greaterphysiological activity in dry soils than do plantsadapted to moist habitats; and, in the process,they transfer more water to the atmosphereunder dry conditions.

There are important differences amongspecies in stomatal conductance. Stomatal con-ductance under favorable conditions is highestin rapidly growing plants adapted to moistfertile soils (Körner et al. 1979, Schulze et al.1994) (see Chapter 5).

Water Losses from Ecosystems

Water inputs are the major determinant ofwater outputs from ecosystems. The water lossfrom ecosystems equals the inputs in precipita-tion (P) adjusted for any changes in waterstorage (DS). The major avenues of loss areevapotranspiration (E) and runoff (R).

P ± DS = E + R (4.8)

Just as in the case of carbon and energy, thechanges in water storage are generally smallrelative to inputs and outputs, when averagedover long time periods.

P ª E + R (4.9)

where DS is small (i.e., input ª output).Consequently, the quantity of water entering

the ecosystem largely determines water output,just as gross primary production (GPP; carboninput) is the major determinant of ecosystemrespiration (carbon output) (see Chapters 6and 7). The route by which water leaves anecosystem depends on the partitioning between evapotranspiration and runoff. Thispartitioning has a critical effect on regionalhydrologic cycles because water that returns to the atmosphere is available to support precipitation in the same or other ecosystems.Runoff supplies the water input to aquaticecosystems and provides most of the water usedby humans. In a sense, runoff is the “leftovers”of water that entered in precipitation and was not transferred to the atmosphere by evapotranspiration. In summary, controls over evapotranspiration largely determine thepartitioning between evapotranspiration andrunoff.

Evaporation from Wet Canopies

Evaporation of water intercepted by thecanopy is most important in ecosystems with ahigh surface roughness. Forests have high ratesof evaporation from wet canopies, primarilybecause the effective mixing that occurs inrough forest canopies promotes rapid evapora-tion from each leaf (Kelliher and Jackson 2001).The large water-storage capacity of forestcanopies is less important in explaining the


quantity of water evaporated from wetcanopies. The evaporation rate from a wetcanopy depends primarily on the climatic con-ditions that drive evaporation (primarily vaporpressure deficit) and the degree to which envi-ronmental conditions in the canopy are coupledby turbulence to conditions in the atmosphere.Turbulence, in turn, is greatest in ecosystemswith a tall aerodynamically rough canopy. Inforests, which are tightly coupled to atmos-pheric conditions, wet canopy evaporation islargely independent of net radiation and issimilar during the day and night. In grasslands,which are less tightly coupled to the atmos-phere, wet canopy evaporation depends on netradiation as well as vapor pressure deficit andis greater during the day than at night. Due todifferences in canopy roughness, forests havegreater wet canopy evaporation than do shrub-lands or grasslands, and conifer forests evapo-rate more water from wet canopies than dodeciduous forests.

Climate is the other factor that governs evap-oration from wet canopies. Climate determinesthe frequency with which the canopy interceptsprecipitation or dew and the conditions thatdrive evaporation. Ecosystems in wet climatesgenerally have greater canopy evaporationbecause of the more frequent capture of rain-fall by the canopy, even though the low vaporpressure deficit of wet climates causes thisevaporation to occur slowly. The frequency ofrainfall and dew formation is probably at leastas important as total precipitation in governingthe annual flux of wet canopy evaporation.Canopy evaporation increases exponentiallywith temperature because of the temperatureeffects on vapor pressure deficit (McNaughton1976), so ecosystems generally lose more intercepted water through canopy evaporationin warm than in cold climates. Despite thesegeneralizations, the interactions among multi-ple controls over wet canopy evaporation areso complex that they are best addressedthrough physically based models that considerall these factors simultaneously (Monteith andUnsworth 1990, Waring and Running 1998).

Canopies that intercept precipitation as snowor ice frequently store twice as much waterequivalent as when precipitation is received inliquid form. Snow interception and subsequent

sublimation (vaporization of a solid) is greatestin ecosystems with a high leaf area index (LAI),the quantity of leaf area per unit ground area.Most snow usually falls to the ground, wherelow net radiation and low wind speeds mini-mize sublimation. In tundra, however, wherethere is no canopy in winter to shade the snowand in continental boreal forests with low precipitation and low wind speeds, sublimationcan account for 30% and 50%, respectively,of winter precipitation (Liston and Sturm 1998,Pomeroy et al. 1999, Sturm et al. 2001).

Evapotranspiration from Dry Canopies

Soil moisture directly limits evapotranspirationrate in dry soils. Plant water potential and transpiration rate are surprisingly insensitive to water availability until plants have depletedabout 75% of the plant-available soil water(Fig. 4.16). Evapotranspiration is therefore relatively insensitive to precipitation in moistenvironments (Fig. 4.17). As soils dry, however,their hydraulic conductivity declines. Thiscreates a relatively abrupt threshold of soilmoisture, below which the rate of water supplyto roots declines and plants experience waterstress (lower water potential) (Fig. 4.16). Underthese circumstances, stomatal conductancedeclines below its physiological maximum,causing a decline in evapotranspiration, just asdescribed earlier for individual leaves. Evapo-transpiration rates are therefore generally lowin deserts, even though climatic conditionscould support a high evapotranspiration rate, ifmoisture were available.

When soil moisture is adequate, vegetationstructure and climate govern evapotranspira-tion rate. There are two conductances, bound-ary layer conductance and surface conductance,that govern ecosystem effects on evapotranspi-ration. The boundary layer conductance (alsotermed aerodynamic conductance) is a measureof the physical controls over water vapor trans-fer from the ecosystem to the atmosphere.It depends on wind speed and the size andnumber of roughness elements, such as trees.Boundary layer conductance is greatest whensurface turbulence mixes large quantities of airfrom the bulk atmosphere with air inside the


canopy. Turbulent mixing couples the evapora-tion at the leaf or soil surfaces with the mois-ture content of the bulk air of the atmosphere.Boundary layer conductance is thereforehigher in ecosystems such as forests with tall,aerodynamically rough canopies than in grass-lands or crops. Forests therefore transportwater more effectively to the atmosphere andreduce soil moisture more rapidly than dograsslands.

Surface conductance is a measure of thepotential of leaf and soil surfaces in the eco-system to lose water. Under moist conditions,

-2.0

-1.6

-1.2

-0.8

-0.4

0 00 20 40 50 6030

Pre

daw

n pl

ant w

ater

pot

entia

l (M

Pa)

Available soil water(% of capacity)

0.2

0.4

0.6

0.8

1.0

Tran

spira

tion

rate

(cm

d-1

)

Transpiration

Plant water potential

Str

eam

flow

or

evap

otra

nspi

ratio

n (L

m-2

yr-1

)

Precipitation (L m-2 yr-1) (mm yr-1)

Evapotranspiration

Stream

flow

1500 1800900 1200

400

600

800

1000

1200

1400

Figure 4.17. Relationship between annual waterinput (precipitation) and output (evapotranspirationand stream flow) from a temperate forest watershed(Hubbard Brook in the United States) over a 19-yrperiod. In this moist forest, evapotranspiration varieslittle among years, whereas stream flow is quite sensitive to the quantity of precipitation. (Redrawnwith permission from Springer-Verlag; Bormann andLikens 1979.)

Figure 4.16. Response of plant water poten-tial and transpiration to soil moisture (Sucoff1972, Gardner 1983, Waring and Running1998). Soil moisture has little effect on plantwater potential or transpiration until about75% of the available water has been removedfrom the rooting zone.

surface conductance is surprisingly insensitiveto vegetation properties (Kelliher et al. 1995).In sparse vegetation, evaporation from the soilsurface is the major avenue of water loss. AsLAI increases, transpiration increases (moreleaf area to transpire); this is counteracted by adecrease in soil evaporation (more shading andless turbulent exchange at the soil surface).Consequently, surface conductance is relativelyinsensitive to LAI.

Vegetation affects surface conductance pri-marily through its effects on stomatal conduc-tance (Kelliher et al. 1995). Even this influence,


however, is often relatively small. Maximumstomatal conductance of individual leaves isrelatively similar among natural ecosystems(Körner 1994, Kelliher et al. 1995). Woody and herbaceous ecosystems, for example, havesimilar stomatal conductance of individualleaves (Körner 1994) and similar surface con-ductance of entire ecosystems (Kelliher et al.1995). Crops, however, which have about 50%higher stomatal conductance than does naturalvegetation, also have about 50% higher surfaceconductance (Schulze et al. 1994, Kelliher et al.1995). There are currently insufficient data toknow whether ecological variation in stomatalconductance associated with gradients in soilfertility causes similar variation in surface conductance and therefore evapotranspirationfrom ecosystems.

In summary, aerodynamic roughness, whichdepends on plant height and the number ofroughness elements, is the main way in whichvegetation influences evapotranspiration fromdry canopies under conditions of adequatewater supply. Stomatal conductance exertsadditional control in some ecosystems. It alsohas an increasingly important control overevapotranspiration as soil moisture declines.In other words, stomatal conductance accountsfor temporal variation in evapotranspiration inresponse to variation in soil moisture, butsurface roughness is the major factor explain-ing differences in evapotranspiration amongmoist ecosystems.

Vegetation structure also influences the rela-tive importance of different climatic variablesin regulating evapotranspiration. In aerody-namically rough, well-mixed canopies such asopen-canopied forests, the moisture content ofthe air within the canopy is similar to that of thebulk air above the canopy. Under these well-coupled conditions, evapotranspiration is deter-mined more by the moisture content of the air (and the accompanying stomatal response)than by net radiation (Waring and Running1998). In short canopies, by contrast, the airadjacent to the leaves is mixed less readily withthe bulk air above the canopy, allowing evapo-transpiration to increase the moisture contentwithin the canopy environment. In other words,the canopy air becomes decoupled from condi-

tions in the bulk atmosphere. In these smoothcanopies, evapotranspiration is determinedmore by net radiation than by the moisturecontent of the bulk air, just as when canopiesare wet.The decoupling coefficient is a measureof the degree to which a canopy is decoupledfrom the atmosphere (Table 4.5) (Jarvis andMcNaughton 1986). It is determined primarilyby canopy height. In summary, net radiation isthe dominant environmental control over evap-otranspiration in short canopies, whereas thevapor pressure deficit is the dominant controlin tall canopies, when water is freely available(Waring and Running 1998).

Changes in Storage

Water inputs that exceed outputs replenishwater that is stored in soil and groundwater.Water that enters the soil is retained until thesoil reaches field capacity. Additional watermoves downward to groundwater. In cold cli-mates in winter, most of the precipitation inputis stored above ground in the snowpack. Thesnowpack substantially increases the quantityof water that an ecosystem can store and the residence time of water in the ecosystem.Stored water supports evapotranspiration at times when evapotranspiration exceeds pre-cipitation; the declines in soil moisture during

Table 4.5. Decoupling coefficient of vegetationcanopies in the field under conditions of adequatemoisture supply.

Vegetation Decoupling coefficienta

Alfalfa 0.9Strawberry patch 0.85Permanent pasture 0.8Grassland 0.8Tomato field 0.7Wheat field 0.6Prairie 0.5Cotton 0.4Heathland 0.3Citrus orchard 0.3Forest 0.2Pine woods 0.1

a A completely smooth surface has a decoupling coefficientof 1.0, and a canopy in which the air is identical to that inthe atmosphere has a decoupling coefficient of zero.Data from Jarvis and McNaughton (1986) and Jones (1992).


periods of dry weather draw down waterstorage. The seasonal recharge and depletion of stored water are important controls overevapotranspiration and net primary production(NPP) in many ecosystems.

Groundwater—the water beneath therooting zone—is a large pool that is inaccessi-ble to plants in many ecosystems. The size ofthis pool depends on the depth to impermeablelayers and the porosity of materials in this layer.Porosity governs the pore volume available to hold water and the resistance to lateraldrainage of water. The groundwater pool has arelatively constant size, so when new waterenters groundwater from the top, it displacesolder water that drains laterally to streams,rivers and oceans. The time lag between inputsto groundwater and outputs can be substantial(months to millennia) because of the large sizeof this pool. Groundwater therefore generallyhas an isotopic composition quite differentfrom that of precipitation or soil water (Fig.4.10).

People modify groundwater pools by chang-ing the vegetation and associated rooting depthand by tapping groundwater to support humanactivities. Introduction of deep-rooted exoticspecies in arid regions increases the pool ofwater available to support evapotranspirationby vegetation. This can cause the water table todrop. The introduction of deep-rooted Tamarixin North American deserts, for example, causedthe water table to drop enough that desertponds have dried, endangering endemic fishspecies (Berry 1970).

Removal of vegetation causes the water tableto rise because surface water is no longertapped to support evapotranspiration. Theclearing of heathlands for agriculture inwestern Australia, for example, reduced thedepth of the rooting zone, causing salinegroundwater to rise close to the surface. Thisreduced the productive potential of the crops,further reducing evapotranspiration and thedepth to groundwater. Finally, evaporationfrom the soil surface increased soil salinity tothe point that soils no longer supported crop growth in many areas, nor could they berecolonized by native heath vegetation (Nulsenet al. 1986). In this way human modification of

vegetation permanently altered the hydrologiccycle and all aspects of ecosystem structure andfunctioning.

Expansion of human populations into aridregions is frequently subsidized by tappinggroundwater that would otherwise be unavail-able to surface organisms. Wells that providewater to animals in African savannas, forexample, attract high densities of animals,which overgraze and alter the composition ofnearby vegetation. Often 80 to 90% of thewater in populated arid areas is used to supportirrigated agriculture. Irrigated agriculture ishighly productive because warm temperaturesand high solar radiation support a high pro-ductivity, when the natural constraints of waterlimitation are removed. The substantial cost ofirrigated agriculture often requires that it beintensively managed with fertilizers and pesti-cides (see Chapter 16).These irrigated lands areimportant sources of fruits, vegetables, cotton,rice, and other high-value crops. Conversion ofarid regions to irrigated agriculture, however,reduces the amount of water available forrunoff. Human use of water in the arid south-western United States, for example, convertedthe Rio Grande River from a major river to asmall stream with intermittent flow duringsome times of year. Irrigation also increases soilevaporation, which increases soil salinity in afashion similar to that described for westernAustralia.

In cases where evapotranspiration of irri-gated agriculture exceeds precipitation, there isnot only a decrease in runoff but also a deple-tion of the groundwater pool. The Ogallalaaquifer in the north-central United States, forexample, accumulated water when the climatewas much wetter than today. Tapping of this“fossil” water has increased the depth to thewater table substantially. Continued draw-down of this aquifer cannot be sustained indef-initely, because current water sources cannotreplenish it as rapidly as it is being depleted tosupport irrigation.

Runoff

Runoff is the difference between precipitationinputs, changes in storage, and losses to evapo-


transpiration (Eq. 4.8). Average runoff from anecosystem depends primarily on precipitationand evapotranspiration because long-termchanges in storage are usually negligible.Runoff responds to variation in precipitationmuch more strongly than does evapotranspira-tion (Fig. 4.17) because it constitutes the left-overs after the water demands for evapo-transpiration and groundwater recharge havebeen met. Runoff is therefore greater in wetthan in dry climates or seasons. Over hours toweeks, runoff generally increases after rainfallevents and decreases during dry periods.Changes in water storage buffer this linkagebetween precipitation and runoff. The rechargeof soil moisture in grasslands, shrublands, anddry forests, for example, may prevent largeincreases in flow from occurring after rainfallwhen soils are dry. In ecosystems with a smallcapacity to store water such as deserts withcoarse-textured soils and a calcic layer orecosystems underlain by permafrost, runoffresponds almost immediately to precipitation,and rainstorms can cause flash floods.Conversely, slowly draining groundwater pro-vides a continued source of water to streams (base flow) even at times when there is no precipitation.

In ecosystems that develop a snowpack inwinter, precipitation inputs are stored in theecosystem during winter, causing winter streamflows to decline, regardless of the seasonality ofprecipitation. Much of the water stored in the snowpack can move directly to streams,when the snow melts, causing large springrunoff events. Glacial rivers, for example,have greatest runoff in midsummer, when temperatures are highest, whereas nonglacialrivers in the same climate zone have peak flow in early spring following spring snow melt(Fig. 4.18).

Flow in streams and rivers integrates the precipitation, evapotranspiration, and changesin storage throughout the watershed. In largerivers, the seasonal variations in flow oftenreflect patterns of precipitation and evapotran-spiration that occur upstream, hours to weekspreviously. These integrative effects of runofffrom large watersheds make this an effectiveindicator of temporal changes in the hydrologiccycle.

Seasonal variations in stream flow are amajor determinant of the structure and season-ality of ecosystem processes in streams andrivers. Periods of high flow in small streams,for example, scour stream channels, removing

2000

1600

1200

800

400

0

140

120

100

80

60

40

20

0 Non

-gla

cial

riv

er m

ean

daily

dis

char

ge (

m3 s

ec-1

)

Gla

cial

riv

er m

ean

daily

dis

char

ge (

m3 s

ec-1

)

Glacialriver

Non-glacialriver

J F M A M J J A S O N DMonth

Figure 4.18. Average daily discharge from a glacialand a clearwater river. Runoff from the clearwaterriver peaks at snow melt, whereas the glacial river

has peak discharge when temperatures are warmestin midsummer. (Redrawn with permission fromPhyllis Adams; Adams 1999.)

Summary 95

or redistributing sediments, algae, and detritus(Power 1992a). In larger rivers, high flow eventsmay lead to predictable patterns of bankerosion and deposition. Dams that reduce theintensity of high-flow events dramatically alterthe natural disturbance regime and functioningof freshwater ecosystems (see Chapter 14).

Vegetation strongly influences the quantityof runoff. Because evapotranspiration is such alarge component of the hydrologic budget of anecosystem, any vegetation change that altersevapotranspiration inevitably affects runoff.Deforestation of a watershed, for example, candouble runoff (see Fig. 13.13). As vegetationregrows during succession, runoff returns topreharvest levels. Regional changes in landcover can have long-term effects on regionalhydrology. Watersheds that lose forest coverexhibit increased runoff, whereas those thatgain forest cover through reforestation showless runoff (Trimble et al. 1987) (Fig. 4.19).More subtle vegetation changes also alterrunoff. Conifer forests produce less runoff thandeciduous forests because of their greater leafarea their higher rates and longer season forevapotranspiration (Swank and Douglass1974).

Summary

The energy and water budgets of ecosystemsare inextricably linked because net radiation isan important driving force for evapotranspira-tion, and evapotranspiration is a large com-ponent of both water and energy flux fromecosystems. Net radiation is the balancebetween incoming and outgoing shortwave andlongwave radiation. Ecosystems affect net radi-ation primarily through albedo (shortwavereflectance), which depends on the reflectanceof individual leaves and other surfaces and oncanopy roughness, which depends primarily oncanopy height and complexity. Most absorbedenergy is released to the atmosphere as latentheat flux (evapotranspiration) and sensibleheat flux. Latent heat flux cools the surface and transfers water vapor to the atmosphere,whereas sensible heat flux warms the surfaceair. The Bowen ratio, the ratio of sensible tolatent heat flux, determines the strength of thecoupling of the water cycle to the energybudget.

Water enters terrestrial ecosystems primarilyas precipitation and leaves as evapotranspi-ration and runoff. Water moves through eco-systems in response to gradients in waterpotential, which is determined by pressurepotential, osmotic potential, and matric poten-tial. Water enters the ecosystem and movesdown through the soil in response to gravity.Available water in the soil moves along a film of liquid water through the soil-plant-atmosphere continuum in response to agradient in water potential that is driven bytranspiration (evaporation from the cell sur-faces inside leaves). Evapotranspiration fromcanopies depends on the driving forces forevaporation (net radiation and vapor pressuredeficit of the air) and two conductance terms,boundary layer and surface conductances.Boundary layer conductance depends on thedegree to which the canopy is coupled to theatmosphere, which varies with canopy heightand aerodynamic roughness. Surface conduc-tance is mainly influenced by the average stom-atal conductance of leaves in the canopy.Stomatal and surface conductances are rela-tively similar among natural ecosystems but are

100

200

300

400

500

600

700

Incr

ease

in s

trea

mflo

w (

mm

yr-1

)

Area deforested (% of watershed)

10 20 30 40 50 60 70 80 90 100

Figure 4.19. Influence of deforestation on changesin stream flow in the southeastern United States.Stream flow increases linearly with the proportion of the watershed that is deforested. Also included in this dataset are watersheds that show reducedstream flow in response to increases in forest cover.(Redrawn with permission from Water ResourceResearch; Trimble et al. 1987.)


somewhat higher in crop systems. Climate influ-ences evapotranspiration by determining thedriving forces for evapotranspiration and thewater availability in the soil, which determinesstomatal conductance. Vegetation influencesevapotranspiration through plant height andaerodynamic roughness, which govern bound-ary layer conductance, and through stomatalconductance, which influences surface conduc-tance and the plant response to soil moisture.

The partitioning of water loss between evap-otranspiration and runoff depends primarily onwater storage in the rooting zone and the rateof evapotranspiration. Runoff is the leftoverwater that drains from the ecosystem at timeswhen precipitation exceeds evapotranspirationplus increases in water storage. Human activi-ties alter the hydrologic cycle primarily throughchanges in land cover and use, which affectevapotranspiration and soil water storage.

Review Questions

1. What climatic and ecosystem propertiesgovern energy input to an ecosystem?

2. What are the major avenues by whichenergy absorbed by an ecosystem isexchanged with the atmosphere? Whatdetermines the total energy exchange? Whatdetermines the relative importance of thepathways by which energy is exchanged?

3. What are the consequences of transpirationfor ecosystem energy exchange and for thelinkage between energy and water budgetsof an ecosystem?

4. How might global changes in climate andland use alter the components of energyexchange in an ecosystem?

5. What are the major pathways of watermovement in an ecosystem? What deter-mines the balance among these pathways,for example, between evaporation, transpi-ration, and runoff? How do climate, soils,and vegetation influence the pools andfluxes of water in an ecosystem?

6. What are the mechanisms driving wateruptake and loss from plants? How do plantproperties influence water uptake and loss?

7. How do the controls over water loss fromplant canopies differ from the controls at thelevel of individual leaves?

8. Describe how grassland and forests differ in properties that influence wet canopyevaporation, transpiration, soil evaporation,infiltration, and runoff. What will be the consequences for runoff and for regionalclimate of a policy that encourages thereplacement of grasslands with forests so asto increase terrestrial carbon storage?

Additional Reading

Campbell, G.S., and J.M. Norman. 1998. An Intro-duction to Environmental Biophysics. Springer-Verlag, New York.

Dawson, T.E. 1993. Water sources of plants as deter-mined from xylem-water isotopic composition:Perspectives on plant competition, distribution,and water relations. Pages 465–496 in J.R.Ehleringer, A.E. Hall, and G.D. Farquhar, editors.Stable Isotopes and Plant Carbon-Water Relations.Academic Press, San Diego, CA.

Jarvis, P.G., and K.G. McNaughton. 1986. Stomatalcontrol of transpiration: Scaling up from leaf toregion. Advances in Ecological Research 15:1–49.

Jones, H.G. 1992. Plants and Microclimate: A Quan-titative Approach to Environmental Plant Physiol-ogy. Cambridge University Press, Cambridge, UK.

Kelliher, F.M., R. Leuning, M.R. Raupach, and E.-D.Schulze. 1995. Maximum conductances for evapo-ration from global vegetation types. Agriculturaland Forest Meteorology 73:1–16.

Monteith, J.L., and M. Unsworth. 1990. Principles ofEnvironmental Physics. 2nd ed. Arnold, London.

Oke, T.R. 1987. Boundary Layer Climates. 2nd ed.Methuen, London.

Schulze, E.-D., F.M. Kelliher, C. Körner, J. Lloyd, andR. Leuning. 1994. Relationship among maximumstomatal conductance, ecosystem surface conduc-tance, carbon assimilation rate, and plant nitrogennutrition: A global ecology scaling exercise.Annual Review of Ecology and Systematics25:629–660.

Sperry, J.S. 1995. Limitations on stem water transportand their consequences. Pages 105–124 in B.L.Gartner, editor. Plant Stems: Physiology and Func-tional Morphology. Academic Press, San Diego,CA.

Waring, R.H., and S.W. Running. 1998. Forest Ecosys-tems: Analysis at Multiple Scales. Academic Press,New York.

Introduction

The energy fixed by photosynthesis directlysupports plant growth and produces organicmatter that is consumed by animals and soilmicrobes. The carbon derived from photosyn-thesis makes up almost half of the organicmatter on Earth; hydrogen and oxygen accountfor most of the remainder. Human activitieshave radically modified the rate at whichcarbon enters the terrestrial biosphere bychanging most of the controls over this process.We have increased by 30% the quantity ofatmospheric CO2 to which all terrestrial plantsare exposed. On a regional scale we havealtered the availability of water and nutrients,the major resources that determine the capac-ity of plants to use atmospheric CO2. Finally,through changes in land cover and the intro-duction and extinction of species, we havechanged the regional distribution of thecarbon-fixing potential of the terrestrial bios-phere. Because of the central role that carbonplays in the climate system (see Chapter 2) andthe biosphere, it is critical that we understandthe factors that regulate its cycling through vegetation and ecosystems. We address carboninputs to terrestrial ecosystems through photo-synthesis in this chapter and inputs to aquaticecosystems in Chapter 10. In Chapters 6 and 7,

we explore the carbon losses from plants andecosystems, which, together with photosynthe-sis, govern the patterns of accumulation andloss of carbon in ecosystems.

Overview

The availability of water and nutrients is themajor factor governing carbon input to ecosys-tems. Photosynthesis is the process by whichmost carbon and chemical energy enter ecosys-tems. The proximate controls over photosyn-thesis by a single leaf are the availability ofreactants such as light energy and CO2; tem-perature, which governs reaction rates; and the availability of nitrogen, which is required to produce photosynthetic enzymes (Fig. 5.1).Photosynthesis at the scale of ecosystems istermed gross primary production (GPP). Likephotosynthesis by individual leaves, GPP variesdiurnally and seasonally in response to changesin light, temperature, and nitrogen supply.Differences among ecosystems in annual GPP, however, are determined primarily by thequantity of leaf area and the length of time thatthis leaf area is photosynthetically active. Leafarea and photosynthetic season, in turn, dependon the availability of soil resources (water andnutrients), climate, and time since disturbance.

5Carbon Input to Terrestrial Ecosystems

Photosynthesis by plants provides the carbon and energy that drive most biologicalprocesses in ecosystems. This chapter describes the controls over this carbon input.

97

98 5. Carbon Input to Terrestrial Ecosystems

In this chapter we explore the mechanismsbehind these causal relationships.

Carbon and energy are linked as they movethrough ecosystems because the same pro-cesses govern their entry into ecosystems inphotosynthesis, transfer within ecosystems, andloss from ecosystems. Photosynthesis uses lightenergy (i.e., radiation in the visible portion ofthe spectrum) to reduce CO2 and producecarbon-containing organic compounds. Thisorganic carbon and its associated energy arethen transferred among components within theecosystem and are eventually released to theatmosphere by respiration or combustion.

The energy content of organic matter differsamong carbon compounds, but for whole tis-sues, it is relatively constant at about 20kJg-1

of ash-free dry mass (Golley 1961, Larcher1995). The carbon concentration of organicmatter is also variable but averages about 45%in herbaceous tissues and 50% in wood (Goweret al. 1999). Both the carbon and energy con-

tents of organic matter are greatest in materi-als such as seeds and animal fat that have highlipid content and are lowest in tissues with highconcentrations of minerals or organic acids(Fig. 5.2). Because of the relative constancy ofthe carbon and energy contents of organicmatter, carbon, energy, and biomass have been used interchangeably as currencies of thecarbon and energy dynamics of ecosystems.Thepreferred units differ among fields of ecology,depending on the processes that are of greatestinterest or are measured most directly. Pro-duction studies, for example, typically focus onbiomass; trophic studies, on energy; and gasexchange studies, on carbon.

Photosynthetic Pathways

C3 Photosynthesis

The rate of carbon input to ecosystems dependson the response of photosynthetic biochemistry

STATE FACTORS

Interactivecontrols

Directcontrols

BIOTA

PARENTMATERIAL

TIME

CLIMATE

Plantfunctional

types

Soilresources

Leaf area

Nitrogen

Seasonlength

Temperature

Light

GPP

LONG-TERMCONTROLS

SHORT-TERMCONTROLS

CO2

Figure 5.1. The major factors governing grossprimary production (GPP) in ecosystems. These controls range from the direct controls, which deter-mine the diurnal and seasonal variations in GPP,to the interactive controls and state factors, whichare the ultimate causes of ecosystem differences in GPP. Thickness of the arrows associated with

direct controls indicates the strength of the effect.The factors that account for most of the variation inGPP among ecosystems are leaf area and length ofthe photosynthetic season, which are ultimatelydetermined by the interacting effects of soilresources, climate, vegetation, and disturbanceregime.

Photosynthetic Pathways 99

to environment. A brief overview of the biochemistry of photosynthesis provides amechanistic basis for understanding the envi-ronmental controls over carbon input toecosystems.

There are two major groups of reactions inphotosynthesis. The light-harvesting reactionstransform light energy into a temporary formof chemical energy. The carbon-fixation reac-tions use the products of the light-harvestingreactions to convert CO2 into sugars, a morepermanent form of chemical energy that can bestored, transported, or metabolized. In light,both groups of reactions occur simultaneouslyin the chloroplasts, which are organelles insidethe mesophyll cells (photosynthetic cells) ofgreen leaves (Fig. 5.3). In the light-harvestingreactions, chlorophyll (a light-absorbingpigment) captures energy from visible light.Absorbed radiation is converted to chemicalenergy (NADPH and ATP), and oxygen is pro-duced as a waste product (Fig. 5.3).Visible radi-ation accounts for only 40% of incoming solar

radiation (see Chapter 2), which places anupper limit on the potential efficiency of pho-tosynthesis in converting solar radiation intochemical energy.

The carbon-fixation reactions of photosyn-thesis use the chemical energy (ATP andNADPH) from the light-harvesting reactions toreduce CO2 to sugars. The rate-limiting step inthe carbon-fixation reactions is the reaction of a five-carbon sugar (ribulose-bisphosphate;RuBP) with CO2 to form two three-carbonsugars. Because the initial products of photo-synthesis are three-carbon sugars, this photosynthetic pathway is known as C3 photo-synthesis. The initial attachment of CO2 to acarbon skeleton is catalyzed by the enzymeribulose-bisphosphate carboxylase-oxygenase(Rubisco). The rate of this reaction is generallylimited by the products of the light reaction andby the concentration of CO2 in the chloroplast.A surprisingly high concentration of Rubisco isrequired for carbon fixation. Rubisco accountsfor about 25% of leaf nitrogen, and other photosynthetic enzymes make up an addi-tional 25%. The remaining enzymatic steps inthe carbon-fixation reactions use ATP andNADPH from the light-harvesting reactions to regenerate RuBP as a carbon acceptor tosustain further photosynthesis (Fig. 5.3). Themost notable features of the carbon-fixationreactions are (1) their large nitrogen require-ment for Rubisco and other photosyntheticenzymes; (2) their dependence on the productsof the light-harvesting reactions (ATP andNADPH), which in turn depend on irradiance(the light received by the leaf); and (3) their fre-quent limitation by CO2 supply to the chloro-plast. The basic biochemistry of photosynthesistherefore dictates that this process must be sen-sitive to light and CO2 supplies over time scalesof milliseconds to minutes and sensitive tonitrogen supply over time scales of days toweeks (Fig. 5.1).

Rubisco is both a carboxylase, which initiatesthe carbon-fixation reactions of photosynthesis,and an oxygenase, which catalyzes the reactionbetween RuBP and oxygen (Fig. 5.3). The oxy-genase activity initiates a series of steps thatbreaks down sugars to CO2. This process ofphotorespiration immediately respires away 20

Herb

Leaf Stem Seed Root/rhizome

28

26

24

22

20

18

Ene

rgy

cont

ent (

kJ g

-1)

Broad-leavedtree

Needle-leavedtree

Figure 5.2. Energy content of major tissues inconifer trees, dicotyledonous trees, and dicotyledo-nous herbs. Compounds that contribute to a highenergy content include lipids (seeds), terpenes andresins (conifers), proteins (leaves), and lignin (woodytissues). Values are expressed per gram of ash-freedry mass. (Redrawn with permission from Springer-Verlag; Larcher 1995.)


to 40% of the carbon fixed by C3 photosynthe-sis and regenerates ADP and NADP in theprocess. Why do C3 plants have such an ineffi-cient system of carbon acquisition, by whichthey immediately lose a third of the carbon thatthey acquire from photosynthesis? Althoughwe have no definite answer to this question, themost likely explanation is that photorespirationacts as a safety valve. It provides a supply ofreactants (ADP and NADP) to the light reac-tion under circumstances in which an inade-quate supply of CO2 limits the rate at whichthese reactants can be regenerated by carbon-fixation reactions. In the absence of photores-

piration, continued light harvesting producesoxygen radicals that destroy photosyntheticpigments.

Plants have additional lines of defenseagainst excessive energy capture, which are at least as important as photorespiration. Onesuch photoprotection mechanism involves pig-ments that change from one form to another inthe xanthophyll cycle. When excess excitationenergy is present and cannot be processed togenerate ATP and NADPH, xanthophyllpigment is converted to a form that can re-ceive excess absorbed energy from the excitedchlorophyll (Demming-Adams and Adams

Stroma

Photo-respiration

ATP

NADPH

2 C2 compounds

Sugarexport

Light-harvestingreactions

NADP

ADPThylakoid

Carbon-fixationreactions

2 C3 sugars

C2export

CO2O2

H+ + O2H2O

O2

RuBP(C5 sugar)

H+

Ligh

t

chl

e- transport chain

StarchT

hyla

koid

Thy

lako

id

Figure 5.3. A chloroplast showing the location of themajor photosynthetic reactions. The light-harvestingreactions occur in the thylakoid membranes; chloro-phyll (chl) absorbs visible light and funnels it to reac-tion centers (Photosystems I and II). In PhotosystemII, water is split to H+ and O2, and the resulting elec-trons are then passed down an electron-transportchain inside the thylakoid, ultimately to NADP, pro-ducing NADPH. During this process, protons moveacross the thylakoid membrane to the stroma,and theproton gradient drives the synthesis of ATP.ATP andNADPH provide the energy to regenerate ribulose-bisphosphate (RuBP) within the carbon fixation

reactions. RuBP reacts either with CO2 to produce sugars and starch (carbon-fixation reactions of photosynthesis) or with O2 to produce two-carbonintermediates (photorespiration). These two-carbonintermediates are exported from the chloroplast tomitochondria or peroxisomes, where they are againconverted to sugars, with loss of CO2 and ATP.Through either carbon fixation or photorespiration,ADP and NADP again become reactants available toproduce additional ATP and NADPH.The net effectof photosynthesis is to convert light energy into chem-ical energy (sugars and starches) that is available tosupport plant growth and maintenance.


1996). The energy is then harmlessly dissipatedas heat. The xanthophyll cycle processes much of the energy that is not used for carbonfixation under high light and serves as furtherprotection against photodestruction of photo-synthetic pigments under conditions of highlight.

Net photosynthesis is the net rate of carbongain measured at the level of individual leaves.It is the balance between simultaneous CO2 fix-ation and leaf respiration in the light (both photorespiration and mitochondrial respira-tion). The overall efficiency of converting lightenergy into sugars is about 6% under optimalconditions at low light, but is closer to 1%under most field conditions.

The CO2 used in photosynthesis diffusesalong a concentration gradient from the atmos-phere outside the leaf into the chloroplast.Carbon dioxide first diffuses across a layer ofrelatively still air close to the leaf surface (theplant boundary layer) and then through the stomata (small pores in the leaf surface),the diameter of which is regulated by the plant(Fig. 5.4). Once inside the leaf, CO2 diffusesthrough air spaces between cells, dissolves inwater on the cell surfaces, and diffuses the shortdistance from the cell surface to the chloro-

plast. Stomata are the largest (and most vari-able) component of the total resistance to CO2

diffusion.The thin flat shape of most leaves andthe abundance of air spaces inside leaves max-imize the rate of CO2 diffusion from the bulkair to the chloroplast.

Cell walls inside the leaf are covered with athin film of water, which facilitates the efficienttransfer of CO2 from the air to the interior ofcells. This water readily evaporates, and watervapor diffuses out through the stomata acrossthe boundary layer to the atmosphere. Thisprocess is called transpiration. The openstomata that are necessary for plants to gaincarbon are an avenue for water loss. In otherwords, plants face an inevitable trade-offbetween CO2 uptake (which is necessary todrive photosynthesis) and water loss (whichmust be replaced by absorption of water fromthe soil). This trade-off can be as high as 400moles of water lost for each mole of CO2

absorbed. Plants regulate CO2 uptake andwater loss by changing the size of stomatalopenings, which regulates stomatal conduc-tance, the flux of water vapor or CO2 per unitdriving force (i.e., for a given concentrationgradient). When plants reduce stomatal con-ductance to conserve water, photosynthesis

Leaf

Boundary layer

Bulkair

Stoma

Cuticle

Epidermal cell

Mesophyllcells

CO2 H2O

Guard cell

Figure 5.4. Cross-section of a leaf showing the diffusion pathways of CO2 and H2O into and out ofthe leaf, respectively. Length of the horizontal arrows

outside the leaf is proportional to wind speeds in theboundary layer.


declines, reducing the efficiency with whichplants convert light energy to carbohydrates.Plant regulation of CO2 delivery to the chloro-plast is therefore a compromise between maxi-mizing photosynthesis and minimizing waterloss and depends on the relative supplies ofCO2, light, and mineral nutrients.

C4 Photosynthesis

C4 photosynthesis adds a set of carbon-fixationreactions that enable some plants to increasephotosynthetic water use efficiency in dry envi-ronments. About 85% of vascular plant speciesfix carbon by the C3 photosynthetic pathway,in which Rubisco is the primary carboxylatingenzyme. The first products of C3 photosynthesisare three-carbon sugars. About 5% of theglobal flora photosynthesize by the C4 photo-synthetic pathway. C4 species dominate manywarm high-light environments, particularlytropical grasslands and savannas. C4-dominatedecosystems account for nearly a third of the ice-free terrestrial surface (see Table 6.5) and aretherefore quantitatively important in the globalcarbon cycle. In C4 photosynthesis, phospho-enolpyruvate (PEP) is first carboxylated byPEP carboxylase in mesophyll cells to producefour-carbon organic acids (Fig. 5.5). Theseorganic acids are transported to specializedbundle sheath cells, where they are decarboxy-lated. The CO2 released from the organic acidsthen enters the normal C3 pathway of photo-synthesis to produce sugars that are exportedfrom the leaf. There are three ecologicallyimportant features of the C4 photosyntheticpathway.

First, C4 acids move to the bundle sheathcells, where they are decarboxylated, concen-trating CO2 at the site where Rubisco fixescarbon. This increases the efficiency of car-boxylation by Rubisco because it increases theconcentration of CO2 relative to O2, whichwould otherwise compete for the active site ofthe enzyme. Apparent photorespiration mea-sured at the leaf level is low in C4 plantsbecause most of the Rubisco reacts with CO2

rather than with O2 and because the PEP carboxylase in the mesophyll cells scavengesany photorespired CO2 that escapes from the

bundle sheath cells. The high efficiency ofRubisco in C4 plants reduces the quantity ofRubisco (and therefore nitrogen) required forC4 photosynthesis.

Second, PEP carboxylase is more effectivethan Rubisco in drawing down the concentra-tion of CO2 inside the leaf. This increases theCO2 concentration gradient between the exter-nal air and the internal air spaces of the leaf. AC4 plant can therefore absorb CO2 with moretightly closed stomata than can a C3 plant, thusreducing water loss.

Third, the net cost of regenerating the carbonacceptor molecule (PEP) of the C4 pathway istwo ATPs for each CO2 fixed, a 30% increasein the energy requirement of photosynthesiscompared to C3 plants.

The major advantages of the C4 photosyn-thetic pathway are that less water is lost and lessnitrogen is required to maintain a given rate ofphotosynthesis compared to C3 plants. C4 plantstherefore often have a high CO2 fixationrate under high-light, low-nitrogen conditions.Moreover, due to their lack of photorespira-tion, which is quite temperature sensitive, C4

plants can maintain higher rates of net photo-synthesis at high temperatures than can C3

plants; this explains their success in warm envi-ronments. The main disadvantage of the C4

pathway is the additional energy cost for eachcarbon fixed by photosynthesis.The C4 pathwayis therefore most advantageous in warm, high-light conditions, such as tropical grasslands.The C4 pathway occurs in 18 plant families andhas evolved independently at least 30 times(Kellogg 1999). C4 species first became abun-dant in the late Miocene 6 to 8 million yearsago, probably triggered by the global decline inatmospheric CO2 concentration (Cerling 1999).C4 grasslands expanded during glacial periods,when CO2 concentrations declined, andretracted at the end of glacial periods, whenatmospheric CO2 concentration increased, sug-gesting that the evolution of C4 photosynthesiswas tightly tied to variations in atmosphericCO2 concentration. However, there is little geographic variation in atmospheric CO2

concentration, so the current global geographicdistribution of C4 plants appears to be con-trolled primarily by temperature and by the


availability of light, water, and nitrogen. C4 pho-tosynthesis is absent from most woody plants.

C4 plants have an isotopic signature thatenables us to track their past and present rolein ecosystems. C4 plants incorporate muchmore 13C than do C3 plants during photosyn-thesis (Box 5.1) and therefore have a distinctisotopic signature that remains with anyorganic matter produced by this photosyntheticpathway. Isotopic measurements are important

tools in studying ecological processes in ecosys-tems in which the relative abundance of C3 andC4 plants has changed over time (Ehleringer etal. 1993).

Crassulacean Acid Metabolism Photosynthesis

Crassulacean acid metabolism (CAM) enablesecosystems to gain carbon under extremely dry

Day

C3 plants

C4 plants

CAM plants

C3 Ps

H2O

Sugar

Mes

C4 acid

Sugar

CO2 + C3

C3 Ps

Mes

Sugar

C4 Ps

CO2

C4 Ps

C4 acid

Mes BS

H2O

CO2

C3 Ps

C3

+

Mes

Mes

Mes

BS

CO2

Night

CO2

CO2

CO2

Rmi

RmiRmi

C4 acid

C4 Ps

H2O

Figure 5.5. Cellular location and temporal cycle ofCO2 fixation and water exchange in leaves with C3,C4, and CAM photosynthetic pathways. In C3 andCAM plants, all photosynthesis occurs in mesophyll(Mes) cells. In C4 plants, C4 carbon fixation (C4 Ps)

occurs in mesophyll cells and C3 fixation (C3 Ps)occurs in bundle sheath (BS) cells. Mitochondrial res-piration (Rmi) occurs at night. Exchanges of CO2 andwater vapor with the atmosphere occur during theday in C3 and C4 plants and at night in CAM plants.


conditions. Succulent plant species in dry envi-ronments, including many epiphytes in tropicalforests, gain carbon through CAM photosyn-thesis. CAM accounts for just a small propor-tion of terrestrial carbon gain, because it isactive only under extremely dry conditions.Even in these environments, some CAM plantsswitch to C3 photosynthesis when sufficientwater is available.

In CAM photosynthesis, plants close theirstomata during the day, when high tissue tem-peratures and low relative humidity of the airwould otherwise cause large transpirationalwater loss (Fig. 5.5). At night, they open theirstomata, and CO2 enters the leaf and is fixed byPEP carboxylase. The resulting C4 acids arestored in vacuoles until the next day when theyare decarboxylated, releasing CO2 to be fixed

Box 5.1. Carbon Isotopes

The three isotopic forms of carbon (12C, 13C,and 14C) differ in the number of neutrons buthave the same number of protons. The addi-tional atomic mass causes the heavier iso-topes to react more slowly in some reactions,particularly in the carboxylation of CO2 byRubisco. The lightest of these isotopes (12C) is preferentially fixed by carboxylatingenzymes. C3 plants generally have a rela-tively high CO2 concentration inside the leaf,due to their high stomatal conductance.Under these circumstances, Rubisco dis-criminates against the heavier isotope 13C,causing 13CO2 to accumulate within the leaf.13CO2 therefore diffuses out of the leafthrough the stomata along a concentrationgradient of 13CO2 at the same time that 12CO2

is diffusing into the leaf. In C4 and crassu-lacean acid metabolism (CAM) plants, incontrast, PEP carboxylase has such a highaffinity for CO2 that it reacts with most of theCO2 that enters the leaf, resulting in rela-tively little discrimination against 13CO2.Consequently the 13C concentrations ofCAM and C4 plants are much higher (lessnegative) than those of C3 plants (Table 5.1).

This difference in isotopic compositionamong C3, C4, and CAM plants remains inany organic compounds derived from theseplants. Thus it is possible to calculate the rel-ative proportions of C3 and C4 plants in thediet of animals by measuring the 13C contentof their tissues; this can be done even in fossilbones such as those of early humans.Changes in the isotopic composition of fossilbones clearly indicate changes in diet. In

situations in which vegetation has changedfrom C3 to C4 dominance (or vice versa), theorganic matter in plants differs in its isotopiccomposition from that of the soil. Changes in the carbon isotope composition of soilorganic matter over time therefore providesa tool for estimating the rates of turnover ofsoil organic matter that formed beneath theprevious vegetation.

Table 5.1. Representative 13C concentrations(‰) of atmospheric CO2 and selected plant andsoil materials.

Material d13C (‰)a

Pee Dee limestone standard 0.0Atmospheric CO2 -8Plant material

Unstressed C3 plant -27Water-stressed C3 plant -25Unstressed C4 plant -13Water-stressed C4 plant -13CAM plantb -27 to -11

Soil organic matterDerived from unstressed C3 plants -27Derived from C4 or CAM plants -13

a The concentrations are expressed relative to an inter-nationally agreed-on standard (Pee Dee belemnite):

where d13C is the isotope ratio

in delta units relative to a standard, and Rsam and Rstd

are the isotope abundance ratios of the sample and stan-dard, respectively (Ehleringer and Osmond 1989).b Values of -11 under conditions of CAM photosyn-thesis. Many CAM plants switch to C3 photosynthesisunder favorable moisture regimes, giving an isotopicratio similar to that of unstressed C3 plants.Data from O’Leary (1988) and Ehleringer and Osmond(1989).

d 13 1000 1CRR

stdsam

std

= -ÊË

ˆ¯ ,

Net Photosynthesis by Individual Leaves 105

by normal C3 photosynthesis. Thus, in CAMplants, there is a temporal (day–night) separa-tion of C3 and C4 carbon dioxide fixation,whereas in C4 plants there is a spatial separa-tion of C3 and C4 carbon dioxide fixationbetween bundle sheath and mesophyll cells.CAM photosynthesis is energetically expen-sive, like C4 photosynthesis; it therefore occursprimarily in dry high-light environments, suchas deserts, shallow rocky soils, and canopies oftropical forests. CAM photosynthesis allowssome plants to gain carbon under extremely dry conditions that would otherwise precludecarbon fixation in ecosystems.

Net Photosynthesis by Individual Leaves

Basic Principle of Environmental Control

Plants adjust the components of photosynthesisso physical and biochemical processes co-limitcarbon fixation. Photosynthesis operates mostefficiently when the rate of CO2 diffusion intothe leaf matches the biochemical capacity of theleaf to fix CO2. Plants regulate the componentsof photosynthesis to achieve this balance, asseen from the response of photosynthesis to theCO2 concentration inside the leaf (Fig. 5.6).When the internal CO2 concentration is low,photosynthesis increases linearly with increas-ing CO2 concentration. Under these circum-stances, the leaf has more carbon fixationcapacity than it can effectively use, and photo-synthesis is limited by the rate of diffusion ofCO2 into the leaf. The plant can increase photo-synthesis only by opening stomatal pores.Alter-natively, if CO2 concentration inside the leaf ishigh, photosynthesis shows little response tovariation in CO2 concentration.In this case,pho-tosynthesis is limited by the rate of carboxyla-tion of RuBP (the asymptote in Fig. 5.6), andchanges in stomatal opening have little influ-ence on photosynthesis. At high internal CO2

concentrations, carboxylation may be limited by(1) insufficient light (or light-harvesting pig-ments) to provide energy; (2) insufficient nitro-gen invested in photosynthetic enzymes to

process the ATP, NADPH, and CO2 present inthe chloroplast; or (3) insufficient phosphate orsugar phosphates to synthesize RuBP.

Under a wide variety of circumstances, plantsadjust the components of photosynthesis soCO2 diffusion and biochemistry are aboutequally limiting to photosynthesis (Farquharand Sharkey 1982). Plants make this adjustmentby altering stomatal conductance, which occurswithin minutes, or by changing the concentra-tions of light-harvesting pigments or photosyn-thetic enzymes, which occurs over days toweeks. The general principle of co-limitation ofphotosynthesis by biochemistry and diffusionprovides the basis for understanding most ofthe adjustments by individual leaves to mini-mize the environmental limitations of photo-synthesis. Stomatal conductance is regulated sophotosynthesis occurs near the break point ofthe CO2-response curve (Körner et al. 1979)(Fig. 5.6), at which CO2 supply and carbon-fixation capacity are about equally limiting tophotosynthesis.

Light Limitation

Leaves adjust stomatal conductance and pho-tosynthetic capacity to maximize carbon gain in

0

Net

pho

tosy

nthe

sis

(µm

ol m

-2 s

-1)

350 700

CO2compensation

point

Internal concentration (ppmv)CO2

Biochemicallimitation

Limitation byCO2 diffusion

Figure 5.6. Relationship of the net photosyntheticrate to the CO2 concentration inside the leaf. Photo-synthetic rate is limited by the rate of CO2 diffusioninto the chloroplast in the linear portion of the CO2-response curve and by biochemical processes at higher CO2 concentrations. The CO2 compensa-tion point is the minimal CO2 concentration at whichthe leaf shows a net gain of carbon.


different light environments. Leaves experi-ence large fluctuations (10- to 1000-fold) inincident light due to changes in sun angle,cloudiness, and the location of sunflecks(patches of direct sunlight that penetrate aplant canopy) (Fig. 5.7). Leaf chloroplasts

respond to changes in light availability overminutes by changing both the levels of metabo-lites, which influence the activity of photosyn-thetic enzymes, and the stomatal conductance,which influences CO2 supply and water loss(Pearcy 1990, Chazdon and Pearcy 1991). Stom-atal conductance increases in high light, whenCO2 demand is high, and decreases in low light,when photosynthetic demand for CO2 is low.These stomatal adjustments result in a rela-tively constant CO2 concentration inside theleaf, as expected from our hypothesis of co-limitation of photosynthesis by biochemistryand diffusion. It allows plants to conserve waterunder low light and to maximize carbon uptakeat high light.

At low light, when the supply of ATP andNADPH from the light-harvesting reactionslimits the rate of carbon fixation, net photosyn-thesis increases linearly with increasing light(Fig. 5.8). The slope of this line (the quantumyield of photosynthesis) is a measure of the effi-ciency with which plants use absorbed light toproduce sugars. The quantum yield is similaramong all C3 plants at low light in the absenceof environmental stress. In other words, all C3

plants have a relatively constant photosyntheticlight use efficiency (LUE) (about 6%) of con-verting absorbed visible light (photosyntheti-cally active radiation, PAR) into chemical

1200

0

1200

0June July

Irra

dian

ce (

µmol

m-2

s-1)

Cloud

Sunfleck

Abovecanopy Mid-

canopy

1200

11:00 11:05 11:10

Dawn

12:00 24:00

Frontal system

Time

Clouds

0

0

Dusk

Figure 5.7. Hypothetical time course of photosyn-thetically active radiation above and below thecanopy of a temperate forest over minutes, hours,and months. Over the course of a few minutes, lightat the top of the canopy varies with cloudiness.Below the canopy, light also varies due to the pres-ence of sunflecks of direct irradiance, which can lasttenths of seconds to minutes. During a day, there arelarge changes in light due to changes in solar angle,with smaller fluctuations caused by passing clouds.Convective activity often increases cloudiness in theafternoon. During the growing season, the majorcauses of variation in light are seasonal changes inthe solar angle and the passage of frontal systems.Some times of year have greater frequency of cloudi-ness than others due to changes in directions of theprevailing winds and the passage of frontal systems.

Lightsaturation

Lightcompensation

point

Irradiance (µmol m-2 s-1)Net

pho

tosy

nthe

sis

(µm

ol m

-2 s

-1)

01000 2000

Lightlimitation

Photo-oxidation

Figure 5.8. Relationship of net photosynthetic rate to photosynthetically active radiation and theprocesses that limit photosynthesis at different irradiance. The linear increase in photosynthesis in response to increased light (in the range of lightlimitation) indicates relatively constant light use efficiency.


energy under these low-light conditions. Athigh irradiance, photosynthesis becomes lightsaturated—that is, it no longer responds tochanges in light supply, due to the finite capac-ity of the light-harvesting reactions to capturelight. As a result, light energy is converted lessefficiently into sugars at high light. Photosyn-thesis may decline at extremely high light as aresult of photo-oxidation of photosyntheticenzymes and pigments.

Over longer time scales (days to months)plants respond to variations in light availabilityby producing leaves with different photosyn-thetic properties. This physiological adjustmentby an organism in response to a change in someenvironmental parameter is known as acclima-tion. Leaves at the top of the canopy (sunleaves) have more cell layers, are thicker, andtherefore have greater photosynthetic capacityper unit leaf area than do shade leaves pro-duced under low light (Terashima andHikosaka 1995, Walters and Reich 1999). Therespiration rate of a tissue depends on itsprotein content (see Chapter 6), so the low pho-tosynthetic capacity and protein content ofshade leaves are associated with a lower respi-ration rate per unit area than in sun leaves. Forthis reason, shade leaves maintain a more pos-itive carbon balance (photosynthesis minus res-piration) under lower light than do sun leaves(Fig. 5.9). The net effect of acclimation to vari-ation in light availability is to extend the rangeof light availability over which vegetation main-tains a relatively constant LUE—that is, a rela-tively constant relationship between absorbedPAR and net photosynthesis.

Plants can also produce shade leaves as aresult of adaptation, the genetic adjustment bya population to maximize performance in a par-ticular environment. Species that are adapted tohigh light and are intolerant of shade typicallyhave a higher photosynthetic capacity per unitmass or area than do shade-tolerant species,even in the shade (Walters and Reich 1999).The main disadvantage of the high protein andphotosynthetic rate of shade-intolerant speciesis that these species also have a higher respira-tion rate than do shade-tolerant species, due totheir higher protein content. In addition, shade-intolerant species produce short-lived leaves, so

they must continuously produce new leaves tomaintain their leaf area.The net carbon balanceof individual leaves of shade-tolerant andshade-intolerant species in the shade is similar,but shade-intolerant species often have a lessfavorable whole-plant carbon balance in theshade owing to their more rapid leaf turnover.Shade-intolerant species have a higher carbon-gaining capacity at high light.

Variations in leaf angle also increase the effi-ciency with which a plant canopy uses light.At high light, plants produce leaves that aresteeply angled, so they absorb less light (seeChapter 4). This is advantageous because itreduces the probability of overheating orphoto-oxidation of photosynthetic pigments atthe top of the canopy. At the same time, itallows more light to penetrate to lower leaves.Leaves at the bottom of the canopy, on theother hand, are more horizontal in orientationto maximize light capture.The variations in leafangle and photosynthetic properties optimizethe performance of individual leaves, whichexplains why these patterns have evolved. Thisoptimization at the level of individual plantstranslates into a maximization of the LUE and

Species C (sun)

Species B(intermediate)

Species A (shade)

C

B

A

CO

2 up

take

(µm

ol m

-2 s

-1)

Total vegetation

Irradiance (µmol m-2 s-1)

0

Figure 5.9. Light-response curves of net photosyn-thesis in plants acclimated to low, intermediate, andhigh light. Horizontal arrows show the range of irra-diance over which net photosynthesis is positive andresponds linearly to irradiance for each species andfor the vegetation as a whole. Acclimation increasesthe range of irradiance over which net photosynthe-sis responds linearly to light (i.e., has a constantLUE).


carbon fixation by the canopy as a whole. Thisis one of many examples in ecosystem ecologyin which selection for optimal performance atthe level of individuals gives rise to predictablepatterns at the level of ecosystems.

Leaf area is the major factor governing thelight environment experienced by individualleaves within the canopy. There is a maximumleaf area that plants in an ecosystem can attain,because light is a directional resource that isgreatest at the top of the canopy and decreasesexponentially within the canopy, according tothe following equation:

I = I0e-kL (5.1)

where I is the irradiance (the quantity ofradiant energy received at a surface per unittime) at any point in the canopy, Io is the irra-diance at the top of the canopy; k is the extinc-tion coefficient, and L is the projected leaf areaindex (LAI; the leaf area per unit of groundarea) above the point of measurement.

LAI is a key parameter governing ecosystemprocesses because it determines the light atten-uation through a canopy and strongly influ-ences the capacity of vegetation to gain carbonand transfer water and energy to the atmos-phere (see Chapter 4). LAI has been defined intwo ways: (1) Projected LAI is the leaf areaprojected onto a horizontal plane. (2) TotalLAI is the total surface area of leaves, includ-ing the upper and lower surface of flat leavesand the cylindrical surface of conifer needles.Total LAI is approximately twice the value ofprojected LAI, except in the case of coniferneedles, for which the projected leaf area ismultiplied by p (3.14) to get total leaf area.Total LAI is particularly useful in describingthe effective leaf area of conifer forests, inwhich leaves are more cylindrical than flat. Aflat leaf cannot absorb more photons thanmove through a horizontal plane, because lightis a directional resource. Conifer needles can,however, absorb considerably more light perunit of projected leaf area than flat leaves.Conifer needles are particularly effective inabsorbing diffuse light, which provides a moreuniform illumination of the overall canopy.Diffuse light makes up a larger proportion oftotal irradiance at low sun angles and under

cloudy conditions. This may explain the pre-dominance of conifers in high-latitude forests,where sun angles are low and in temperate rain forests, where conditions are usuallycloudy. Unfortunately, there is no consistentagreement on whether data should beexpressed as projected LAI or total LAI, andmany review papers do not specify clearlywhich definition of LAI is being used. Microm-eteorologists measuring radiation transfer andecologists working with broad-leaved foreststend to use projected LAI, whereas ecophysi-ologists interested in carbon exchange andecologists working in conifer forests tend to usetotal LAI. Each definition of LAI has advan-tages for addressing particular questions. Theimportant thing is to specify which definition isbeing used.

Projected LAI varies widely among ecosys-tems but typically has values of 1 to 8m2 leafm-2 ground for ecosystems with a closedcanopy. The extinction coefficient is a constantthat describes the exponential decrease in irra-diance through a canopy. It is low for verticallyinclined or small leaves (e.g., 0.3 to 0.5 forgrasses), allowing substantial light penetrationinto the canopy, but high for near-horizontalleaves (0.7 to 0.8). Clumping of leaves aroundstems, as in conifers, and variable leaf anglesresult in intermediate values for k. Equation 5.1indicates that light is distributed unevenly in anecosystem and that the leaves near the top ofthe canopy capture most of the available light.Irradiance at the ground surface of a forest, forexample, is often only 1 to 2% of that at the topof the canopy. At very low irradiance, leaf respiration completely offsets photosyntheticcarbon gain, resulting in zero net photosynthe-sis, the light compensation point of the leaf (Fig.5.8). A mature shaded leaf typically does notimport carbon from the rest of the plant, so theleaf senesces and dies if it falls below the lightcompensation point for extended periods oftime. This puts an upper limit on the leaf areathat an ecosystem can support, regardless ofhow favorable the climate and the supply of soilresources may be.

Do differences in light availability explainthe differences among ecosystems in carbongain? In midsummer, when plants of most


ecosystems are photosynthetically active, thedaily input of visible light is nearly as great inthe Arctic as in the tropics but is spread overmore hours and is more diffuse at high latitudes(Billings and Mooney 1968). The greater dailycarbon gain in the tropics than at high latitudesis therefore unlikely to be a simple function of the light available to drive photosynthesis.Neither can variation in light availability due tocloudiness explain differences among ecosys-tems in energy capture. The most productiveecosystems on Earth, the tropical and temper-ate rain forests, have a high frequency of cloudi-ness, whereas arid grasslands and deserts, whichare less cloudy and receive nearly 10-fold morelight annually, are less productive. Seasonal and interannual variations in irradiance can,however, contribute to temporal variation incarbon gain by ecosystems. Aerosols emitted by volcanic eruptions and fires, for example,can reduce solar irradiance and photosynthesisover large areas in particular years. In summary,light availability strongly influences daily andseasonal patterns of carbon input and the distribution of photosynthesis within thecanopy, but it is only a minor factor explainingregional variations in carbon inputs to ecosys-tems (Fig. 5.1).

CO2 Limitation

Changes in stomatal conductance by leavesminimize the effects of CO2 supply on photo-synthesis. The free atmosphere is sufficientlywell mixed that its CO2 concentration variesglobally by only 4%—not enough to cause sig-nificant regional variation in photosynthesis. Indense canopies, photosynthesis reduces CO2

concentration somewhat. The shade leaves thatexperience this low CO2, however, tend to belight limited and therefore are relatively unre-sponsive to CO2 concentration. Consequently,even this decline in CO2 concentration withinthe canopy causes little variation in whole-ecosystem photosynthesis (Field 1991). In otherwords, CO2 differs from other resourcesrequired by plants in that plant uptake of CO2

does not greatly deplete the availability of thisresource to other plants (Rastetter and Shaver1992).

Although spatial variation in CO2 concen-tration does not explain much of the globalvariation in photosynthetic rate, the continuedworldwide increases in atmospheric CO2 con-centration (see Fig. 6.11) could cause a generalincrease in carbon gain by ecosystems. A dou-bling of the CO2 concentration to which leavesare exposed, for example, leads to a 30 to 50%increase in photosynthetic rate (Curtis andWang 1998). Enhancement of photosynthesisby addition of CO2 is most likely to occur inecosystems dominated by C3 plants, whose pho-tosynthetic rate is not CO2 saturated at currentatmospheric concentrations (Fig. 5.6). Themagnitude of this stimulation of photosynthe-sis by rising atmospheric CO2 concentration is,however, uncertain. Herbaceous plants anddeciduous trees (but not conifers) sometimesacclimate to increased CO2 concentration byreducing photosynthetic capacity and stomatalconductance (Ellsworth 1999, Mooney et al.1999), as expected from our hypothesis of co-limitation of photosynthesis by biochemistryand diffusion. In other cases, acclimation has no effect on photosynthetic rate and stomatalconductance (Curtis and Wang 1998). The down regulation of CO2 uptake in response toelevated CO2 enables plants to gain similaramounts of carbon while minimizing water loss.It also causes photosynthesis to respond lessstrongly to elevated CO2 than we might expectfrom a simple extrapolation of a CO2-responsecurve of photosynthesis (Fig. 5.6).

Over the long term, indirect effects of ele-vated CO2 may become important. In dry envi-ronments, for example, the reduced stomatalconductance caused by elevated CO2 leads to a decline in transpiration, which reduces therate at which water is lost from the soil andincreases soil moisture. Elevated CO2 often hasa greater effect on plant growth through thereduction of moisture limitation than through a direct stimulation of photosynthesis by ele-vated CO2. These indirect effects complicatethe prediction of long-term responses ofecosystem carbon gain to elevated CO2. Giventhat the atmospheric CO2 concentration hasincreased 30% (by 90 parts per million byvolume; ppmv) since the beginning of theIndustrial Revolution, it is important to under-


stand and predict these indirect effects of ele-vated CO2 on carbon gain by ecosystems.

Photosynthesis in C4 plants is relatively unre-sponsive to CO2 concentration in the short termbecause PEP carboxylase is highly effective indrawing down CO2 concentration inside theleaf. This leads to the prediction that C4 plantsmight be displaced by C3 plants if atmosphericCO2 concentration continues to increase, just asoccurred in tropical grasslands when CO2 con-centration rose at the end of the last glaciation(Cerling 1999). C4 plants are, however, oftenjust as sensitive to the indirect effects of CO2

(e.g., increased soil moisture) as are C3 plants,so the long-term effects of elevated CO2 on thecompetitive balance of C3 and C4 plants are dif-ficult to predict (Mooney et al. 1999).

Nitrogen Limitation andPhotosynthetic Capacity

Plant species differ 10- to 50-fold in their pho-tosynthetic capacity. Photosynthetic capacity isthe photosynthetic rate per unit leaf mass mea-sured under favorable conditions of light, mois-ture, and temperature. It is a measure of thecarbon-gaining potential per unit of biomassinvested in leaves. Photosynthetic capacity cor-relates strongly with leaf nitrogen concentra-tion (Fig. 5.10) (Field and Mooney 1986, Reichet al. 1997, Reich et al. 1999) because photosyn-thetic enzymes account for a large proportion ofthe nitrogen in leaves (Fig. 5.1). Many ecologi-cal factors can lead to a high leaf nitrogen con-centration and therefore a high photosyntheticcapacity. Plants growing in high-nitrogen soils,for example, have higher tissue nitrogen con-centrations and photosynthetic rates than dothe same species growing on less fertile soils.

This acclimation of plants to a high nitrogensupply contributes to the high photosyntheticrates in agricultural fields and other ecosystemswith a rapid nitrogen turnover. Many speciesdiffer in their nitrogen concentration, evenwhen growing in the same soils. Speciesadapted to productive habitats usually produceleaves that are short lived and have high tissuenitrogen concentrations and high photosyn-thetic rates. Nitrogen-fixing plants also typicallyhave high leaf nitrogen concentrations and cor-

respondingly high photosynthetic rates. Envi-ronmental stresses that cause plants to produceleaves with a low leaf nitrogen concentrationresult in low photosynthetic capacity. In sum-mary, regardless of the cause of variation in leafnitrogen concentration, there is always a strongpositive correlation between leaf nitrogen con-centration and photosynthetic capacity (Fieldand Mooney 1986) (Fig. 5.10).

Plants with a high photosynthetic capacityhave a high stomatal conductance, in theabsence of environmental stress (Fig. 5.11), asexpected from our hypothesis of co-limitationof photosynthesis by biochemistry and diffu-sion. This enables plants with a high photosyn-thetic capacity to gain carbon rapidly, at thecost of high rates of water loss. Conversely,species with a low photosynthetic capacity con-serve water as a result of their lower stomatalconductance.

There appears to be an unavoidable trade-offbetween traits that maximize photosyntheticrate and traits that maximize leaf longevity(Fig. 5.12) (Reich et al. 1997, Reich et al. 1999).Many species of plants that grow in low-nutrient environments produce long-lived

1000

100

10

17 21 63

Leaf nitrogen (mg g-1)

Net

pho

tosy

nthe

sis

(nm

ol g

-1 s

-1)

Figure 5.10. Relationship between leaf nitrogenconcentration and maximum photosynthetic capac-ity for plants from Earth’s major biomes. Opencircles and the solid regression line are for 11 speciesfrom six biomes using a common methodology. Exesand the dashed regression line are data from the lit-erature. (Redrawn with permission from Proceedingsof the National Academy of Sciences U. S. A., Vol. 94© 1997 National Academy of Sciences, USA; Reichet al. 1997.)


leaves because there are insufficient nutrientsto support rapid leaf turnover (Chapin 1980).Shade-tolerant species also produce longer-lived leaves than do shade-intolerant species(Walters and Reich 1999). Long-lived leavestypically have a low leaf nitrogen concentrationand a low photosynthetic capacity; they musttherefore photosynthesize for a relatively longtime to break even in their lifetime carbonbudget (Chabot and Hicks 1982, Gulmon andMooney 1986, Reich et al. 1997). To survive,long-lived leaves must have sufficient structuralrigidity to withstand drought and/or winter des-iccation. These structural requirements causeleaves to be dense—that is, to have a smallsurface area per unit of biomass, termed specificleaf area (SLA). Long-lived leaves must also be well defended against herbivores andpathogens, if they are to persist. This requires

substantial allocation to lignin, tannins, andother compounds that deter herbivores, butalso contribute to tissue mass and a low SLA.Many woody plants in dry environments also produce long-lived leaves. For the samereasons, these leaves typically have a low SLAand a low photosynthetic capacity (Reich et al.1999).df

tu

td

dcmo

co

gl

trscte

ce

bc

15

Max

imum

sto

mat

al c

ondu

ctan

ce (

mm

s-1

)

10

5

00 10 20 30 40

Leaf nitrogen concentration (mg g-1)

Figure 5.11. Relationship between leaf nitrogenconcentration and maximal stomatal conductance of plants from Earth’s major biomes. Each point and its standard error represent a differentbiome. bc, Broad-leaved crops; ce, cereal crops; co,evergreen conifer forest; dc, deciduous conifer forest;df, tropical deciduous forest; gl, grassland; mo, mon-soonal forest; sc, sclerophyllous shrub; td, temperatedeciduous broad-leaved forest; te, temperate evergreen broad-leaved forest; tr, tropical rain forest; tu, herbaceous tundra. (Redrawn with per-mission from the Annual Review of Ecology and Systematics, Vol. 25 © 1994 by Annual Reviews,www.AnnualReviews; Schulze et al. 1994.)

1000

1000

100

100

100

1010

101

1N

et p

hoto

synt

hesi

s (n

mol

g-1

s-1

)Le

af n

itrog

en (

mg

g-1)

Leaf lifespan (months)

Spe

cific

leaf

are

a (c

m2 g

-1)

110 100

Figure 5.12. The effect of leaf life span on photo-synthetic capacity (photosynthetic rate measuredunder favorable conditions), leaf nitrogen concen-tration, and specific leaf area. Symbols same as inFigure 5.10. (Redrawn with permission from Pro-ceedings of the National Academy of Sciences U. S. A., Vol. 94 © 1997 National Academy of Sciences, USA; Reich et al. 1997.)


The broad relationship among species withrespect to photosynthetic rate and leaf life span is similar in all biomes; a 10-fold decreasein leaf life span gives rise to about a 5-foldincrease in photosynthetic capacity (Reich et al. 1999). Species with long-lived leaves, lowphotosynthetic capacity, and low stomatal con-ductance are common in all low-resource envi-ronments, including those that are dry, infertile,or shaded.

Plants in productive environments, in con-trast, produce short-lived leaves with a hightissue nitrogen concentration and a high pho-tosynthetic capacity; this allows a large carbonreturn per unit of biomass invested in leaves, ifsufficient light is available. These leaves have ahigh SLA, which maximizes the quantity of leafarea displayed and the light captured per unitof leaf mass. The resulting high rates of carbongain support a high maximum relative growthrate in the absence of environmental stress or competition from other plants (Fig. 5.13)(Schulze and Chapin 1987). Many early succes-sional habitats, such as recently abandonedagricultural fields or postfire sites, have suffi-

cient light, water, and nutrients to support highgrowth rates and are characterized by specieswith short-lived leaves, high tissue nitrogenconcentration, high SLA, and high photosyn-thetic rates (see Chapter 13). Even in late succession, environments with high water and nutrient availability are characterized byspecies with relatively high nitrogen concentra-tions and photosynthetic rates. Plants in thesehabitats can grow quickly to replace leavesremoved by herbivores or to fill canopy gapsproduced by death of branches or individuals.

In summary, plants produce leaves with acontinuum of photosynthetic characteristics,ranging from short-lived thin leaves with a highnitrogen concentration and high photosyn-thetic rate to long-lived dense leaves with a lownitrogen concentration and low photosyntheticrate. These correlations among traits are so consistent that SLA is often used in ecosystemcomparisons as an easily measured index ofphotosynthetic capacity (Fig. 5.14).

There is only modest variation in photosyn-thetic capacity per unit leaf area because leaveswith a high photosynthetic capacity per unitleaf biomass also have a high SLA. Photosyn-thetic capacity calculated per unit leaf area(Parea) is a measure of the capacity of leaves to

1

23

45

6

7

0

2.0

1.5

1.0

0.5

05 10 15 20 3025

Maximum photosynthetic rate (µmol m-2 s-1)

Max

imum

rel

ativ

e gr

owth

rat

e (g

g-1

wee

k-1)

Figure 5.13. Relationship between photosyntheticrate and relative growth rate for major plant growthforms. 1, Agricultural crop species; 2, herbaceous sunspecies; 3, grasses and sedges; 4, summer deciduoustrees; 5, evergreen and deciduous dwarf shrubs; 6,herbaceous shade species and bulbs; 7, evergreenconifers. (Redrawn with permission from Springer-Verlag; Schulze and Chapin 1987.)

1000

100

101

Net

pho

tosy

nthe

sis

(nm

ol g

-1 s

-1)

70 490

Specific leaf area (cm2 g-1)

10

Figure 5.14. The relationship between SLA andphotosynthetic capacity. The consistency of this rela-tionship makes it possible to use SLA as an easilymeasured index of photosynthetic capacity. Symbolssame as in Figure 5.10. (Redrawn with permissionfrom Proceedings of the National Academy of Sci-ences U. S. A., Vol. 94 © 1997 National Academy ofSciences USA; Reich et al. 1997.)


capture a unit of incoming radiation. It is cal-culated by dividing photosynthetic capacity perunit leaf mass (Pmass) by SLA.

(5.2)

where (gcm-2 s-1) = (gg-1 s-1)/(cm2 g-1).There is relatively little variation in Parea

among plants from different ecosystems(Lambers and Poorter 1992). In productivehabitats, both mass-based photosynthesis andSLA are high (Fig. 5.12). In unproductive habi-tats both of these parameters are low, resultingin modest variation in area-based photosyn-thetic rate.To the extent that Parea varies amongplants, it is highest in species with short-livedleaves (Reich et al. 1997). There is, however, nostrong pattern of photosynthetic rate per unitarea among ecosystems (Lambers and Poorter1992). Mass-based photosynthetic capacity is agood measure of the physiological potential forphotosynthesis (the photosynthetic rate perunit of biomass invested in leaves). Area-basedphotosynthetic capacity is a good measure ofthe effectiveness of these leaves at the ecosys-tem scale (photosynthetic rate per unit of avail-able light). Variation in soil resources has amuch greater effect on the quantity of leaf areaproduced than on the photosynthetic capacityper unit leaf area.

Water Limitation

Water limitation reduces the capacity of indi-vidual leaves to match CO2 supply with lightavailability. Water stress is often associated withhigh light because sunny conditions correlatewith low precipitation (low water supply) andwith low humidity (high rate of water loss).High light also leads to an increase in leaf tem-perature and water vapor concentration insidethe leaf and therefore greater water loss bytranspiration (see Chapter 4). The high-lightconditions in which a plant would be expectedto increase stomatal conductance to minimizeCO2 limitations to photosynthesis are thereforeoften the same conditions in which the result-ing transpirational water loss is greatest andmost detrimental to the plant. This trade-offbetween a response that maximizes carbon gain

PP

areamass

SLA=

(stomata open) and one that minimizes waterloss (stomata closed) is typical of the physio-logical compromises faced by plants whosephysiology and growth may be limited by morethan one environmental resource (Mooney1972). When water supply is abundant, leavestypically open their stomata in response to highlight, despite the associated high rate of waterloss. As leaf water stress develops, stomatal conductance declines to reduce water loss. Thisdecline in stomatal conductance reduces pho-tosynthetic rate and the efficiency of using lightto fix carbon (i.e., LUE) below levels found inunstressed plants.

Plants that are acclimated and adapted to dryconditions reduce their photosynthetic capacityand leaf nitrogen content toward a level thatmatches the low stomatal conductance that isnecessary to conserve water in these environ-ments (Wright et al. 2001). A high photosyn-thetic capacity provides little benefit if the plantmust maintain a low stomatal conductance tominimize water loss. Conversely, low nitrogenavailability or other factors that constrain leaf nitrogen concentration result in leaves with low stomatal conductance (Fig. 5.11). Thisstrong correlation between photosyntheticcapacity and stomatal conductance maintainsthe balance between photosynthetic capacityand CO2 supply—that is, the co-limitation of photosynthesis by diffusional and biochemi-cal processes. In addition to their low photosynthetic capacity and low stomatal conductance, plants in dry areas minimize water stress by reducing leaf area (by sheddingleaves or producing fewer new leaves). Somedrought-adapted plants produce leaves thatminimize radiation absorption; their leavesreflect most incoming radiation or are steeplyinclined toward the sun (see Chapter 4)(Ehleringer and Mooney 1978). High radiationabsorption is a disadvantage in dry environ-ments because it increases leaf temperature,which increases respiratory carbon loss (seeChapter 6) and transpirational water loss (seeChapter 4).

Thus there are several mechanisms by whichplants in dry environments reduce radiationabsorption and photosynthetic capacity to con-serve water and carbon. The low leaf area, the


reflective nature of leaves, and the steep angleof leaves are the main factors accounting forthe low absorption of radiation and low carboninputs in dry environments. In other words,plants adjust to dry environments primarily byaltering leaf area and radiation absorptionrather than by altering photosynthetic capacityper unit leaf area.

Water use efficiency (WUE) of photosynthe-sis is defined as the carbon gain per unit ofwater lost. Water use is quite sensitive to thesize of stomatal openings, because stomatalconductance has slightly different effects on therates of CO2 entry and water loss.Water leavingthe leaf encounters two resistances to flow: thestomata and the boundary layer of still air onthe leaf surface (Fig. 5.4). Resistance to CO2

diffusion from the bulk air to the site of photosynthesis includes the same stomatal andboundary layer resistances plus an additionalinternal resistance associated with diffusion ofCO2 from the cell surface into the chloroplastand any biochemical resistances associated withcarboxylation. Because of this additional resis-tance to CO2 movement into the leaf, anychange in stomatal conductance has a propor-tionately greater effect on water loss than oncarbon gain. In addition, water diffuses morerapidly than does CO2 because of its smallermolecular mass and because of the steeper con-centration gradient that drives diffusion acrossthe stomata. For all these reasons, as stomataclose, water loss declines to a greater extentthan does CO2 absorption. The low stomatalconductance of plants in dry environmentsresults in less photosynthesis per unit of timebut greater carbon gain per unit of water loss—that is, greater WUE. Plants in dry environ-ments also enhance WUE by maintaining asomewhat higher photosynthetic capacity thanwould be expected for their stomatal conduc-tance, thereby drawing down the internal CO2

concentration and maximizing the diffusiongradient for CO2 entering the leaf (Wright et al.2001). Carbon isotope ratios in plants providean integrated measure of WUE during plantgrowth because the 13C concentration of newlyfixed carbon increases under conditions of lowinternal CO2 concentration (Ehleringer 1993)(Box 5.1). C4 and CAM photosynthesis are

additional adaptations that augment WUE ofecosystems.

Temperature Effects

Extreme temperatures limit carbon uptake.Photosynthetic rate is typically highest nearleaf temperatures commonly experienced onsunny days (Fig. 5.15). Leaf temperature maydiffer substantially from air temperature due tothe cooling effects of transpiration, the effectsof leaf surface properties on light absorption,and the influence of adjacent surfaces on thethermal and radiation environment of the leaf(see Chapter 4). At low temperatures, photo-synthesis is limited directly by temperature, asare all chemical reactions. At high tempera-tures, photosynthesis also declines, due toincreased photorespiration and, under extremeconditions, enzyme inactivation and des-truction of photosynthetic pigments. Tempera-ture extremes often have a greater effect onphotosynthesis than does average temperaturebecause of damage to photosynthetic machin-ery (Waring and Running 1998).

There are several factors that minimize thesensitivity of photosynthesis to temperature.The enzymatically controlled carbon-fixationreactions are typically more sensitive to low

Tissue temperature (οC)

100

80

60

40

20

0-5 0 5 10 15 20 25 30 35 40

NeuropogonAtriplex

Tidestromia Ambrosia

Pho

tosy

nthe

sis

(% o

f max

imum

)

Antarctic Desert Coastal dune

Figure 5.15. Temperature response of photosynthe-sis in plants from contrasting temperature regimes.Species include antarctic lichen (Neuropogonacromelanus), a cool coastal dune plant (Ambrosiachamissonis), an evergreen desert shrub (Atriplexhymenelytra), and a summer-active desert perennial(Tidestromia oblongifolia). (Redrawn with permis-sion from Blackwell Science, Ltd; Mooney 1986.)

Gross Primary Production 115

temperature than are the biophysically con-trolled light-harvesting reactions. Carbon fixation reactions therefore tend to limit pho-tosynthesis at low temperature. Plants adaptedto cold climates compensate for this by pro-ducing leaves with high concentrations of leafnitrogen and photosynthetic enzymes, whichenable carboxylation to keep pace with theenergy supply from the light-harvesting reac-tions (Berry and Björkman 1980). This explainswhy arctic and alpine plants typically have high leaf nitrogen concentrations despite lowsoil nitrogen availability (Körner and Larcher1988). Plants in cold environments also havehairs and other morphological traits that raise leaf temperature above air temperature(Körner 1999). In hot environments with anadequate water supply, plants produce leaveswith high photosynthetic rates. The associatedhigh transpiration rate can cool the leaf, so leaftemperatures are much lower than air temper-atures. In hot, dry environments, plants closestomata to conserve water, and the coolingeffect of transpiration is reduced. Plants inthese environments often produce small leavesthat shed heat effectively and maintain tem-peratures close to air temperature (see Chapter4). In summary, despite the sensitivity of pho-tosynthesis to short-term variation in tempera-ture, leaf properties minimize the differences inleaf temperature among ecosystems, and plantsacclimate and adapt so there is no clear rela-tionship between temperature and averagephotosynthetic rate in the field, when ecosys-tems are compared.

Pollutants

Pollutants reduce carbon gain primarily byreducing leaf area or photosynthetic capacity.Many pollutants, such as sulfur dioxide (SO2)and ozone, reduce photosynthesis through theireffects on growth and the production of leafarea. Pollutants also directly reduce photosyn-thesis by entering the stomata and damagingthe photosynthetic machinery, thereby reduc-ing photosynthetic capacity (Winner et al.1985). Plants then reduce stomatal conductanceto balance CO2 uptake with the reduced capac-ity for carbon fixation.This reduces the entry of

pollutants into the leaf, reducing the vulnera-bility of the leaf to further injury. Plantsgrowing in low-fertility or dry conditions arepreadapted to pollutant stress because theirlow stomatal conductance minimizes the quan-tity of pollutants entering leaves. These plantsare therefore less affected by pollutants thanare rapidly growing crops and other plants withhigh stomatal conductance.

Gross Primary Production

Gross primary production is the sum of thephotosynthesis by all leaves measured at theecosystem scale. It is typically integrated overtime periods of days to a year (gCm-2 ofground yr-1) and is the process by which carbonand energy enter ecosystems. GPP is generallyestimated from simulation models rather thanmeasured directly, because it is impossible tomeasure the net carbon exchange of all theleaves in the canopy in isolation from otherecosystem components (e.g., respiration bystems and soil) and without modifying the ver-tical gradient in environment. The results ofthese modeling studies suggest that most con-clusions derived from leaf-level measurementsof net photosynthesis can be extended to theecosystem scale, when the vertical profiles ofphotosynthetic capacity and environment areconsidered. Measurement of whole-ecosystemcarbon exchange provides another way to esti-mate GPP (see Chapter 6).

Canopy Processes

The vertical profile of leaf photosynthetic properties in a canopy maximizes GPP. In mostclosed-canopy ecosystems, photosyntheticcapacity decreases exponentially through thecanopy in parallel with the exponential declinein irradiance (Eq. 5.1) (Hirose and Werger1987).This matching of photosynthetic capacityto light availability is the response we wouldexpect from individual leaves within thecanopy, because it maintains the co-limitationof photosynthesis by diffusion and biochemicalprocesses in each leaf. It also serves to maxi-mize GPP in closed-canopy ecosystems. The


matching of photosynthetic capacity to lightavailability occurs through the preferentialtransfer of nitrogen to leaves at the top of thecanopy. At least three processes cause this tohappen. (1) Sun leaves at the top of the canopydevelop more cell layers than shade leaves andtherefore contain more nitrogen per unit leafarea. (2) New leaves are produced primarily atthe top of the canopy, causing nitrogen to betransported to the top of the canopy (Field1983, Hirose and Werger 1987). (3) Leaves atthe bottom of the canopy senesce when theybecome shaded below their light compensationpoint. Much of the nitrogen resorbed fromthese senescing leaves (see Chapter 8) is trans-ported to the top of the canopy to support theproduction of young leaves with high photo-synthetic capacity. The accumulation of nitrogen at the top of the canopy is most pro-nounced in dense canopies, which developunder circumstances of high water and nitrogenavailability (Field 1991). In environments inwhich leaf area is limited by water, nitrogen, ortime since disturbance, there is less advantageto concentrating nitrogen at the top of thecanopy, because light is abundant throughoutthe canopy. In these canopies, light availability,nitrogen concentrations, and photosyntheticrates are more uniformly distributed throughthe canopy.

Canopy-scale relationships between lightand nitrogen appear to occur even in multi-species communities. In a single individual,there is an obvious selective advantage to opti-mizing nitrogen distribution within the canopybecause this provides the greatest carbonreturn per unit of nitrogen invested in leaves.We know less about the factors governingcarbon gain in multispecies stands. In suchstands, the individuals at the top of the canopyaccount for most of the photosynthesis and maybe able to support greater root biomass toacquire more nitrogen, compared to smallersubcanopy or understory individuals (Hikosakaand Hirose 2001). This specialization amongindividuals probably contributes to the verticalscaling of nitrogen and photosynthesis that isobserved in multispecies stands.

Vertical gradients in other environmentalvariables often reinforce the maximization of

carbon gain near the top of the canopy. Thecanopy modifies not only light availability but also other variables that influence photo-synthetic rate, including wind speed, tempera-ture, relative humidity, and CO2 concentration(Fig. 5.16). The most important of these effectsis the decrease in wind speed from the free atmosphere to the ground surface. The friction of air moving across Earth’s surfacecauses wind speed to decrease exponentiallyfrom the free atmosphere to the top of thecanopy. In other words, Earth’s surface createsa boundary layer similar to that which developsaround individual leaves (Fig. 5.4). Wind speedcontinues to decrease from the top of thecanopy to the ground surface in ways thatdepend on canopy structure. Smooth canopies,characteristic of crops or grasslands, show agradual decrease in wind speed from the top ofthe canopy to the ground surface, whereasrough canopies, characteristic of many forests,create more friction and turbulence, whichincrease the vertical mixing of air within thecanopy (see Chapter 4) (McNaughton andJarvis 1991). For this reason, gas exchange inrough canopies is more tightly coupled to con-ditions in the free atmosphere than in smoothcanopies.

Tem

pera

ture

Vapo

rpr

essu

re

Win

dsp

eed

Leaf area

1o C 0.1 kPa 1 m s-1

Tree

hei

ght

Figure 5.16. Typical vertical gradients in tempera-ture, vapor pressure, and wind speed in a forest.Temperature is highest in the midcanopy wheremost energy is absorbed. The increase in wind speedat the bottom of the canopy occurs in open forests,where there is little understory. (Redrawn with permission from Academic Press; Landsberg andGower 1997.)


Wind speed is important because it reducesthe thickness of the boundary layer of still airaround each leaf, producing steeper gradientsin temperature and in concentrations of CO2

and water vapor from the leaf surface to theatmosphere. This speeds the diffusion of CO2

into the leaf and the loss of water from the leaf,enhancing both photosynthesis and transpira-tion.A reduction in thickness of the leaf bound-ary layer also brings leaf temperature closer toair temperature. The net effect of wind on pho-tosynthesis is generally positive at moderatewind speeds and adequate moisture supply,enhancing photosynthesis at the top of thecanopy. When low soil moisture or a longpathway for water transport from the soil to the top of the canopy reduces water supply tothe uppermost leaves, as in tall forests, theuppermost leaves reduce their stomatal con-ductance, causing the zone of maximum photo-synthesis to shift farther down in the canopy.Although multiple environmental gradientswithin the canopy have complex effects on pho-tosynthesis, they probably enhance photosyn-thesis near the top of canopies in ecosystemswith sufficient water and nutrients to developdense canopies.

Canopy properties extend the range of lightavailability over which the LUE of the canopyremains constant. The light-response curve ofcanopy photosynthesis, measured in closedcanopies (total LAI greater than about 3), sat-urates at higher irradiance than does photo-synthesis by a single leaf (Fig. 5.17) (Jarvis andLeverenz 1983). This increase in the efficiencyof converting light energy into fixed carbonoccurs for several reasons. The less horizontalangle of leaves at the top of the canopy reducesthe probability of light saturation of the upperleaves and increases the light penetration intothe canopy. The clumped distribution of leavesin shoots, branches, and crowns also increaseslight penetration into the canopy. Conifercanopies are particularly effective in distribut-ing light through the canopy due to the clump-ing of needles around stems. This could explainwhy conifer forests frequently support a higherLAI than deciduous forests. The light compen-sation point also decreases from the top to thebottom of the canopy (Fig. 5.9), so lower leaves

maintain a positive carbon balance, despite the relatively low light availability. In cropcanopies, where water and nutrients are highlyavailable, the linear relationship betweencanopy carbon exchange and irradiance (i.e.,constant LUE) extends up to irradiance typicalof full sunlight. In other words, there is no evi-dence of light saturation, and LUE remainsconstant over the full range of natural lightintensities (Fig. 5.18) (Ruimy et al. 1996).

Satellite Estimates of GPP

The similarity among ecosystems of canopyLUE provides a basis for estimating carboninput to ecosystems globally, using remotesensing. An important conclusion of leaf- andcanopy-level studies of photosynthesis is thatthere are many factors that cause ecosystems toconverge toward a relatively similar efficiency ofconverting light energy into carbohydrates. (1)All C3 plants have a similar quantum yield(LUE) at low to moderate irradiance. (2) Pene-tration of light and vertical variations in photo-

0

10

1500

0

-10

20

30

40

1000500

Leaf

Pho

tosy

nthe

sis

(µm

ol m

-2s-1

)

Can

opy


Figure 5.17. Light-response curve of a single leafand a forest canopy. Canopies maintain a constantLUE (linear response of photosynthesis to light)over a broader range of light availability than doindividual leaves. (Reprinted with permission fromAdvances in Ecological Research; Ruimy et al.1996.)


synthetic properties through a canopy extendthe range of irradiance over which the LUEremains constant. (3) Light use efficiency isreduced primarily by short-term environmentalstresses that cause plants to reduce stomatalconductance. Over the long term, however,plants respond to such stresses by reducing the concentrations of photosynthetic pigmentsand enzymes so photosynthetic capacitymatches stomatal conductance and by reducingleaf area. In other words, plants in low-resourceenvironments reduce the amount of lightabsorbed more strongly than they reduce theefficiency with which absorbed light is converted to carbohydrates. Modeling studiessuggest that light use efficiency varies about twofold among ecosystems (Field 1991),although this is difficult to demonstrate conclusively because GPP cannot be directlymeasured.

If LUE is indeed similar among ecosystems,GPP could be estimated by determining thequantity of light absorbed by ecosystems, whichcan be measured from satellites. Leaves at thetop of the canopy have a disproportionatelylarge effect on the light that is both absorbed

and reflected by the ecosystem. The reflectedradiation can also be measured by satellites.This similarity in bias between the vertical dis-tribution of absorbed and reflected radiationmakes satellites an ideal tool for estimatingcanopy photosynthesis.The challenge, however,is to estimate the fraction of absorbed radiationthat has been absorbed by leaves rather than bysoil or other nonphotosynthetic surfaces. Vege-tation has a different spectrum of absorbed andreflected radiation than does the atmosphere,water, clouds, and bare soil.This occurs becausechlorophyll and associated light-harvesting pig-ments or accessory pigments, which are con-centrated at the canopy surface, absorb visiblelight (VIS) effectively. The optical propertiesthat result from the cellular structure of leaves,however, makes them highly reflective in thenear infrared (NIR) range. Ecologists haveused these unique properties of vegetation togenerate an index of vegetation “greenness”:the normalized difference vegetation index(NDVI).

(5.3)NDVINIR VISNIR VIS

=-( )+( )

0 1000 1500 500 2000 0 1000 1500 500 2000

10

0

-10

20

30

40

50


Net

eco

syst

em e

xcha

nge

(µm

ol m

-2s-1

)

A B

Figure 5.18. Effect of vegetation type and irradi-ance on net ecosystem exchange in forests (A) andcrops (B). Forests maintain a relatively constantLUE up to 30 to 50% of full sun, although there is

considerable variability. Crops maintain a constantlight LUE over the entire range of naturally occur-ring irradiance. (Redrawn with permission fromAdvances in Ecological Research; Ruimy et al. 1996.)


Sites with a high rate of carbon gain gener-ally have a high NDVI because of their highchlorophyll content (low reflectance of VIS)and high leaf area (high reflectance of NIR).Species differences in leaf structure also influence infrared reflectance (and thereforeNDVI). Conifer forests, for example, generallyhave a lower NDVI than deciduous forestsdespite their greater leaf area. Consequently,NDVI must be used cautiously when compar-ing ecosystems dominated by different types ofplants (Verbyla 1995).

NDVI is ecologically useful because it is proportional to the amount of light energyabsorbed by photosynthetic tissues, when struc-turally similar stands of vegetation are com-pared (Fig. 5.19). This index of absorbedradiation is most useful in ecosystems with low to moderate NDVI, because the relation-ship saturates at high NDVI. To the extent thatlight use efficiency is similar among ecosystemtypes, the summation of absorbed radiationthrough the season of active plant growthshould be a reasonable index of carbon input toecosystems.

Controls over GPP

Ecosystem differences in GPP are determinedprimarily by leaf area index and the duration of the photosynthetic season and secon-darily by the environmental controls over photosynthesis.

Leaf Area

Variation in soil resource supply accounts formuch of the spatial variation in leaf area andGPP among ecosystem types. Analysis of satel-lite imagery shows that about 70% of the ice-free terrestrial surface has relatively opencanopies (Graetz 1991) (Fig. 5.20). GPP corre-lates closely with leaf area below a total LAI ofabout 8 (projected LAI of 4) (Schulze et al.1994), suggesting that leaf area is a criticaldeterminant of GPP on most of Earth’s terres-trial surface (Fig. 5.1). GPP is less sensitive to

FPAR from NDVI

Mea

sure

d F

PA

R

00

0.2

0.2

0.4

0.4

0.6

0.6

0.8

0.8

1.0

1.0

Figure 5.19. Relationship between the fraction of photosynthetically active radiation (FPAR)absorbed by vegetation estimated from satellitemeasurements of NDVI (x-axis) and FPAR mea-sured in the field (y-axis). Data were collected froma wide range of ecosystems, including temperate andtropical grasslands and temperate and boreal coniferforests. Satellites provide an approximate measure of the photosynthetically active radiation absorbedby vegetation and therefore the carbon inputs toecosystems. (Redrawn with permission from Journalof Hydrometeorology; Los et al. 2000.)

50

30

10

0.25

10 30 50 70 100 250 600 1000

Hei

ght o

f tal

lest

str

atum

(m

)

Projected foliage cover (%) of tallest stratum[log scale]

Opencanopies (70%)

Closedcanopies (30%)

Tropicalforests (12%)

Savanna(18%)

Grass/Shrublands(12%)

Crops(12%)Tundra

(7%)Desert (27%)

Borealforest (7%)

Temperateforest (7%)

2

Figure 5.20. Projected foliage cover and canopyheight of the major biomes. Typical values for thatbiome and the percentage of the terrestrial surfacethat it occupies are shown. The vertical line shows100% canopy cover. (Reprinted with permissionfrom Climatic Change, Vol. 18 © 1991 Kluwer Academic Publishers; Graetz 1991.)


LAI in dense canopies, because the leaves inthe middle and bottom of the canopy con-tribute relatively little to GPP. The availabilityof soil resources, especially water and nutrientsupply, is a critical determinant of LAI for two reasons: Plants in high-resource environ-ments produce a large amount of leaf biomass,and leaves produced in these environmentshave a high SLA—that is, a large leaf area perunit of leaf biomass. As discussed earlier, a highSLA maximizes light capture and thereforecarbon gain per unit of leaf biomass (Fig. 5.12)(Lambers and Poorter 1992, Reich et al. 1997).In low-resource environments, plants producefewer leaves, and these leaves have a lowerSLA. Ecosystems in these environments have alow LAI and therefore a low GPP.

Disturbances, herbivory, and pathogensreduce leaf area below levels that resources cansupport. Soil resources and light extinctionthrough the canopy determine the upper limitto the leaf area that an ecosystem can support.However, many factors regularly reduce leafarea below this potential LAI. Drought andfreezing are climatic factors that cause plants toshed leaves. Other causes of leaf loss includephysical disturbances (e.g., fire and wind) andbiotic agents (e.g., herbivores and pathogens).After major disturbances the remaining plantsmay be too small, have too few meristems, orlack the productive potential to produce theleaf area that could potentially be supported bythe climate and soil resources of a site. For thisreason, LAI tends to increase with time afterdisturbance to an asymptote and then (at leastin forests) often declines in late succession (seeChapter 13).

Human activities increasingly affect the leafarea of ecosystems in ways that cannot be predicted from climate. Overgrazing by cattle,sheep, and goats, for example, directly removesleaf area and causes shifts to vegetation typesthat are less productive and have less leaf areathan would otherwise occur in that climatezone. Acid rain and other pollutants also causeleaf loss. Nitrogen deposition can stimulate leafproduction above levels that would be pre-dicted from climate and soil type. Because ofhuman activities, LAI cannot be estimatedsimply from correlations with climate. Fortu-

nately, satellites provide the opportunity to estimate LAI directly. This information is animportant input to global models that calculateregional patterns of carbon input to terrestrialecosystems.

Length of the Photosynthetic Season

The length of the photosynthetic seasonaccounts for much of the ecosystem differencesin GPP. Most ecosystems experience times thatare too cold or too dry for significant photo-synthesis to occur. During winter in cold cli-mates and times with negligible soil water indry climates, plants either die (annuals), losetheir leaves (deciduous plants), or becomephysiologically dormant (some evergreenplants). During these times, there is negligiblecarbon uptake by the ecosystem, regardless oflight availability and CO2 concentration. In asense, the nonphotosynthetic season is simply acase of extreme environmental stress. Condi-tions are so severe that plants gain negligiblecarbon. At high latitudes and altitudes and indry ecosystems, this is probably the major con-straint on carbon inputs to ecosystems (Fig. 5.1;see Chapter 6) (Körner 1999). For annuals anddeciduous plants, the lack of leaf area is suffi-cient to explain the absence of photosyntheticcarbon gain in the nongrowing season. Lack of water or extremely low temperatures can,however, prevent even evergreen plants from gaining carbon. Some evergreen species partially disassemble their photosyntheticmachinery during the nongrowing season.These plants require some time following thereturn of favorable environmental conditionsto reassemble their photosynthetic machinery(Bergh and Linder 1999), so not all early-season irradiance is used efficiently to gaincarbon. In tropical ecosystems, however, whereconditions are more continuously favorable for photosynthesis, leaves maintain their photosynthetic machinery from the time they are fully expanded until they are shed.Models that simulate GPP often define thelength of the photosynthetic season in terms ofthresholds of minimum temperature or mois-ture below which plants do not produce leavesor do not photosynthesize.

Review Questions 121

Environmental Controls over Growing-Season GPP

Environmental controls over GPP during thegrowing season are identical to those describedfor net photosynthesis of individual leaves. Asdescribed earlier, variations in light, tempera-ture, and moisture account for much of thediurnal and seasonal variations in net photo-synthesis. These factors have a particularlystrong effect on leaves at the top of the canopy,which account for most GPP. The thinnerboundary layer and greater distance for watertransport from roots, for example, makes theuppermost leaves particularly sensitive to vari-ation in temperature, soil moisture, and relativehumidity. Consequently, the response of GPP todiurnal and seasonal variations in environmentis similar to that described for net photosyn-thesis by individual leaves. Variation in theseenvironmental factors explains much of thediurnal and seasonal variations in GPP. Soilresources (nutrients and moisture) influenceGPP through their effects on both photosyn-thetic potential and leaf area. Photosyntheticpotential is closely matched with the light-harvesting capacity of leaves. Soil resourcestherefore influence GPP primarily by deter-mining the capacity of canopies to capture light(through variations in leaf area and photosyn-thetic potential) rather than through variationsin the efficiency of converting light to carbohy-drates. The seasonal changes in GPP are sensi-tive to variations in the photosynthetic rate ofindividual leaves, due to variations in light and temperature; this causes variation in lightuse efficiency. In contrast, differences in GPPamong ecosystem types and among successionalstages may be determined more strongly by dif-ferences in the quantity of light absorbed as aresult of differences in leaf area and photosyn-thetic potential.

Summary

Most carbon enters terrestrial ecosystemsthrough photosynthesis mediated by plants.Thelight-harvesting reactions of photosynthesistransform light energy into chemical energy,

which is used by the carbon-fixation reactionsto convert CO2 to sugars. The enzymes thatcarry out these reactions account for about halfof the nitrogen in the leaf. Plants adjust thecomponents of photosynthesis so physical andbiochemical processes co-limit carbon fixation.At low light, for example, plants reduce thequantity of photosynthetic machinery per unitleaf area by producing thinner leaves. Asatmospheric CO2 concentration increases,plants reduce stomatal conductance. The neteffect of these adjustments is that ecosystemcarbon gain is relatively insensitive to differ-ences among ecosystems in light or CO2 avail-ability. The major environmental factors thatexplain differences among ecosystems incarbon gain are the length of time during whichconditions are suitable for photosynthesis andthe soil resources (water and nutrients) avail-able to support the production and mainte-nance of leaf area. Environmental stresses suchas inadequate water supply, extreme tempera-tures, and pollutants reduce the efficiency withwhich plants use light to gain carbon. Plantsrespond to these stresses by reducing leaf areaand nitrogen content so as to maintain a rela-tively constant efficiency in the use of light tofix carbon.

Review Questions

1. How do plants with different photosyntheticpathways differ in their photosyntheticresponses to water and nitrogen? What is thebiochemical basis for these differences inresponse to environment?

2. How does each major environmental vari-able (CO2, light, nitrogen, water, tempera-ture, pollutants) affect photosynthetic rate inthe short term? How do the photosyntheticproperties of individual leaves change inresponse to changes in these factors to opti-mize photosynthetic performance?

3. How does the response of photosynthesis toone environmental variable (e.g., water or nitrogen) affect the response to otherenvironmental variables (e.g., light, CO2, orpollutants)? Considering these interactionsamong environmental variables, how might


anthropogenic increases in nitrogen inputsaffect the response of Earth’s ecosystems torising atmospheric CO2?

4. How do environmental stresses affect lightuse efficiency in the short term? How does vegetation adjust to maximize LUE in stressful environments over the longterm?

5. What factors are most important in explain-ing differences among ecosystems in GPP?Over what time scale does each of thesefactors have its greatest impact on GPP?Explain your answers.

6. What factors most strongly affect leaf areaand photosynthetic capacity of vegetation?

7. How do the factors regulating photosynthe-sis in a forest canopy differ from those inindividual leaves? How does availability ofsoil resources (water and nutrients) and thestructure of the canopy influence the impor-tance of these canopy effects?

Additional Reading

Ehleringer, J.R., and C.B. Field, editors. 1993. ScalingPhysiological Processes: Leaf to Globe. AcademicPress, San Diego, CA.

Field, C., and H.A. Mooney. 1986. The photosynthesis-nitrogen relationship in wildplants. Pages 25–55 in T.J. Givnish, editors. On theEconomy of Plant Form and Function. CambridgeUniversity Press, Cambridge, UK.

Lambers, H., F.S. Chapin III, and T. Pons. 1998.Plant Physiological Ecology. Springer-Verlag,Berlin.

Larcher, W. 1995. Physiological Plant Ecology.Springer-Verlag, Berlin.

Mooney, H.A. 1972. The carbon balance of plants.Annual Review of Ecology and Systematics3:315–346.

Reich, P.B., M.B. Walters, and D.S. Ellsworth. 1997.From tropics to tundra: Global convergence inplant functioning. Proceedings of the NationalAcademy of Sciences U. S. A. 94:13730–13734.

Ruimy,A., P.G. Jarvis, D.D. Baldocchi, and B. Saugier.1996. CO2 fluxes over plant canopies and solarradiation: A review. Advances in EcologicalResearch 26:1–68.

Sage, R.F., and R.K. Monson, editors. 1999. C4 PlantBiology. Academic Press, San Diego, CA.

Schulze, E.-D., F.M. Kelliher, C. Körner, J. Lloyd, andR. Leuning. 1994. Relationship among maximumstomatal conductance, ecosystem surface conduc-tance, carbon assimilation rate, and plant nitrogennutrition: A global ecology scaling exercise.Annual Review of Ecology and Systematics25:629–660.

Introduction

The carbon balance of vegetation and ecosys-tems governs the productivity of the biosphereand the impact of ecosystems on the EarthSystem. Plant production determines theamount of energy available to sustain all organ-isms, including humans. We depend on plantproduction directly for food and fiber and indi-rectly because of the critical role of vegetationin all ecosystem processes (see Chapter 16).Much of the carbon produced by plants even-tually moves to the soil, where it influences thecapacity of soils to retain water and nutrientsand therefore to support plant production (seeChapter 3). Carbon cycling through ecosystemsalso directly affects Earth’s climate by modify-ing the concentration of atmospheric CO2 (seeChapter 2). Because of the many critical rolesof carbon balance in the biosphere and theEarth System, the recent rapid change incarbon cycling of plants and ecosystems is anissue of fundamental societal importance.

Overview

Carbon that enters ecosystems as gross primaryproduction (GPP) accumulates within the eco-system, returns to the atmosphere via respira-

tion or disturbance, or is transported laterallyto other ecosystems. About half of GPP isrespired by plants to provide the energy thatsupports their growth and maintenance(Schlesinger 1997, Waring and Running 1998).Net primary production (NPP) is the netcarbon gain by vegetation and equals the dif-ference between GPP and plant respiration.Plants lose carbon through several pathwaysbesides respiration (Fig. 6.1). The largest ofthese releases is typically the transfer of carbonfrom plants to the soil. This occurs through lit-terfall (the shedding of plant parts and death ofplants), root exudation (the secretion of solubleorganic compounds by roots into the soil), andcarbon transfers to microbes that are symbiot-ically associated with roots (e.g., mycorrhizaeand nitrogen-fixing bacteria). These carbontransfers from plants to soil eventually give riseto soil organic matter (SOM), which is typicallythe largest pool of ecosystem carbon. Herbi-vores also remove carbon from plants. Her-bivory often accounts for 5 to 10% of NPP in terrestrial ecosystems but can be less than1% in some forests or greater than 50% insome grasslands (see Chapter 11). Herbivoresaccount for most of the carbon loss from plantsin aquatic ecosystems (see Chapter 10). Plantsalso release carbon to the atmosphere throughemission of volatile organic compounds or by

6Terrestrial Production Processes

The balance between gross primary production and ecosystem respiration deter-mines the net carbon accumulation by the biosphere. This chapter describes thefactors that regulate the carbon balance of terrestrial vegetation and ecosystems.

123

124 6. Terrestrial Production Processes

combustion in fires. Volatile emissions typicallyaccount for less than 1% of NPP but give plantstheir distinctive smells, which govern the behav-ior of many herbivores and are an importantcomponent of atmospheric chemistry. Fire israre in some ecosystems but can be the majoravenue of carbon release from plants in many

ecosystems in some years. Finally, carbon can beremoved from vegetation by human harvest orother disturbances.

The carbon balance of ecosystems dependsnot only on the carbon balance of vegetationbut also on the respiration of heterotrophs,organisms that eat live or dead organic matter.

Plants

Soil organic matter and microbes

Fleach

Litterfall Animals

Fdisturb

GPP

Rplant

Rheterotr

Emissions

NPP = GPP - Rplant

NEP ~ GPP - (Rplant + Rheterotr + Fdisturb + Fleach )~

Figure 6.1. Overview of the major carbon fluxes ofan ecosystem. Carbon enters the ecosystem as grossprimary production (GPP), through photosynthesisby plants. Roots and aboveground portions of plantsreturn about half of this carbon to the atmosphereas plant respiration (Rplant). Net primary production(NPP) is the difference between carbon gain by GPPand carbon loss through Rplant. Most NPP is trans-ferred to soil organic matter as litterfall, root death,root exudation, and root transfers to symbionts;some NPP is eaten by animals and sometimes is lostfrom the ecosystem through disturbance. Animalsalso transfer some carbon to soils through excretionand mortality. Most carbon entering the soil is lostthrough microbial respiration (which, together withanimal respiration, is termed heterotrophic respira-

tion; Rheterotr). Additional carbon is lost from soilsthrough leaching and disturbance. Net ecosystemproduction (NEP) is the net carbon accumulation byan ecosystem; it equals the carbon inputs from GPPminus the various avenues of carbon loss: respira-tion, leaching, and disturbance. If an ecosystem wereat steady state, in the absence of disturbance, carboninputs in GPP would approximately balance thecarbon outputs in plant respiration (about 50% ofGPP), heterotrophic respiration (40 to 50% of GPP),and leaching (0 to 10% of GPP). Most ecosystems,however, generally show either a net gain or net lossof carbon (i.e., positive or negative NEP, respec-tively), due to an imbalance between GPP and thevarious avenues of carbon loss.

Plant Respiration 125

Heterotrophic respiration by microbes andanimals converts organic matter to CO2, whichis lost from the ecosystem to the atmosphere.In some ecosystems fire transforms additionalorganic matter to CO2, which moves to theatmosphere. Finally carbon leaches from eco-systems in dissolved and particulate forms andmoves laterally through erosion and depositionof soil, movement of animals, etc. These lateralfluxes of carbon from terrestrial ecosystems arecritical energy subsidies to aquatic ecosystemsand constitute a significant component of thecarbon budgets of many ecosystems.

Plant Respiration

Physiological Basis of Respiration

Respiration provides the energy for a plant toacquire nutrients and to produce and maintainbiomass. At the ecosystem scale, plant respira-tion includes mitochondrial respiration of non-photosynthetic organs at all times and therespiration of leaves at night (Schlesinger1997). Leaf respiration in the light is includedwithin GPP (see Chapter 5). Plant respirationis not “wasted” carbon. It serves the essentialfunction in plants of providing energy forgrowth and maintenance, just as it does inanimals and microbes. We can separate total

plant respiration (Rplant) into three functionalcomponents: growth respiration (Rgrowth), main-tenance respiration (Rmaint), and the respiratorycost of ion uptake (Rion).

Rplant = Rgrowth + Rmaint + Rion (6.1)

Each of these components of respirationinvolves mitochondrial oxidation of carbo-hydrates to produce ATP. They differ only inthe functions for which ATP is used by theplant. Separation of respiration into these com-ponents allows us to understand the ecologicalcontrols over plant respiration.

All plants are similar in their efficiency ofconverting sugars into new biomass. Growthof new tissue requires biosynthesis of manyclasses of chemical compounds, including cellu-lose, proteins, nucleic acids, and lipids (Table6.1). The carbon cost of synthesizing each com-pound includes the carbon that is incorporatedinto that compound plus the carbon oxidized toCO2 to provide the ATPs that drive biosynthe-sis. These carbon costs can be calculated foreach class of compound from a knowledge oftheir biosynthetic pathways (Penning de Vrieset al. 1974). The cost of producing 1g tissue canthen be calculated from the concentration ofeach class of chemical compound in a tissue andits carbon cost of synthesis.

There is a threefold range in the carbon costof synthesis of the major classes of chemical

Concentration Cost Total costComponent (%) (mg C g-1 product) (mg C g-1 tissue)a

Sugar 11.9 438 52Nucleic acid 1.2 409 5Polysaccharide 9.0 467 42Cellulose 21.6 467 101Hemicellulose 31.0 467 145Amino acid 0.9 468 4Protein 9.7 649 63Tannin 4.8 767 37Lignin 4.2 928 39Lipid 5.7 1212 69

Total cost 557

a The four most expensive constituents account for 37% of the cost of synthesis butonly 24% of the mass of the tissue. The total cost of production (557 mg C g-1 tissue)is equivalent to 1.23 g carbohydrate per gram of tissue, indicating growth respira-tion consumes about 25% more carbon than accumulates in biomass.Reprinted with permission from American Naturalist; Chapin (1989).

Table 6.1. Concentration andcarbon cost of major chemicalconstituents in a sedge leaf.


compounds found in plants (Table 6.1). Themost expensive compounds in plants are pro-teins, tannins, lignin, and lipids. In general,metabolically active tissues, such as leaves, havehigh concentrations of proteins, tannins, andlipids. The tannins and lipophilic substancessuch as terpenes serve primarily to defendprotein-rich tissues from herbivores andpathogens (see Chapter 11). Structural tissueshave high lignin and low protein, tannin, andlipid concentrations. Leaves of rapidly growingspecies with high protein concentration havehigher tannin and lower lignin concentrationsthan leaves with low protein concentrations.Consequently, most plant tissues contain someexpensive constituents, although the nature ofthese constituents differs among plant partsand species. In fact, the carbon cost of plantgrowth is surprisingly similar across species,tissue types, and ecosystems (Chapin 1989,Poorter 1994) (Fig. 6.2). On average, growthrespiration is about 25% of the carbon incor-porated into new tissues (Table 6.1) (Waringand Running 1998). The rates of growth andtherefore of growth respiration measured at theecosystem scale (grams of carbon per squaremeter per day) increase when temperature and

moisture favor growth, but growth respirationis always a nearly constant fraction of NPP,regardless of environmental conditions.

The total respiratory cost of ion uptake prob-ably correlates with NPP. Ion transport acrossmembranes is energetically expensive and mayaccount for 25 to 50% of root respiration(Lambers et al. 1998). Several factors cause thiscost of ion uptake to differ among ecosystems.The quantity of nutrients absorbed is greatestin productive environments, although the res-piratory cost per unit of absorbed nutrientsmay be greater in slow-growing plants fromunproductive environments (Lambers et al.1998). The respiratory cost of nitrogen uptakeand use depends on the form of nitrogenabsorbed, because nitrate must be reduced to ammonium (an exceptionally expensiveprocess) before it can be incorporated into pro-teins or other organic compounds. The cost ofnitrate reduction is also variable among plantspecies and ecosystems, depending on whetherthe nitrate is reduced in the roots or the leaves(see Chapter 8). In general, we expect Rion tocorrelate with the total quantity of ions ab-sorbed and therefore to show a positive rela-tionship with NPP.

Maintenance respiration: How variable is the cost of maintaining plant biomass? All livecells, even those that are not actively growing,require energy to maintain ion gradients acrosscell membranes and to replace proteins, mem-branes, and other constituents. Maintenancerespiration provides the ATP for these mainte-nance and repair functions. Laboratory experi-ments suggest that about 85% of maintenancerespiration is associated with the turnover ofproteins (about 6% turnover per day), explain-ing why there is a strong correlation betweenprotein concentration and whole-tissue respira-tion rate in nongrowing tissues (Penning deVries 1975). We therefore expect maintenancerespiration to be greatest in ecosystems withhigh tissue nitrogen concentrations and/or alarge plant biomass and thus to be greatest in productive ecosystems. Simulation modelssuggest that maintenance respiration mayaccount for about half of total plant respiration;the other half is associated with growth and ion uptake (Lambers et al. 1998). These pro-

440 600 760 920

Leaf

Stem

Root

Seed/fruit

Construction costs (mg g-1)

Figure 6.2. Range of construction costs for a surveyof leaves (n = 123), stems (n = 38), roots (n = 35), andseeds or fruits (n = 31). Values are means with 10thand 90th percentiles. The carbon cost of producingnew biomass differs little among plant parts, exceptfor seeds and fruits that store lipid. These tissueshave a higher cost of synthesis than do other plantparts. (Redrawn with permission from SPB Aca-demic; Poorter 1994.)

Net Primary Production 127

portions may vary with environment and plantgrowth rate and are difficult to estimate precisely.

Maintenance respiration depends on envi-ronment as well as tissue chemistry. It increaseswith temperature because proteins and mem-brane lipids turn over more rapidly at high tem-peratures. Drought also imposes short-termmetabolic costs associated with synthesis ofosmotically active organic solutes (see Chapter4). These effects of environmental stress onmaintenance respiration are the major factorsthat alter the partitioning between growth andrespiration and therefore are the major sourcesof variability in the efficiency of convertingGPP into NPP. Maintenance respiration in-creases during times of environmental changebut, following acclimation, maintenance respi-ration returns to values close to those predictedfrom biochemical composition (Semikhatova2000). Over the long term, therefore, mainte-nance respiration is not strongly affected byenvironmental stress.

Plant respiration is a relatively constant proportion of GPP, when ecosystems are compared. Although the respiration rate of any given plant increases exponentially withambient temperature, acclimation and adapta-tion counterbalance this direct temperatureeffect on respiration. Plants from hot environ-ments have lower respiration rates at a giventemperature than do plants from cold places(Billings and Mooney 1968). The net result of these counteracting temperature effects isthat plants from different thermal environ-ments have similar respiration rates when mea-sured at their mean habitat temperature(Semikhatova 2000).

In summary, studies of the basic componentsof respiration associated with growth, ionuptake, and maintenance suggest that totalplant respiration should be a relatively constantfraction of GPP. These predictions are consis-tent with the results of model simulations ofplant carbon balance. These modeling studiesindicate that total plant respiration is about half(48 to 60%) of GPP when a wide range ofecosystems is compared (Ryan et al. 1994,Landsberg and Gower 1997) (see Fig. 6.6).Vari-ation in maintenance respiration is the most

likely cause for variability in the efficiency ofconverting GPP into NPP. There are too fewdetailed studies of ecosystem carbon balance toknow how variable this efficiency is amongseasons, years, and ecosystems. The currentview that the efficiency of converting GPP toNPP is relatively constant may reflect insuffi-cient data or could emerge as an importantecosystem generalization.

Net Primary Production

What Is NPP?

Net primary production is the net carbon gainby vegetation. It is the balance between the carbon gained by gross primary productionand carbon released by plant mitochondrialrespiration.

NPP = GPP - Rplant (6.2)

Like GPP, NPP is generally measured at theecosystem scale, usually over relatively longtime intervals, such as a year (grams biomass orgrams carbon per square meter per year). NPPincludes the new biomass produced by plants,the soluble organic compounds that diffuse orare secreted by roots into the soil (root exuda-tion), the carbon transfers to microbes that are symbiotically associated with roots (e.g.,mycorrhizae and nitrogen-fixing bacteria), andthe volatile emissions that are lost from leavesto the atmosphere (Clark et al. 2001a). Mostfield measurements of NPP document only thenew plant biomass produced and thereforeprobably underestimate the true NPP by atleast 30% (Table 6.2). Root exudates arerapidly taken up and respired by microbes adja-cent to roots and are generally measured infield studies as a portion of root respiration.Volatile emissions are also rarely measured butare generally a small fraction (less than 1 to5%) of NPP and thus are probably not a majorsource of error (Guenther et al. 1995). Somebiomass aboveground and belowground dies oris removed by herbivores before it can be mea-sured, so even the new biomass measured infield studies is an underestimate of biomassproduction. For some purposes, these errors


may not be too important.A frequent objectiveof measuring NPP, for example, is to estimatethe rate of biomass increment. Root exudates,transfers to symbionts, losses to herbivores, andvolatile emissions are lost from plants andtherefore do not directly contribute to biomassincrement. Consequently, failure to measurethese components of NPP does not bias esti-mates of biomass accumulation. However,these losses of NPP from plants fuel otherecosystem processes such as herbivory, de-composition, and nutrient turnover and so areimportant components of the overall carbondynamics of ecosystems and a critical carbonsource for microbes.

Some components of NPP, such as root pro-duction, are particularly difficult to measureand have sometimes been assumed to be someconstant ratio (e.g., 1 : 1) of aboveground pro-duction (Fahey et al. 1998). Fewer than 10% ofthe studies that report ecosystem NPP actuallymeasure components of belowground produc-tion (Clark et al. 2001a). Estimates of above-ground NPP sometimes include only largeplants (e.g., trees in forests) and exclude under-story shrubs or mosses, which can account for asubstantial proportion of NPP in some ecosys-tems. Most published summaries of NPP do notexplicitly state which components of NPP havebeen included (or sometimes even whether theunits are grams of carbon or grams of biomass).For these reasons, considerable care must beused when comparing data on NPP or biomassamong studies.

Physiological Controls over NPP

Photosynthesis, NPP, and respiration: Who is in charge? NPP is the balance of carbon gainedby GPP and the carbon lost by respiration of all plant parts (Fig. 6.1). However, this simpleequation (Eq. 6.2) does not tell us whether the conditions governing photosynthesis dictatethe amount of carbon that is available tosupport growth or whether conditions influ-encing growth rate determine the magnitude ofphotosynthesis. On short time scales (seconds to days), environmental controls over photo-synthesis (e.g., light and water availability)strongly influence photosynthetic carbon gain.However, on weekly to annual time scales,plants appear to adjust leaf area and photosyn-thetic capacity so carbon gain matches the soilresources that are available to support growth(see Fig. 5.1). Plant carbohydrate concentra-tions are usually lowest when environmentalconditions favor rapid growth (i.e., carbo-hydrates are drawn down by growth) and tendto accumulate during periods of drought ornutrient stress or when low temperature con-strains NPP (Chapin 1991b). If the products ofphotosynthesis directly controlled NPP, wewould expect high carbohydrate concentrationsto coincide with rapid growth or to show no consistent relationship with growth rate.

Results of growth experiments also indicatethat growth is not simply a consequence of the controls over photosynthetic carbon gain.Plants respond to low availability of water,nutrients, or oxygen in their rooting zone byproducing hormones that reduce growth rate.The decline in growth subsequently leads to adecline in photosynthesis (Gollan et al. 1985,Chapin 1991b, Davies and Zhang 1991). Thegeneral conclusion from these experiments isthat plants actively sense the resource supply intheir environment and adjust their growth rateaccordingly. These changes in growth rate thenchange the sink strength (demand) for carbo-hydrates and nutrients, leading to changes inphotosynthesis and nutrient uptake (Chapin1991b, Lambers et al. 1998). The resultingchanges in growth and nutrition determine the leaf area index (LAI) and photosyntheticcapacity, which, as we have seen, largely

Table 6.2. Major components of NPP and repre-sentative values of their relative magnitudes.

Components of NPPa % NPP

New plant biomass 40–70Leaves and reproductive parts (fine litterfall) 10–30Apical stem growth 0–10Secondary stem growth 0–30New roots 30–40

Root secretions 20–40Root exudates 10–30Root transfers to mycorrhizae 10–30

Losses to herbivores and mortality 1–40Volatile emissions 0–5

a Seldom, if ever, have all of these components been mea-sured in a single study.


account for ecosystem differences in carboninput (Gower et al. 1999) (see Fig. 5.1).

The feedbacks to photosynthesis are not100% effective: Leaf carbohydrate concentra-tions increase during the day and decline atnight, allowing plants to maintain a relativelyconstant supply of carbohydrates to nonphoto-synthetic organs. Similarly, carbohydrate con-centrations increase during periods (hours toweeks) of sunny weather and decline undercloudy conditions. Over these short time scales,the conditions affecting photosynthesis are the primary determinants of the carbohydratesavailable to support growth. The short-termcontrols over photosynthesis by environmentprobably determine the hourly to weekly pat-terns of NPP, whereas soil resources governannual carbon gain and NPP.

Environmental Controls over NPP

The climatic controls over NPP are mediatedprimarily through the availability of below-ground resources. At a global scale, the largestecosystem differences in NPP are associated

with variation in climate. NPP is greatest inwarm moist environments, where tropical rainforests occur, and is least in climates that aredry (e.g., deserts) or cold (e.g., tundra) (seeplates 1 to 3). NPP correlates most stronglywith precipitation; NPP is highest at 2 to 3myr-1 of precipitation (typical of rain forests)and declines at extremely high precipitation(Gower 2002, Schuur et al. 2001). When dryecosystems (i.e., deserts) are excluded, NPPalso increases exponentially with increasingtemperature (Fig. 6.3). The largest differencesin NPP reflect biome differences in bothclimate and vegetation structure. When ecosys-tems are grouped into biomes, there is a 14-foldrange in average NPP (Table 6.3; Fig. 6.4). Dothese correlations of NPP with climate reflect asimple direct effect of temperature and mois-ture on plant growth, or are other factorsinvolved?

Extensive research in temperate grasslandsillustrates some of the complexities of environ-mental controls over NPP. When grassland sitesare compared across precipitation gradients,the average NPP of a site increases with

˙˙˙

˙

˙

˙

˙

˙˙̇

˙

˙˙

˙

˙

˙̇

˙

˙˙̇̇̇

˙

˙

˙

˙

˙

˙

˙˙

˙

˙

˙

˙

˙

˙

˙

˙

˙̇

˙

˙

˙

˙

˙

˙˙

˙

˙

˙

˙˙

˙ ˙

˙

˙

˙

˙̇

˙

˙˙

˙

˙

˙

˙

˙

˙̇

˙̇

˙

˙˙

˙˙

˙

˙

˙

˙

˙

˙

˙

˙

˙˙˙

˙

˙

˙

˙

˙

˙

˙

˙

0

1000

2000

3000

4000

5000

0 2000 4000 6000 8000

˙̇

˙̇

˙

˙

˙˙

˙˙

˙

˙

˙

˙

˙ ˙ ˙

˙

˙

˙̇

˙̇

˙

˙

˙̇

˙

˙

˙

˙

˙

˙

˙

˙

˙

˙˙

˙˙

˙

˙ ˙

˙

˙˙̇

˙

˙˙

˙

˙

˙̇˙

˙̇ ˙̇

˙˙

˙̇

˙

˙

˙

˙˙˙ ˙˙̇

˙̇

˙˙

˙

˙̇

˙˙

-15 -10 -5 0 5 10 15 20 25 30

NP

P (

gm

-2yr

-1)

Precipitation (mm yr-1) Temperature (oC)

Figure 6.3. Correlation of NPP (in units of biomass)with temperature and precipitation. NPP is greatestin warm moist environments such as tropical forestsand lowest cold or dry ecosystems such as tundra anddeserts. In tropical forests, NPP declines at extremely

high precipitation (>3 m yr-1), due to indirect effectsof excess moisture, such as low soil oxygen and lossof nutrients through leaching. (Figure modified from Schuur Unpublished; data from Lieth 1975, Clark etal. 2001b, and Schuur et al. 2001.)


TEDBL

TEDBL

10

8

6

4

2

0

TRDBL

TEENL

TEENL

TEEBL

TREBL

BOENL

BODBL

BODBL

WDLND

WDLND

TUNDR

TUNDR

DESRT

DESRT

GRSTE

GRSTE

BODNL

BODNL

TEEBL

TREBL

TRDBL

GRSTR

GRSTR

BOENL

2000

1500

1000

500

00

1000 2000 3000

Tota

l LA

I (m

2 m-2

)A

bove

grou

nd N

PP

(g

m-2

yr-1

)

Precipitation (mm yr-1)

Figure 6.4. LAI and abovegroundnet primary production of majorbiome types as a function of precip-itation. DESRT, desert; TUNDR,tundra; BOENL, boreal evergreenneedle-leafed; WDLND, woodland;BODBL, boreal deciduous broad-leaved; BODNL, boreal deciduousneedleleaved; GRSTE, temperategrassland; TEDBL, temperate deciduous broadleaved; TEENL,temperate evergreen needle-leafed;TRDBL, tropical deciduous broad-leafed; TREBL, tropical evergreenbroadleaved; GRSTR, tropicalgrassland; and TEEBL, temperateevergreen broad-leafed. (Redrawnwith permission from Blackwell Scientific; Gower 2002.)

Table 6.3. Net primary production of the major biome types based on biomass harvestsa.

Aboveground BelowgroundNPP Belowground NPP NPP Total NPP

Biome (g m-2 yr-1) (g m-2 yr-1) (% of total) (g m-2 yr-1)

Tropical forests 1400 1100 0.44 2500Temperate forests 950 600 0.39 1550Boreal forests 230 150 0.39 380Mediterranean shrublands 500 500 0.50 1000Tropical savannas and grasslands 540 540 0.50 1080Temperate grasslands 250 500 0.67 750Deserts 150 100 0.40 250Arctic tundra 80 100 0.57 180Crops 530 80 0.13 610

a NPP is expressed in units of dry mass. NPP estimated from harvests excludes NPP that is not available to harvest as aresult of consumption by herbivores, root exudation, transfer to mycorrhizae, and volatile emissions.Data from Saugier et al. (2001).


increasing precipitation (Fig. 6.5) (Sala et al.1988, Lauenroth and Sala 1992), just asobserved across biomes (Fig. 6.3 and 6.4). Inany single grassland site, NPP also increases inyears with high precipitation and responds toexperimental addition of water, demonstratingthat grassland NPP is water limited (Lauenrothet al. 1978). However, part of the water limita-tion reflects the effects of water on moisture-limited decomposition and therefore nutrientsupply (see Chapters 7 and 8). Thus at least tworesources (water and nutrients) limit the NPPof temperate grasslands, and the relative impor-tance of these resources depends on climateand soil type. What about other resources? No one has tested whether addition of lightwould stimulate the productivity of any naturalecosystem. A doubling of atmospheric CO2

stimulates grassland NPP by 10 to 30%, butmost of this stimulation reflects the effects ofCO2 on water and nutrient availability ratherthan the direct effects of CO2 on photosynthe-sis. Finally, species composition and biomassinfluence the response of grassland NPP toclimate.Arid grasslands are never as productive

in wet years as grasslands that regularly receivehigh moisture inputs, presumably because arid grasslands lack the plant species, biomass,or soil fertility to exploit effectively the years of high moisture (Fig. 6.5) (Lauenroth and Sala 1992). In grasslands, therefore, waterappears to be the factor that most strongly controls NPP, but soil moisture determines NPP in at least three ways: through its directstimulation of NPP, through its effects on nutri-ent supply, and through its effect on the speciescomposition and productive capacity of theecosystem.

The controls over NPP in deserts are similarto those in grasslands: Desert NPP correlatesclosely with precipitation among sites, amongyears, and in response to water addition(Gutierrez and Whitford 1987). Even in deserts,however, NPP is greatest in patches with highnutrient availability (Schlesinger et al. 1990)and responds to added nitrogen, especially inexperiments that also add water (Gutierrez andWhitford 1987), indicating a secondary limita-tion of desert NPP by nutrient supply.

In the tundra, where the climate correlationssuggest that NPP should be temperaturelimited, NPP increases more in response toadded nitrogen than to experimental increasesin temperature (Chapin et al. 1995, McKane etal. 1997). Thus, in tundra, the climate–NPP cor-relation probably reflects the effects of tem-perature on nitrogen supply (see Chapter 9) orlength of growing season more than a directtemperature effect on NPP (Chapin 1983). Sim-ilarly, NPP in the boreal forest correlatesclosely with soil temperature, but soil warmingexperiments demonstrate that this effect ismediated primarily by enhanced decomposi-tion and nitrogen supply (Van Cleve et al.1990).

Thus in ecosystems in which climate–NPPcorrelations suggest a strong climatic limitationof NPP, experiments and observations indicatethat this is mediated primarily by climaticeffects on belowground resources. What con-strains NPP in warm, moist climates where temperature and moisture appear optimal forgrowth?

Tropical forests typically have higher NPPthan other biomes (Fig. 6.4). Among tropical

1000

800

600

400

200

00 600 1000 1500A

bove

grou

nd N

PP

(g

m-2

yr-1

)

Precipitation (mm yr-1)

Spatial va

riatio

n

Temporal variation

Figure 6.5. Correlation of grassland NPP with pre-cipitation across grassland sites (spatial variation)and through time for a single site (temporal varia-tion). A single site responds less to interannual vari-ation in precipitation than would be expected fromthe relationship between average precipitation andaverage NPP across all sites, because a single sitelacks the species and productive potential capable ofexploiting high moisture availability. (Redrawn withpermission from Ecological Applications; Lauenrothand Sala 1992.)


forests, litter production tends to correlate withthe supply of nutrients, especially phosphorus(Vitousek 1984), suggesting that NPP in tropical forests may also be limited by thesupply of belowground resources. NPP in tropical forests is maximal at intermediatelevels of precipitation (Schuur et al. 2001).NPP in dry tropical forests is moisture limited,but in extremely wet climates (greater than 2 to 3myr-1 of precipitation) NPP declines inresponse to increasing precipitation, probablydue to oxygen limitation to roots and/or soilmicrobes and to leaching loss of essential nutri-ents. NPP in tropical forests is therefore prob-ably also limited by the supply of belowgroundresources, including nutrients and sometimeswater or oxygen.

In a temperate salt marsh, where water andapparently nutrients are abundant, NPPresponds directly to increases in CO2 (Drakeet al. 1996), as do crops that receive a high nutrient supply. However, NPP is enhanced bynutrient additions even in the most fertile agri-cultural systems (Evans 1980), indicating thewidespread occurrence of nutrient limitation toNPP (see Fig. 8.1).

In summary, experiments and observations ina wide range of ecosystems provide a relativelyconsistent picture. NPP is generally constrainedby the supply of belowground resources. Thefactors determining the supply and acquisitionof belowground resources are the major directcontrols over NPP and therefore the carboninput to ecosystems. Only in the most produc-tive ecosystems do other factors, such as CO2

concentration, directly enhance the NPP ofecosystems.

The importance of belowground resources incontrolling NPP is consistent with our earlierconclusion that GPP is governed more by leafarea and length of the photosynthetic seasonthan by the direct effects of temperature andCO2 on photosynthesis (see Chapter 5). In fact,modeling studies suggest that NPP is a surpris-ingly constant fraction (40 to 52%) of GPPacross broad environmental gradients (Fig. 6.6)(Landsberg and Gower 1997, Waring andRunning 1998). This is consistent with our con-clusion that GPP and NPP are controlled by thesame factors.

Allocation

Allocation of NPP

Patterns of biomass allocation minimizeresource limitation and maximize resourcecapture and NPP. Our discussion of the con-trols over NPP suggests an interesting paradox:A high leaf area is necessary to maximize NPP,yet the major factors that constrain NPP arebelowground resources. The plant is faced witha dilemma of how to distribute biomassbetween leaves (to maximize carbon gain) androots (to maximize acquisition of belowgroundresources). Plants exhibit a consistent patternof allocation—the distribution of growthamong plant parts—that maximizes growth inresponse to the balance between abovegroundand belowground resource supply rates(Enquist and Niklas 2001).

In general, plants allocate production to min-imize limitation by any single resource. Plantsallocate new biomass preferentially to rootswhen water or nutrients limit growth.They allo-cate new biomass preferentially to shoots whenlight is limiting (Reynolds and Thornley 1982).

2000

1500

1000

1000

500

00 2000 3000

NP

P (

g C

m-2

yr-1

)

GPP (g C m-2 yr-1)

Figure 6.6. Relationship between GPP and NPP in11 forests of the United States, Australia, and NewZealand (Williams et al. 1997). These forests wereselected from a wide range of moisture and temper-ature conditions. GPP and NPP were estimated usinga model of ecosystem carbon balance. The simula-tions suggest that all these forests show a similar par-titioning of GPP between plant respiration (53%)and NPP (47%), despite large variations in climate.(Redrawn with permission from Academic Press;Waring and Running 1998.)

Allocation 133

Plants can increase acquisition of a resource byproducing more biomass of the appropriatetissue, by increasing the activity of each unit of biomass, or by retaining the biomass for alonger time. A plant can, for example, increasecarbon gain by increasing leaf area or photo-synthetic rate per unit leaf area or by retainingthe leaves for a longer time before they areshed. Similarly, a plant can increase nitrogenuptake by altering root morphology or byincreasing root biomass, root longevity, nitro-gen uptake rate per unit root, or extent of mycorrhizal colonization. Changes in allocationand root morphology have a particularly strongimpact on nutrient uptake. It is the integratedactivity (mass multiplied by acquisition rate perunit biomass multiplied by time) that must bebalanced between shoots and roots to maxi-mize growth and NPP (Garnier 1991). Theseallocation rules are key features of all simula-tion models of NPP.

Observations in ecosystems are generallyconsistent with allocation theory. Tundra, grass-lands, and shrublands, for example, allocate alarger proportion of NPP belowground than doforests (Table 6.3) (Gower et al. 1999, Saugieret al. 2001).

Allocation Response to Multiple Resources

NPP in most ecosystems is limited moststrongly by a single resource, but it alsoresponds to other resources. If plants were perfectly successful in allocating biomass toacquire the most limiting resource, they wouldbe equally limited by all resources (Bloom et al.1985, Rastetter and Shaver 1992). As we haveseen, this is seldom the case. NPP in mostecosystems responds most strongly to a partic-ular resource, for example to water in deserts,grasslands, and arid shrublands; to nitrogen intundra, boreal forests, and temperate forests;and to phosphorus in tropical wet and dryforests. Thus, as a first approximation, desertsare water-limited ecosystems, and temperateforests are nitrogen-limited ecosystems. Inmany ecosystems, however, NPP does respondto increased availability of more than oneresource. Why does this occur?

The simplest view of environmental limita-tion is that growth is limited by a singleresource at any moment in time. Anotherresource becomes limiting only when thesupply of the first resource increases above the point of limitation (Liebig’s law of theminimum).At least four processes contribute tothe multiple resource limitation observed inmany ecosystems: (1) Plants adjust resourceacquisition to maximize capture of (and mini-mize limitation by) the most limiting resource.(2) Changes in the environment (e.g., rainstorms, pulses of nutrient supply) change therelative abundance of resources so differentfactors limit NPP at different times. (3) Plantsexhibit mechanisms that increase the supply ofthe most limiting resource. (4) Differentresources limit different species in an ecosys-tem, so ecosystem-scale NPP responds to theaddition of more than one resource. Each ofthese processes contributes to the response ofecosystems to multiple resources.

Plants adjust resource acquisition to maxi-mize capture of (and minimize limitation by)the most limiting resource. As discussed earlier,plants adjust allocation of new production to roots vs. shoots to minimize limitation by belowground vs. aboveground resources,respectively. Plants also alter allocation withinthe root system to maximize capture of themost limiting belowground resource (Rastetterand Shaver 1992). In deserts, for example, nu-trient availability is greatest close to the soilsurface, whereas water supplies are generallymore consistently available at depth. Theamount of nutrient or water that a new rootacquires therefore depends on the depth atwhich it is produced. To acquire water, somedesert plants produce coarse, deep water roots,which effectively conduct water but are rela-tively ineffective in absorbing nutrients. Otherplants produce only shallow roots and remainactive only when surface water is available. Thebiochemical investment by roots is specific foreach nutrient. Nitrogen uptake, for example,requires synthesis of specific enzymes to absorbnitrogen, reduce nitrate, and assimilate reducednitrogen into amino acids, whereas differentenzymes are required to absorb phosphorus(see Chapter 8).


Changes in the environment (e.g., rainstorms, pulses of nutrient supply) change therelative abundance of resources so differentfactors limit NPP at different times. Mostecosystems probably experience temporalchanges in the factor that most limits NPPbecause essential resources do not becomeequally available at the same time. Light, forexample, decreases but water increases duringrainy periods. Many ecosystems experience apulse of nutrient availability at the beginning ofthe growing season, when temperatures may besuboptimal for growth. Because all the majorfactors that determine NPP change dramati-cally over several time scales, it would be sur-prising if there were not corresponding changesin the relative importance of these factors inlimiting NPP.

Temporal changes in the limitation of NPPare buffered by storage. Plants accumulate carbohydrates or nutrients during times whentheir availability is high and use their stores to support growth when the supply declines(Chapin et al. 1990). Over seasonal time scales,plants use stored carbohydrates and nutrientsto support their burst of spring growth andreplenish these stores at other times when photosynthesis and nutrient uptake exceed thedemands for growth (see Chapter 8). Plantsvary substantially in their capacity to storewater (see Chapter 4). In summary, storageenables plants to acquire resources when theyare readily available and use them at times oflow supply, thus reducing temporal variation inthe nature of resource limitation.

Plants exhibit several mechanisms thatincrease the supply of the most limitingresource. Plants that have symbiotic associa-tions with nitrogen-fixing microbes directlypromote nitrogen inputs to ecosystems (seeChapter 8). Some ericoid and ectomycorrhizalassociates of other plant species break down proteins and transport the resultingamino acids to plants (Read 1991). Some plants enhance the supply of phosphorusthrough the production of organic chelates that solubilize mineral phosphorus or throughthe production of phosphatases that cleaveorganic phosphates. Plants also exude carbo-

hydrates that enhance mineralization near the root (see Chapter 9). These mechanisms will be described in more detail in later chapters.

Different resources limit different species, soecosystem-scale NPP responds to the additionof more than one resource. Every species in anecosystem has slightly different environmentalrequirements and therefore will be limited bydifferent resource combinations (Tilman 1988).Tundra species, for example, differ in theirresponse to temperature, light, and nutrients(Chapin and Shaver 1985) and in some cases tothe addition of nitrogen vs. phosphorus. Thesedifferences among plant species in the factorslimiting growth contribute to the co-existenceof species in a variable environment (Tilman1988). This may be particularly important inexplaining why species differ in their produc-tivity among years and why the productivity of ecosystems varies less among years than does the productivity of any of the componentspecies (Chapin and Shaver 1985). Spatial heterogeneity in the supply of potentially limiting resources is another important reasonwhy different plants may be limited by differ-ent resources.

Diurnal and Seasonal Cycles of Allocation

Photosynthesis and growth are highly resilientto daily and seasonal variations in the environ-ment. Daily and seasonal variations in the environment are two of the most predictableperturbations experienced by ecosystems.Many organisms adjust their physiology andbehavior based on innate circadian (about24-h) rhythms that lead to 24-h cycles. Forexample, stomatal conductance and carbongain show a circadian rhythm even under constant conditions because stomata have an innate 24-h cycle of stomatal opening andclosing. Plants store starch in the leaves duringthe day and break it down at night, so the rateof carbohydrate transport to roots is nearlyconstant (Lambers et al. 1998). Thus below-ground processes, such as root exudation andcarbon transport to mycorrhizae, are buffered

Allocation 135

from diurnal variations in photosyntheticcarbon gain.

Organisms adjust seasonally in response tochanging photoperiod (day length). Many tem-perate plants, for example, exhibit a relativelypredictable pattern of phenology, the seasonaltiming of production and loss of leaves, flowers,fruits, etc. Plant leaves begin to senesce andreduce their rates of photosynthesis when daylength or other environmental cues signal thecharacteristic onset of winter. Before senes-cence, plants transport carbohydrates andnutrients from leaves to storage organs toprevent their loss during senescence (Chapin etal. 1990). These stores provide resources tosupport plant growth the next spring, so NPPdoes not depend entirely on acquisition of newresources at times when no leaves are present.Other ecosystem processes change as eitherdirect consequences of changes in environment(e.g., the decline in decomposition duringwinter due to lower temperatures) or indirectconsequences of changes in other processes(e.g., the pulse of litter input to soil after leafsenescence). Ecosystem processes largelyrecover after each period of the cycle due to thepredictable nature of diurnal and seasonal per-turbations and the resilience of most processesto these changes. It is therefore unnecessary toconsider explicitly the physiological basis of cir-cadian and photoperiodic controls to predictecosystem processes over longer time scales(see Chapter 13). In contrast to temperateecosystems, tropical wet forests exhibit a lesswell-defined seasonality. Individual species fre-quently shed their leaves synchronously, butspecies differ in their timing of senescence,so the ecosystem as a whole shows no strongseasonal pulse of production and senescence.

The seasonality of plant growth depends onthe seasonality of leaf area and factors regulat-ing photosynthesis. Spring growth of plants isinitially supported by stored reserves of carbonand nutrients that were acquired in previousyears. Leaves quickly become a net source ofcarbon for the rest of the plant, and growthduring the remainder of the growing season islargely supported by the current year’s photo-synthate. There is often competition among

plant parts for allocation of a limited carbo-hydrate supply early in the growing season,resulting in a seasonal progression of produc-tion of different plant parts, for example, withleaves produced first, followed by roots, andthen by wood (Kozlowski et al. 1991). Plantsspecies differ, however, in their seasonal pat-terns of allocation. Plants with evergreen leavesmay allocate NPP to root growth earlier thanwould deciduous plants, because they alreadyhave a leaf canopy that can provide carbon (Kummerow et al. 1983). Ring-porous temper-ate trees must first allocate carbon to xylemproduction in spring to develop a functionalwater transport system. The water columns in their large-diameter vessels cavitate (break)during winter freezing, so xylem vessels remain functional for only a single growingseason. This large carbon requirement torebuild xylem vessels each spring may explainthe northern boundary of ring-porous speciessuch as oaks (Zimmermann 1983). Seedlings in dry environments often depend entirely ontheir cotyledons for photosynthesis during thefirst weeks of growth and allocate all NPP toroot growth to provide a dependable watersupply. The allocation calendar of a plant pro-vides a general seasonal framework for alloca-tion. Variations in environment cause plants to modify this allocation calendar to achievethe appropriate balance of carbon, water, andnutrients.

In ecosystems with short growing seasons,such as arctic tundra, a substantial proportionof the “current” year’s production is actuallysupported by resources that were acquired inprevious years. In late summer, carbon andnutrient stores are replenished to support thenext year’s production. This seasonal pattern ofstorage buffers plant production from seasonaland interannual variations in the environment(Chapin et al. 1990). The seasonality of plantgrowth is constrained by the availability of leafarea early and late in the growing season butotherwise follows seasonal patterns of factorsthat govern photosynthesis (temperature, light,and moisture); the relative importance andtiming of these seasonal controls differ amongspecies and ecosystems.


Tissue Turnover

The balance between NPP and biomass lossdetermines the annual increment in plantbiomass. Plants retain only part of the biomassthat they produce. Some of this biomass loss is physiologically regulated by the plant, forexample the senescence of leaves and roots.Senescence occurs throughout the growingseason in grasslands and during autumn or atthe beginning of the dry season in many trees.Other losses occur with varying frequency andpredictability and are less directly controlled bythe plant, such as the losses to herbivores andpathogens, wind-throw, and fire. The plant alsoinfluences these tissue loss rates through thephysiological and chemical properties of thetissues it produces. Still other biomass transfersto dead organic matter result from mortality of individual plants. Given the substantial,although incomplete, physiological control overtissue loss, why do plants dispose of the biomassin which they have invested so much carbon,water, and nutrients to produce?

Tissue loss is an important mechanism bywhich plants balance resource requirementswith resource supply from the environment.Plants depend on regular large inputs ofcarbon; water; and, to a lesser extent, nutrientsto maintain vital processes. For example, oncebiomass is produced, it must have continuedcarbon inputs to support maintenance respira-tion. If the plant (or organ) cannot meet these carbon demands, the plant (or organ)dies. Similarly, if the plant cannot absorb suffi-cient water to replace the water that isinevitably lost during photosynthesis, it mustshed transpiring organs (leaves) or die. Theplant must therefore shed biomass wheneverresources decline below some threshold neededfor maintenance. Senescence is just as impor-tant as production in adjusting to changes inresource supply and is the only mechanism by which plants can reduce biomass whenresources decline in abundance.

Senescence is the programmed breakdown oftissues. The location of senescence is physio-logically controlled to eliminate tissues that areleast useful to the plant. Grazing of above-ground tissues, for example, can cause a pulse

of root mortality (Ruess et al. 1998), whereasgrazing of belowground tissues reduces thelongevity of leaves (Detling et al. 1980).Although the controls over senescence andmortality of belowground tissues are poorlyunderstood, these patterns of senescenceappear to maintain the functional balancebetween leaves and roots in response to achanging environment (Garnier 1991).

Growth and senescence together enableindividual plants to explore new territory. Leafand shoot growth generally occur at the top ofthe canopy or in canopy gaps, where light avail-ability is highest.This is balanced by senescenceof leaves and stems in less favorable light envi-ronments (Bazzaz 1996). This balance betweenbiomass production and loss allows trees andshrubs to grow toward the light. Similarly, rootsoften proliferate in areas of nutrient enrich-ment or where there is minimal competitionfrom other roots, and root death is greatest inzones of local water or nutrient depletion (seeChapter 8). This exploration of unoccupiedhabitat by plants would occur much less effec-tively if there were not senescence and loss of tissues in less favorable habitats to reducemaintenance requirements and to providenutrient capital to produce new tissues. Theexploration of new territory through synchro-nized growth and senescence reduces spatialvariability in ecosystems by filling canopy gapsand exploiting nutrient-rich patches of soil.

Senescence causes tissue loss at times whenmaintenance costs greatly exceed resourcegain. In seasonally variable environments, thereare extended periods of time when temperatureor moisture is predictably unfavorable. In theseecosystems, the cost of producing tissues thatcan withstand the rigors of this unfavorableperiod and of maintaining tissues when theyprovide negligible benefit to the plant mayexceed the cost of producing new tissues whenconditions again become favorable (Chabotand Hicks 1982). Arctic, boreal, and temperateecosystems, for example, predictably experi-ence seasons that are too cold for effectivegrowth or resource acquisition. There is a pulseof autumn senescence of leaves and roots, oftentriggered by some combination of photoperiodand low temperature (Ruess et al. 1996). Dry

Global Distribution of Biomass and NPP 137

ecosystems experience similar pulses of leafand root senescence with the onset of drought.Senescence and tissue loss are therefore highlypulsed in most ecosystems and occur just beforethe period when conditions are least favorablefor resource acquisition and growth. These sea-sonal pulses of senescence cause the greatesttissue loss in highly seasonal environments.

Leaf longevity varies among plant speciesfrom a few weeks to several years or decades.In general, plants in high-resource environ-ments produce short-lived leaves with a highspecific leaf area (SLA) and a high photo-synthetic rate per leaf mass, but they have littleresistance to environmental stresses and arepoorly defended against herbivores. These disposable leaves are typically shed when con-ditions become unfavorable (winter or dryseason) and are replaced the next spring. Bothroot and leaf longevity are greater in low-resource environments (Berendse and Aerts1987) and lower at high latitudes than in thetropics (Fig. 6.7).The greater longevity of leavesfrom low-resource environments reduces thenutrient requirement by plants to maintain leafarea (see Chapter 8).

Senescence enables plants to shed parasites,pathogens, and herbivores. Because leaves andfine roots represent relatively large packets of nutrients and organic matter, they are con-stantly under attack from pathogens, parasites,

and herbivores. Phyllosphere fungi, forexample, begin colonizing and growing onleaves shortly after budbreak, initially as para-sites and later as part of the decomposer com-munity, when the leaf is shed (see Chapter 7).These fungi account for some of the mottledappearance of older leaves. Pathogenic rootfungi are a major cause of reduced yields inagroecosystems and are common in naturalecosystems. Plants have a variety of mecha-nisms for detecting natural enemies andrespond initially through the production ofinduced chemical defenses (see Chapter 11)and, in the case of severe attack, by sheddingtissues.

Large unpredictable biomass losses occur inmost ecosystems. Wind storms, fires, herbivoreoutbreaks, and epidemics of pathogens fre-quently cause large tissue losses that are unpre-dictable and occur before any programmedsenescence of tissues and associated nutrientresorption can occur. These unpredictablebiomass losses incur approximately twice thenutrient loss to the plant as that occurring after senescence (see Chapter 8). They oftenincrease spatial heterogeneity of light andnutrient resources in the ecosystem throughcreation of gaps, which range in scale from theloss of individual leaves to the destruction ofbiomass over large regions. Most ecosystemsare at some stage in the regrowth after suchbiomass losses.

Global Distribution of Biomassand NPP

Biome Differences in Biomass

The plant biomass of an ecosystem is thebalance between NPP and tissue turnover. NPPand tissue loss are seldom in perfect balance.NPP tends to exceed tissue loss shortly afterdisturbance; at other times tissue loss exceedsNPP. As ecosystems or landscapes approachsteady state (see Chapter 14), however, there isoften a consistent relationship between plantbiomass and the climate or biome type thatcharacterizes that ecosystem. Average plantbiomass varies 50-fold among Earth’s major

High latitude Temperate Tropical

Turn

over

(yr

-1)

1.0

0.8

0.6

0.4

0.2

0

Trees: complete system

Trees: fine roots

Grasslands

Shrublands

Wetlands

Figure 6.7. Synthesis of information on rootturnover in major ecosystem types along a latitudi-nal gradient. (Redrawn with permission from NewPhytologist; Gill and Jackson 2000.)


terrestrial biomes (Table 6.4). Forests have themost biomass. Among forests, average biomassdeclines 5-fold from the tropics to the low-stature boreal forest, where NPP is low andstand-replacing fires frequently remove bio-mass. Deserts and tundra have only 1% asmuch aboveground biomass as do tropicalforests. In any biome, disturbance frequentlyreduces plant biomass below levels that theclimate and soil resources could support. Crops,for example, have a biomass similar to that of tundra or desert, despite more favorablegrowing conditions; regular removal of cropbiomass by harvest prevents it from accumu-lating to levels that climate and soil resourcescould potentially support. When disturbancefrequency declines, for example through fireprevention in grasslands and savannas, biomassoften increases through invasion of shrubs andtrees.

Patterns of biomass allocation reflect thefactors that most strongly limit plant growth inecosystems (Table 6.4). Between 70 and 80% ofthe biomass in forests is aboveground becauseforests characterize sites with relatively abun-dant supplies of water and nutrients, so lightoften limits the growth of individual plants. Inshrublands, grasslands, and tundra, however,water or nutrients more severely limit produc-tion, and most biomass occurs belowground.Because of favorable water and nutrientregimes, crops maintain a smaller proportion ofbiomass as roots than do most unmanagedecosystems.

Tropical forests account for about half ofEarth’s total plant biomass, although they occur

on only 13% of the ice-free land area; otherforests contribute an additional 30% of globalbiomass (Table 6.5). Nonforest biomes there-fore account for less than 20% of total plantbiomass, although they occupy 70% of the ice-free land surface. Crops for example, accountfor only 1% of terrestrial biomass, althoughthey occupy more than 10% of the ice-free landarea. Thus most of the terrestrial surface hasrelatively low biomass (see Fig. 5.20). Thisobservation alone raises concerns about tropi-cal deforestation, independent of the associatedspecies losses.

Biome Differences in NPP

The length of the growing season is the majorfactor explaining biome differences in NPP.Most ecosystems experience times that are toocold or too dry for significant photosynthesis orfor plant growth to occur. When NPP of eachbiome is adjusted for the length of the growingseason, all forested ecosystems have similarNPP (about 5gm-2 d-1), and there is only abouta threefold difference in NPP between desertsand tropical forests (Table 6.6). These calculations suggest that the length of thegrowing season accounts for much of the biomedifferences in NPP (Gower et al. 1999, Körner1999).

Leaf area accounts for much of the biomedifferences in carbon gain during the growingseason. Average total LAI varies about sixfoldamong biomes; the most productive ecosystemsgenerally have the highest LAI (Table 6.6;Fig. 6.4). When NPP is adjusted for differences

Shoot Root Root TotalBiome (g m-2) (g m-2) (% of total) (g m-2)

Tropical forests 30,400 8,400 0.22 38,800Temperate forests 21,000 5,700 0.21 26,700Boreal forests 6,100 2,200 0.27 8,300Mediterranean shrublands 6,000 6,000 0.5 12,000Tropical savannas and grasslands 4,000 1,700 0.3 5,700Temperate grasslands 250 500 0.67 750Deserts 350 350 0.5 700Arctic tundra 250 400 0.62 650Crops 530 80 0.13 610

a Biomass is expressed in units of dry mass.Data from Saugier et al. (2001).

Table 6.4. Biomass distribu-tion of the major terrestrialbiomesa.

Global Distribution of Biomass and NPP 139

in both length of growing season and leaf area,unproductive ecosystems, such as tundra ordesert, do not differ consistently in NPP frommore productive ecosystems. If anything, theless-productive ecosystems may have higherNPP per unit of leaf area and growing seasonlength than do crops and forests. On average,plants in most biomes produce 1 to 3g biomassm-2 leaf d-1 during the growing season. This isequivalent to a GPP of 1 to 3g carbon m-2 leafd-1, because NPP is about half of GPP andbiomass is about 50% carbon. Apparent differ-ences among biomes in these values reflect asubstantial uncertainty in the underlying data.At this point, there is little evidence for strong

ecological patterns in NPP per unit leaf areaand length of growing season.

LAI is both a cause and a consequence of dif-ferences in NPP. LAI is determined largely bythe availability of soil resources (mainly waterand nutrients), climate, and time since distur-bance (see Chapter 5; Fig. 6.4). Tropical rainforests, for example, occur in a warm, moistclimate that provides adequate water and nutrient release to support a large leaf area.These leaves remain photosynthetically activethroughout the year, because there are no long periods of unfavorable weather causingmassive leaf loss, and plants can tap stores of deep groundwater during dry months

Total C pool Total NPPBiome Area (106 km2) (Pg C) (Pg C yr-1)

Tropical forests 17.5 340 21.9Temperate forests 10.4 139 8.1Boreal forests 13.7 57 2.6Mediterranean shrublands 2.8 17 1.4Tropical savannas and grasslands 27.6 79 14.9Temperate grasslands 15.0 6 5.6Deserts 27.7 10 3.5Arctic tundra 5.6 2 0.5Crops 13.5 4 4.1Ice 15.5

Total 149.3 652 62.6

a Biomass is expressed in units of carbon, assuming that plant biomass is 50%carbon.Data from Saugier et al. (2001).

Table 6.5. Global distributionof terrestrial biomes and theirtotal carbon in plant biomassa.

Table 6.6. Productivity per day and per unit leaf areaa.

Daily NPP per Daily NPPSeason lengthb ground area Total LAIc per leaf area

Biome (days) (g m-2 d-1) (m2 m-2) (g m-2 d-1)

Tropical forests 365 6.8 6.0 1.14Temperate forests 250 6.2 6.0 1.03Boreal forests 150 2.5 3.5 0.72Mediterranean shrublands 200 5.0 2.0 2.50Tropical savannas and grasslands 200 5.4 5.0 1.08Temperate grasslands 150 5.0 3.5 1.43Deserts 100 2.5 1.0 2.50Arctic tundra 100 1.8 1.0 1.80Crops 200 3.1 4.0 0.76

a Calculated from Table 6.3. NPP is expressed in units of dry mass.b Estimated.c Data from Gower (2002).


(Woodward 1987). Deserts, in contrast, producelittle leaf area because of inadequate precipita-tion and water storage, and arctic tundra is coldand supplies nitrogen too slowly to produce a large leaf area. In both deserts and tundra,the short growing season gives little time for leaf production, and unfavorable conditionsbetween growing seasons limit leaf survival.Theresulting low leaf area that generally character-izes these ecosystems is a major factor account-ing for their low productivity (Table 6.6).

Disturbances modify the relationshipbetween climate and NPP. There is substantialvariability in NPP among sites within a biome.Some of this variability reflects variation instate factors such as climate and parent mate-rial. However, disturbance also affects NPPsubstantially, in part through changes inresource supply and LAI. Forest NPP, forexample, frequently increases after disturbanceuntil the canopy closes, and the available lightis fully used (see Fig. 13.8) (Ryan et al. 1997).In later successional forests, NPP declines for avariety of reasons (see Chapter 13).

About 60% of the NPP of the biosphereoccurs on land; the rest occurs in aquaticecosystems (see Chapter 15). When summed at the global level, tropical forests account for about a third of Earth’s terrestrial NPP;all forests account for about half of terrestrialNPP (Table 6.5). Grasslands and savannasaccount for an additional third of terrestrialNPP; these ecosystems are much more im-portant in their contribution to terrestrial production than to biomass. Crops are 10-fold more productive than the global average;they account for about 10% of terrestrial pro-duction and occupy 1% of the ice-free landsurface.

Net Ecosystem Production

Ecosystem Carbon Storage

Net ecosystem production (NEP) is the netaccumulation of carbon by an ecosystem. It isthe balance between carbon entering andleaving the ecosystem. Most carbon enters theecosystem as gross primary production and

leaves through several processes, includingplant and heterotrophic respiration, leaching,plant volatile emissions, methane flux, and disturbance. Lateral transfers, such aserosion/deposition or animal movements, canbe either carbon inputs to or outputs from theecosystem (Fig. 6.8; Box 6.1). Alternatively, wecan express the carbon input to ecosystems asNPP (the balance between GPP and plant res-piration) and the respiratory carbon loss asheterotrophic respiration from microbes andanimals (Rheterotr) (Box 6.1). This enables us toconsider separately the fluxes associated withplants (carbon inputs) and with heterotrophs(carbon outputs).

NEP is ecologically important because it represents the increment in carbon stored by anecosystem. When integrated globally, NEPdetermines the impact of the terrestrial bio-sphere on the quantity of CO2 in the atmo-sphere, which strongly affects climate and theamount of carbon transferred to oceans (seeChapter 15). The components of NEP showlarge temporal variation. Disturbances such asfire or forest harvest are episodic events thatdominate the carbon exchange of an ecosystemwhen they occur, but are less important at othertimes. GPP is an important process during thegrowing season, when conditions are suitablefor photosynthesis but is negligible at othertimes of year. Heterotrophic respiration andleaching loss of carbon also vary seasonally and through succession (see Chapter 13). Inecosystems that have not recently experienceddisturbance, NEP is a small net differencebetween two large fluxes: (1) photosyntheticcarbon gain and (2) carbon loss through respi-ration (primarily plants and microbes) andleaching (Fig. 6.9). During the season of peakplant growth, NEP is positive because photo-synthesis exceeds respiration. In winter, whenphotosynthesis is low, NEP is negative and is mainly due to heterotrophic respiration.Thus carbon uptake seldom balances carbonloss from ecosystems at any point in the sea-sonal cycle.

There is a necessary functional linkagebetween NPP and heterotrophic respiration.NPP provides the organic material that fuelsheterotrophic respiration, and heterotrophic

Net Ecosystem Production 141

respiration releases the minerals that supportNPP (Harte and Kinzig 1993). For thesereasons, NPP and heterotrophic respirationtend to be closely matched in ecosystems atsteady state. At steady state, by definition, NEPequals zero, regardless of carbon input orclimate. In fact, peat bogs, which are quiteunproductive, are the ecosystems with thegreatest long-term carbon storage, because theanaerobic soil conditions characteristic of thesebogs restrict decomposition more strongly thanthey restrict NPP.

Leaching

Leaching of dissolved organic carbon (DOC)and dissolved inorganic carbon (DIC) togroundwater and streams is a quantitativelyimportant avenue of carbon loss from someecosystems (Fig. 6.8). Groundwater is generallysupersaturated with respect to CO2 because ofthe high CO2 concentration in the soil atmo-sphere. Some dissolved CO2 leaches out of theecosystem to groundwater and then moves tostreams and lakes, where the excess CO2 is

Atmospheric inputs Atmospheric outputs

Net ecosystem production

Animalconsumption

Net primaryproduction

Gross primaryproduction

Microbialrespiration

Animalrespiration

Plant litterfall,mortality, and

exudation

D Animal biomassD Plant biomass D Soil organic matter

Plantrespiration

Soil organic matterinputs and

microbial biomass

C export

Leaching loss

Volatileemissions

Animalmortality

Harvest

Fire

Assimilation

Excretion

Net secondaryproduction

C im

port CO2

CO2

CO2

CO2

Leaching

CH4

Figure 6.8. Overview of the carbon fluxes of an ecosystem. The large box represents the ecosystem, whichexchanges carbon with the atmosphere, other ecosystems, and groundwater.


Box 6.1. Components of an Ecosystem Carbon Budget

Plant biomass accumulation is the balancebetween GPP and the losses of carbon fromplants. An important motivation for study-ing primary production and decomposition (see Chapter 7) is to understand how these processes influence the net carbon gain or loss by an ecosystem. Here we sum-marize, in an ecosystem context, the majorprocesses that were described previously(Fig. 6.8).

GPP is the net carbon gain by photo-synthesis in the light and is the best measureof the carbon that enters ecosystems. Non-photosynthetic organs (and leaves at night)respire (plant respiration; Rplant) some of thecarbon fixed by GPP to support growth andmaintenance. Rplant returns about half of thecarbon fixed by GPP to the atmosphere. NPPis the net annual carbon gain by vegetationand equals the difference between GPP andplant respiration (Eq. 6.2).

The second largest avenue of carbon lossfrom plants, after respiration, is the carbonflux to the soil (Fpl-soil) through litterfall (theshedding of plant parts and plant mortality),secretion of soluble organic compounds by roots into the soil (root exudation), andcarbon fluxes to microbes that are symbioti-cally associated with roots (e.g., mycorrhizaeand nitrogen fixers). These carbon fluxesfrom plants to soil are the largest inputs tosoil organic matter (SOM). Another impor-tant pathway of carbon loss from plants isherbivory (Fherbiv), the consumption of plantsby animals (see Chapter 11). Plants also losesmall amounts of carbon to the atmosphereby emission of volatile organic compounds(Femiss). Finally, carbon can be lost fromplants by fire (Fpl-fire) or human harvest(Fharv). Carbon losses due to fire or harvestdiffer from the other components of theplant carbon budget because they are highlyepisodic. In a typical year, these fluxes havevalues of zero, but, when they occur, they canbe the dominant pathway of carbon lossfrom vegetation. The annual accumulation

of plant biomass (DBplant) depends on thebalance between carbon gain through NPP(or GPP) and the various pathways ofcarbon loss over a given time (t) interval.

(B6.1)

or

As with plants, the accumulation of animalbiomass (DBanimal) equals inputs minusoutputs—that is, the plant biomass eaten (Fherbiv) minus losses to animal respiration (Ranimal) and fluxes from animalsto soils (Fanim-soil) as a result of excretion andmortality. Soil animals feed primarily onmicrobial biomass (Fmicro-anim). When animalsare eaten by other animals (includinghumans), some of the carbon remains in theanimal biomass pool (but transferred to adifferent organism). The rest of the carbon is transferred to the soil as unconsumedbiomass or feces or is transferred into or outof the ecosystem (Flateral) (see Chapter 11).Transfers of carbon within the animal boxare subsumed in the overall carbon budgetequations.

(B6.3)

SOM accumulates when the annualcarbon inputs to SOM from plants (Fpl-soil)and animals (Fanim-soil) exceed the respirationof soil microorganisms, including soilanimals (Rmicrob), and carbon losses due to consumption of microbes by soil animals(Fmicro-anim), methane emissions to the atmo-sphere (FCH4

), leaching of organic and inorganic carbon to groundwater (Fleach),and fire (Fsoil-fire). Leaching of dissolved

DD

Bt

F F

R F

animal = +

- +( )

herbiv microb-anim

animal anim-soil

DD

Bt

R F F

F F F

plantplant pl-soil herbiv

emiss pl-fire harv

GPP= - - +(

+ + + )

DD

Bt

F F

F F F

plantpl-soil herbiv

emiss pl-fire harv

NPP= - +(

+ + + )

(B6.2)


organic and inorganic carbon in ground-water can be a significant loss from ecosys-tems, although it is often poorly quantified.There can be lateral carbon fluxes (Flateral)into or out of the ecosystem due to deposi-tion or erosion.

(B6.4)

The pool sizes of plants, animals, and SOMget larger when inputs exceed outputs andget smaller when outputs exceed inputs.Fire, for example, causes an instantaneousdecrease in the pool size of plant and soilcarbon, whereas these pools generally in-crease in size during succession after fire (seeChapter 13).

Ecosystem carbon accumulation dependsprimarily on the balance between carboninputs through photosynthesis and carbonlosses through respiration and disturbance.Net ecosystem production is the net annualcarbon accumulation by the ecosystem andis the sum of the net carbon accumulation inplants, animals, and the soil plus lateral trans-fers of carbon among ecosystems (Flateral)(Olsen 1963, Bormann et al. 1974, Aber andMelillo 1991).

(B6.5)

NEP is positive when carbon inputs to theecosystem exceed carbon losses and is negative when losses exceed inputs. It is difficult, however, to measure NEP accu-rately from changes in the carbon pools inplants, animals and SOM, because thechanges in these pools over short time inter-vals are small relative to measurementerrors. NEP is therefore often estimatedfrom changes in those fluxes by whichcarbon enters or leaves the ecosystem, ignor-ing the fluxes of carbon that occur within theecosystem.

NEPSOMplant anim

lateral=+ +( )

±D D D

DB B

tF

DD

SOMpl-soil anim-soil

microb microb-anim

CH leach soil-fire4

tF F

R F

F F F

= +

- ( ++ + + )

When carbon fluxes are aggregated at theecosystem scale, some fluxes cancel out,because a loss from one component repre-sents a gain by another component andtherefore does not affect the total quantityof carbon in the ecosystem. Important fluxesthat cancel out are consumption of plantsand microbes by animals and the carbonfluxes from plants or animals to soil. Thuslitterfall and herbivory alter ecosystemcarbon budgets primarily by altering thelocation of the carbon within the ecosystem,not by altering directly the carbon inputs toor losses from the ecosystem. Even largechanges in these fluxes such as occur duringinsect outbreaks or hurricanes are simplyinternal transfers of carbon within theecosystem and need not be represented asseparate terms in an overall budget ofchanges in carbon pools.

Some ecosystem fluxes can be aggregatedat the ecosystem scale. Ecosystem respira-tion (Recosyst) is the combined respiration ofplants, animals and microbes. It can be partitioned into plant respiration (Rplant),also termed autotrophic respiration, and heterotrophic respiration (Rheterotr)—that is,the respiration by organisms that gain their carbon by consuming organicmatter rather than producing it them-selves. Heterotrophs include animals andmicrobes.

Rheterotr = Ranimal + Rmicrob (B6.6)

Recosyst = Rplant + Rheterotr (B6.7)

It is useful to treat separately those dis-turbances, such as fire and harvest, thatdirectly remove carbon from ecosystems,because these disturbances are episodic innature and frequently involve large fluxesthat occur over a few hours to days.

Fdisturb = Fpl-fire + Fsoil-fire + Fharv (B6.8)

Based on Equations B6.1 to B6.5 we candescribe NEP in terms of carbon fluxesrather than changes in pool sizes.


Alternatively, we can use Equations 6.2and B6.1 to B6.6 to express the carbon inputas NPP and the respiratory carbon loss asheterotrophic respiration from microbes andanimals (Rheterotr). This enables us to considerseparately the fluxes associated with plants(carbon inputs) and with heterotrophs(carbon outputs).

(B6.10)

Some of the terms in Equation B6.11 areusually small and can be ignored in someecosystems some of the time. For example,emission of volatile organic compoundsfrom plants are 0 to 5% of GPP (see Chapter5), and net CH4 flux from soils is substantialonly in sites with anaerobic soils (0 to 8% of GPP). In ecosystems where all thesefluxes are small, the NEP equation can besimplified.

(B6.11)

NEP NPP

in some ecosystemslateral ecosyst disturb

leach

= ± - ( ++ ) [ ]

F R F

F

NEP NPP lateral heterotr disturb

leach emiss CH4

= ± - ( ++ + + )

F R F

F F F

NEPSOM

GPP

+ F

GPP

plant animallateral

plant pl-soil herbi

emiss pl-fire harv

herbiv microb-anim

animal anim-siol

pl-soil anim-soil microb

microb-anim CH

leach soil-fire lateral

lateral

4

=+ +( )

±

= - - +([+ + + ) ]

+[- +( )]+ + - ([+ ++ + ) ] ±

= ±

D D DD

B Bt

F

R F F

F F F

F

R F

F F R

F F

F F F

F

v

-- +(+ ) - + +( )- + +( )

= ± - +(+ + + )

R R

R F F F

F F F

F R F

F F F

plant animal

microb pl-fire soil-fire harv

leach emiss CH

lateral ecosyst disturb

leach emiss CH

4

4

GPP

or

(B6.12)

The net ecosystem exchange (NEE) ofCO2 can be estimated from measurements ofCO2 exchange above the canopy, furthersimplifying the terms that must be measuredto estimate NEP.

(B6.13)

Over short time scales, in the absence oferosion and deposition, between episodes offire or harvest, NEP simply reflects the netCO2 exchange with the atmosphere and theleaching loss of carbon to groundwater.

(B6.14)

Over long time scales and over large regions,the fluxes associated with disturbance,erosion, and deposition cannot be ignored,even though we seldom measure these fluxesdirectly (Harden et al. 2000). Net biome pro-duction (NBP) is NEP integrated at largespatial scales. It explicitly includes carbonremovals by fire and harvest. NBP is the best measure of ecosystem carbon balance at the regional scale (Schulze et al. 2000,Randerson et al., in press), because it explic-itly includes fluxes associated with distur-bance, erosion, and deposition.

(B6.15)

NBP NEP NEE

at regional scales

lateral

disturb leach

= = ±- +( )[ ]

F

F F

NEP NEE in the

absence of fire, deposition,

and erosion, harvest

leach= - [

]

F

NEP NEE

in some ecosystemslateral disturb leach= ± - ( + )

[ ]F F F

NEP NPP

in some ecosystemslateral heterotr disturb

leach

= ± - ( ++ ) [ ]

F R F

F

(B6.9)


released to the atmosphere (see Chapter 10).Approximately 20% of the CO2 produced inarctic soils, for example, leaches to ground-water and is released from lakes and streams(Kling et al. 1991). Dissolved organic carbon is also lost from ecosystems by leaching togroundwater. Despite their importance, leach-ing losses of carbon to groundwater are seldommeasured and therefore frequently ignored inecosystem carbon budgets.

Lateral Transfers

Lateral transfer of carbon into or out of ecosys-tems can be important to the long-term carbonbudgets of ecosystems. Carbon can move later-ally in ecosystems through erosion and deposi-tion by wind or water or by movement byanimals (Fig. 6.8). These lateral transfers areusually so small that they are undetectable inany single year. Over long time periods orduring extreme events, such as floods or land-slides, these lateral transfers can, however, bequantitatively important. These transfers aretypically more important for elements that aretightly cycled within ecosystems and have smallannual inputs and losses (e.g., phosphorus) thanthey are for carbon.

Disturbance

Disturbance is an episodic cause of carbon lossfrom many ecosystems. Disturbances such asfire and harvest of plants or peat can be thedominant avenues of carbon losses from ecosys-tems in the years when they occur. In manycases the carbon losses with disturbance are solarge that they become significant componentsof long-term carbon budgets (Fig. 6.8; Box 6.1).Carbon losses during fires in the Canadianboreal forest, for example, are equivalent to 10to 30% of average NPP (Harden et al. 2000).

Controls over Net Ecosystem Production

NEP is determined by factors that cause animbalance between carbon gain and loss. Anecosystem is never at equilibrium at anymoment in time. NEP varies with season, timesince disturbance, interannual variation inweather, and long-term trends in environment.High-latitude ecosystems, for example, are a netcarbon source in warm years and a carbon sinkin cool years (Goulden et al. 1998) because het-erotrophic respiration responds to temperaturemore strongly than does photosynthesis in coldclimates. We expect NEP to change in responseto long-term changes in any factor that differs inits effects on GPP and the various avenues ofcarbon loss from ecosystems (e.g., plant respira-tion, heterotrophic respiration, disturbance,or leaching loss). Increased concentrations ofatmospheric CO2 or nitrogen inputs to ecosys-tems, for example, have greater direct effects onGPP than on decomposition, whereas a reduc-tion in the soil moisture of poorly drained wetlands increases decomposition and fireprobability more strongly than it affects NPP.Human activities are currently having greatestimpact on precisely those environmental factorsthat we expect to affect plants and decomposersdifferentially and therefore to affect global ter-restrial carbon storage (see Chapter 15).

Net carbon accumulation by an ecosystemdepends more strongly on time since distur-bance than on climate. The greatest causes ofvariation in NEP among ecosystems are cycles

Figure 6.9. Representative seasonal pattern of grossprimary production, ecosystem respiration, and netecosystem production of an ecosystem. NEP is thedifference between two large fluxes (carbon inputsin GPP and carbon losses, of which Recosyst and Fleach

are generally greatest).Annual CO2 flux in this graphis at steady state because the NEP summed over theannual cycle is close to zero. Here, carbon losses dueto disturbance are assumed to be zero.


of disturbance and succession (see Chapter 13).Most disturbances initially cause a large nega-tive NEP. Fire, for example, releases carbondirectly by combustion. Fire also removes vegetation that transpires water and shades the ground surface. The warm, moist soils inrecently burned sites promote decompositionand leaching loss and reduce plant biomass and therefore NPP. The net effect is a negativeNEP during and in the first years after fire(Kasischke et al. 1995, Harden et al. 2000) (seeFig. 13.11). Agricultural tillage breaks up soilaggregates and increases access of soil microbesto soil organic matter. The resulting increase indecomposition can lead to a negative NEP afterconversion of natural ecosystems to agriculture.During succession after disturbance there istypically a rapid increase in plant biomass andsoil organic matter, because NPP increasesmore rapidly during succession than doesdecomposition (see Chapter 13). The NEP of agiven stand therefore fluctuates dramaticallythrough cycles of disturbance and succession.The spatial scale of these imbalances betweenNPP and decomposition ranges from individualgopher mounds or treefall gaps to large distur-bances such as those caused by fire, forestharvest, or regional programs of land conver-sion to agriculture.

Because of the sensitivity of NEP to succes-sional status, NEP estimated at the regionalscale depends on the relative abundance ofstands of different ages. NEP at the regionalscale is termed net biome production (NBP)(Schulze et al. 2000). At times of increasing dis-turbance frequency, NBP is likely to be nega-tive. Conversely, areas that have experiencedwidespread abandonment of agricultural landsin the last century, as in Europe or the north-eastern United States, may experience a posi-tive NBP. Inadequate information on theregional variation in disturbance frequency andNBP is one of the greatest causes of uncertaintyin explaining recent changes in the globalcarbon cycle (see Chapter 15).

Net Ecosystem Exchange

Net ecosystem exchange (NEE) provides adirect measure of the net CO2 exchange

between ecosystems and the atmosphere. Oneof the greatest impediments to understandingthe carbon dynamics of ecosystems is that wecannot directly measure most of the componentprocesses at the ecosystem scale. GPP (net photosynthesis) cannot be readily separatedfrom respiration by nonphotosynthetic plantparts.We can directly measure only some of thecomponents of NPP, such as the accumulationof plant biomass. Decomposition is not easilyseparated from root respiration at the ecosys-tem scale. Many of the most pressing societalissues surrounding ecosystem carbon dynamics,however, revolve around NEP, the balance ofcarbon inputs and outputs. An important toolin improving our estimates of NEP has been anenhanced ability to measure NEE, which is thenet exchange of CO2 between the ecosystemand the atmosphere. NEE is the balancebetween GPP and ecosystem respiration(Recosyst), the sum of plant and heterotrophicrespiration—that is, the total respiration by anecosystem.

(6.3)

NEE, which excludes fluxes associated with dis-turbance and leaching, is the largest componentof NEP in most ecosystems most of the time(Box 6.1).

GPP (net photosynthesis) is zero in the dark,so NEE is a direct measure of ecosystem respi-ration (Recosyst) under these conditions.

NEEdark = -Recosyst [in the dark] (6.4)

The total diurnal Recosyst can be estimatedfrom simple models of Recosyst as an exponentialfunction of temperature (see Fig. 7.4). Duringthe day, NEE is approximately equal to the sumof GPP and ecosystem respiration.

NEElight ª GPP - Recosyst [in the light]

(6.5)

or

GPP ª NEElight + Recosyst (6.6)

The results of this calculation are only approx-imate, because mitochondrial respiration inleaves declines in the light, when much of the energy for metabolism comes directly

NEE GPP

GPPplant heterotr

ecosyst

= - +( )= -

R R

R


from carbon fixation (a component of GPP).Nonetheless, it is the closest thing to a direct measurement of GPP that is currentlyavailable.

A global network of sites measures NEEcontinuously in many of the world’s ecosys-tems. These measurements show that, in theabsence of disturbance, most temperate ecosys-tems that have been measured are net sinks forCO2 (Fig. 6.10) (Valentini et al. 2000).There areat least four possible explanations for thisimportant finding: (1) Ecosystems may typicallybe carbon sinks between episodes of distur-bance, and disturbance may be the factor thatbrings NEP into balance at the regional scale.(2) Recent environmental changes, such asincreased atmospheric CO2 and nitrogen depo-sition, may have stimulated photosynthesismore than respiration. (3) Midsuccessionalecosystems with high NEP may have been over-represented in the sampling network relative tothe rest of the world. Many western Europeanforests, where these studies were done, are pro-ductive midsuccessional sites that are develop-ing after agricultural abandonment. (4) Carbon

loss through leaching and other transfers maybe an important component of the regionalcarbon balance.These nongaseous losses wouldnot be detected in measurements of NEE.

A second striking result of this study is thatlatitudinal variation in NEE reflects variationsin ecosystem respiration rather than in GPP.Recent high-latitude warming could contributeto the greater respiration observed at high lat-itudes. At the few sites where there are long-term measurements of NEE, both respirationand photosynthesis contribute to interannualvariations in NEE (Goulden et al. 1996, 1998).Only recently has NEE been measured inenough ecosystems to begin to identify regionalpatterns in NEE and their likely causes.

Global Patterns of NEE

Seasonal and latitudinal variations in the CO2

concentration of the atmosphere provide a clearindication of global-scale patterns of NEE(Fung et al. 1987, Keeling et al. 1996b). At highnorthern latitudes, conditions are warm duringsummer, and photosynthesis exceeds total respiration (positive NEE), causing a decline in the concentration of atmospheric CO2

(Fig. 6.11). Conversely, in winter, when photo-synthesis is reduced by low temperature andshedding of leaves, respiration becomes thedominant carbon exchange (negative NEE),causing an increase in atmospheric CO2. Theseseasonal changes in the balance between photo-synthesis and respiration occur synchronouslyover broad latitudinal bands, giving rise toregular annual fluctuations in atmospheric CO2,literally the breathing of the biosphere (i.e., alllive organisms on Earth) (Fung et al. 1987).

Latitudinal variations in climate modifythese patterns of annual carbon exchange. Incontrast to the striking seasonality of NEE atnorth temperate and high latitudes, the con-centration of atmospheric CO2 remains nearlyconstant in the tropics, because carbon uptakeby photosynthesis is balanced by approximatelyequal carbon loss by respiration throughout the year. In other words, NEE is close to zerothroughout the year. There is also relativelyweak seasonality of atmospheric CO2 at highsouthern latitudes where oceans occupy most

-1

-0.5

0

0.5

1

1.5

40 45 50 55 60 65

NEE

Respiration

GPP

Car

bon

exch

ange

(kg

m-2

yr-1

)

Latitude

C loss

C uptake

Figure 6.10. Latitudinal variation in annual grossprimary production, net ecosystem exchange, andecosystem respiration (Recosyst) among 12 naturallyoccurring European forests (Valentini et al. 2000).The greater NEE of low-latitude forests reflectedtheir lower rates of ecosystem respiration. There wasno latitudinal trend in GPP.


of Earth’s surface. Carbon exchange in theoceans is largely determined by physicalfactors, such as wind, temperature, and CO2

concentration in the surface waters (seeChapter 15), which show less seasonal varia-tion. In summary, the global patterns of varia-tion in atmospheric CO2 concentration provideconvincing evidence that carbon exchange byterrestrial ecosystems is large in scale and sensitive to climate.

The final general pattern evident in theatmospheric CO2 record is a gradual increase inCO2 concentration from one year to the next(Fig. 6.11), primarily a result of fossil fuel inputsto the atmosphere that began with the Indus-trial Revolution in the nineteenth century (seeChapter 15). The rising concentration of atmo-spheric CO2 is an issue of international concernbecause CO2 is a greenhouse gas that con-tributes to climate warming. Note that theinterannual variation in CO2 concentrationcaused by biospheric exchange is much largerthan the annual CO2 increase. If there weresome way to increase net carbon uptake by

ecosystems over the long term, this mightreduce the rate of climate warming. There aretherefore important societal reasons for under-standing the controls over NEP in terrestrialecosystems.

Summary

Plant respiration provides the energy to acquirenutrients and to produce and maintain biomass.All plants are similar in their efficiency of con-verting sugars into biomass. Therefore, ecosys-tem differences in plant respiration largelyreflect differences in the amount and nitrogencontent of biomass produced and, secondarily,in the effects of environmental stress, particu-larly temperature and moisture, on mainte-nance respiration. Most ecosystems appear to exhibit a similar efficiency of converting photosynthate (GPP) into NPP; about half of the carbon gain becomes NPP, and the other half is returned to the atmosphere asplant respiration.

380

370

360

350

340

60oN

30oN

0o

60oS30oS

90oS93 94 95 96 97 98 99

90 91 92Year

Latitude

CO

2(p

pmv)

Figure 6.11. Seasonal and latitudinal variations inthe concentration of atmospheric CO2. Seasonal andlatitudinal variations in CO2 concentration reflectprimarily the balance of terrestrial photosynthesisand respiration. The upward trend in concentration

across years results from anthropogenic CO2 inputsto the atmosphere. (Redrawn with permission fromthe National Oceanic and Atmospheric Administra-tion, Climate Monitoring and Diagnostics Labora-tory, Carbon Cycle-Greenhouse Gases.)


Net primary production is the net carbongained by vegetation. It includes new plant biomass produced, root exudation, carbontransfers to root symbionts, and the emission ofvolatile organic compounds by plants. Biomedifferences in NPP correlate with climate at theglobal scale largely because temperature andprecipitation determine the availability of soilresources required to support plant growth.Plants actively sense the availability of theseresources and adjust photosynthesis and NPPto match this resource supply. For this reason,NPP is greatest in environments with highavailability of belowground resources. Afterdisturbance, NPP is often reduced below levelsthat the environment can support. Plants maximize production by allocating new growthto tissues that acquire the most limitingresources. Constantly shifting patterns of allo-cation reduce the degree of limitation of NPPby any single resource and make NPP in mostecosystems responsive to more than oneresource.

Tissue loss is just as important as NPP inexplaining changes in plant biomass. Pro-grammed loss of tissues provides a supply ofplant resources that supports new production.Biomass and NPP are greatest in warm,moist environments and least in environmentsthat are cold or dry. The length of the photo-synthetic season and leaf area are the twostrongest determinants of the global patterns in NPP. Most ecosystems have a similar (1 to 3g biomass m-2 of leaf d-1) daily NPP per unitleaf area.

Net ecosystem production is a measure of the rate of carbon accumulation in ecosystems.It correlates more strongly with time since disturbance than with environment. NEP isgenerally greatest in midsuccession, whenecosystems accumulate plant biomass andSOM. NEP is greater under conditions thatpromote NPP (e.g., elevated CO2, N deposition)than under conditions that promote decompo-sition. Net biome production integrates NEP atthe regional scale, taking account of regional patterns of disturbance and stand age. Humanactivities are altering most of the major controlsover NEP at a global scale in ways that arelikely to affect global climate.

Review Questions

1. What controls the partitioning of carbonbetween growth and respiration? Explainwhy the efficiency of converting sugars intonew biomass is relatively constant.

2. What factors influence the variability inmaintenance respiration?

3. Describe the multiple ways in which climateaffects the NPP of grasslands or tundra.

4. There is generally a close correlationbetween GPP and NPP. Describe themechanisms that account for short-termvariations in GPP and NPP (e.g., diurnaland seasonal variations).

5. Describe the mechanisms that account for the relationship between GPP and NPPwhen ecosystems from different climaticregimes are compared.

6. How does allocation to roots vs. shootsrespond to shade, nutrients, CO2, grazing, orwater?

7. How does variation in allocation influenceresource limitation, resource capture, andNPP?

8. Why do plants senesce tissues in which they have invested carbon and nutrientsrather than retaining tissues until they areremoved by disturbance or herbivory?How does this physiologically programmedsenescence influence NPP?

9. Describe the carbon budget of a plant andof an ecosystem in terms of GPP, respira-tion, and production. How would youexpect each of these parameters to respondto changes in temperature, water, light, andnitrogen?

10. How do the controls over NEP differ fromthe controls over GPP and decomposition.Why are these controls different?

Additional Reading

Chapin, F.S. III. 1991. Integrated responses of plantsto stress. BioScience 41:29–36.

Chapin, F.S. III, E.-D. Schulze, and H.A. Mooney.1990. The ecology and economics of storage inplants. Annual Review of Ecology and Systematics21:423–448.


Clark, D.A., S. Brown, D.W. Kicklighter, J.Q. Cham-bers, J.R. Thomlinson, and J. Ni. 2001. Measuringnet primary production in forests: Concepts andfield methods. Ecological Applications 11:356–370.

Lieth, H. 1975. Modeling the primary productivity ofthe world. Pages 237–263 in H. Lieth and R.H.Whittaker, editors. Primary Productivity of theBiosphere. Springer-Verlag, Berlin.

Poorter, H. 1994. Construction costs and paybacktime of biomass: A whole plant perspective. Pages111–127 in J. Roy and E. Garnier, editors. A Whole-

Plant Perspective on Carbon-Nitrogen Interactions.SPB Academic, The Hague.

Rastetter, E.B., and G.R. Shaver. 1992. A model ofmultiple element limitation for acclimating vege-tation. Ecology 73:1157–1174.

Schlesinger, W.H. 1977. Carbon balance in terrestrialdetritus. Annual Review of Ecology and Systemat-ics 8:51–81.


Introduction

Decomposition is the physical and chemicalbreakdown of detritus (i.e., dead plant, animal,and microbial material). Decomposition causesa decrease in detrital mass, as materials areconverted from dead organic matter into inorganic nutrients and CO2. If there were nodecomposition, ecosystems would quickly accu-mulate large quantities of detritus, leading to asequestration of nutrients in forms that areunavailable to plants and a depletion of atmos-pheric CO2. Depletion of these resources innondecomposing detritus would eventuallycause many biological processes to grind to ahalt. Although this has never occurred, therehave been times such as the Carboniferousperiod (see Fig. 2.12) when decomposition didnot keep pace with primary production, leadingto vast accumulations of carbon-containing coaland oil. The balance between NPP and decom-position therefore strongly influences carboncycling at ecosystem and global scales.

If the climate warming associated withanthropogenic CO2 emissions were to causeeven small changes in the balance between netprimary prouction (NPP) and decomposition,the CO2 concentration in the atmospherewould be greatly altered and therefore so

would the rate of climate warming. Under-standing the impacts of decomposition oncarbon cycling is thus critical for making pro-jections about the future state of Earth’sclimate.

Overview

The leaching, fragmentation, and chemicalalteration of dead organic matter by decompo-sition produces CO2 and mineral nutrients anda remnant pool of complex organic compoundsthat are resistant to further microbial break-down. Decomposition is a consequence ofinteracting physical and chemical processesoccurring inside and outside of living organ-isms. Decomposition results from three types of processes with different controls and con-sequences. (1) Leaching by water transferssoluble materials away from decomposingorganic matter into the soil matrix. Thesesoluble materials either are absorbed by organ-isms, react with the mineral phase of the soil, orare lost from the system in solution. (2) Frag-mentation by soil animals breaks large piecesof organic matter into smaller ones, whichprovide a food source for soil animals andcreate fresh surfaces for microbial colonization.

7Terrestrial Decomposition

Decomposition breaks down dead organic matter, releasing carbon to the atmos-phere and nutrients in forms that can be used for plant and microbial production.This chapter describes the key controls over decomposition and soil organic matteraccumulation by ecosystems.

151

152 7. Terrestrial Decomposition

Soil animals also mix the decomposing organicmatter into the soil. (3) Chemical alteration ofdead organic matter is primarily a consequenceof the activity of bacteria and fungi, althoughsome chemical reactions also occur sponta-neously in the soil without microbial mediation.

Dead plant material (litter) and animalresidues are gradually decomposed until theiroriginal identity is no longer recognizable, atwhich point they are considered soil organicmatter (SOM). Litter consists primarily of compounds that are too large and insoluble topass through microbial membranes. Microbestherefore secrete exoenzymes (extracellularenzymes) into their environment to initiatebreakdown of litter. These exoenzymes convertmacromolecules into soluble products that canbe absorbed and metabolized by microbes.Microbes also secrete products of metabolism,such as CO2 and inorganic nitrogen, andproduce polysaccharides that enable them toattach to soil particles.When microbes die, theirbodies become part of the organic substrateavailable for decomposition.

The controls over organic matter breakdownchange radically once soil organic matterbecomes incorporated into mineral soil. Thesoil moisture and thermal regimes of mineralsoil are quite different from those in the litterlayer. In the mineral soil, SOM can complexwith clay minerals or undergo nonenzymaticchemical reactions to form more complex compounds. Humus, for example, is a complexmixture of chemical compounds with highlyirregular structure containing abundant aro-matic rings. Humus tends to accumulate in soilsbecause exoenzymes cannot easily degrade itsirregular structure (Oades 1989).

Decomposition is largely a consequence ofthe feeding activity of soil animals (fragmenta-tion) and heterotrophic microbes (chemicalalteration). The evolutionary forces that shapedecomposition are those that maximize thegrowth, survival, and reproduction of soilorganisms. Controls over decomposition aretherefore best understood in terms of the controls over the activities of these organisms.The ecosystem consequences of decompositionare the mineralization of organic matter toinorganic components (CO2, mineral nutrients,

and water) and the transformation of organicmatter into complex organic compounds thatare recalcitrant (i.e., resistant to further micro-bial breakdown). In other words, decomposi-tion occurs to meet the energetic andnutritional demands of decomposer organisms,not as a community service for the carbon cycle.

Leaching of Litter

Leaching is the rate-determining step for massloss of litter when it first falls to the ground.Leaching is the physical process by whichmineral ions and small water-soluble organiccompounds dissolve in water and move throughthe soil. During leaf senescence, many of thecompounds in a leaf are broken down andtransported to other plant parts (see Chapter8).This resorption process is still actively occur-ring when the leaf is shed, so the senesced leafcontains relatively high concentrations ofwater-soluble breakdown products that arereadily leached. Leaching begins when tissuesare still alive and is most important duringtissue senescence and when litter first falls tothe ground. Leaching losses from litter are pro-portionally more important for nutrients thanfor carbon. Leaching losses from fresh litter aregreatest in environments with high rainfall andare negligible in dry environments. Compoundsleached from leaves include sugars, aminoacids, and other compounds that are labile(readily broken down) or are absorbed intactby soil microbes. Leachates frequently supporta pulse of microbial growth and respirationduring periods of high litterfall.

Litter Fragmentation

Fragmentation creates fresh surfaces for micro-bial colonization and increases the proportionof the litter mass that is accessible to microbialattack. Fresh detritus is initially covered by aprotective layer of cuticle or bark on plants orof skin or exoskeleton on animals. These outercoatings are designed, in part, to protect tissuesfrom microbial attack. Within plant tissues, thelabile cell contents are further protected from

Chemical Alteration 153

microbial attack by lignin-impregnated cellwalls. Fragmentation of litter greatly enhancesmicrobial decomposition by piercing these pro-tective barriers and by increasing the ratio oflitter surface area to mass.

Animals are the main agents of litter fragmentation, although freeze–thaw andwetting–drying cycles can also disrupt the cel-lular structure of litter. Animals fragment litteras a by-product of their feeding activities. Bears,voles, and other mammals tear apart wood or mix the soil as they search for insects,plant roots, and other food. Soil invertebratesfragment the litter to produce particles that are small enough to ingest. Enzymes in animalguts digest the microbial “jam” that coats thesurface of litter particles, providing energy and nutrients to support animal growth andreproduction.

Chemical Alteration

Fungi

Fungi are the main initial decomposers of ter-restrial dead plant material and, together withbacteria, account for 80 to 90% of the totaldecomposer biomass and respiration. Fungihave networks of hyphae (i.e., filaments thatenable them to grow into new substrates andtransport materials through the soil over dis-tances of centimeters to meters). Hyphal net-works enable fungi to acquire their carbon inone place and their nitrogen in another, muchas plants gain CO2 from the air and water andnutrients from the soil. Fungi that decomposelitter on the forest floor, for example, mayacquire carbon from the litter and nitrogenfrom the mineral soil. Fungi are the principaldecomposers of fresh plant litter, because theysecrete enzymes that enable them to penetratethe cuticle of dead leaves or the suberized exte-rior of roots to gain access to the interior of adead plant organ. Here they proliferate withinand between dead plant cells.At a smaller scale,some fungi gain access to the nitrogen andother labile constituents of dead cells by break-ing down the lignin in cell walls. This energyinvestment in lignin-degrading enzymes serves

primarily to gain access to the relatively labilecontents of the interior of cells.

Fungi produce hyphae with a dense concen-tration of cytoplasm when there is adequatesubstrate to support growth. The hyphaecontain more vacuoles (and proportionally lesscytoplasm) when resources are scarce.This flex-ible growth strategy enables fungi to grow intonew areas to explore for substrate, even whencurrent substrates are exhausted. A substantialproportion (perhaps 25%) of the carbon andnitrogen used to support fungal growth aretransported from elsewhere in the hyphalnetwork, rather than being absorbed from theimmediate environment where the fungalgrowth occurs (Mary et al. 1996).

Fungi have enzyme systems capable ofbreaking down virtually all classes of plantcompounds.They have a competitive advantageover bacteria in decomposing tissues with lownutrient concentrations because of their abilityto import nitrogen and phosphorus. White-rotfungi specialize on lignin degradation in logs,whereas brown-rot fungi cleave some of theside-chains of lignin but leave the phenol unitsbehind (giving the wood a brown color).White-rot fungi are generally outcompeted by morerapidly growing microbes when nitrogen isabundant, so nitrogen additions have littleeffect (or sometimes a negative effect) onwhite-rot fungal decomposition of wood.

Fungi account for 60 to 90% of the microbialbiomass in forest soils, where litter frequentlyhas a high lignin and low nitrogen concentra-tion. They have a competitive advantage at lowpH, which is also common in forest soils. Fungimake up about half the microbial biomass ingrassland soils, where pH is higher and wood isabsent. Most fungi lack a capacity for anaero-bic metabolism and are therefore absent from or dormant in anaerobic soils and aquaticsediments.

Mycorrhizae are a symbiotic associationbetween plant roots and fungi in which theplant gains nutrients from the fungus in returnfor carbohydrates (see Chapter 8). Althoughmycorrhizal fungi get most of their carbon fromplant roots, they can also play a role in decom-position by breaking down proteins into aminoacids, which are absorbed; amino acids both


support fungal growth and are transferred totheir host plants (Read 1991). Mycorrhizalfungi also produce cellulases to gain entry intoplant roots, but it is uncertain whether thesecellulases participate in decomposition of deadorganic matter.

Bacteria

The small size and large surface to volume ratioof bacteria enable them to absorb soluble sub-strates rapidly and to grow and divide quicklyin substrate-rich zones. This opportunist strat-egy explains the bacterial dominance in the rhizosphere (the zone of soil directly influencedby plant roots) and in dead animal carcasses,where labile substrates are abundant. Bacteriaare also important in lysing and breaking downlive and dead bacterial and fungal cells. Themajor functional limitation resulting from theirsmall size is that each bacterium completelydepends on the substrates that move to it. Someof these substrates are products of bacterialexoenzymes. These products diffuse to the bac-terium along a concentration gradient createdby the activity of the exoenzymes (whichproduce soluble substrates), and by the uptakeof substrates by the bacterium (which re-duces substrate concentrations at the bacterialsurface). Other soluble substrates flow past thebacterium in water films moving through thesoil. This water movement is driven by gradi-ents in water potential associated with planttranspiration, evaporation at the soil surface,and gravitational water movement after rain(see Chapter 4). Water movement (and there-fore the supply rate of substrates) is most rapidin macropores (relatively large air or waterspaces between aggregates). Bacteria thereforeoften line the macropore surfaces and absorbsubstrates from the flowing water, just as fish-ermen net salmon migrating up a stream or an intertidal filter-feeder extracts organic par-ticles from the water column. Macropores are also preferentially exploited by roots becauseof the reduced physical resistance to root elongation, providing an additional source oflabile substrates to bacteria. Bacteria attachedto the exposed surfaces of macropores are vul-nerable to predation by protozoa and nema-todes, which use the water films in macropores

as highways to move through the soil.This leadsto rapid bacterial turnover on exposed particlesurfaces.

There is a wide range of bacterial types insoils. Rapidly growing gram-negative bacteriaspecialize on labile substrates secreted by roots. Actinomycetes are slow-growing, gram-positive bacteria that have a filamentous struc-ture similar to that of fungal hyphae. Like fungi,actinomycetes produce lignin-degradingenzymes and can break down relatively recal-citrant substrates. They often produce fungi-cides to reduce competition from fungi.

The bacterial communities that coat soilaggregates exhibit a surprisingly complex struc-ture. They are often present as biofilms, amicrobial community embedded in a matrix of polysaccharides secreted by bacteria. Thismicrobial “slime” protects bacteria fromgrazing by protozoa and reduces bacterialwater stress by retaining water. The matrix alsoincreases the efficiency of bacterial exoenzymesby preventing them from being swept away inmoving water films. The bacteria in biofilmsoften act as a consortium—that is, a group ofgenetically unrelated bacteria, each of whichproduces only some of the enzymes required tobreak down complex macromolecules. Thebreakdown of these molecules to the point thatsoluble products are released requires the coor-dinated production of exoenzymes by severaltypes of bacteria.This is analogous to an assem-bly line, in which the final product, such as a caror a television set, depends on the coordinatedaction of several consecutive steps; as with anassembly line, no bacterium benefits unless allthe steps are in place to produce the finalproduct. The evolutionary forces and popula-tion interactions that shape the composition ofmicrobial consortia are virtually unknown.Consortia are particularly important in thebreakdown of pesticides and other organicresidues that humans have added to the environment.

Most bacteria are immobile and move pas-sively through the soil, carried by soil water oranimals. An important consequence of immo-bility is that a bacterial colony eventuallyexhausts the substrates in its immediate envi-ronment, especially in microenvironmentswithin soil particles that have restricted water

Chemical Alteration 155

movement. When bacteria exhaust their sub-strate, they become inactive and reduce theirrespiration to negligible rates. Bacteria mayremain inactive for years. Live bacteria havebeen recovered from permafrost that is tens ofthousands of years old. Between 50 and 80% ofthe bacteria in soils are metabolically inactive(Norton and Firestone 1991). Inactive bacteriareactivate in the presence of labile substrates,for example, when a root grows through the soiland exudes carbohydrates. The inactive bacte-ria in soils represent a reservoir of decomposi-tion potential analogous to the buried seedpool, which is an important source of plant col-onizers after a disturbance. Like the buriedseed pool, the enzymatic potential of theseinactive bacteria may be different from theenzymes produced by the active bacterial com-munity. Consequently, DNA probes or microbi-ological culturing techniques are better indicesof what the soil could do (its metabolic poten-tial) than of its actual metabolic activity at anygiven time.

Soil Animals

Soil animals influence decomposition by frag-menting and transforming litter, grazing popu-lations of bacteria and fungi, and altering soil structure. The microfauna is made up of the smallest animals (less than 0.1mm).They include nematodes; protozoans, such asciliates and amoebae; and some mites (Fig. 7.1)(Wallwork 1976, Lousier and Bamforth 1990).Protozoans consist of a single cell and ingesttheir prey primarily by phagocytosis—that is, byenclosing them in a membrane-bound structurethat enters the cell. Protozoans are usuallymobile and are voracious predators of bacteriaand other microfauna species (Lavelle et al.1997). Nematodes are an abundant and trophi-cally diverse group in which each species specializes on bacteria, fungi, roots, or other soil animals. Bacterial-feeding nematodes inforest litter, for example, can consume about 80gm-2 yr-1 of bacteria, resulting in the miner-alization of 2 to 13gm-2 yr-1 of nitrogen—

Figure 7.1. Representative types and sizes of soilfauna. Microfauna are most important as predators;mesofauna, as organisms that fragment litter; and

macrofauna, as ecosystem engineers. (Redrawn withpermission from Blackwell Scientific; Swift et al.1979.)


a substantial proportion of the nitrogen thatannually cycles through the soil (Anderson etal. 1981). Protozoans are particularly importantpredators in the rhizosphere and other soilmicrosites that have rapid bacterial growthrates (Coleman 1994). The preferential grazingby protozoa on bacteria (even on particularspecies of bacteria), alters the microbial com-munity composition, tending to reduce bacter-ial to fungal ratios in these soils compared tosoils from which protozoa are excluded. Proto-zoans and nematodes are aquatic animals thatmove through water films on the surface of soilparticles and are therefore more sensitive towater stress than are fungi and the mesofaunaand macrofauna species that fragment soil particles. Their populations fluctuate dramati-cally, both spatially and temporally, due todrying–wetting events and to predation (Beareet al. 1992). When protozoans die, their bodiesare rapidly broken down by soil microbes, espe-cially by bacteria.

The mesofauna includes a taxonomicallydiverse group of soil animals 0.1 to 2mm inlength (Fig. 7.1). They are the animals that havethe greatest effect on decomposition. Meso-fauna species fragment and ingest litter coatedwith microbial biomass, producing largeamounts of fecal material that has greatersurface area and moisture-holding capacitythan the original litter (Lavelle et al. 1997).Thisaltered litter environment is more favorable fordecomposition. These organisms selectivelyfeed on litter that has been conditioned bymicrobial activity. Collembola are small insectsthat feed primarily on fungi, whereas mites(Acari) are a more trophically diverse group ofspiderlike animals that consume decomposinglitter or feed on bacteria and/or fungi.

Large soil animals (the macrofauna), such asearthworms and termites, are ecosystem engi-neers that alter resource availability by modi-fying the physical properties of soils and litter(Jones et al. 1994). Some of them fragmentlitter, like the mesofauna species (Lavelle et al.1997). Others burrow or ingest soil, reducingsoil bulk density, breaking up soil aggregates,and increasing soil aeration and the infiltrationof water (Beare et al. 1992). The passagescreated by earthworms create channels in the

soil through which water and roots readily pen-etrate. They create patterns of soil structurethat promote or constrain the activities of soilmicrobes and other soil animals. In temperatepastures, earthworms may process 4kgm-2 yr-1

of soil, moving 3 to 4mm of new soil to theground surface each year (Paul and Clark1996). This is a geomorphic force that is, onaverage, orders of magnitude larger than land-slides or surface soil erosion (see Table 3.1). Soilmixing by earthworms tends to disrupt the for-mation of distinct soil horizons. Once the soilenters the digestive tract of an earthworm,mixing and secretions by the earthworm stimu-late microbial activity, so soil microbes act asgut mutualists. Many of the soil organisms arelysed and digested during passage through thegut; the resulting products are absorbed by theearthworm. Earthworms are most abundant in the temperate zone, whereas termites are the dominant ecosystem engineers in tropicalsoils. Termites eat plant litter directly, digest thecellulose with the aid of mutualistic protozoansin their guts, and mix the organic matter intothe soil. Dung beetles in tropical grasslandsperform a similar function with mammaliandung. This burial of surface organic matterplaces it in a humid environment where decom-position occurs more rapidly.

The soil fauna is critical to the carbon andnutrient dynamics of soils. Microbes contain 70 to 80% of the labile carbon and nitro-gen in soils, so variations in predation rates of microbes by animals dramatically altercarbon and nitrogen turnover in soils. Soilanimals have high respiration rates and metab-olize much of the microbial carbon from theirfood to CO2 to support their high energeticcosts of movement. As a result, the microbialnitrogen and phosphorus acquired by soilanimals generally exceeds their requirementsfor growth and reproduction. These nutrientsare therefore excreted and become availablefor plant uptake (see Chapter 8). Soil animalsaccount for only about 5% of soil respiration,so their major effect on decomposition is theirenhancement of microbial activity throughfragmentation (Wall et al. 2001), rather thantheir own processing of energy derived fromdetritus. Soil food webs are complex (see

Temporal and Spatial Heterogeneity of Decomposition 157

Chapter 11), so many of the effects of soilanimals on decomposition are indirect. Loss orexclusion of soil invertebrates can reducedecomposition rate (and therefore nutrientcycling) substantially, indicating the importantrole of animals in the decomposition process(Swift et al. 1979, Verhoef and Brussaard 1990).

Temporal and SpatialHeterogeneity of Decomposition

Temporal Pattern

The predominant controls over decompositionchange with time. Decomposition is the conse-quence of the interactions of fragmentation,chemical alteration, and leaching. As soon as aleaf unfolds, it is colonized by aerially bornebacteria and fungal spores that begin breakingdown the cuticle and leaf surfaces that areexposed by herbivores, pathogens, or physicalbreakage (Haynes 1986). This phyllospheredecomposition of live leaves is generallyignored because it is not readily separated fromplant-controlled changes in leaf mass andchemistry. It does, however, provide a microbialinnoculum that rapidly initiates decompositionof labile substrates when the leaf falls to theground. Similarly, breakdown of the root cortexbegins while the conducting tissues of roots stillfunction in water and nutrient transport.

As litter decomposes, its mass decreasesapproximately exponentially with time. Leaflitter frequently loses 30 to 70% of its mass inthe first year and another 20 to 30% of its massin the next 5 to 10 years (Haynes 1986). Anexponential decline in litter mass implies that aconstant proportion of the litter is decomposedeach year.

Lt = L0e-kt (7.1)

(7.2)

where L0 is the litter mass at time zero, and Lt

is the mass at time t. The decomposition con-stant, k, is an exponent that characterizes thedecomposition rate of a particular material.Themean residence time, or the time required for

lnLL

ktt

0= -

the litter to decompose under steady-state con-ditions, equals 1–

k. Residence time of litter can

also be estimated as the average pool size oflitter divided by the average annual input.

(7.3)

The estimation of residence times from poolsand fluxes assumes that the ecosystem is insteady state, which is often not the case (seeChapter 1). Midsuccessional ecosystems, forexample, generally receive more litterfall inputthan would occur at steady state, leading to an overestimate of k. Year-to-year variation inweather or directional changes in climate causemore rapid changes in litterfall than in the litterpool, also creating biases in estimates of resi-dence time. The decomposition constant varieswidely with substrate composition. Sugars, forexample, have a residence time of hours to days,whereas lignin has a residence time of months todecades, depending on the ecosystem. Plant andanimal tissues differ substantially in their chem-ical composition and therefore in their decayconstants. Taken as a whole, litter generally hasa residence time of months to years, and organicmatter mixed with mineral soil has a residencetime of years to centuries.

The exponential model of decomposition(Eq. 7.1), which implies a constant decomposi-tion rate, is only a rough approximation of thepattern of decline in litter mass with time. Theprocess is more accurately described by a curvewith at least three phases (Fig. 7.2). During thefirst phase, leaching of cell solubles is the pre-dominant process. Fresh litter can lose 5% of itsmass in 24h due to leaching alone. The secondphase of decomposition occurs more slowly andinvolves a combination of fragmentation by soilanimals, chemical alteration by soil microbes,and leaching of decay products from the litter.Decomposition during this second phase isoften measured as mass loss from dead leaves(Aerts 1997), roots (Berg et al. 1998), or twigsthat are tethered on threads or placed in meshlitter bags and weighed periodically (Vogt et al.1986, Robertson and Paul 2000). The exponen-tial model of decomposition has been applied

lk

k= =litter poollitterfall

orlitterfall

litter pool


primarily to this second phase. The final phaseof decomposition occurs quite slowly andinvolves the chemical alteration of organicmatter that is mixed with mineral soil and theleaching of breakdown products to other soillayers. Decomposition during this final phase isoften estimated from measurements of soil res-piration or isotopic tracers (Schlesinger 1977,Trumbore and Harden 1997). The decomposi-tion rate and decomposition constant (k in Eq.7.1) gradually decline through these threephases of decomposition.

In seasonal environments, microbial respira-tion often occurs over a longer time period andpeaks later in the season than does plantgrowth. Like plant growth, microbial respira-tion is favored by warm, moist conditions andis therefore greatest during the season ofmaximum plant growth. Heterotrophic respira-tion, however, typically begins earlier and endslater than does plant growth for at least three

reasons: (1) Microbial respiration typicallyoccurs over a broader range of temperatures(e.g., -10° to 40°C) and soil moistures thanplant growth. (2) The soil is buffered from tem-perature extremes that aboveground parts ofplants must cope with. (3) Because soils warmmore slowly than the air, heterotrophic respi-ration generally lags behind gross primary production (GPP), with relatively low rates inearly spring, when leaf growth is most active.Heterotrophic respiration continues in autumnand winter, long after leaf senescence. Micro-bial activity is also influenced by the seasonal-ity of plant activity. Root turnover andexudation are often greatest in midseason whenphotosynthesis is high, contributing to the mid-season peak in soil respiration. Autumn senes-cence provides an additional input of substratesthat contributes to late-season soil respiration.

Spatial Pattern

Most decomposition occurs near the soilsurface, where litter inputs are concentrated.Gravity carries most aboveground litter to theground surface, where the initial decompositionand nutrient release occur. Roots thereforetend to grow in surface soils to access thesenutrients. Thus root litter is also produced pri-marily in surface soils, reinforcing the surfacelocalization of most decomposition. Deep rootsare, however, not negligible, especially in dryenvironments. These “water roots” can befound to depths of 10 to 100m, depending onthe depth of the water table. Soil mixing byanimals, especially termites and earthworms,and leaching of dissolved organic matter alsotransfer surface carbon to depth. About half ofthe soil organic carbon therefore is typicallybelow 20cm depth, even though only a third of the roots are below that depth (Fig. 7.3)(Jobbágy and Jackson 2000). Deep-soil decom-position therefore cannot be ignored.The deep-soil carbon is often older, more recalcitrant, andmore strongly protected by complexes with soilminerals than is surface carbon (Trumbore andHarden 1997).

Decomposition rate is spatially heteroge-neous at several scales. The litter layer abovethe mineral soil exhibits large daily changes intemperature and moisture. Decomposition in

Cellulose and hemicellulose

Microbial products

Lignin

Tropics: 0

Arctic:

Mas

s re

mai

ning

(%

of o

rigi

nal)

Cellsolubles

Phase 1

0

0

1 2

5 20

3

100

50

100

Time (yr)

Phase 2 Phase 3

Figure 7.2. Representative time course of leaf-litterdecomposition showing the major chemical con-stituents (cell solubles, cellulose and hemicellulose,microbial products, and lignin), the three majorphases of litter decomposition, and the time scalescommonly found in warm (tropical) and cold (arctic)environments. Leaching dominates the first phase of decomposition. Substrate composition of litterchanges during decomposition because labile sub-strates, such as cell solubles, are broken down morerapidly than are recalcitrant compounds, such aslignin and microbial cell walls.

Factors Controlling Decomposition 159

this layer is dominated by fungi that importnitrogen from below. This is a radically differ-ent environment from the mineral soil, wheretemperature and moisture are more stable,some of the organic matter is humified andrecalcitrant, and mineral soil surfaces bind deadorganic matter and microbial enzymes. At afiner scale, the rhizosphere around roots is acarbon-rich microenvironment that supportsmuch higher microbial activity than the bulksoil, which is virtually a nutritional desert.Finally, the interior of soil aggregates is morelikely to be anaerobic than are the surfaces ofsoil pores. Movement within the soil of roots,water, and soil animals is constantly changingthe spatial arrangement of these different envi-ronments for decomposition.

In some ecosystems, such as tropical forests,significant quantities of aboveground litter arecaught on epiphytes and branches of thecanopy. In these wet ecosystems, substantialdecomposition, nutrient release, and nutrientuptake by rooted epiphytes occur in the canopyand short-circuit the soil phase (Nadkarni1981). Some terrestrial litter and dissolvedorganic carbon (DOC) also enter streams andlakes, where they become energy sources foraquatic food webs (see Chapter 10). In un-

productive ecosystems, the DOC that entersstreams is so recalcitrant that it remains largelyunprocessed, leading to the “black-water”rivers that characterize many tropical andboreal forests and temperate swamps.

Factors ControllingDecomposition

Decomposition is controlled by three types offactors: the physical environment, the quantityand quality of substrate available to decom-posers, and the characteristics of the microbialcommunity (Swift et al. 1979).

The Physical Environment

Temperature

Temperature affects decomposition directly bypromoting microbial activity and indirectly byaltering soil moisture and the quantity andquality of organic matter inputs to the soil.Rising temperature causes an exponentialincrease in microbial respiration over a broadtemperature range (Fig. 7.4), speeding up the mineralization of organic carbon to CO2.This temperature response is similar to thatobserved in respiration of most organisms. Atmoderate temperatures, most of the respiratoryenergy supports microbial growth. As tempera-ture increases, however, an increasing propor-tion of the energy is used for maintenance andmay not lead to a corresponding increase inmicrobial production. Microbial communitycomposition also changes in response to tem-perature toward a community dominated byindividuals that are adapted and acclimated to higher temperatures. There are thereforeseveral physiological and community changesthat account for the deceptively simpleresponse of microbial respiration to tempera-ture. Continuously high temperature, andtherefore rapid decomposition, explain whymany tropical forests have a small litter pooldespite their high productivity (Fig. 7.5).

Temperature also affects decompositionthrough freeze–thaw events. Freezing killsmany of the microbes present in decomposinglitter and SOM, releasing soluble organic mate-

0

10

20

30

40

50

60

70

0 20 40 60 80 100

Car

bon

pool

(%

of t

otal

)

Soil depth (cm)

Root carbon

Soil carbon

Figure 7.3. Globally averaged depth profiles of soilorganic matter and roots in the first meter of soil.(Redrawn with permission from Ecological Applica-tions; Jobbágy and Jackson 2000.)


rials into the soil. This pulse of available sub-strate can support rapid decomposition andnitrogen mineralization the following spring(Lipson et al. 1999). Freezing and thawing alsostimulates decomposition by physically disrupt-ing soil aggregates and the cellular structure of litter, thereby exposing fresh surfaces todecomposition. In some arctic ecosystems thedecomposition that occurs during autumn,

winter, and spring accounts for most of theannual litter mass loss (Hobbie and Chapin1996).

Temperature has many indirect effects ondecomposition (Fig. 7.6). High temperaturereduces soil moisture by increasing evaporationand transpiration. Soil drying reduces decom-position in dry climates but accelerates it wheresoils are wet enough to restrict oxygen supply.

10

20 30 400

2

4

6

8

12R

espi

ratio

n (µ

L O

2g-

1hr

-1)

Temperature (oC)

60

140

100

220

260

300

340

380

-10 -5 0 5 10 15 20 25 30 35 40 0 10

Temperature (oC)

Soi

l res

pira

tion

(µm

ol C

O2

m-2

s-1 )

20

180

A B

Figure 7.4. Relationship between temperature andsoil respiration in (A) laboratory incubations oftundra soils and (B) field measurements of soil res-piration in 15 studies, where data have been fitted to

Temperate conifer

Forest floor (g C m-2)

Temperatedeciduous

1400

1200

1000

800

600

400

200

00 1000 2000 3000 4000 5000 6000

Abo

vegr

ound

litte

rfal

l (g

C m

-2yr

-1)

Tropicalforest

k = 4 k = 1 k = 0.25

k = 0.12

k = 0.1

Figure 7.5. Forest-floor biomass and abovegroundlitter inputs for selected evergreen forests. Linesshow the relationship between aboveground litter-

have the same respiration rate at 10°C. (A, Flanaganand Veum 1974. B, Redrawn with permission fromFunctional Ecology; Lloyd and Taylor 1994.)

fall and forest floor mass for selected decompositionconstants. (Redrawn with permission from Ecology;Olsen 1963.)


The stimulation of microbial activity by warmtemperatures also initiates a series of feedbackloops that influence decomposition. The con-sumption of oxygen by microbial and root res-piration constrains decomposition in wet soilsor wet microsites (e.g., the interior of soil aggre-gates). On the other hand, the nutrientsreleased by decomposition at high tempera-tures increase the quantity and quality of litterproduced by plants, altering the substrate avail-able for decomposition. High temperatures alsoincrease the rate of chemical weathering, whichin the short term enhances nutrient supply. Incold climates, low temperature leads to a layerof permanently frozen soils (permafrost) thatrestricts drainage and therefore decomposition.Most of the indirect temperature effectsenhance soil respiration at warm temperaturesand contribute to the more rapid decomposi-tion observed in warm climates.

Moisture

Carbon accumulation is greatest in wet soilsbecause decomposition is more restricted byhigh soil moisture but is less restricted by lowsoil moisture than is NPP. Decomposers, likeplants, are most productive under warm moistconditions, provided sufficient oxygen is avail-able. This accounts for the high decomposition

rates in tropical forests (Gholz et al. 2000). Thedecomposition rate of mineral soil generallydeclines at soil moistures less than 30 to 50%of dry mass (Haynes 1986), due to the reduc-tion in thickness of moisture films on soil surfaces and therefore the rate of diffusion of substrates to microbes (Stark and Fire-stone 1995). Osmotic effects further restrict theactivity of soil microbes under conditions ofextremely low soil moisture or salt accumula-tion. Bacteria function at lower water availa-bility than do plant roots, so decompositioncontinues in soils that are too dry to supportplant activity. The high concentrations ofosmotic metabolites synthesized by microbes indry or saline conditions create severe osmoticgradients after soil wet-up, causing many micro-bial cells to burst. This results in pulses of nutri-ent availability after the first rains. Evenshort-term drying–wetting cycles, such as rainstorms or the daily formation and evaporationof dew, can strongly influence decomposition inthe litter layer and surface soils. The net effectof drying–wetting cycles is the stimulation ofdecomposition, if the cycles are infrequent (asgenerally occurs in soils). Frequent moisturefluctuations, as in the litter layer, can, however,reduce microbial population numbers to anextent that decomposition rates are reduced(Clein and Schimel 1994). Drying–wetting

Water availability

Oxygen

Microbialactivity

Plantgrowth

Litterquantity and

quality

Nutrient availability

Soilrespiration

Evaporation Weathering

Temperature

(+/-)

Moisture-mediatedeffects

Direct effects Nutrient-mediatedeffects

Rootrespiration

Figure 7.6. Direct and indirect effects of temperature on soil respiration.


cycles tend to stimulate the decomposition oflabile substrates (e.g., hemicellulose), which arebroken down largely by rapidly growing bacte-ria, and to retard the decomposition of recalci-trant ones (e.g., lignin) (Haynes 1986), whichare broken down by slow-growing fungi.

Decomposition is also reduced at high soilmoisture contents (e.g., greater than 100 to150% of soil dry mass in mineral soils) (Haynes1986). Oxygen diffuses 10,000 times moreslowly through water than through air, so wateracts as an effective barrier to oxygen supply todecomposers in wet soils or in wet micrositeswithin well-drained soils. Oxygen limitation todecomposition can occur for many reasons,including topographic controls over drainage,presence of hardpans or permafrost, high claycontent, or compaction by animals and agricul-tural equipment. Irrigation or rain events canlead to short-term oxygen depletion. In warmenvironments, the solubility of oxygen in wateris low, and oxygen is rapidly depleted by rootand microbial respiration, making decomposi-tion particularly sensitive to high soil moisture.Decomposition is also frequently oxygenlimited in bogs and wetlands and in arctictundra, where permafrost prevents drainageover a wide range of topographic situations.NPP is frequently less limited by high soil mois-ture than is decomposition, because manyplants that are adapted to these conditionstransport oxygen from leaves to the roots. Thelarge accumulations of SOM in histosol soils ofswamps and bogs at all latitudes clearly indicatethe importance of oxygen to decomposition.

Decaying logs create their own uniquemicroenvironment and generally have a highermoisture content than does the adjacentsurface litter. Log decomposition rate maytherefore be limited by oxygen supply at timeswhen microbes in neighboring surface litter aremoisture limited. The decomposition rate oflogs generally decreases with increasing logdiameter, because large logs generally havemore moisture and less oxygen than small ones.

Soil Properties

All else being equal, decomposition occursmore rapidly in neutral than in acidic soils due

to a variety of interacting factors. Fungi tend to predominate in acidic soils (Haynes 1986).The increase in bacterial abundance and theoverall increase in decomposition rate at higherpH probably reflects a complex of interactingfactors, including changes in plant species composition and associated changes in thequantity and quality of litter. Many factors can acidify soils, including cation leaching, aciddeposition, and the accumulation of organicacids in soil organic matter during succession.Alternatively, pH can increase in response todust input (Walker et al. 1998), particularly in deserts, braided river valleys of glacial landscapes, and degraded agricultural lands.Regardless of the cause of the change in acidityand associated plant species composition,low pH tends to be associated with low decom-position rates.

Clay minerals reduce the decomposition rateof soil organic matter, thereby increasing soilorganic content. Clays alter the physical envi-ronment of soils by increasing water-holdingcapacity (see Chapter 3). The resulting restric-tion in oxygen supply can reduce decomposi-tion in wet clay soils. Even at moderate soilmoisture, clays enhance organic accumulationby binding soil organic matter (making it lessaccessible to microbial enzymes); bindingmicrobial enzymes (reducing their effective-ness in breaking down substrates); and bindingthe soluble products of exoenzyme activity(making these products less available forabsorption by soil microbes). This binding oforganic matter to clays occurs because the highdensity of negatively charged sites on clay min-erals attract the positive charges on the organicmatter (amine groups) or form bridges withpolyvalent cations (Ca2+, Fe3+, Al3+, Mn4+) thatbind to negative groups (e.g., carboxyl groups)on organic matter (Fig. 7.7). The net effect ofthis binding by clay minerals is to protect soilorganic matter and reduce its decompositionrate. SOM protection by clay minerals is mostimportant in ecosystems such as grasslands andtropical forests, in which decomposition is rela-tively rapid and where soil animals rapidly mixfresh litter with mineral soil. Mineral protectionof SOM is less important in conifer forests andtundra in which much of the decomposi-


tion occurs above the mineral soil in a well-developed organic mat (O horizon).

Both the type and the quantity of clay influ-ence decomposition. Many tropical clay miner-als have a high aluminum concentration thatbinds tightly to organic matter through cova-lent bonds. Clays with a multilayered latticestructure bind organic compounds between thesilicate layers, making them particularly effec-tive in SOM protection (see Chapter 3).

Soil Disturbance

Soil disturbance increases decomposition bypromoting aeration and exposing new surfacesto microbial attack. The mechanism by whichdisturbance stimulates decomposition is basi-cally the same at all scales, ranging from themovement of earthworms through soils totillage of agricultural fields. Disturbance dis-rupts soil aggregates so the organic matter con-tained within them becomes more exposed tooxygen and microbial colonization. This distur-bance effect explains why the introduction ofEuropean earthworms to the northeasternUnited States considerably speeded forestdecomposition rates and why plowing causesrapid organic matter loss from grassland orforest soils after conversion to agriculture (seeFig. 14.12). This disturbance effect is most pro-nounced in warm wet soils, where the increased

aeration has greatest effect on decomposition.A soil converted to irrigated cotton, forexample, lost half its organic content in 3 to 5years (Haynes 1986), reversing a period of cen-turies to millennia that were required to accu-mulate soil organic matter. The loss of organicmatter and disruption of aggregates by plowingeventually impedes the drainage of water, thegrowth of roots, and the mineralization of soilnutrients.

Substrate Quality and Quantity

Litter

Carbon quality of substrates may be the pre-dominant chemical control over decomposi-tion. There is a 5-fold to 10-fold range indecomposition rate of litter in a given climate,due to differences in substrate quality—thatis, susceptibility of a substrate to decomposi-tion measured under standardized condi-tions. Animal carcasses decompose morerapidly than plants; leaves decompose morerapidly than wood; deciduous leaves decom-pose more rapidly than evergreen leaves; andleaves from high-nutrient environmentsdecompose more rapidly than leaves frominfertile sites (Figs. 7.8 and 7.9). These differ-ences in decomposition rate are a logical con-sequence of the types of chemical compounds

H

HH HOHH

HHHHHHH

H

CH

COOH

CH2

CH

CH

CH

C

O

R

OH2 NH

M

M MM

M

OH

OH

OH2

OH

NH

O

O

O

O

O

O

OO

O

O

O

OO

O

O

O

O

OO

O

O

O

O

O

O

O

O

C C

C

C(sugar)

N

N

R

(peptide)

H2O

HO

OH2

O

Clay mineral

Figure 7.7. The interactions between soil organic matter and clay particles, as mediated by water (H—O—H) and metal ions (M). (Redrawn with permission from Human Chemistry, 2nd edition by F.J.Stevenson, © John Wiley & Sons, Inc.)


present in litter. These compounds can be cate-gorized roughly as labile metabolic compounds,such as sugars and amino acids; moderatelylabile structural compounds, such as celluloseand hemicellulose; and recalcitrant structuralmaterial, such as lignin and cutin. Rapidlydecomposing litter generally has higher con-centrations of labile substrates and lower con-centrations of recalcitrant compounds thandoes slowly decomposing litter.

Five interrelated chemical properties oforganic matter determine substrate quality (J.Schimel, personal communication, 2001): thesize of molecules, the types of chemical bonds,the regularity of structures, the toxicity, and thenutrient concentrations. (1) Large moleculescannot pass through microbial membranes so they must be processed extracellularly byexoenzymes. This limits the degree of controlthat a given microbe can exert over the detec-tion of substrate availability, the delivery ofenzymes in response to substrate supply, andthe efficient use of breakdown products. Due todifferences in molecular size, sugars and amino

0

50

100

0 12 24Time (mo)

Mas

s re

mai

ning

(%

of o

rigin

al)

Pine branch

Pine needle

Pin-cherry leaf

Figure 7.8. Time course of decomposition of adeciduous leaf, a conifer needle, and wood in a Canadian temperate forest (MacLean and Wein1978).

20

0

10

20

30

40

50

60

70

80

Dry

mas

s lo

ss(%

of o

rigi

nal)

Vin

es

Her

bs

Her

bs

Shr

ubs

Tree

s

Gra

min

oids

Gra

min

oids

Sub

shru

bs

Tree

s

Shr

ubs

Sub

shru

bs

Deciduous leaves Deciduous leaves OtherEvergreen leaves

0

20

40

60

80

100

Tree

s &

shr

ubs

Suc

cule

nts

Bro

mel

iads

Aph

yllo

usTemperate climate Semi-arid climate

A B

Figure 7.9. A, Decomposition rate of leaves ofBritish deciduous and evergreen plant species.(Redrawn with permission from Journal of Ecology;

Cornelissen 1996). B, Decomposition rate of de-ciduous plants and aridzone plants in Argentina.(N. Perez; Perez-Harguindeguy et al. 2000.)


acids are metabolized more readily than cellu-lose and proteins, respectively. (2) Some chem-ical bonds are easier to break than others. Esterlinkages that bind phosphate to organic skele-tons and peptide bonds that link amino acids toform proteins, for example, are easier to breakthan the double bonds of aromatic rings. Forthese reasons, the nitrogen in proteins is muchmore available to microbes than the nitrogencontained in aromatic rings. (3) Compoundslike lignin that have a highly irregular structuredo not fit the active sites of most enzymes,so they are broken down much more slowly than are compounds like cellulose, whichconsist of chains of regularly repeating glucoseunits. (4) Some soluble compounds such as phe-nolics and alkaloids are toxic and kill or reducethe activity of microbes that absorb them. (5)Organic nitrogen and phosphorus are the majorsources of nutrients for supporting microbialgrowth, so organic matter, such as straw,that contains low concentrations of these ele-ments may not provide sufficient nutrients toallow microbes to use fully the carbon presentin the litter.

All of these chemical properties influencedecomposition, but their relative importance isnot well understood. Nonetheless, any of theseproperties can serve as a predictor of decom-position rate because the properties tend to bestrongly correlated with one another. The ratioof carbon concentration to nitrogen concentra-tion (C:N ratio), for example, has frequentlybeen used as an index of litter quality, becauselitter with a low C :N ratio (high nitrogen concentration) generally decomposes quickly(Enríquez et al. 1993, Gholz et al. 2000).However, neither the nitrogen concentration of the litter nor the nitrogen availability in thesoil directly influences the decomposition ratein most natural ecosystems (Haynes 1986,Prescott 1995, Prescott et al. 1999, Hobbie andVitousek 2000); this suggests that C :N ratio isnot the chemical property that directly controlsdecomposition in these ecosystems. This con-trasts with agricultural residues such as straw,which have a low nitrogen concentration and ahigh concentration of moderately labile carbonsources like cellulose and hemicellulose. Nitro-

gen concentration appears to limit directly thedecomposition rate of organic matter primarilywhen labile carbon substrates are available tosupport microbial growth (Haynes 1986). Thisis more likely to occur in the rhizosphere thanin fresh litter. Under other circumstances,carbon lability rather than nitrogen may be the primary control over decomposition rate(Hobbie 2000). Despite our uncertainty of themechanistic role of C :N ratio in decomposi-tion, many biogeochemical models use thisratio as a predictor of decomposition rate whendifferent ecosystem types are compared (seeChapter 9).

In recalcitrant litter, the concentration oflignin or the lignin :N ratio is often a good pre-dictor of decomposition rate (Berg and Staaf1980, Melillo et al. 1982, Taylor et al. 1989) (Fig. 7.10), again suggesting an important roleof carbon quality in determining decompositionrates of litter. The carbon quality of litter isprobably best defined in terms of the classes oforganic compounds present and the enzymaticpotential of the decomposer community, asdescribed later.This information is available forso few ecosystems, however, that more readilymeasured properties, such as C :N ratio or

Initial lignin:nitrogen ratio 10 20 30 40 60 70 80 90

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

50

Dec

ompo

sitio

n co

nsta

nt

Figure 7.10. Relationship between the decomposi-tion constant and the lignin : N ratio of litter.(Redrawn with permission from Ecology; Mellillo etal. 1982.)


lignin :N ratio, are frequently used as predictorsof decomposition rate.

The effects of litter quality on decompositionrate often depend on the age of the litter. High-quality litter, for example, loses its labile carbonso quickly that the remaining old litter mayhave a lower decomposition potential thanlitter that initially had a low litter quality andslow decomposition rate (Berg and Ekbohm1991).

The availability of belowground resources isthe major ecological control over litter quality.Rapidly growing plants from high-resourcesites typically produce litter that decomposesquickly because the same morphological andchemical traits that promote NPP also regulatedecomposition (Hobbie 1992). Both NPP anddecomposition are enhanced by a high alloca-tion to leaves and by the production of leaveswith a short life span. These tissues decomposerapidly because they have high concentrationsof labile compounds such as proteins and low concentrations of recalcitrant cell-wallcomponents such as lignin (Reich et al.1997). Consequently, species from productivesites produce litter that decomposes rapidly(Cornelissen 1996) (Fig. 7.9). Species differ-ences in litter quality make up an importantmechanism by which plant species affectecosystem processes (see Chapter 12) (Hobbie1992) and are excellent predictors of landscapepatterns of litter decomposition (Flanagan andVan Cleve 1983).

Soil Organic Matter

Both the age and the initial quality of SOMinfluence its rate of decomposition. As litterdecomposes, its decomposition rate declines,because microbes first consume the more labilesubstrates, leaving progressively more recalci-trant compounds in the remaining litter (Fig.7.2). Through fragmentation by soil inverte-brates and these chemical alterations, the litterbecomes converted to soil organic matter. Asmicrobes die, chitin and other recalcitrant com-ponents in their cell walls comprise an increas-ing proportion of the litter mass (actually litterplus microbial mass), and nonenzymatic reac-tions produce recalcitrant humic compounds.

All these processes contribute to a gradualreduction in organic matter quality as SOMages. The C :N ratio also declines as decompo-sition proceeds, because carbon is respiredaway, and some of the mineralized nitrogen isincorporated into humus. The decline in C :Nratio is not, however, an indicator of increasednitrogen availability, because the nitrogenbecomes incorporated into aromatic rings andother chemical structures that are recalcitrant.In summary, in SOM, as in litter, the carbonquality is a better predictor of decompositionrate than is the C :N ratio or the nitrogen concentration of SOM (Berg and Staaf 1980,Melillo et al. 1982).

Site differences in nutrient availability influ-ence SOM decomposition primarily throughtheir effects on the carbon quality of litter andSOM, rather than through direct nutrienteffects on SOM decomposition. Sites with highproductivity and litter quality typically producea low-lignin SOM that decomposes readily(Van Cleve et al. 1983). As in the case of freshlitter, SOM decomposition rate does not showa consistent response to nutrient addition(Haynes 1986, Fog 1988), suggesting that nutri-ents seldom directly regulate SOM decomposi-tion. Decomposition of SOM increases inresponse to nitrogen addition primarily whenthe organic matter consists of labile carbon sub-strates, for example when straw is plowed intoagricultural soils (Mary et al. 1996) or whenroot exudation is enhanced by elevated CO2

(Hu et al. 2001) (see Chapter 9).The heterogeneous nature of SOM makes it

difficult to identify the chemical controls overits decomposition. It is a mixture of organiccompounds of different ages and chemical com-positions. Components of SOM include frag-ments of recently shed root and leaf litter,together with soil organic matter that is thou-sands of years old (Oades 1989). These differ-ent aged components of SOM can be separatedby density centrifugation, because recently pro-duced particles are less dense than older onesand are less likely to be bound to mineral par-ticles. Soils in which a large proportion of theSOM is in the light fraction generally havehigher decomposition rates (Robertson andPaul 2000).Alternatively, soil can be chemically


separated into distinct fractions, such as water-soluble compounds, humic acids, and fulvicacids, that differ in average age and ease ofbreakdown. SOM as a whole typically has a res-idence time of 20 to 50 years, although this canrange from 1 to 2 years in cultivated fields tothousands of years in environments with slowdecomposition rates. Even in a single soil, dif-ferent chemical fractions of SOM have resi-dence times ranging from days to thousands ofyears. Computer simulations of decompositionrate capture ecosystem carbon dynamics moreeffectively when they distinguish among thesedifferent soil carbon pools (Parton et al. 1993,Clein et al. 2000).

Decomposition in the rhizosphere is morerapid than in bulk soil for reasons that arepoorly understood. The rhizosphere makes upvirtually all the soil in fine-rooted grasslands,where the average distance between roots isabout 1mm, whereas forests are less denselyrooted (often 10mm between roots) (Newman1985). Roots alter the chemistry of the rhizos-phere by secreting carbohydrates and absorb-ing nutrients.These processes are most active inthe region behind the tips of actively growingroots (Fig. 7.11) (Jaeger et al. 1999b). Thegrowth of bacteria in the zone of exudation(Norton and Firestone 1991) is supported byabundant carbon availability (20 to 40% ofNPP; see Table 6.2) and is therefore limitedmost strongly by nutrients (Cheng et al. 1996).Bacteria must acquire their nutrients forgrowth by breaking down SOM. In other words,plant roots use carbon-rich exudates to “prime”the decomposition process in the rhizosphere,just as you might use water to prime a pump.Microbial immobilization of nutrients in therhizosphere benefits the plant only if thesenutrients are subsequently released andbecome available to the root. Two processesmay contribute to the release of nutrients fromrhizosphere microbes: First, protozoa andnematodes may graze the populations of rhizosphere bacteria, using bacterial carbon tosupport their high energetic demands andexcreting the excess nutrients (Clarholm 1985).Second, as the root matures and exudation ratedeclines, those bacteria that survive predationmay become energy limited and break down

nitrogen-containing compounds to meet theirenergy demands, releasing the nitrogen into therhizosphere as ammonium. The relative con-tribution of grazing and starvation in theseprocesses is unknown, but net nitrogen miner-alization in the rhizosphere has been estimatedto be 30% higher than in bulk soil. Rhizospheredecomposition occurs most readily in soils withrelatively labile soil carbon and low soil lignin(Bradley and Fyles 1996) and therefore mayoccur to a greater extent in grasslands or earlysuccessional communities than in matureforests. Rhizosphere decomposition may bemore sensitive to factors influencing plant car-bohydrate status (e.g., light and grazing) than to soil environment (Craine et al. 1999), so thenature of controls over decomposition (soilenvironment vs. plant carbohydrate status)could differ substantially among ecosystems.However, the extent of rhizosphere decompo-sition and the nature of its ecological controlsare not well characterized under field condi-tions, so it is difficult to evaluate its ecologicalimportance.

Mycorrhizal fungi are functionally an extension of the root system, allowing theroot–fungal symbiosis to absorb nutrients at adistance from the root. The mycorrhizosphere

Rootprocesses

Microbialprocesses

Bacterialstarvation

Nitrogenuptake

Rootexudation

Predation byprotozoans

Rapidbacterial growth

N m

in.

N e

xcr.

N im

mob

.

Sloughing ofroot cap

SO

M b

reak

dow

n

RO

OT

Figure 7.11. Root and microbial proces-ses in the rhizosphere and the resulting effects on soilorganic matter breakdown and nitrogen dynamics inthe rhizosphere. N immob., N immobilization by rhizosphere microbes; N excr., N excretion by protozoa; N min., N miniralization by carbon-starvedmicrobes.


around mycorrhizal fungal hyphae rapidlymoves plant carbon into the bulk soil througha combination of hyphal turnover and exuda-tion (Norton et al. 1990). This may prime thedecomposition process, just as occurs in the rhi-zosphere of roots, although nothing is knownabout this process under field conditions.

Microbial Community Compositionand Enzymatic Capacity

Soil enzyme activity depends on microbial com-munity composition and the nature of the soilmatrix. We have seen that soil animals stronglyaffect decomposition through their effects onsoil structure, litter fragmentation, and micro-bial community composition. The compositionof the microbial community is important in turnbecause it influences the types and rates ofenzyme production and thus the rates at whichsubstrates are broken down. Enzymes thatbreak down common substrates like proteinsand cellulose are produced by so many types ofmicrobes that these enzymes occur universallyin soils (Schimel 2001). Enzymes involved inprocesses that occur only in specific environ-ments, such as denitrification or methane pro-duction and oxidation, appear more sensitive tomicrobial community composition (Gulledge et al. 1997, Schimel 2001).

Soil enzyme activity is also influenced by therates at which enzymes are inactivated in soils,either by degradation by soil proteases or bybinding to soil minerals. Binding of an enzymeto the external surface of roots or microbes frequently prolongs enzyme activity in the soil,whereas binding to mineral particles can alterthe enzyme configuration or block the activesite of the enzyme, thereby reducing its activity.A brief description of a few soil enzymesystems illustrates some of the microbial andsoil controls over exoenzyme activity.

Most soil microbes, including ericoid andectomycorrhizal fungi, produce enzymes (pro-teases and peptidases) that break down pro-teins into amino acids. These breakdownproducts are readily absorbed by microbes andused either to produce microbial protein or toprovide respiratory energy. Because proteasesare subject to attack by other proteases, their

lifetime in the soil is short, and soil proteaseactivity tends to mirror microbial activity. Phos-phatases, which cleave phosphate from organicphosphate compounds, are, however, more longlived, so their activity in soil is correlated morestrongly with the availability of organic phos-phate in soil than with microbial activity(Kroehler and Linkins 1991).

Cellulose is the most abundant chemical con-stituent of plant litter. It consists of chains ofglucose units, often thousands of units in length;but none of this glucose is available until actedon by exoenzymes. Cellulose breakdownrequires three separate enzyme systems (Pauland Clark 1996): Endocellulases break downthe internal bonds to disrupt the crystallinestructure of cellulose. Exocellulases then cleaveoff disaccharide units from the ends of chains,forming cellobiose, which is then absorbed bymicrobes and broken down intracellularly toglucose by cellobiase. Some soil microbes,including most fungi, can produce the entiresuite of cellulase enzymes. Other organisms,such as some bacteria, produce only some cellulase enzymes and must function as part ofmicrobial consortia to gain energy from cellu-lose breakdown.

Lignin is degraded slowly because only someorganisms (primarily fungi) produce the neces-sary enzymes, and these microbes produceenzymes only when other more labile sub-strates are unavailable. Lignin forms nonenzy-matically by condensation reactions withphenols and free radicals, creating an irregularstructure that does not fit the specificityrequired by the active site of most enzymes.For this reason, lignin-degrading enzymes usefree radicals, which have a low specificity forsubstrates. Oxygen is required to generate these free radicals, so lignin breakdown doesnot occur in anaerobic soils. Decomposers generally invest more energy in producinglignin-degrading enzymes than they gain bymetabolizing the breakdown products of lignin(Coûteaux et al. 1995). Lignin is apparentlydegraded primarily to provide access to labilecompounds such as cellulose, hemicellulose,and protein, which actually meet the energeticand nitrogen requirements of the decomposers.Some of the enzymes involved in lignin break-

Long-Term Storage of Soil Organic Matter 169

down may also function in the formation andbreakdown of soil humus.

Long-Term Storage of SoilOrganic Matter

In climates that are favorable for decomposi-tion, humus is the major long-term reservoir ofsoil carbon. Up to this point, we have focusedprimarily on the factors controlling the break-down and loss of soil organic matter. Equallyimportant are the processes that transform soilorganic matter into relatively recalcitranthumus, allowing its accumulation in soils. Soilhumus decomposes slowly for several reasons.As with lignin, its highly irregular structure isnot efficiently attacked by a single enzymesystem. Its large size and highly cross-linkedform make most of the structure inaccessible tosoil enzymes.

Its tendency to bind with soil minerals pro-tects it from enzymatic attack. Much of theSOM in soil is therefore not good “food” formicrobes, despite its high nitrogen content.Humus also constitutes a large reservoir ofnitrogen in many ecosystems. This nitrogenturns over extremely slowly, except when dis-turbance increases the rate of humus decom-position. As the carbon in humus is respiredaway, the nitrogen is released, providing animportant nutrient source to support ecosystemrecovery after disturbance (see Chapter 13).The sensitivity of humus to breakdown follow-ing disturbance makes ecosystems with largehumus accumulations, such as tropical forestsor grasslands, particularly vulnerable to carbonloss after such changes.

The formation of humus by humificationoccurs through a combination of biotic andabiotic processes (Zech and Kogel-Knabner1994). The following five steps have been impli-cated in humus formation (Fig. 7.12), althoughthe relative importance of factors governingthese steps is poorly understood.

1. Selective preservation. Decompositionselectively degrades labile compounds in detri-tus, leaving behind recalcitrant materials likewaxes, cutins, suberin, lignin, chitin, and

microbial cell walls. Partial microbial break-down of these recalcitrant leftovers often produces compounds with reactive groups andside chains that are common reactants in thenonspecific soil reactions that occur duringhumification.

2. Microbial transformation. Enzymaticbreakdown of SOM produces low molecularweight water-soluble products, some of whichparticipate in humus formation. Amino com-pounds, such as amino acids from proteinbreakdown and sugar amines from degradationof microbial cell walls, are particularly impor-tant in humification (see Step 5).

3. Polyphenol formation. Soluble phenoliccompounds are important reactants in humusformation.They come from at least three sources(Haynes 1986): microbial degradation of plantlignin, the synthesis of phenolic polymers by soilmicrobes from simple nonlignin plant precur-sors, and polyphenols produced by plants asdefenses against herbivores and pathogens.

4. Quinone formation. The polyphenoloxidase and peroxidase enzymes produced by fungi to break down lignin and other phenolic compounds also convert polyphenolsinto highly reactive compounds called quinones(Fig. 7.13).

5. Abiotic condensation. The quinones spon-taneously undergo condensation reactions withmany soil compounds, especially compounds

Labilecompounds

Phenolics Lignin,waxes, etc.

Poly-phenols

Quinones

Humus

Microbialbiomass

Aminocompounds

2

3 1

3

1

1

5

1

2

3

4

Selective preservation

Microbial transformation

Polyphenol formation Abiotic condensation

Plant litter

4

5

Quinone formation

Figure 7.12. Principle pathways of humus forma-tion. See text for details.


that have amino groups (with which they reactmost readily) or that are abundant (such asrecalcitrant compounds that accumulate insoils).

The chemical nature of humus differs amongecosystems due to differences in the raw mate-rials available for humification (Haynes 1986,Paul and Clark 1996). Forests produce a pre-dominance of humic acids, which are large,relatively insoluble compounds with extensivenetworks of aromatic rings and few side chains.Phenolic-rich plants, which are common inmany forests, provide many of the phenolic pre-cursors for humic acids. Humin contains morelong-chain nonpolar groups derived from cutinand waxes than do humic acids and are also rel-atively insoluble. Fulvic acids are more watersoluble because of their extensive side chainsand many charged groups. Their more openstructure binds readily to other organic andinorganic materials. Grasslands have a morebalanced mixture of fulvic and humic acids thando forests, perhaps because grassland plantsproduce fewer polyphenolic precursors forhumus formation. The nitrogen content ofhumic acids (4%) is fivefold greater than infulvic acids, but most of this nitrogen is in ringstructures that are not readily broken down.

Environmentally protected organic matteraccumulates in cold and wet environments. Inenvironments in which low oxygen availabilityor low temperature inhibits decomposition,organic matter accumulates in a relatively non-decomposed state.This organic matter accumu-lates, not because it is recalcitrant, but becauseconditions constrain the activity of decom-posers more strongly than they constraincarbon inputs by plants. Ecosystems with some

of the largest carbon stores, such as wetlandsand tundra, have soil organic matter that ishighly labile and decomposes quickly, once theenvironmental limitations to decompositionare released. This makes carbon balance ofthese ecosystems quite vulnerable to globalenvironmental change.

Decomposition at the Ecosystem Scale

Aerobic Heterotrophic Respiration

Aerobic heterotrophic respiration is the majoravenue of carbon loss from ecosystems. It is thesum of aerobic respiration by soil microbes,which is equivalent to stand-level decomposi-tion (discussed above), and the respiration byanimals. Microbes and animals are groupedtogether as heterotrophs because they derivetheir energy and carbon from organic matterproduced by plants (see Chapter 11). Decom-position accounts for most heterotrophic respiration, but animal respiration is also a sig-nificant avenue of carbon loss from someecosystems (see Fig. 6.1). We discuss the factorsregulating the consumption of plants andmicrobes by animals in Chapter 11.

The controls over stand-level decompositionare similar to the controls over GPP and NPP.As in the case of GPP and NPP, we cannotdirectly measure stand-level decompositionunder undisturbed field conditions. Soil respi-ration includes the CO2 respired by soilmicrobes, soil animals, and roots, and thesecannot be separated by direct measurements.Isotopic tracers and ecosystem models are twotools that have proven particularly valuable

OH

OHR1OHR1

OH

OR1

N - R2

R2 - NH2

Polyphenoloxidase

Condensationreaction

Amine

Phenolic Quinone Humus

O

Figure 7.13. Reactions that occur during humus formation.

Decomposition at the Ecosystem Scale 171

in estimating decomposition at the ecosystemscale (Box 7.1). Both of these approaches indicate that stand-level decomposition ratedepends not only on environment, as discussedearlier, but also on the amount and quality ofrecent carbon inputs to soils (Fig. 7.14). Thequantity of carbon input to soils, in turn, gen-erally depends on NPP. Carbon quality is alsohighest in productive stands.

Since GPP and NPP are important determi-nants of stand-level decomposition rate, it is notsurprising that the controls over stand-leveldecomposition are similar to those for GPP andNPP. In other words, decomposition is ulti-mately controlled by the availability of soilresources, disturbance regime, and climate (Fig.7.14). Measurements of soil respiration, whichincludes both heterotrophic and root respira-tion, are consistent with this generalization. Soil

respiration correlates closely with NPP (Raichand Schlesinger 1992) (Fig. 7.15). Carbon lossthrough soil respiration is about 25% higherthan carbon inputs through NPP, suggestingthat about 25% of soil respiration derives fromroots, and the rest comes from decomposition(Raich and Schlesinger 1992). Both NPP anddecomposition are higher in the tropics than inthe arctic and higher in rain forests than indeserts, due to similar environmental sensitivi-ties of plants and decomposers. Likewise, plantspecies that are highly productive produce litterof higher quality than do species of low poten-tial productivity. Habitats dominated by pro-ductive species are therefore characterized byhigh rates of litter decomposition (Hobbie1992), high concentrations of labile carbon, andhigh microbial biomass (Zak et al. 1994), allcontributing to the high stand-level decomposi-

Box 7.1. Isotopes and Soil Carbon Turnover

The quantity of soil carbon differs dramati-cally among ecosystems (Post et al. 1982).The total quantity of carbon in an ecosystem,however, gives relatively little insight into its dynamics. Tropical forests and tundra,for example, have similar quantities of soil carbon, despite their radically differentclimates and productivities. The simplestmeasure of soil carbon turnover is its residence time estimated from the pool sizeand carbon inputs (Eq. 7.3). These measure-ments show that, even though tropicalforests and arctic tundra have similar sizesoil carbon pools, the turnover may be 500times more rapid in the tropical forest. Moresophisticated approaches to estimating soil carbon turnover using carbon isotopes(Ehleringer et al. 2000) lead to a similar conclusion. In the tropics, 85% of the 14Cthat entered ecosystems during the era ofnuclear testing in the 1960s has been con-verted to humus, whereas this proportion isonly 50% in temperate soils and approxi-mately 0% in boreal soils (Trumbore 1993,Trumbore and Harden 1997). This compari-son clearly indicates more rapid turnover of

soil organic matter in the tropics than at highlatitudes.

Carbon isotopes can also be used to esti-mate the impacts of land use change oncarbon turnover in situations in which thevegetation change is associated with a changein carbon isotopes. In Hawaii, for example,replacement of C3 forests by pastures domi-nated by C4 grasses causes a gradual changein the carbon isotope ratio of soil organicmatter from values similar to C3 plantstoward values similar to C4 plants (Townsandet al. 1995). This information can be used toestimate the quantity of the original forestcarbon that remains in the ecosystem:

(B7.1)

where %CS1 is the percentage of soil derivedfrom the initial ecosystem type, CS2 is the 13Ccontent of soil from the second soil type, CV2

is the 13C content of soil from the second vegetation type, and CV1 is the 13C content ofvegetation from the initial ecosystem type.

%CC CC CS

S V

V V1

2 2

1 2100=

--

¥


STATE FACTORS

Interactivecontrols

Directcontrols

LONG-TERMCONTROLS

SHORT-TERMCONTROLS

Plantfunctional

types

Soilresources

Litter quantity

Carbon quality

Litter C:N

Oxygen

TemperatureH2O

NPP

DECOMPOSITION

Indirectcontrols

BIOTA

PARENTMATERIAL

CLIMATE

TIME

Figure 7.14. The major factors governing decompo-sition at the ecosystem scale. These controls rangefrom proximate controls that determine the seasonalvariations in decomposition to the state factors and interactive controls that are the ultimate causes ofecosystem differences in decomposition. Thicknessof the arrows indicates the strength of the direct and

D

M

W FSA

B

T

G

1400

1000

600

200

200 400 600 800 10000

Soi

l res

pira

tion

(g C

m-2

yr-1

)

NPP (g C m-2 yr-1)

Figure 7.15. Relationship between mean annual soilrespiration rate and mean annual NPP for Earth’smajor biomes. Root respiration probably accountsfor the 25% greater soil respiration than NPP at anypoint along this regression line. A, agricultural lands;B, boreal forest and woodland; D, desert scrub; F,temperate forest; G, temperate grassland; M, moisttropical forest; S, tropical savanna and dry forest;T, tundra; W, mediterranean woodland and heath.(Redrawn with permission from Tellus; Raich andSchlesinger 1992.)

tion rates of productive sites. The relativeimportance of the direct effects of climate ondecomposition vs. its indirect effects mediatedby availability of soil resources and the quan-tity and quality of litter inputs remains to bedetermined.

Decomposition and carbon inputs to soils areseldom precisely in balance. Disturbances, her-bivore outbreaks, and other events periodicallycause organic matter inputs to soils to differsubstantially from NPP (see Chapter 13). Afterhurricanes, for example, large inputs of plantmaterial to the soil cause a pulse of decompo-sition to coincide with a sharp dip in NPP. Inaddition, decomposition is generally less sensi-tive to drought and more sensitive to lowoxygen and to warm temperatures than is NPP,so seasonal or interannual variations in weathercause stand-level decomposition to be greateror less than NPP at any moment in time.

indirect effects. The factors that account for most ofthe variation in decomposition among ecosystemsare the quantity and carbon quality of litter inputs,which are ultimately determined by the interactingeffects of soil resources, climate, vegetation, and dis-turbance regime.

Decomposition at the Ecosystem Scale 173

Stand-level decomposition shows little rela-tionship with the total quantity of organicmatter in soils because most soil carbon iseither relatively recalcitrant or decomposesslowly because of an unfavorable soil environ-ment (low temperature or low oxygen avail-ability). Most decomposition derives fromrelatively recent litter inputs rather than fromolder soil organic matter. Consequently, totalsoil organic content is not a good measure ofthe food available for microbes or a good pre-dictor of stand-level decomposition (Clein et al.2000). In fact, the largest soil carbon accumula-tions frequently occur in ecosystems with slowdecomposition, such as peat bogs.

The activity of soil microbes is more impor-tant than microbial biomass in determiningdecomposition rate. In boreal forests, forexample, greatest decomposition occurs in soilswith large inputs of high-quality litter. Micro-bial biomass is a relatively constant proportion(about 2%) of total soil carbon and thereforehas the largest pool size (in grams per squaremeter) in those stands with the largest quanti-ties of soil carbon; these are the stands withlowest productivity and slowest decomposition(Vance and Chapin 2001). In agricultural soils,microbial biomass also tends to be higher inextremely wet or dry soils, where decomposi-tion is slow, than in moderately moist soils withhigher decomposition rate (Insam 1990). Sincemost microbial biomass is inactive, it is proba-bly more important as a reservoir of nutrients(see Chapter 9) than as a predictor of decom-position rate.This differs from the controls overcarbon inputs to ecosystems, where the quanti-ties of plant biomass and leaf area areextremely important determinants of GPP.Microbial processes like nitrification, which areconducted by a restricted number of microbialgroups, on the other hand, appear to be sensi-tive to the population sizes of these groups (seeChapter 9).

Anaerobic Heterotrophic Respiration

Decomposition in anaerobic environmentsoccurs slowly and produces energy inefficiently.Most of this chapter has focused on aerobicdecomposition, which consumes oxygen and

returns carbon to the atmosphere as CO2. Wet-lands, estuaries, and sediments beneath lakesand oceans, however, occupy vast areas ofEarth’s surface. These are environments whereoxygen supply frequently limits decompositionrate, so organisms must use other electronacceptors to derive energy from organic matter.Oxygen is the preferred electron acceptor,when it is available, because it provides themost energy return per unit of organic matteroxidized. Progressively less energy is releasedwith transfer to each of the following electronacceptors (see Chapter 3):

O2 > NO3- > Mn4+ > Fe3+ > SO4

2- > CO2 > H+

(7.4)

where NO3- is nitrate ion and SO4

2- is sulfateion. As oxygen becomes depleted by aerobicdecomposition, denitrifiers gain a competitiveadvantage. They use most of the metabolicmachinery associated with aerobic respirationto transfer electrons from organic matter tonitrate, producing the gases nitrous oxide(N2O) and di-nitrogen (N2), as waste products(see Chapter 9). The availability of nitrate isgenerally limited in anaerobic environmentsbecause nitrification, which produces nitrate, isan aerobic process. As the supply of nitratebecomes depleted, other bacteria, using otherelectron acceptors, gain a competitive advan-tage. Decomposition shifts to fermenters thatbreak down labile organic compounds toacetate, other simple organic compounds, andhydrogen. These fermentation products arethen used by sulfate reducers or methanogens,depending on the availability of sulfate, whichtransfer electrons to sulfate or CO2 to producehydrogen sulfide or methane, respectively.Estuaries, salt marshes, and ocean sedimentsfrequently have enough marine-derived sul-fate to make sulfate reduction the dominantpathway of anaerobic metabolism. The supplyof sulfate is limited in many terrestrial envi-ronments, however, so the production ofmethane by a specialized group of Archaeaknown as methanogens becomes quantitativelyimportant. Wetlands, such as swamps and ricepaddies, are therefore important methanesources.


Methane emission from soils to the atmos-phere is of global concern. Methane is 20-foldmore effective in absorbing infrared radiationthan is CO2. Moreover, its concentration in theatmosphere has risen dramatically in recentdecades, in part as a result of increased area ofrice paddies and reservoirs (see Fig. 15.3). Evenin wetlands, methane accounts for only 5 to15% of the carbon released to the atmosphereby decomposers. Methane is thus quantitativelymore important in its role as a greenhouse gasthan as a path of carbon loss from ecosystems(see Fig. 6.8).

Methane is even more highly reduced thanare carbohydrates, so it is an effective energysource for organisms that have access tooxygen. Another group of bacteria (methan-otrophs) that occur in the surface soils of wet-lands use this methane as an energy source andconsume much of the methane before it dif-fuses to the atmosphere. The enzyme systemthat converts ammonium to nitrate also reactswith methane, causing well-aerated soils to bea net sink for methane. There are thereforeimportant transfers between methane produc-ers and consumers that occur both verticallywithin poorly drained ecosystems and horizon-tally from lowland to upland ecosystems.

Summary

Decomposition is the conversion of deadorganic matter into CO2 and inorganic nutri-ents through the action of leaching, fragmenta-tion, and chemical alteration. Leachingremoves soluble materials from decomposingorganic matter. Fragmentation by soil animalsbreaks large pieces of organic matter intosmaller ones that provide a food source for soilanimals and create fresh surfaces for microbialcolonization. Fragmentation also mixes thedecomposing organic matter into the soil.Chemical alteration of dead organic matter isprimarily a consequence of the activity of bac-teria and fungi, although some chemical reac-tions occur spontaneously in the soil withoutmicrobial mediation.

Decomposition rate is regulated by physicalenvironment, substrate quality, and the compo-

sition of the microbial community (includingsoil animals). Carbon chemistry is a strongdeterminant of litter quality; labile substrates,such as sugars and proteins, decompose more rapidly than recalcitrant ones, such aslignin and microbial cell walls. Nitrogen andphosphorus supply can also constrain thedecomposition of labile carbon substrates, suchas agricultural residues and root exudates.Plants in high-resource environments producelitter with high litter quality and therefore rapiddecomposition rates. Decomposition ratedeclines with time, as recalcitrant substrates aredepleted. Soil animals strongly influencedecomposition by fragmenting litter, consum-ing soil microbes, and mixing the litter intomineral soil. The environmental factors thatfavor NPP (warm, moist, fertile soils) alsopromote decomposition so there is no clearrelationship between the amount of carbon thataccumulates in soils with either NPP or decom-position rate.

Review Questions

1. What is decomposition, and why is it impor-tant to the functioning of ecosystems?

2. What are the three major processes thatgive rise to decomposition? What are the major controls over each of theseprocesses? Which of these processes isdirectly responsible for most of the massloss from decomposing litter?

3. What are the major similarities and differ-ences between bacteria and fungi in theways in which they decompose deadorganic matter? How do these two groupsof decomposers differ in their response tomoisture and nutrients? Why?

4. What roles do soil animals play in decom-position? How does this role differ betweenprotozoans and earthworms?

5. Why do decomposer organisms secreteenzymes into the soil rather than breakingdown dead organic matter inside theirbodies?

6. What chemical traits determine the qualityof soil organic matter? How do carbonquality and the C :N ratio differ between


the litter of plants growing on fertile vs.infertile soils?

7. Describe the mechanisms by which tem-perature and moisture affect decomposi-tion rate.

8. How do roots influence decompositionrate? How does decomposition in the rhi-zosphere differ from that in the bulk soil?Why?

9. How do soil properties and disturbanceaffect decomposition rate?

10. How is humus formed? What are its pre-cursors? How long does humus remain inthe soil of undisturbed ecosystems? Why ishumus formation important to the func-tioning of ecosystems?

Additional Reading

Anderson, J.M. 1991. The effects of climate changeon decomposition processes in grassland andconiferous forests. Ecological Applications 1:326–347.

Beare, M.H., R.W. Parmelee, P.F. Hendrix, W. Cheng,D.C. Coleman, and D.A. Crossley Jr. 1992. Micro-bial and faunal interactions and effects on litternitrogen and decomposition in agroecosystems.Ecological Monographs 62:569–591.

Coûteaux, M.-M., P. Bottner, and B. Berg. 1995. Litterdecomposition, climate and litter quality. Trends inEcology and Evolution 10:63–66.

Fog, K. 1988.The effect of added nitrogen on the rateof decomposition of organic matter. BiologicalReview 63:433–462.

Haynes, R.J. 1986. The decomposition process: Min-eralization, immobilization, humus formation, anddegradation. Pages 52–126 in R.J. Haynes, editors.Mineral Nitrogen in the Plant-Soil System. Acade-mic Press, Orlando, FL.

Mary, B., S. Recous, D. Darwis, and D. Robin. 1996.Interactions between decomposition of plantresidues and nitrogen cycling in soil. Plant and Soil181:71–82.

Oades, J.M. 1989. An introduction to organic matterin mineral soils. Pages 89–159 in J.B. Dixon,and S.B. Weed, editors. Minerals in Soil Environ-ments. Soil Science Society of America,Madison, WI.

Paul, E.A., and F.E. Clark. 1996. Soil Microbiologyand Biochemistry. 2nd ed. Academic Press, SanDiego, CA.

Swift, M.J., O.W. Heal, and J.M. Anderson. 1979.Decomposition in Terrestrial Ecosystems.Blackwell Scientific, Oxford, UK.

Zech, W., and I. Kogel-Knabner. 1994. Patterns andregulation of organic matter transformation insoils: Litter decomposition and humification.Pages 303–335 in E.-D. Schulze, editor. FluxControl in Biological Systems: From Enzymes toPopulations and Ecosystems. Academic Press, SanDiego, CA.

Introduction

Nutrient cycling in ecosystems involves highlylocalized exchanges between plants, soil, andsoil microbes. In contrast to carbon, which isexchanged with a well-mixed atmospheric pool,nutrients are absorbed by plants and returnedto the soil largely within the extent of the rootsystem of an individual plant. More than 90%of the nitrogen and phosphorus absorbed byplants of most ecosystems comes from the re-cycling of nutrients that were returned fromvegetation to soils in previous years (Table 8.1).The controls over nutrient uptake and use must therefore be examined at a more localscale than for carbon. Individual ecosystems,and indeed individual plants, have strong localeffects on nutrient supply (Hobbie 1992, VanBreemen and Finzi 1998). The patterns of nu-trient cycling beneath a deep-rooted oak thatabsorbs calcium from depth and produces acation-rich litter, for example, may be quite different from those beneath a shallow-rootedpine that absorbs less cations and producesmore organic acids (Andersson 1991).

Nutrient supply constrains the productivityof the terrestrial biosphere. Experimental ad-dition of nutrients increases productivity of virtually every ecosystem, indicating the

widespread importance of nutrients in con-straining terrestrial production. Within any cli-matic zone, there is usually a strong positivecorrelation between soil fertility and plant production. Even in agricultural ecosystems,where nutrients are added regularly, produc-tion usually responds to nutrient addition. Themajor factors responsible for increased produc-tivity of agriculture during the Green Rev-olution were increased rates of fertilizer,water, and pesticide application and breedingof crops capable of using water and nutrients,mostly through greater allocation to har-vestable tissues (Fig. 8.1) (Evans 1980). Therehave been no major changes in photosyntheticrate during crop evolution. Given the wide-spread occurrence of nutrient limitation(Vitousek and Howarth 1991), an understand-ing of the controls over acquisition, use, andloss of nutrients by vegetation is essential to characterizing the controls over plant production and other ecosystem processes.

Overview

The quantity of nutrients that cycle throughvegetation depends on the dynamic balancebetween nutrient supply from the soil and

8Terrestrial Plant Nutrient Use

Nutrient uptake, use, and loss by plants are key steps in the mineral cycling ofecosystems. This chapter describes the factors that regulate nutrient cycling throughvegetation.

176

Nutrient Movement to the Root 177

nutrient demand by vegetation. The balance ofnutrients required to support maximal growthis similar for most plants (Ingestad and Ågren1988). Any nutrient present in less than theoptimal balance is likely to limit growth, soplants invest preferentially in absorption of the nutrients that most strongly limit growth.Nutrients that accumulate in excess of plantrequirements are absorbed more slowly. Nutri-

ent ratios in plants therefore converge towarda common ratio. This pattern was first observedin the ocean (Redfield 1958) but also tends tooccur on land. The consequence of this conver-gence toward a common nutrient ratio is thatthe nutrient that most strongly limits growthdetermines cycling rates of all nutrients. Thiselement stoichiometry (Elser and Urabe 1999)defines patterns of cycling of most nutrients inecosystems. The key to understanding nutrientcycling is thus to determine the factors control-ling the cycling of the most strongly limitingelement. These cycling rates may be con-strained by either the supply of that nutrientfrom soil or its demand by vegetation. Supplyrate of the growth-limiting nutrient could beconstrained, for example, by climatic factors or by the chemical nature of parent material.Vegetation demand for the most limiting nu-trient could be constrained, for example, bywater limitation of growth or by the amount ofbiomass present after disturbance. In thischapter we first explore the controls over nutrient uptake by vegetation, then the rela-tionship between nutrient content and produc-tion, and finally the controls over nutrient lossfrom vegetation.

Nutrient Movement to the Root

Nutrients contact the root surface by threemechanisms: diffusion, mass flow, and rootinterception. Roots absorb only those nutrientsthat are in direct contact with live cells. Becauseroots constitute only a small proportion (much

Source of plant nutrient (% of total)

Nutrient Deposition/fixation Weathering Recycling

Temperate forest(Hubbard Brook)

Nitrogen 7 0 93Phosphorus 1 <10? >89Potassium 2 10 88Calcium 4 31 65

Tundra (Barrow)Nitrogen 4 0 96Phosphorus 4 <1 96

Data from Whittaker et al. (1979) and Chapin et al. (1980b).

Table 8.1. Major sources ofnutrients that are absorbed byplants.

Japan

United Kingdom

UnitedStates

Australia

India

1974

Gra

in y

ield

(g

m-2

)

Fertilizer use (g m-2)

1963

1969

1950

1966

00

10 20 30 40 50

100

200

300

400

500

600

Figure 8.1. Response of grain yield of cereal cropsto fertilizer addition. These studies were conductedduring the Green Revolution. Yield is most respon-sive to low nutrient addition rates, often saturatingwith further nitrogen additions. (Redrawn with per-mission from American Scientist; Evans 1980.)

178 8. Terrestrial Plant Nutrient Use

less than 1%) of the belowground volume,nutrients must first move from the bulk soil(i.e., the soil that is not in direct contact withroots) to the root surface before plants canabsorb them.

Diffusion

Diffusion is the process that delivers mostnutrients to plant roots. It is the movement ofmolecules or ions along a concentration gr-adient. Nutrient uptake and mineralizationprovide the driving forces for diffusion to theroot surface by reducing nutrient concentrationat the root surface (uptake) and increasing theconcentration elsewhere in the soil (mineral-ization). Mineralization and other inputs to the pool of soluble nutrients are the main con-trols over the quantity of nutrients available todiffuse to the root surface (see Chapter 9).

Cation exchange capacity (CEC) of soils alsoinfluences the pool of nutrients available todiffuse to the root and the volume of soil thatthe root exploits. Soils with a high CEC storemore available cations per unit soil volume—that is, they have a high buffering capacity—butretard the rate of nutrient movement to theroot surface through exchange reactions (seeChapter 3). The root therefore draws morenutrients from a given volume of soil in soilswith a high CEC but acquires these nutrientsmore slowly. The net effect of a high cation ex-change capacity is to increase the supply ofnutrients available to a root under conditionsof high base saturation—that is, where theexchange complex has abundant cations. Anionexchange capacity is generally much lower thancation exchange capacity (see Chapter 3), somost anions, like nitrate, diffuse more rapidly insoils than do cations. Some anions, like phos-phate, however, tend to precipitate readily fromthe soil solution and therefore diffuse slowly tothe root surface.

Rates of diffusion differ strikingly amongions, due to differences in charge density (i.e.,the charge per unit hydrated volume of theion). Charge density, in turn, depends on thenumber of charges per ion and the hydratedradius of the ion. Divalent cations, like calciumand magnesium, are bound more tightly to the

exchange complex and diffuse more slowlythan do monovalent cations, such as ammoniumand potassium. Ions of a given charge alsodiffer slightly in diffusion rates because of differences in radius and number of water molecules that are loosely bound to the ion.Soil particle size and moisture determine the path length of diffusion from the bulk soil tothe root surface. Ions diffuse through waterfilms that coat the surface of soil particles. Thehigher the water content and the smaller theparticle size, the more direct the diffusion pathfrom the bulk soil to the root surface. Moistsoils therefore permit more rapid diffusion thandry soils, and soils with a high clay content allowmore rapid diffusion than do coarse-texturedsandy soils.

Each absorbing root creates a diffusion shell,or a cylinder of soil that is depleted in the nutri-ents absorbed by the root. This diffusion shellconstitutes the zone of soil directly influencedby plant uptake. The root accesses a relativelylarge volume of soil for those ions that diffuserapidly. Nitrate, for example, which diffusesrapidly, is typically depleted in a shell 6 to 10mm in radius around each absorbing root,whereas ammonium is depleted over a radiusof less than 1 to 2mm, and phosphate isdepleted over a radius of less than 1mm. Ittherefore takes a higher root density to fullyexploit the soil for phosphate or ammoniumthan for nitrate. The root densities in manyecosystems are high enough to exploit most ofthe soil volume for nitrate but only a small pro-portion of the soil volume for ammonium orphosphate. The major way in which a plant canenhance uptake of ions that diffuse slowly is toincrease root length and therefore the propor-tion of the soil that it exploits.

Mass Flow

Mass flow of nutrients to the root surface aug-ments the supply of ions. Mass flow is the move-ment of dissolved nutrients to the root surfacein flowing soil water. Transpirational water lossby plants is the major mechanism that causesmass flow of soil solution to the root surface.Mass flow can be an important mechanism forsupplying nutrients that are abundant in the

Nutrient Movement to the Root 179

soil solution or that the plant needs in smallquantities. Calcium, for example, is present insuch a high concentration in many soils that theplant demands for calcium are completely metby mass flow of calcium from the bulk soil tothe root surface (Table 8.2). Corn, for example,receives fourfold more calcium by mass flow tothe root than is acquired by the plant. Plantsthat receive too much calcium by mass flowactively secrete calcium from roots into the soilsolution, creating a diffusion gradient awayfrom the root surface toward the bulk soil.Other nutrients are required in such smallquantities by plants (micronutrients) that theneeds of the plant can be completely met bymass flow (Table 8.2). However, mass flow isnot sufficient for supplying the nutrients thatare required by plants in large quantities butare present in low concentrations in the soilsolution, such as nitrogen, phosphorus, andpotassium. These macronutrients (i.e., nutrientsrequired in large quantities) are supplied pri-

marily by diffusion. The quantity of nutrientscarried to the root surface in the transpirationstream in natural ecosystems, for example, is a small proportion of the nutrients required to support primary production (Table 8.2), somost of the nutrients must be supplied by someother mechanism—primarily diffusion. Even inagricultural soils, in which soil solution concen-trations are much higher, mass flow suppliesless than 10% of those nutrients that typicallylimit plant production. Diffusion therefore,rather than mass flow, is the major mechanismthat supplies potentially limiting nutrients (ni-trogen, phosphorus, and usually potassium) toplants. Diffusion becomes proportionatelymore important in supplying nutrients as soilfertility declines (Table 8.2).

Saturated flow of water through soils sup-plies additional nutrients and replenishes diffu-sion shells. Saturated flow is the movement ofwater through soil in response to gravity (seeChapter 3).After a heavy rain, water drains ver-

Table 8.2. Mechanisms by which nutrients move to the root surface.

Mechanism of nutrient supply

Quantity (% of total absorbed)

absorbed by the RootNutrient plant (g m-2) interception Mass flow Diffusion

Sedge tundra (natural ecosystem)Nitrogen 2.2 0.5 99.5Phosphorus 0.14 0.7 99.3Potassium 1.0 6 94Calciuma 2.1 250 0Magnesium 4.7 83 17

Corn crop (agricultural ecosystem)Nitrogen 19 1 79 20Phosphorus 4 2 4 94Potassium 20 2 18 80Calciuma 4 150 413 0Magnesiuma 4.5 33 244 0Sulfur 2.2 5 95 0Iron 0.2 53Manganesea 0.03 133 0Zinc 0.03 33Borona 0.02 350 0Coppera 0.01 400 0Molybdenuma 0.001 200 0

a Mass flow of these elements is sufficient to meet the total plant requirement, so no additional nutrients must be sup-plied by diffusion. The amount supplied by mass flow is calculated from the concentration of the nutrients in the bulk soilsolution multiplied by the rate of transpiration. The amount supplied by diffusion is calculated by difference; other formsof transport to the root (e.g., mycorrhizae) may also be important but are not included in these estimates.Data from Barber (1984), Chapin (1991a), and Lambers et al. (1998).


tically through the soil by saturated flow when-ever the water content exceeds the soil’s water-holding capacity. Because nutrient availabilityand mineralization rates are generally highestin the uppermost soils, this vertical flow ofwater redistributes nutrients and replenishesdiffusion shells surrounding roots. Both rootgrowth and vertical soil water movement occurpreferentially in soil cracks, quickly eliminatingdiffusion shells around these roots. Saturatedflow is also important in ecosystems wherethere is regular horizontal flow of ground wateracross an impermeable soil layer. Deep-rootedspecies in tundra underlain by permafrost, forexample, have 10-fold greater nutrient uptakeand productivity in areas of rapid subsurfaceflow than in areas without lateral groundwaterflow (Chapin et al. 1988). The high productivityof trees and shrubs in riparian ecosystemsresults in part because their roots often extendto the water table and to groundwater beneaththe stream (the hyporrheic zone), where rootstap the saturated flow of nutrients through therooting zone.

Root Interception

Root interception is not an important mecha-nism for directly supplying nutrients to roots.As roots elongate into new soil, they interceptavailable nutrients in this unoccupied soil.The quantity of available nitrogen, phosphorus,and potassium per unit soil volume is, how-ever, always less than the quantity of nutrientsrequired to construct the root, so root inter-ception can never be an important mechanismof nutrient supply to the shoot. Root growth iscritical, not because it intercepts nutrients, butbecause it explores new soil volume and createsnew root surface to which nutrients can moveby diffusion and mass flow.

Nutrient Uptake

Nutrient uptake. Who is in charge? Threefactors govern nutrient uptake by vegetation:nutrient supply rate from the soil, root length,and root activity. Just as with photosynthesis,

several factors influence nutrient uptake at the ecosystem scale. Our main conclusions inthis section are as follows: (1) Nutrient supplyrate is the major factor accounting for differ-ences among ecosystems in nutrient uptake at steady state. In other words, nutrient sup-ply by the soil rather than plant traits deter-mines biome differences in nutrient uptake by vegetation. (2) Plant traits such as root lengthand root activity strongly influence total nu-trient uptake by vegetation in ecosystems inwhich biomass is increasing rapidly after dis-turbance. (3) Root length is the major factorgoverning which plants in an ecosystem aremost successful in competing for a limitedsupply of nutrients.

Nutrient Supply

Nutrient uptake by vegetation at steady state isdriven primarily by nutrient supply. There areunresolved debates about the relative impor-tance of soil and plant characteristics in deter-mining stand-level rates of nutrient absorption.There are several lines of evidence, however,suggesting that nutrient supply exerts primarycontrol over nutrient uptake by vegetation atsteady state. The most direct evidence for thecontrolling role of nutrient supply in drivinguptake by vegetation is that most ecosystemsrespond to nutrient addition with increaseduptake and net primary production (NPP) (Fig.8.1). This differs strikingly from the controlsover photosynthesis, where ecosystem dif-ferences in carbon uptake are determined primarily by capacity of vegetation to acquirecarbon (leaf area and photosynthetic capacity),rather than by the supplies of CO2 or light (see Chapter 5).

Simulation models support the conclusionthat plant uptake is more sensitive to nutrientsupply and to the volume of soil exploited byroots than to the kinetics of nutrient uptake,particularly for immobile ions like phosphate.At low nutrient supply rates, for example,variation in factors affecting diffusion (diffu-sion coefficient and buffering capacity) androot length (elongation rate) have a muchgreater effect on nutrient uptake than do kinetics (maximum and minimum capacity for

Nutrient Uptake 181

uptake or affinity of roots for nutrients) orfactors influencing mass flow (transpirationrate) (Fig. 8.2).

Development of Root Length

Root biomass differs less among ecosystemsthan does aboveground biomass at steady state,due to counteracting effects of production and allocation. In favorable environments, totalplant biomass is large, but most of this biomassis allocated aboveground to compete for light.In dry or infertile environments, on the otherhand, there is less biomass but a larger propor-tion of it is allocated belowground (see Chapter6). The net result of these counteracting effectsis that fine root biomass is probably moresimilar among ecosystems at steady state than

is aboveground biomass. Thus the large differ-ences among ecosystems in nutrient uptake andNPP reflect differences in nutrient supply morethan differences in root biomass. Production ofnew root biomass enhances nutrient uptake byvegetation only if there are zones in the soilthat are not yet exploited by roots—that is,where total root length is low enough that dif-fusion shells among adjacent roots do notoverlap. This is most likely to occur after dis-turbance, just as new leaf production above-ground enhances carbon gain primarily whenthe leaf canopy is sparse (Craine et al., in press).Even with a fully developed “root canopy,”increased root growth by an individual plantmay be advantageous because it increases theproportion of the total nutrient supply capturedby that plant.

P u

ptak

e (µ

mol

pot

-1)

Change in parameter

e (Root elongation rate)

c (Bulk soil P concentration)

b (Buffer capacity of soil)

D (Diffusion coefficient of P)

Imax (Maximum uptake rate)

T (Transpiration rate)

R* (P uptake threshold)

Km (Affinity of roots for P)

0.4

0.5

0.3

0.2

0.10.5 1.0 1.5 2.0

e

c

b

D

Imax

TR*

Km

Figure 8.2. Effect of changing parameter values(from 0.5 to 2.0 times the standard value) in a modelthat simulates phosphate uptake by roots of soybean.The factors that have greatest influence on phosphate uptake are plant parameters that determine the quantity of roots (e) and soil parame-

ters that influence phosphate supply from the soil (c, b, and D). (Redrawn with permission fromAnnual Review of Plant Physiology, Volume 36 ©1985 by Annual Reviews www.AnnualReviews.org;Clarkson 1985.)


Roots grow preferentially in resource hotspots. Root growth in the soil is not random.Roots that encounter microsites of high nutri-ent availability branch profusely (Hodge et al.1999), allowing plants to exploit preferentiallyzones of high nutrient availability. This explainswhy root length is greatest in surface soils (Fig.7.3), where nutrient inputs and mineraliza-tion are greatest, even though roots tend to begeotropic (i.e., grow vertically downward).This exploitation of nutrient hot spots ensuresthat plants maximize the nutrient return for agiven investment in roots and reduces the fine-scale heterogeneity in soil nutrient con-centrations. At a finer scale, root hairs, the elongate epidermal cells of the root that extendout into the soil, increase in length (e.g., from0.1 to 0.8mm) in response to a reduction in thesupply of nitrate or phosphate (Bates andLynch 1996). Both of these responses increasethe length and surface area of roots availablefor nutrient uptake. Exploitation of hot spotsdoes not always occur, however (Robinson1994), and may be more pronounced in fast-growing than in slow-growing species (Huanteet al. 1998).

Root length is a better predictor of nutrientuptake than is root biomass. Root length cor-relates closely with nutrient acquisition inshort-term studies of nutrient uptake by plantsfrom soils. Roots with a high specific root length(SRL; i.e., root length per unit mass) maximizetheir surface area per unit root mass and there-fore the volume of soil that can be explored by a given investment in root biomass. We know much less about the morphology andphysiology of roots in soil than of leaves. Thelimited available data suggest, however, thatherbaceous plants (especially grasses) oftenhave a greater SRL than woody plants and thatthere is a wide range in SRL among roots in anyecosystem. Much of the variation in SRLreflects the multiple functions of belowgroundorgans. Roots can have a high SRL eitherbecause they have a small diameter or becausethey have a low tissue density (mass per unitvolume). Some belowground stems and coarseroots have large diameters to store carbohy-drates and nutrients or to transport water andnutrients and play a minor role in nutrient

uptake. There may also be a tradeoff betweenSRL and longevity among fine roots, with high-density roots being less prone to desiccationand herbivory than low-density roots. Both theleaves and roots of slowly growing species oftenhave high tissue density, low rates of resourceacquisition (carbon and nutrients, respectively)but greater longevity than do leaves and rootsof more rapidly growing species (Craine et al.2001).

Mycorrhizae

Mycorrhizae increase the volume of soilexploited by plants. Mycorrhizae are symbioticrelationships between plant roots and fungalhyphae, in which the plant acquires nutrientsfrom the fungus in return for carbohydratesthat constitute the major carbon source for thefungus. About 80% of angiosperm plants, allgymnosperms, and some ferns are mycorrhizal(Wilcox 1991). These mycorrhizal relationshipsare important across a broad range of environ-mental and nutritional conditions, includingfertilized crops (Allen 1991, Smith and Read1997). With respect to nutrient uptake, mycor-rhizal hyphae basically serve as an extension ofthe root system into the bulk soil, often pro-viding 1 to 15m of hyphal length per centime-ter of root—that is, an increase in absorbinglength of two to three orders of magnitude.Because the nutrient transport through hyphaeoccurs more rapidly than by diffusion along atortuous path through soil water films, mycor-rhizae reduce the diffusion limitation of uptakeby plants. The small diameter of mycorrhizalhyphae (less than 0.01mm) compared to roots(generally 0.1 to 1mm) enables plants to exploitmore soil with a given biomass investment inmycorrhizal hyphae than for the same biomassinvested in roots. Plants typically invest 4 to20% of their gross primary production (GPP)in supporting mycorrhizal hyphae (Lambers etal. 1996). Most of this carbon supports mycor-rhizal respiration rather than fungal biomass,so a given carbon investment in mycorrhizalbiomass can represent a large carbon cost to the plant. Mycorrhizae are most important insupplementing the nutrients that diffuse slowly through soils, particularly phosphate and

Nutrient Uptake 183

potentially ammonium in ecosystems with low rates of nitrification. Although laboratoryexperiments show that plants consistentlyexclude mycorrhizae from roots under high-nutrient conditions, the extensive distributionof mycorrhizae across a wide range of soil fer-tilities, including most crop ecosystems, sug-gests that mycorrhizae continue to provide anet benefit to plants even in relatively fertilesoils.

There are a range of mycorrhizal types, butthe most common are arbuscular mycorrhizae(AM; also termed vesicular arbuscular mycor-rhizae, VAM) and ectomycorrhizae. AM fungigrow through the cell walls of the root cortex(i.e., the layers of root cells involved in nutri-ent uptake), much as does a root pathogenicfungus. In contrast to root pathogens, AMproduce arbuscules, which are highly branchedtreelike structures produced by the fungus andsurrounded by the plasma membrane of theroot cortical cells. Arbuscules are the structuresthrough which nutrients and carbohydrates areexchanged between the fungus and the plant.AM are most common in herbaceous commu-nities, such as grasslands, and in phosphorus-limited tropical forests and early successionaltemperate forests. Many AM associations arerelatively nonspecific and can occur even withecotmycorrhizal plant species shortly after dis-turbance. AM are generally eliminated afterectomycorrhizae colonize the roots of thesespecies.

In a given ecosystem type, AM associationsare best developed under conditions of phos-phorus limitation, where they short-circuit thediffusion limitation of uptake (Allen 1991,Read 1991). Their effectiveness in overcomingphosphorus limitation may contribute to thenitrogen limitation in many temperate ecosys-tems (Grogan and Chapin 2000). The AM sym-biosis is a dynamic interaction between plantand fungus, in which both roots and hyphaeturn over rapidly. Under conditions in whichplant growth is carbon limited, as in youngseedlings or in shaded or highly fertile condi-tions, mycorrhizae may act as parasites andreduce plant growth (Koide 1991). Under theseconditions, the plant reduces the number ofinfection points in new roots.As older roots die,

this reduces the proportion of colonized roots,thus decreasing the carbon drain from theplant. AM associations might be viewed as abalanced parasitism between root and fungusthat is carefully regulated by both partners.

Ectomycorrhizae are relatively stable associ-ations between roots and fungi that occur pri-marily in woody plants. The exchange organ is a mantle or sheath of fungal hyphae that surround the root plus additional hyphae thatgrow through the cell walls of the cortex (theHartig net). Roots respond to ectomycorrhizalcolonization by reducing root elongation andincreasing branching, forming short, highlybranched rootlets. Fungal tissue accounts forabout 40% of the volume of these root tips. Aswith AM, ectomycorrhizae involve an exchangeof nutrients and carbohydrates between thefungus and the plant. In contrast to AM, ecto-mycorrhizae generally prolong root longevity.Ectomycorrhizae also differ from AM in thatthey have proteases and other enzymes thatattack organic nitrogen compounds. The fungusthen absorbs the resulting amino acids andtransfers them to the plant (Read 1991). Ecto-mycorrhizae therefore enhance both nitrogenand phosphorus uptake by plants.

There are other mycorrhizal associations thatdiffer functionally from AM and ectomycor-rhizae. Fine-rooted heath plants in the familiesEricaceae and Epacridaceae, for example, formmycorrhizae in which the fungal tissue accountsfor 80% of the root volume.These mycorrhizae,like ectomycorrhizae, hydrolyze organic nitrogen and transfer the resulting amino acidsto their host plants. Many nonphotosyntheticorchids totally depend on their mycorrhizae forcarbon as well as nutrients. Their mycorrhizalfungi generally form links between the orchidand some photosynthetic plant species, espe-cially conifers. In this case, the plant is clearlyparasitic on the fungus.

As with the orchid–fungal association, ecto-mycorrhizae and AM often attach to severalhost plants, often of different species. Carbonand nutrients can be transferred among plantsthrough this fungal network, although rela-tively few studies have shown a net transfer ofcarbon among plants (Simard et al. 1997). Ifthese fungal connections among plants cause


large net transfers of carbon and nutrients, theycould alter competitive interactions, withpotentially important ecosystem consequences.The net carbon transfer from canopy plants to shaded seedlings (Simard et al. 1997), forexample, could promote the establishment ofshade-tolerant tree seedlings, which wouldincrease total ecosystem carbon demand andperhaps net photosynthesis (GPP). The quanti-tative significance of these transfers in naturalecosystems is totally unknown.

Root Uptake Properties

Active transport is the major mechanism bywhich plants absorb potentially limiting nu-trients from the soil solution at the root sur-face. Plant roots acquire nutrients from the soilsolution primarily by active transport, anenergy-dependent transport of ions across cellmembranes against a concentration gradient.Due to the high concentrations of ions andmetabolites inside plant cells, there is a constant

leakage out of the root along a concentrationgradient. Phosphate, for example, leaks fromroots at about a third of the rate at which it isabsorbed from the soil. This passive leakage ofions, sugars, and other metabolites probablyaccounts for much of the exudation from fineroots. Ions that enter the root move passivelyby diffusion and mass flow through the cellwalls of the cortex toward to the interior of theroot (Fig. 8.3). As nutrients move through thecortical cell walls toward the center of the root,adjacent cortical cells absorb these nutrients byactive transport. Nutrients can move throughthe cell walls only as far as the endodermis,a suberin-coated (wax-coated) layer of cellsbetween the cortex and the xylem. Once nutrients have been absorbed by cortical cells,they move through a chain of interconnectedcells to the endodermis, where they are secretedinto the dead xylem cells that transport thenutrients to the shoot in the transpirationstream. As much as 30 to 50% of the carbonbudget of the root goes to supporting nutrient

Root &rhizosphere

Cortex Corticalcell

PlasmodesmataXylem

Endodermis

Diffusion shell

Phloem

EpidermisRoot hair

Transporter

Cell wall

Plasma membrane

Figure 8.3. Cross-section of a root at three scales.The rhizosphere (or diffusion shell) is the zone of soilinfluenced by the root. The cortex has an outer layerof cells (the epidermis), some of which are elongatedto form root hairs. The cortex is separated from thetransport tissues (xylem and phloem) by a layer ofwax-impregnated cells (the endodermis). Each corti-cal cell absorbs ions that diffuse through the pore

spaces in the cell wall to the cell membrane. Mem-brane-bound proteins (transporters) transport ionsacross the cell membrane by active transport. Ionsmove from the outermost cortical cells toward theendodermis either through the cell walls or throughthe cytoplasmic connections between adjacent corti-cal cells (plasmodesmata).

Nutrient Uptake 185

absorption, indicating the large energetic costof nutrient uptake. Elements required in smallquantities are often absorbed simply by massflow or diffusion into the root cortical cells(Table 8.2).

Some plant species tap pools of nutrients thatare unavailable to other plants. Although allplants require the same suite of nutrients insimilar proportions, nitrogen is available inseveral forms (nitrate, ammonium, amino acids,etc.) that differ in availability among ecosys-tems. Many species preferentially absorbammonium (and perhaps amino acids) overnitrate, when all nitrogen forms are equallyavailable (Table 8.3). Species differ, however, intheir relative preference for these nitrogenforms, frequently showing a high capacity toabsorb the forms that are most abundant in the ecosystems to which they are adapted.Many species that occupy highly organic soilsof tundra and boreal forest ecosystems, forexample, preferentially absorb amino acids(Chapin et al. 1993, Näsholm et al. 1998), al-though even agricultural species use amino acidnitrogen (Näsholm et al. 2000). An importantcommunity consequence of species differencesin nitrogen preference is that nitrogen re-presents several distinct resources for whichspecies can compete. Species in the same com-munity often have quite different isotopic sig-natures of tissue nitrogen, because they acquire

nitrogen from different sources—either dif-ferent chemical fractions (nitrate, ammonium,organic nitrogen) or different soil depths(Nadelhoffer et al. 1996) (Fig. 8.4). Changes inspecies composition could therefore alter thenitrogen pools that are used to support primaryproduction. In most cases, the species presentare capable of using the prevailing forms ofavailable nitrogen. If human activities alter the prevailing form of available nitrogen, forexample through nitrate deposition in conifer-ous forests, this novel form of nitrogen may beused less effectively by the extant vegetation.

Table 8.3. Preference ratios for plant absorption of different forms of nitrogena when all forms are equallyavailable.

NH4+ : NO3

- Glycine : NH4+

Species preferenceb preferenceb References

Arctic vascular plants 1.1 2.1 ± 0.6 (12) Chapin et al. (1993), Kielland (1994)Arctic nonvascular plants 5.0 ± 1.5 (2) Kielland (1997)Boreal trees 19.3 ± 5.8 (4) 1.3 Chapin et al. (1986b), Kronzucker

et al. (1997), Näsholm et al. (1998)Alpine sedges 3.9 ± 1.3 (12) 1.5 ± 0.4 (11) Raab et al. (1999)Temperate heath 1.0 Read and Bajwa (1985)Salt marsh 1.3 Morris (1980)Mediterranean shrub 1.2 Stock and Lewis (1984)Barley 1.0 Chapin et al. (1993)Tomato 0.6 Smart and Bloom (1988)

a Assumes all forms of nitrogen are equally available.b A preference ratio >1 indicates that the first form of nitrogen is absorbed preferentially over the second. Number ofspecies studied in parentheses. Research shows that many plants preferentially absorb glycine (a highly mobile aminoacid) over ammonium and preferentially absorb ammonium over nitrate, when all forms are equally available.

Tissue δ15N

Woody evergreen

Woody deciduous

Graminoids,cryptogams

Aquatic

Pla

nt fu

nctio

nal t

ype

-10 -8 -6 -4 -2 0 2 4

Figure 8.4. Concentration of 15N in tissues from dif-ferent growth forms of boreal plants. (Figure pro-vided by K. Kielland; Kielland 1999.)


Spruces, for example, preferentially use ammo-nium over nitrate (Kronzucker et al. 1997), sonitrate entering in acid rain might leach fromthe ecosystem and pollute groundwater, lakes,and streams rather than being effectivelyabsorbed by plants.

The three major forms of nitrogen differ intheir carbon cost of incorporation into biomass.The carbon cost of incorporating amino acids isminimal, whereas ammonium must be attachedto a carbon skeleton (the process of assimila-tion) before it is useful to the nitrogen economyof the plant. Finally, nitrate must be reduced toammonium, which is energetically expensive,before it can be assimilated. Most plants reducesome of the nitrate in leaves, using excess reduc-ing power from the light reaction. In this case,the high energy cost of nitrate reduction doesnot detract from energy available for other plantprocesses. High availability of light or nitrateusually increases the proportion of nitratereduced in leaves. Species also differ in theircapacity to reduce nitrate in leaves, with speciesadapted to high-nitrate environments usuallyhaving a higher capacity to reduce nitrate in theleaves. Tropical and subtropical perennials andmany annual plants typical of disturbed habitats,for example, reduce a substantial proportion oftheir nitrate in leaves,whereas temperate peren-nials reduce most nitrate in the roots (Lamberset al. 1998). Nitrogen availability is usually solimited in temperate and high-latitude terres-trial environments that the relative availabilityof nitrogen forms is more important than cost ofassimilation in determining which forms ofnitrogen are used by plants. Plants usuallyabsorb whatever they can get.

Plant species also differ in the pools of phos-phorus they can tap. Roots of some plantspecies produce phosphatase enzymes thatrelease inorganic phosphate for absorption byplant roots. The dominant sedge in arctictussock tundra, for example, meets about 75%of its phosphorus requirement by absorbing theproducts of its phosphatase enzymes (Kroehlerand Linkins 1991). Other plants, particularlythose in dry environments, secrete chelates suchas citrate that diffuse from the root into thebulk soil. These chelates bind iron from insolu-ble iron phosphate complexes, which solubilizes

phosphate; soluble phosphate then diffuses tothe root (Lambers et al. 1998). Some plants,particularly Australian and South African heathplants in the Proteaceae, produce dense clustersof roots (proteoid roots) that are particularlyeffective in secreting chelates and solubilizingiron phosphates. There are many classes ofchelates (siderophores) produced by plantroots, although the benefit to the plant fromthese secretions is poorly known. Plants there-fore differ in the soil phosphorus pools they can exploit, but we have only a rudimentaryunderstanding of the ecosystem consequencesof these species differences.

Species differences in rooting depth anddensity influence the pool of nutrients that canbe absorbed by vegetation. Grasslands andforests growing adjacent to one another on the same soil often differ greatly in annual nu-trient uptake and productivity, because themore deeply rooted forest trees exploit a largersoil volume and therefore a larger pool of waterand available nutrients than do shallow-rootedspecies (see Chapter 12). In summary, there areseveral mechanisms by which species composi-tion influences the quantity and form of nutri-ents acquired by vegetation.

Root uptake capacity increases in response toplant demand for nutrients. When the above-ground environment favors rapid growth andassociated high demand for nutrients, plantroots respond by synthesizing more transportproteins in root cortical cells, thus increasing thecapacity of the root to absorb nutrients. Speciesthat have an inherently high relative growthrate or experience conditions that support rapidgrowth therefore have a high capacity to absorbnutrients (Chapin 1980). High light and warmair temperatures, for example, increase rootuptake capacity, whereas shade, drought, andphenologically programmed periods of reducedgrowth lead to a low uptake capacity. The ratesof nutrient uptake by vegetation are thereforeinfluenced both by soil factors that determinenutrient supply and by plant factors that deter-mine nutrient demand. There is a close correla-tion in the field between nutrient uptake and NPP (Fig. 8.5). It is difficult, however, to separate cause from effect in explaining this correlation.

Nutrient Uptake 187

Changes in root uptake kinetics fine-tune thecapacity of plants to acquire specific nutrients.Ion-transport proteins are specific for parti-cular ions. In other words, ammonium, nitrate,phosphate, potassium, and sulfate are eachtransported by a different membrane-boundprotein that is individually regulated (Clarkson1985). Plants induce the synthesis of additionaltransport proteins for those ions that specifi-cally limit plant growth. Roots of a phosphorus-limited plant therefore have a high capacity toabsorb phosphate, whereas roots of a nitrogen-limited plant have a high capacity to absorbnitrate and ammonium (Table 8.4). Nitratereductase, the enzyme that reduces nitrate toammonium (the first step before nitrate-nitrogen can be incorporated into amino acidsfor biosynthesis) is also specifically induced bypresence of nitrate. There are therefore several adjustments that plants make toimprove resource balance. Plants first alter theroot to shoot ratio to improve the balancebetween acquisition of belowground andaboveground resources. Plants then regulatethe location of root growth to exploit hot spotsof nutrient availability. Finally, plants adjusttheir capacity to absorb specific nutrients, whichbrings the plant nutrient ratios closer to valuesthat are optimal for growth.

Nutrient ratios define a stoichiometry ofnutrient cycles in ecosystems. The oceanogra-pher Redfield (1958) noted that algae with aratio of nitrogen to phosphorus (N :P ratio)greater than 14 :1 tend to respond to phospho-rus addition, whereas algae with a lower N :Pratio are nitrogen limited. This ratio is nowreferred to as the Redfield ratio. The importantimplication of nutrient ratios is that, if they are constant, the element that most stronglyconstrains production by vegetation defines the quantities of all elements that are cycledthrough vegetation. Marine algae with high N :P ratios, for example, preferentially absorbphosphorus. They absorb nitrogen and othernutrients in proportion (the Redfield ratio) to the phosphorus that they are able to ac-quire. Algae with low N:P ratios preferentiallyabsorb nitrogen and absorb phosphorus andother nutrients in proportion to the nitrogenthat they are able to acquire. The most stronglylimiting element therefore determines thecycling rates of all elements. Experiments interrestrial ecosystems suggest that terrestrialplants also adjust their mineral nutrition to con-verge on the same Redfield ratio. Heath plantswith N :P ratios in leaves of less than 14 :1 generally respond to experimental additions of nitrogen, whereas plants with N :P ratiosgreater than 16 :1 generally respond to addedphosphorus but not to nitrogen (Fig. 8.6). This

0 500 1000 1500

5

10

15

0

Production (g m-2 yr-1)

Nitr

ogen

upt

ake

(g m

-2 y

r-1) Evergreen forest

Deciduous forest

Figure 8.5. Relationship between nitrogen uptakeof temperate and boreal coniferous and deciduousforests and NPP. (Redrawn with permission fromAcademic Press; Chapin 1993b.)

Table 8.4. Effect of environmental stresses on rateof nutrient absorption by barley.

Uptake rate bystressed plants

Stress Ion absorbed (% of control)

Nitrogen ammonium 209nitrate 206phosphate 56sulfate 56

Phosphorus phosphate 400nitrate 35sulfate 70

Sulfur sulfate 895nitrate 69phosphate 32

Water phosphate 32Light nitrate 73

Data from Lee (1982), Lee and Rudge (1987), and Chapin(1991a).


principle of optimal element ratios is used in agriculture to determine which nutrientslimit crop growth so nutrient additions can bematched to plant requirements.

Element ratios are more variable among ter-restrial plants they are among phytoplanktonbecause of the greater capacity for nutrientstorage. Terrestrial plants have storage organs(e.g., stems) and organelles (e.g., vacuoles) inwhich they store nutrients that are nonlimitingto growth. In this way, terrestrial plants can takeadvantage of short-term pulses of nutrientsupply. Nitrogen and phosphorus, for example,often show an autumn pulse of availability,when leaves are shed and leached by rain,and a spring pulse, after a winter season whendecomposers are more active than plants (seeChapter 6). Plants absorb these nutrients attimes of abundant supply, altering the ratios ofelements in their tissues. These nutrients arethen drawn out of storage at times when thedemands for growth exceed uptake from thesoil. In arctic tundra, for example, each year’sproduction is supported primarily by nutrientsabsorbed in previous years, and uptake serves

primarily to replenish these stores (Chapin et al. 1980a). In one field experiment, tundraplants that were provided with only distilledwater grew just as rapidly as did plants with free access to soil nutrients, indicating that nu-trient stores were sufficient to support an entireseason’s production (Jonasson and Chapin1985). Even in the ocean and freshwater ecosys-tems, element ratios of algae can be variabledue to storage of nonlimiting nutrients in vac-uoles (see Chapter 10).

Nutrient uptake alters the chemical proper-ties of the rhizosphere. Nutrient absorption by plant roots reduces the concentration ofnutrients adjacent to the root. This depletion of soluble nutrients by root uptake can be substantial for nutrients that diffuse readily and create large diffusion shells. The pool sizesof dissolved nutrients in the soil solution aretherefore a poor indicator of nutrient availabil-ity; dissolved nutrient pools can be smallbecause of low mineralization rates or rapiduptake. Plant nutrient uptake is a criticalcontrol over ecosystem retention of mobilenutrients such as nitrate. Forest clearing or crop removal, for example, makes soils moreprone to nitrate leaching into groundwater andstreams (Bormann and Likens 1979).

A second major consequence of plant nutri-ent absorption is a change in rhizosphere pH.Whenever a root absorbs an excess of cations,it secretes hydrogen ions (H+) into the rhizos-phere to maintain electrical neutrality. This H+

secretion acidifies the rhizosphere. Except for nitrogen, which can be absorbed either as a cation (NH4

+) or an anion (NO3-), the ions

absorbed in greatest quantities by plants arecations (e.g., Ca2+, K+, Mg2+), with phosphateand sulfate being the major anions (Table 8.2).When plants absorb most nitrogen as NH4

+,their cation uptake greatly exceeds anionuptake, and they secrete H+ into the rhizos-phere to maintain charge balance, causing acidification of the rhizosphere. When plantsabsorb most nitrogen as NO3

-, theircation–anion absorption is more nearly bal-anced, and roots have less effect on rhizospherepH.Ammonium tends to be the dominant formof inorganic nitrogen in acidic soils, whereasnitrate makes up a larger proportion of inor-

00

1

2

3

4

251510 205

P c

onte

nt (

mg

g-1

)

N content (mg g -1)

N:P <14

N:P >16

P-limitation

co-limitation by N and P

N-limitation

Figure 8.6. Relationship between nitrogen andphosphorus concentration of leaves in heath plants.Each datum point represents a site where nutrientaddition experiments show that plant growth islimited by nitrogen, phosphorus, or both. Plants withan N : P ratio less than 14 respond primarily to nitro-gen, whereas plants with an N : P ratio greater than16 respond primarily to phosphorus. (Redrawn withpermission from Journal of Applied Ecology;Koerselman and Mueleman 1996.)

Nutrient Use 189

ganic nitrogen in basic soils (see Chapter 9).The uptake process therefore tends to makeacidic soils more acidic.

Roots also alter the nutrient dynamics of therhizosphere through large carbon inputs fromroot death, the sloughing of mucilaginous car-bohydrates from root caps, and the exudationof organic compounds by roots. These carboninputs to soil may account for 10 to 30% of theGPP (see Chapter 6). Root exudation providesa labile carbon source that stimulates thegrowth of bacteria, which acquire their nitrogenby mineralizing organic matter in the rhizos-phere (see Chapter 7). This nitrogen becomesavailable to plant roots when bacteria aregrazed by protozoans or bacteria becomeenergy starved due a reduction in root exuda-tion (see Fig. 7.12). Plants are sometimes ef-fective competitors with microbes for soilnutrients, for example, when plant carbon statusis enhanced by added CO2 (Hu et al. 2001). Weknow relatively little, however, about factorsthat govern competition for nutrients betweenplants and microbes.

Nutrient Use

Nutrients absorbed by plants are used primar-ily to support the production of new tissues(NPP). Plants store nutrients only when nutrient uptake exceeds the requirements forproduction (Chapin et al. 1990). Nutrientsabsorbed by roots move upward in the xylemand are then recirculated in the phloem to siteswhere production or storage occurs. The hor-monal balance of the plant governs the patternsof carbon and nutrient transport in the phloem(and therefore the allocation of nutrientswithin the plant).

Nitrogen is incorporated primarily into proteins with lesser amounts in nucleic acidsand lipoproteins (Chapin and Kedrowski 1983,Chapin et al. 1986a). Phosphorus is incor-porated preferentially into sugar phosphatesinvolved in energy transformations (photosyn-thesis and respiration), nucleic acids, and phos-pholipids (Chapin et al. 1982). Calcium is animportant component of cell walls (calciumpectate). Potassium is important in osmotic

regulation. Other cations (e.g., magnesium andmanganese) serve as cofactors for enzymes(Marschner 1995).

The highest concentrations of nitrogen, phos-phorus, and potassium typically occur in leaves(and to a lesser extent in fine roots) because ofthe importance of these elements in metabo-lism. Wood, in contrast, has low concentrationsof all elements, (especially nitrogen and phos-phorus) because of the large proportion ofxylem, which consists of dead cells. Calcium,which is associated with cell walls, forms ahigher proportion of total plant nutrients inwood than other tissues. Roots are intermedi-ate in their tissue concentrations. The quantityof elements allocated to each tissue depends on tissue concentrations and on biomass allocation.

Nutrient supply affects growth more than itaffects nutrient concentration. Plants respondto increased supply of a limiting nutrient in laboratory experiments primarily by increasingplant growth, giving a linear relationshipbetween rate of nutrient accumulation andplant growth rate (Fig. 8.7) (Ingestad andÅgren 1988). Plants also respond to increasednutrient supply in the field primarily throughincreased NPP (Fig. 8.5), with proportionately

0.08 0.09 0.10 0.11 0.12 0.13

0.08

0.09

0.10

0.11

0.12

0.13

Root RGR (g g-1 d-1)

Rel

ativ

e up

take

rat

e (m

g m

g-1d-1

)

Figure 8.7. The rate of nitrogen uptake in tobaccoas a function of the relative growth rate of roots(RGR). (Redrawn with permission from BotanicalGazette,Vol. 139 © 1978 University of Chicago Press;Raper et al. 1978.)


less increase in tissue nutrient concentration.Tissue nutrient concentrations increase sub-stantially only when other factors begin to limitplant growth. The sorting of species by habitatalso contributes to the responsiveness of nutri-ent uptake and NPP to variations in nutrientsupply observed across habitats. Species such as trees that use large quantities of nutrientsdominate sites with high nutrient supply rates, whereas infertile habitats are dominatedby species with lower capacities for nutrientabsorption and growth. Despite these physio-logical and species adjustments, tissue nutrientconcentrations in the field generally increasewith an increase in nutrient supply.

Nutrient use efficiency is greatest where pro-duction is nutrient limited. Differences amongplants in tissue nutrient concentration provideinsight into the quantity of biomass that anecosystem can produce per unit of nutrient.Nutrient use efficiency (NUE) is the ratio ofnutrients to biomass lost in litterfall (i.e., theinverse of nutrient concentration in plant litter)(Vitousek 1982). This ratio is highest in unpro-ductive sites (Fig. 8.8), suggesting that plantsare more efficient in producing biomass perunit of nutrient acquired and lost when nu-trients are in short supply. Several factors contribute to this pattern (Chapin 1980). First,tissue nutrient concentration tends to decline as soil fertility declines, as described earlier.

Individual plants that are nutrient limited alsoproduce tissues more slowly and retain these tissues for a longer period of time, re-sulting in an increase in average tissue age.Older tissues have low nutrient concentrations,causing a further decline in concentration (i.e.,increased NUE). Finally, the dominance ofinfertile soils by species with long-lived leavesthat have low nutrient concentrations furthercontributes to the high NUE of ecosystems oninfertile soils.

Plants maximize NUE in infertile soils byreducing nutrient loss more than by increasingnutrient productivity. There are at least twoways in which a plant might maximize biomassgained per unit of nutrient (Berendse and Aerts 1987): through a high nutrient produc-tivity (an)—that is, a high instantaneous rate of carbon uptake per unit nutrient, and through a long residence time (tr)—that is, theaverage time that the nutrient remains in theplant.

(8.1)

Species characteristic of infertile soils have a long residence time of nutrients but a lownutrient productivity (Chapin 1980, Lambersand Poorter 1992), suggesting that the highNUE in unproductive sites results primarily

NUE g biomass gN

g biomass gN yr yr

n r= ¥ ( )[= ( ) ]

-

- -

a t1

1 1 *

2 4 6 8 10 20

240

12 14 16 18

40

80

120

160

200

280

N in litterfall (g m-2 yr-1)

Bio

mas

s:N

rat

io o

f litt

erfa

ll C

C

C

C

C

CCC

C

CC

CC

C

C

C

C

CC

C

CCC

T

TT

T TTT

T

T

TT

T

T

T

TT T T

TTT

TTT

TT

T

M

DD

DDD

DD

D

DM

M

N

MMM

N

D

CCC

C

C

C

D

DD

D

DD

D

DDD T

DDD

D

D

D N

N NN

N

N N

T

00

Figure 8.8. Relationship betweenthe amount of nitrogen in litterfalland nitrogen use efficiency (ratio of dry mass to nitrogen in that litterfall). C, conifer forests; D,temperate deciduous forests; M,mediterranean-type ecosystems; N,temperate sites dominated by sym-biotic, nitrogen fixers; T, evergreentropical forests. (Redrawn with per-mission from American Naturalist;Vol. 119 © 1982 University ofChicago Press; Vitousek 1982.)

Nutrient Loss from Plants 191

from traits that reduce nutrient loss rather thantraits that promote a high instantaneous rate ofbiomass gain per unit of nutrient (Table 8.5).Shading also reduces tissue loss more stronglythan it reduces the rate of carbon gain (Waltersand Reich 1999).

There is an innate physiological trade-offbetween nutrient residence time and nutrientproductivity. This occurs because the traits that allow plants to retain nutrients reduce their capacity to grow rapidly (Chapin 1980,Lambers and Poorter 1992). Plants with a highnutrient productivity grow rapidly and havehigh photosynthetic rates, which is associatedwith thin leaves, a high specific leaf area, and ahigh tissue nitrogen concentration (see Chapter5). Conversely, a long nutrient residence time is achieved primarily through slow rates ofreplacement of leaves and roots. Leaves that survive a long time have more structuralcells to withstand unfavorable conditions andhigher concentrations of lignin and other secondary metabolites that deter pathogensand herbivores. Together these traits result indense leaves with low tissue nutrient concen-trations and therefore low photosynthetic ratesper gram of biomass (see Chapter 5). The highNUE of plants on infertile soils thereforereflects their capacity to retain tissues for a longtime rather than a capacity to use nutrientsmore effectively in photosynthesis.

Little is known about the trade-offs betweenroot longevity and nutrient uptake rate. Nutri-

ent uptake declines as roots age, lose root hairs,and become suberized; so trade-offs betweenphysiological activity and longevity that havebeen well documented for leaves probably alsoexist for roots.

The trade-off between NUE and rate ofresource capture explains the diversity of planttypes along resource gradients. Low-resourceenvironments are dominated by species thatconserve nutrients through low rates of tissueturnover, high NUE, and the physical andchemical properties necessary for tissues topersist for a long time. These stress-tolerantplants outcompete plants that are less effectiveat nutrient retention in environments that aredry, infertile, or shaded (Chapin 1980, Waltersand Field 1987). A high NUE and associatedtraits constrain the capacity of plants to cap-ture carbon and nutrients. In high-resourceenvironments, therefore, species with high rates of resource capture, rapid growth rates,rapid tissue turnover, and consequently lowNUE, outcompete plants with high NUE. Inother words, neither a rapid growth rate nor ahigh NUE is universally advantageous, be-cause there are inherent physiological trade-offs among these traits. The relative benefits tothe plant of efficiency vs. rapid growth dependson environment.

Nutrient Loss from Plants

The nutrient budget of plants, particularly long-lived plants, is determined just as much bynutrient loss as by nutrient uptake. The poten-tial avenues of nutrient loss from plants includetissue senescence and death, leaching of dis-solved nutrients from plants, consumption oftissues by herbivores, loss of nutrients to para-sites, exudation of nutrients into soils, and cat-astrophic loss of nutrients from vegetation byfire, wind-throw, and other disturbances. Nu-trient loss from plants is an internal transferwithin ecosystems (the transfer from plants tosoil). After this transfer to soil, nutrients arepotentially available for uptake by microbes or plants or may be lost from the ecosystem.Nutrient loss from plants to soil therefore hasvery different consequences from nutrient loss

Table 8.5. Nitrogen use efficiency and its physio-logical components in a heathland evergreen shruband a grassa.

Process Evergreen shrub Grass

Nitrogen productivity 77 110(gbiomassg-1 Nyr-1)

Mean residence time (yr) 1.2 0.8NUE (gbiomassg-1 N) 90 89

a Species are an evergreen shrub (Erica tetralix) and a co-occurring deciduous grass (Molinia caerulea) that isadapted to higher soil fertility. These two species havesimilar NUE, which is achieved by a high nitrogen produc-tivity in the high-nutrient-adapted species and by a highmean residence time in the low-nutrient-adapted species.Data from Berendse and Aerts (1987).


from the ecosystem to the atmosphere orgroundwater.

Senescence

Tissue senescence is the major avenue of nutri-ent loss from plants. Plants reduce senescenceloss of nutrients primarily by reducing tissueturnover, particularly in low-resource environ-ments. The leaves of grasses and evergreenwoody plants, for example, show greater leaflongevity in low-nutrient or low-water environ-ments than in high-resource environments(Chapin 1980). Species differences in tis-sue turnover strengthen this pattern of hightissue longevity in low-resource environments.The proportion of evergreen woody speciesincreases with decreasing soil fertility, reducingthe rate of leaf turnover at the ecosystem level.All else being equal, a reduction in tissueturnover causes a corresponding reduction inthe turnover of the associated tissue nutrients.This reduction in tissue turnover is probablythe single most important adaptation for nutri-ent retention in low-nutrient habitats (Chapin1980, Lambers and Poorter 1992).

Nutrient resorption is the transfer of solublenutrients out of a senescing tissue through thephloem. It plays a crucial, but poorly under-stood, role in nutrient retention by plants.Plants resorb, on average, about half of theirnitrogen, phosphorus, and potassium fromleaves before leaves are shed at senescence(Table 8.6), so nutrient resorption is quantita-tively important to plant nutrient budgets.

Although there is a large variation (0 to 90%)in the proportion of leaf nutrients resorbedboth among species and across sites, there is noconsistent relationship between the proportionof nutrients resorbed and plant nutrient status(Chapin and Kedrowski 1983,Aerts 1995). Effi-cient nutrient resorption is promoted by pres-ence of an active sink, for example when newleaf production coincides with leaf senescencein evergreens. The relatively high resorptionefficiency of graminoids may also reflect theirpattern of growth, in which production of newleaves is linked to senescence of older leaves(Table 8.6). Drought reduces the efficiency ofnutrient resorption (Pugnaire and Chapin 1992,Aerts and Chapin 2000). One possible explana-tion for the lack of clear environmental controlsover nutrient resorption is that nutrient resorp-tion may be influenced by so many factors thatno single environmental control is readily iden-tified. Nutrient resorption may also be such animportant trait that it is equally expressed in allplants (i.e., there may be no strong trade-offsthat make effective resorption disadvantageousin high-resource sites). There have been too few studies of nutrient resorption from roots or wood to draw firm conclusions about factorscontrolling nutrient resorption from thesetissues, although resorption from roots andwood appears to be much less than from leavesand may not occur at all. Resorbed nutrientsare transferred to other plant parts (e.g., seeds,storage organs, or leaves at the top of thecanopy) to support growth at other times orparts of the plant. Some nutrients, such as

Resorption efficiency(% of maximum pool)a

Growth form Nitrogen Phosphorus

All data 50.3 ± 1.0 (287) 52.2 ± 1.5 (226)Evergreen trees and shrubs 46.7 ± 1.6 (108) (a) 51.4 ± 2.3 (88) (a)Deciduous trees and shrubs 54.0 ± 1.5 (115) (b) 50.4 ± 2.0 (98) (a)Forbs 41.4 ± 3.7 (33) (a) 42.4 ± 7.1 (18) (a)Graminoids 58.5 ± 2.6 (31) (b) 71.5 ± 3.4 (22) (b)

a Means ± SE. Number of species in parentheses. Letters indicate statistical differ-ence between growth forms (P < .05).Data from Aerts (1995).

Table 8.6. Nitrogen and phos-phorus resorption efficiency ofdifferent growth forms.

Nutrient Loss from Plants 193

calcium and iron are immobile in the phloemso plants cannot resorb these nutrients fromsenescing tissues. Because these nutrientsseldom limit plant growth, their lack of resorp-tion has little direct nutritional impact onplants, except where acid rain greatly reducestheir availability (Aber et al. 1998, Driscoll et al. 2001).

Leaching Loss from Plants

Leaching of nutrients from leaves is an impor-tant secondary avenue of nutrient loss fromplants. Leachates account for about 15% of theannual nutrient return from aboveground plantparts to the soil. Rain dissolves nutrients on leafand stem surfaces and carries these to the soil asthroughfall (water that drops from the canopy)or stem flow (water that flows down stems).Stem flow typically has high concentrations ofnutrients due to leaching of the stem surface;however, only a small amount of water movesby this pathway. Throughfall typically accountsfor 90% of the nutrients leached from plants.Although plants with high nutrient status losemore nutrients per leaf, the proportion of nutri-ents recycled by leaching is surprisingly similaracross a wide range of ecosystems (Table 8.7).Leaching loss is most pronounced for thosenutrients that are highly soluble or are notresorbed. As much as 50% of the calcium and80% of the potassium in an apple leaf, forexample, can be leached within 24h. Leaching

rate is highest when rain first contacts a leaf andthen declines exponentially with time. Ecosys-tems with very different rainfall regimes maytherefore return similar proportions of nutri-ents to the soil through leaching vs. senescence.Although leaching loss is quantitatively impor-tant to plant nutrient budgets, there are no clearadaptations to minimize leaching loss.The thickcuticle of evergreen leaves was once thought toreduce leaching loss and explain the presence ofevergreen leaves in wet, nutrient-poor forests.There is no evidence, however, that leachingloss is related to cuticle thickness. Like nutrientresorption, leaching loss from plants is a quanti-tatively important term in plant nutrientbudgets that is not well understood. Biologistsunderstand the acquisition of carbon and nutri-ents by plants much better than the loss of theseresources.

Plant canopies can also absorb soluble nutri-ents from precipitation. Canopy uptake fromprecipitation is greatest in ecosystems wherethere is strong growth limitation by a givennutrient. In Germany, for example, nitrogeninputs in precipitation are so high that forestgrowth has switched from nitrogen to phos-phorus limitation. These forest canopies absorbphosphorus directly from precipitation.

Herbivory

Herbivores are sometimes a major avenue of nutrient loss from plants. Herbivores con-sume a relatively small proportion (1 to 10%)of plant production in many ecosystems. Inecosystems such as productive grasslands,however, herbivores regularly eat a large proportion of plant production; and, duringoutbreaks of herbivore population, herbivoresmay consume most aboveground production(see Chapter 11). Herbivory has a much largerimpact on plant nutrient budgets than thebiomass losses suggest, because herbivory pre-cedes resorption, so vegetation loses approxi-mately twice as much nitrogen and phosphorusper unit biomass to herbivores than it doesthrough senescence. Animals also generallyfeed on tissues that are rich in nitrogen andphosphorus, thus maximizing the nutritional

Table 8.7. Nutrients leached from the canopy(throughfall) as a percentage of the total above-ground nutrient return from plants to the soil.

Throughfall (% of annual return)a

Nutrient Evergreen forests Deciduous forests

Nitrogen 14 ± 3 15 ± 3Phosphorus 15 ± 3 15 ± 3Potassium 59 ± 6 48 ± 4Calcium 27 ± 6 24 ± 5Magnesium 33 ± 6 38 ± 5

a Means ± SE, for 12 deciduous and 12 evergreen forests.Data from Chapin (1991b).


impact of herbivory on plants. There has there-fore been strong selection for effective chemi-cal and morphological defenses that deterherbivores and pathogens. These defenses arebest developed in tissues that are long lived and in environments where there is an inade-quate supply of nutrients to readily replacenutrients lost to herbivores (Coley et al. 1985,Gulmon and Mooney 1986, Herms and Mattson1992). Most nutrients transferred from plants toherbivores are rapidly returned to the soil infeces and urine, where they quickly becomeavailable to plants. In this way herbivory speedsup nutrient cycling (see Chapter 11), parti-cularly in ecosystems that are managed forgrazing. Nutrients are susceptible to loss fromthe ecosystem in situations in which overgraz-ing reduces plant biomass to the point thatplants cannot absorb the nutrients returned tothe soil by herbivores.

Other Avenues of Nutrient Loss from Plants

Other avenues of nutrient loss are poorlyknown. Although laboratory studies suggestthat root exudates containing amino acids maybe a significant component of the plant carbonbudget (Rovira 1969), the magnitude of nitro-gen loss from plants by this avenue is unknown.Other avenues of nutrient loss from plantsinclude plant parasites such as mistletoe andnutrient transfers by mycorrhizae from oneplant to another.Although these nutrient trans-fers may be critical to the nutrient distributionamong species in the community, they do notgreatly alter nutrient retention or loss by vege-tation as a whole.

Disturbances cause occasional large pulses of nutrient release. Fire, wind, disease epide-mics, and other catastrophic disturbances causemassive nutrient losses from vegetation whenthey occur (see Chapter 13), but these lossesare small (less than 1% of nutrients cycledthrough vegetation) when averaged over theentire disturbance cycle. Even in fire-pronesavannas and grasslands, fires generally burnduring the dry season after senescence hasoccurred and burn more litter than live plantbiomass. Most plant nutrients in these ecosys-

tems are stored belowground during timeswhen fires are likely to occur.

Summary

Nutrient availability is a major constraint onthe productivity of the terrestrial biosphere.Whereas carbon acquisition by plants is de-termined primarily by plant traits (leaf area and photosynthetic capacity), nutrient uptake isusually governed more strongly by environ-ment (the rate of supply by the soil) rather thanby plant traits. In early succession, however,plant traits can have a significant impact onnutrient uptake by vegetation at the ecosystemlevel. Diffusion is the major process that deliv-ers nutrients from the bulk soil to the rootsurface. Mass flow of nutrients in flowing soilwater augments this nutrient supply for abun-dant nutrients or nutrients that are required insmall amounts by plants. Root biomass differsless among ecosystems than does abovegroundbiomass because those ecosystems that arehighly productive and produce a large above-ground biomass have a relatively low allocationto roots.

Plants adjust their capacity to acquire nutri-ents in several ways. Preferential allocation toroots under conditions of nutrient limitationmaximizes the root length available to absorbnutrients. Root growth is concentrated in hotspots of relatively high nutrient availability,maximizing the nutrient return for roots thatare produced. Plants further increase theircapacity to acquire nutrients through symbioticassociations with mycorrhizal fungi. Plants alterthe kinetics of nutrient uptake in response totheir demand for nutrients. Plants that growrapidly, due either to a favorable environmentor a high relative growth rate, have a highcapacity to absorb nutrients. Plants adjust theabsorption of specific nutrients by maximizingthe capacity to absorb those elements that moststrongly limit growth. In the case of nitrogen,which is frequently the most strongly limit-ing nutrient, plants typically absorb whateverforms are available in the soil. When all formsare equally available, most plants preferentiallyabsorb ammonium or amino acids rather than


nitrate. Nitrate absorption is often important,however, because of its high mobility.

There is an inevitable trade-off between themaximum rate of nutrient investment in newgrowth and the efficiency with which nutrientsare used to produce biomass. Plants producebiomass most efficiently per unit of nutrientunder nutrient-limiting conditions. Nutrient useefficiency is maximized by prolonging tissuelongevity—that is, by reducing the rate at whichnutrients are lost. Senescence is the majoravenue by which nutrients are lost from plants.Plants minimize loss of growth-limiting nutri-ents by resorbing about half of the nitrogen,phosphorus, and potassium from a leaf before itis shed. About 15% of the annual nutrientreturn from aboveground plant parts to the soilcomes as leachates, primarily as through-fallthat drips from the canopy. Herbivores can alsobe important avenues of nutrient loss becausethey preferentially feed on nutrient-rich tissuesand consume these tissues before resorptioncan occur. For these reasons plants lose morethan twice as much nutrients per unit of biomassto herbivores than through senescence. Otherfactors that cause occasional large nutrientlosses from vegetation include disturbances(e.g., fire and wind) and diseases that kill plants.

Review Questions

1. Mass flow, diffusion, and root interceptionare three processes that deliver nutrients tothe root surface. Describe the mechanismby which each process works. What is therelative importance of these threeprocesses in providing nutrients to plants?How does soil fertility influence the relativeimportance of these processes?

2. How do soil and plant properties influencerates of diffusion and mass flow? How can the plant maximize these transportprocesses?

3. What is the major mechanism by whichplants get nutrients into the plant, oncethey have arrived at the root surface?

4. How do plants compensate for (a) lowavailability of all nutrients, (b) spatial vari-ability in nutrients within the soil (localized

hot spots), and (c) imbalance betweennutrients required by plants (e.g., nitrogenvs. phosphorus availability)?

5. What is the rhizosphere? How do plantsinfluence the rhizosphere?

6. How does plant growth rate affect nutrientuptake?

7. What are the major mechanisms by whichmycorrhizae increase nutrient uptake byplants? Under what circumstances are myc-orrhizae most strongly developed?

8. What are the major processes involved inconverting nitrogen from nitrate to a formthat is biochemically useful to the plant?Why is nitrogen acquisition by plants soenergetically expensive?

9. Why are nitrogen and carbon flows inplants so tightly linked? What happens tonutrient uptake when carbon gain isrestricted? What happens to carbon gainwhen nutrient uptake is restricted? Whatare the mechanisms by which these adjust-ments occur?

10. What is nitrogen use efficiency? What are the physiological causes of differences in NUE, and what are the ecosystem consequences?

11. Why do all aquatic and terrestrial plantstend to show a similar balance of nutrients(the Redfield ratio)? How can you use thisinformation to estimate which nutrient ismost strongly limiting to plants in a partic-ular site?

12. What are the major differences in types ofspecies that occur in fertile vs. infertilesoils? What are the advantages and disad-vantages of each plant strategy in each soiltype?

13. What are the major avenues of nutrient lossfrom plants? How do all plants minimizethis nutrient loss? What are the major adap-tations that minimize nutrient loss fromplants that are adapted to infertile soils?

Additional Reading

Aerts, R. 1995. Nutrient resorption from senescingleaves of perennials: Are there general patterns?Journal of Ecology 84:597–608.


Chapin, F.S. III. 1980. The mineral nutrition of wildplants. Annual Review of Ecology and Systematics11:233–260.

Hobbie, S.E. 1992. Effects of plant species on nutri-ent cycling. Trends in Ecology and Evolution7:336–339.

Ingestad, T., and G.I. Ågren. 1988. Nutrient uptakeand allocation at steady-state nutrition. Physiolo-gia Plantarum 72:450–459.

Nye, P.H., and P.B. Tinker. 1977. Solute Movement inthe Soil-Root System. University of CaliforniaPress, Berkeley.

Read, D.J. 1991. Mycorrhizas in ecosystems. Experi-entia 47:376–391.

Vitousek, P.M. 1982. Nutrient cycling and nutri-ent use efficiency. American Naturalist 119:553–572.

Introduction

Human impact on nutrient cycles has funda-mentally changed the regulation of ecosystemprocesses. Rates of cycling of carbon (seeChapters 5 to 7) and water (see Chapter 4) are ultimately regulated by the availability ofbelowground resources, so changes in the avail-ability of these resources fundamentally alterall ecosystem processes. The combustion offossil fuels has released large quantities ofnitrogen and sulfur oxides to the atmosphereand increased their inputs to ecosystems (seeChapter 15). Fertilizer use and the cultivationof nitrogen-fixing crops have further increasedthe fluxes of nitrogen in agricultural ecosystems(Galloway et al. 1995, Vitousek et al. 1997a).Together these human impacts have doubledthe natural background rate of nitrogen inputsto the terrestrial biosphere. The resultingincreases in plant production may be largeenough to affect the global carbon cycle.Anthropogenic disturbances such as forest con-version, harvest, and fire increase the propor-tion of the nutrient pool that is available andtherefore vulnerable to loss from the ecosys-tem. Some of these losses occur by leaching ofdissolved elements to groundwater, causing adepletion of soil cations, an increase in soil

acidity, and increases in nutrient inputs toaquatic ecosystems. Gaseous losses of nitrogeninfluence the chemical and radiative propertiesof the atmosphere, causing air pollution andenhancing the greenhouse effect (see Chapter2). Changes in the cycling of nutrients thereforedramatically affect the interactions amongecosystems as well as the carbon cycle and theclimate of Earth.

Overview

Nutrient cycling involves the entry of nutrientsto ecosystems, their internal transfers betweenplants and soils, and their loss from ecosystems.Nutrients enter ecosystems through the chemi-cal weathering of rocks, the biological fixationof atmospheric nitrogen, and the deposition of nutrients from the atmosphere in rain,wind-blown particles, or gases. Fertilization isan additional nutrient input in managed eco-systems. Internal cycling processes include theconversion of nutrients from organic to inor-ganic forms, chemical reactions that change elements from one ionic form to another,biological uptake by plants and microorgan-isms, and exchange of nutrients on surfaceswithin the soil matrix. Nutrients are lost from

9Terrestrial Nutrient Cycling

Nutrient cycling involves nutrient inputs to, and outputs from, ecosystems and theinternal transfers of nutrients within ecosystems. This chapter describes these nutrient dynamics.

197

198 9. Terrestrial Nutrient Cycling

ecosystems by leaching, trace-gas emission,wind and water erosion, fire, and the removal ofmaterials in harvest.

Most of the nitrogen and phosphorusrequired for plant growth in unmanagedecosystems is supplied by the decomposition ofplant litter and soil organic matter. Inputs andoutputs in these ecosystems are a small fractionof the quantity of nutrients that cycle internally,producing relatively closed systems with con-servative nutrient cycles. Human activities tendto increase inputs and outputs relative to theinternal transfers and make the element cyclesmore open.

There are important differences among ele-ments in their patterns of biogeochemicalcycling. We approach this by first describing the cycling of nitrogen, the element that mostfrequently limits plant production; we thencompare its cycling to that of other elements.

Nitrogen Inputs to Ecosystems

Under natural conditions, nitrogen fixation isthe main pathway by which new nitrogen entersterrestrial ecosystems. Earth’s atmosphere contains an abundant well-distributed pool ofnitrogen; 78% of the atmosphere’s volume is dinitrogen (N2). Thus, although all organismsare literally bathed in nitrogen, it is the elementthat most frequently limits the growth of plantsand animals. This paradox occurs because N2

is unavailable to most organisms. Only certaintypes of bacteria, known as nitrogen fixers, havethe capacity to break the triple bonds of N2 andfix it into ammonium (NH4

+), which they use fortheir own growth.These bacteria occur either asfree-living forms in soils, sediments, and waters;or within nodules formed on the roots of certainvascular plants; or in lichens as symbiotic asso-ciations with fungi. Nitrogen fixed by nitrogen-fixing plants becomes available to other plantsin the community primarily through the pro-duction and decomposition of litter.

Biological Nitrogen Fixation

The characteristics of nitrogenase, the enzymethat catalyzes the reduction of N2 to NH4

+,

dictate much of the biology of nitrogen fixation.The reduction of N2 catalyzed by nitrogenasehas a high-energy requirement and thereforeoccurs only where the bacterium has an abun-dant carbohydrate supply. The enzyme is dena-tured in the presence of oxygen, so organismsmust protect the enzyme from contact withoxygen.

Groups of Nitrogen Fixers

Nitrogen-fixing bacteria in symbiotic associa-tion with plants have the highest rates of nitro-gen fixation. This occurs because plants canprovide the abundant carbohydrates needed tomeet the high energy demand of nitrogen fixa-tion. The most common symbiotic nitrogenfixers are Rhizobium species associated withlegumes (soybeans, peas, etc.) and Frankiaspecies (actinomycete bacteria) associated withalder, Ceanothus, and other nonlegume woodyspecies (Table 9.1).These plant-associated sym-biotic nitrogen-fixing bacteria usually reside inroot nodules, where the nitrogenase enzyme is protected from oxygen. In legumes, this isdone by leghemoglobin. This oxygen-bindingpigment is similar to hemoglobin, which trans-ports oxygen in the bloodstream of vertebrateanimals. Nitrogen-fixing bacteria in nodules areheterotrophic and depend on carbohydratesfrom plants to meet the energy requirements ofnitrogen fixation. The energetic requirementfor nitrogen fixation can be about 25% of thegross primary production (GPP) under labora-tory conditions, two to four times higher than the cost of acquiring nitrogen from soils(Lambers et al. 1998). The relative costs ofnitrogen fixation and nitrogen uptake underfield conditions are more difficult to estimatebecause of the uncertain costs of mycorrhizalassociation and root exudation.When inorganicnitrogen is naturally high or is added to soils,nitrogen-fixing plants generally reduce theircapacity for nitrogen fixation and acquire theirnitrogen from available forms in the soil solution.

Some heterotrophic nitrogen-fixing bacteriaare free living and get their organic carbonfrom the environment (Table 9.1). These bacte-ria have highest nitrogen-fixation rates in soils

Nitrogen Inputs to Ecosystems 199

or sediments with high concentrations oforganic matter, which provide the carbon sub-strate that fuels nitrogen reduction. Other heterotrophic nitrogen fixers occur in the rhizosphere and depend on root exudation androot turnover for their carbon supply. Het-erotrophic nitrogen fixers typically have highestfixation rates in aerobic environments, becauseaerobic respiration yields much more energyper gram of substrate than does anaerobic res-piration. Aerobic heterotrophs have variousmechanisms that reduce oxygen concentrationin the vicinity of nitrogenase. Some have highrespiration rates that scavenge oxygen aroundthe bacterial cells. Others clump together orproduce slime to reduce oxygen diffusion to theenzyme. Some heterotrophs fix nitrogen in low-oxygen situations and therefore require no spe-cialized adaptations to prevent denaturation of nitrogenase by oxygen. These organisms,however, have low rates of nitrogen fixationdue to low energy availability.

There are many free-living nitrogen-fixingphototrophs that produce their own organic

carbon by photosynthesis. These includecyanobacteria (blue-green algae) that occur inaquatic systems and on the surface of manysoils. Many phototrophs have specialized non-photosynthetic cells, called heterocysts, thatprotect nitrogenase from denaturation by theoxygen produced during photosynthesis inadjacent photosynthetic cells.

There are also associative (symbiotic) nitro-gen-fixing phototrophs. For example, nitrogen-fixing lichens are composed of green algae orcyanobacteria as the photosynthetic symbiont,cyanobacteria that fix nitrogen, and fungi thatprovide physical protection. These lichensprovide an important nitrogen input in manyearly successional ecosystems. The small fresh-water fern Azolla and cyanobacteria such asNostoc form a phototrophic association that iscommon in rice paddies and tropical aquaticsystems.

Legumes and other symbiotic nitrogen fixershave the highest rates of nitrogen fixation, typ-ically 5 to 20gm-2 yr-1. Phototrophic symbiontssuch as Nostoc in association with Azolla in rice

Type of associationa Key characteristics Representative genera

Heterotrophic nitrogen bacteriafixers

AssociativeNodulated (symbiotic) legume Rhizobium

nonlegume woody plants FrankiaNon-nodulated rhizosphere Azotobacter, Bacillus

phyllosphere KlebsiellaFree living aerobic Azotobacter, Rhizobium

facultative aerobic Bacillusanaerobic Clostridium

Phototrophic nitrogen cyanobacteriafixers

Associative lichens Nostoc, Calothrixliverworts (Marchantia) Nostocmosses Holosiphongymnosperms (Cycas) Nostocwater fern (Azolla) Nostoc

Free living cyanobacteria Nostoc, Anabaenapurple nonsulfur bacteria Rhodospirilliumsulfur bacteria Chromarium

a Nitrogen-fixing microbes are heterotrophic bacteria if they get their organiccarbon from the environment. They are phototrophic bluegreen algae if theyproduce it themselves through photosynthesis. Among both groups of microbes,some forms are typically associated with plants and others are free living. Note thatthe same microbial genus can have both associative and free-living forms.Data from Paul and Clark (1996).

Table 9.1. Organisms andassociations involved in dinitrogen fixation.


paddies often fix 10gm-2 yr-1. When Nostoc is afree-living phototroph, it fixes about 2.5gm-2

yr-1. In contrast, free-living heterotrophs fixonly 0.1 to 0.5gm-2 yr-1, a quantity similar to theinput from nitrogen deposition in unpollutedenvironments.

Causes of Variation in Nitrogen Fixation

Biotic and abiotic constraints on nitrogen fixa-tion lead to nitrogen limitation in most tem-perate terrestrial and many marine ecosystems.The rate of nitrogen fixation varies widelyamong ecosystems, in part reflecting the typesof nitrogen fixers that are present. Even withina single type of nitrogen-fixing system, how-ever, there are large ranges in fixation rates.What causes this variation? If nitrogen limitsgrowth in many ecosystems, why does nitrogenfixation not occur almost everywhere? Onewould expect nitrogen fixers to have a compet-itive advantage over other plants and microbesthat cannot fix their own nitrogen. Why don’tnitrogen fixers respond to nitrogen limitationby fixing nitrogen until nitrogen is no longerlimiting in the ecosystem? Several factors constrain nitrogen fixation, thereby maintain-ing nitrogen limitation in many ecosystems(Vitousek and Howarth 1991).

Energy availability constrains nitrogen fixa-tion rates in closed-canopy ecosystems. Thecost of nitrogen fixation (3 to 6g carbon g-1 N,not including the cost of nodule production) bysymbiotic and autotrophic nitrogen fixers ishigh relative to that of absorbing ammonium ornitrate. Nitrogen fixation is therefore largelyrestricted to the high-light environments thatoccur in early succession or where nutrients ordrought limit canopy development.As canopiesclose during succession, energy becomes limit-ing to the establishment of nitrogen-fixingplants. These plants could fix nitrogen if theywere in the canopy, but the cost of nitrogen fixation makes it difficult for them to growthrough shade to the canopy. Leguminous treesare common in tropical forests and savannas. Insavannas, where nitrogen cycles are relativelyopen due to large nitrogen losses in fires, legu-minous trees are heavily nodulated and fix substantial quantities of nitrogen (Högberg and

Alexander 1995). Leguminous trees in tropicalforests, however, are seldom nodulated andmay contribute little to the nitrogen inputs totropical forests. Nitrogen fixation in aquaticsystems is most common in shallow waters orwaters with low turbidity where light reachesbenthic cyanobacterial mats. When phosphorusavailability is adequate, these mats have highfixation rates.

Nonsymbiotic heterotrophic nitrogen-fixingbacteria are also limited by the availability oflabile organic carbon. When available carbon is scarce, there is no benefit to heterotrophicnitrogen fixation. Decaying wood, which haslow nitrogen and high levels of organic carbon,often has substantial rates of heterotrophicnitrogen fixation. Heterotrophic nitrogen fixa-tion also occurs in anaerobic sediments, but thegaseous loss of nitrogen by denitrification—that is the conversion of nitrate to gaseousforms, usually exceeds the gains from nitrogenfixation.

Nitrogen fixation in many ecosystems, espe-cially in the tropics, is limited by the availabil-ity of some other nutrient, such as phosphorus.Nitrogen fixers have a high requirement forATP and other phosphorus-containing com-pounds to support the energy transformationsassociated with nitrogen fixation.The growth ofnitrogen fixers therefore often becomes phos-phorus limited before that of other plants.Other elements that can limit nitrogen fixationinclude molybdenum, iron, and sulfur, whichare essential cofactors of nitrogenase. Pasturelegumes are limited by molybdenum in parts ofAustralia, for example, due to low molybdenumavailability in the highly weathered soils, par-ticularly at low pH. Nitrogen fixers may belimited by iron in marine ecosystems, at timeswhen other phytoplankton are nitrogen limited(see Chapter 14). Phosphorus, iron, sulfur, ormolybdenum may, in these cases, be the ulti-mate “master element” that limits production,even though nitrogen is the factor to whichprimary production responds most strongly inshort-term experiments.

Grazing of nitrogen-fixing organisms oftenconstrains their capacity to support continu-ously high nitrogen fixation rates. The highprotein content typical of nitrogen fixers

Nitrogen Inputs to Ecosystems 201

enhances their palatability to many herbivores,although nitrogen-based defenses such as alka-loids, which occur in many nitrogen-fixingplants, deter other herbivores (see Chapter 11).The resulting intense grazing on many nitrogen-fixing plants reduces their capacity tocompete with other plants, causing their abun-dance and nitrogen inputs to the ecosystem todecline. Areas from which grazers are excludedfrequently have more nitrogen-fixing plantsand greater nitrogen inputs to the ecosystemand ultimately more productivity and biomass(Ritchie et al. 1998).

Nitrogen Deposition

Nitrogen is deposited in ecosystems in particu-late, dissolved, and gaseous forms. All ecosys-tems receive nitrogen inputs from atmosphericdeposition. These inputs are smallest (often 0.2to 0.5gm-2 yr-1) in ecosystems downwind frompollution-free open oceans (Hedin et al. 1995).Nitrogen inputs to these coastal ecosystemsderive primarily from organic particulates andnitrate (NO3

-) in sea-spray evaporates andfrom ammonia (NH3) volatilized from seawa-ter. In inland areas, nitrogen derives from thevolatilization of NH3 from soils and vegetationand from dust produced by wind erosion ofdeserts, unplanted agricultural fields, and othersparsely vegetated ecosystems. Lightning alsoproduces nitrate, contributing to atmosphericdeposition.

Human activities are now the major sourceof nitrogen in deposition in many areas of theworld. The application of urea or ammonia fer-tilizer leads to volatilization of NH3, which isthen converted to ammonium (NH4

+) in theatmosphere and deposited in rainfall. Domesticanimal husbandry has also substantially in-creased emissions of NH3 to the atmosphere.The emission of nitric oxides (NO and NO2,together known as NOx) from fossil fuel com-bustion, biomass burning, and volatilizationfrom fertilized agricultural systems havedwarfed natural sources at the global scale:80% of all NOx flux is anthropogenic (Delmaset al. 1997). Nitrogen derived from thesesources can be transported long distancesdownwind from industrial or agricultural areas

before being deposited. “Arctic haze” over the Arctic Ocean and Canadian High Arcticislands, for example, derives primarily from pol-lutants produced in eastern Europe. Inputs ofanthropogenic sources of nitrogen to ecosys-tems can be quite large, for example, 1 to 2gm-2 yr-1 in the northeastern United States and 5to 10gm-2 yr-1 in central Europe—10- to 100-fold greater than background levels of nitrogendeposition. The highest rates are similar to theamounts annually absorbed by vegetation andcycled through litterfall (see Chapter 8). Mostecosystems have a substantial capacity to storeadded nitrogen in soils and vegetation. Oncethese reservoirs become saturated, however,nitrogen losses to the atmosphere and ground-water can be substantial. The nitrogen cycle insome polluted ecosystems has changed frombeing more than 90% closed (see Table 8.1) tobeing just as open as the carbon cycle, in whichthe amount of nitrogen or carbon annuallycycled by vegetation is similar to the amountthat is annually gained and lost from the ecosystem.

Climate and ecosystem structure determinethe processes by which nitrogen is deposited in ecosystems. Deposition occurs by threeprocesses. (1) Wet deposition delivers nutrientsdissolved in precipitation. (2) Dry depositiondelivers compounds as dust or aerosols by sed-imentation (vertical deposition) or impaction(horizontal deposition or direct absorption ofgases such as nitric acid, HNO3, vapor). (3)Cloud-water deposition delivers nutrients inwater droplets onto plant surfaces immersed infog. Although data are most available for wetdeposition because it is most easily measured,wet and dry deposition are often equally impor-tant sources of nitrogen inputs (Fig. 9.1). Wetdeposition of nitrogen is typically greater in wet than in dry ecosystems. Dry deposition ofnitrogen, however, shows no clear correlationwith climate, although arid ecosystems receivea larger proportion of their nitrogen inputs by dry deposition. Cloud-water deposition isgreatest on cloud-covered mountaintops orregions with coastal fog. The relative impor-tance of wet, dry, and cloud-water depositionalso depends on ecosystem structure. Conifercanopies, for example, tend to collect more dry


deposition and cloud-water deposition than dodeciduous canopies because of their greaterleaf surface area. Their rough canopies alsocause moisture-laden air to penetrate moredeeply within the forest canopy and thereforeto contact more leaf surfaces (see Chapter 4).

The form of nitrogen deposition determinesits ecosystem consequences. NO3

- and NH4+ are

immediately available for biological uptake byplants and microbes, whereas some organicnitrogen must first be mineralized. Nitrateinputs such as nitric acid (and ammoniuminputs, if followed by nitrification, the conver-sion of ammonium to nitrate) acidify the soilwhen nitrate accompanied by base cationsleaches from the ecosystem. Sulfuric acid,which often accompanies fossil-fuel sources ofNOx, also acidifies soils. Organic nitrogen com-pounds make up about a third of the totalnitrogen deposition, but their chemical naturevaries among ecosystems (Neff et al. 2002). Incoastal areas, for example, organic nitrogen isdeposited primarily as marine-derived reducedcompounds such as amines. In inland areasaffected by air pollution, most organic nitrogenenters as oxidized organic nitrogen compounds

that result from the reaction of organic com-pounds and NOx in the atmosphere.

Weathering of sedimentary rocks may con-tribute to the nitrogen budgets of some eco-systems. Sedimentary rocks, which make up75% of the rocks exposed on Earth’s surface,sometimes contain substantial nitrogen. In onewatershed underlain by high-nitrogen sedi-mentary rocks, for example, rock weatheringcontributed about 2gNm-2 yr-1(Holloway et al.1998), similar to the quantities that enteredfrom the atmosphere. In most ecosystems,however, rock weathering is thought to provideonly a small nitrogen input to ecosystems.

Internal Cycling of Nitrogen

Overview of Mineralization

In natural ecosystems, most nitrogen absorbedby plants becomes available through thedecomposition of organic matter. Most (morethan 99%) soil nitrogen is contained in deadorganic matter derived from plants, animals,and microbes. As microbes break down thisdead organic matter during decomposition (seeChapter 7), the nitrogen is released as dissolvedorganic nitrogen (DON) through the action ofexoenzymes (Fig. 9.2). Plants and mycorrhizalfungi can absorb some DON, using it to sup-port plant growth. Decomposer microbes alsoabsorb DON, using it to support their nitrogenand/or their carbon requirements for growth.When DON is insufficient to meet this nitrogenrequirement, microbes absorb additional inor-ganic nitrogen (NH4

+ or NO3-) from the soil

solution. Immobilization is the removal of inor-ganic nitrogen from the available pool by micro-bial uptake and chemical fixation. Studies using15N-labeled NH4

+ or NO3- indicate that microbes

take up and immobilize both forms of nitrogen,although NH4

+ uptake is typically greater(Vitousek and Matson 1988, Fenn et al. 1998).Microbial growth is often carbon limited. Underthese circumstances, microbes break down theDON, use the carbon skeleton to support theirenergy requirements for growth and mainte-nance,and secrete NH4

+ into the soil.This processis termed nitrogen mineralization or ammonifi-

0

0.4

0.8

1.2

1.6

2.0

NO

3--

N d

epos

ition

rat

e (g

m-2

yr-1

)

WF HF SC OR CW CD PN AR PW TH

= Wet = Cloud = Dry

Figure 9.1. Wet, dry, and cloud-water deposition ofnitrogen in a variety of ecosystems. WF, WhitefaceMountain, New York; HF, Huntington Forest, NewYork; SC, State College, Pennsylvania; OR, OakRidge, Tennessee; CW, Coweta, North Carolina;CD, Clingman’s Dome, North Carolina; PN, Panola,Georgia; AR, Argonne, Illinois; PW, Pawnee, Col-orado; and TH, Thompson, Washington. (Redrawnwith permission from Ecological Applications;Lovett 1994).

Internal Cycling of Nitrogen 203

cation because ammonium is the immediateproduct of this process. In some ecosystems,some or all NH4

+ is converted to nitrite (NO2-)

and then to nitrate (NO3-).The conversion from

ammonium to nitrate is termed nitrification.

Production and Fate of DissolvedOrganic Nitrogen

The conversion from insoluble organic nitrogento dissolved organic nitrogen is the initial and

typically the rate-limiting step in nitrogen min-eralization (Fig. 9.2). The relatively large poolof nonsoluble organic nitrogen in soils suggeststhat this initial step in nitrogen mineralizationis the rate-limiting step.All of the organic nitro-gen that is eventually mineralized to NH4

+ orNO3

- must first become dissolved before it canbe absorbed by microbes and mineralized. Theflux through the DON pool is therefore usuallylarge, relative to other nitrogen fluxes, even inecosystems in which its concentration is low

Microbes

Soil animals

NO3-

Nitrifiers

Gro

ssm

iner

aliz

atio

n

Den

itrifi

catio

nLe

achi

ng

microbes & exoenzymes

Dead

Imm

obili

zatio

n

Imm

obili

zatio

n

Nitrification

Dead organic matter

NH4+

N depositionN fixation

FireHarvestErosion

Dissolvedorganic nitrogen

Figure 9.2. Simplified terrestrial nitrogen cycle.Both plants and microbes take up dissolved organicnitrogen, NH4

+, and NO3- and release dead organic

matter and DON. Microbes also release ammoniumwhen they absorb more nitrogen than they requirefor growth. Nitrifiers are a specialized microbialgroup that either converts ammonium to nitrite or

nitrite to nitrate. Nitrogen is consumed by animalswhen they eat plants or soil microbes and is returnedto the soil as dead organic matter and DON. Nitro-gen is lost from the ecosystem by denitrification,erosion, harvest, or fire. Nitrogen enters the eco-system through nitrogen deposition or nitrogen fixation.


(Eviner and Chapin 1997). The breakdown ofparticulate organic nitrogen is carried out inparallel with the breakdown and use of partic-ulate organic carbon and is therefore controlledby the same organisms and factors that controldecomposition (see Fig. 7.14). These controlsinclude the quantity and chemical nature of the substrate, the composition of the microbialcommunity, and the environmental factors reg-ulating the activity of soil microbes and animals(see Chapter 7).

Most nitrogen in dead organic matter is con-tained in complex polymers such as proteins,nucleic acids, and chitin (from fungal cell wallsand insect exoskeletons) that are too large topass through microbial membranes. Microbesmust therefore secrete exoenzymes such as pro-teases, ribonucleases, and chitinases to breakdown the large polymers into small water-soluble subunits, such as amino acids andnucleotides that can be absorbed by microbialcells. Urease is another important exoenzymethat breaks down urea from animal urine or fertilizer into CO2 and NH3. The microbialenzymes are themselves subject to attack bymicrobial proteases, so microbes must continu-ally invest nitrogen in exoenzymes to acquirenitrogen from their environment, a potentiallycostly trade-off. Exoenzymes often bind to soilminerals and organic matter.This can inactivatethe enzyme, if the shape of the active site isaltered, or can protect the enzyme againstattack from other exoenzymes, lengthening thetime that the enzyme remains active in the soil(see Chapter 7). Proteases are produced bymycorrhizal and saprophytic fungi and by bacteria.

Plants, mycorrhizal fungi, or decomposermicrobes can absorb DON.When plants absorbDON directly or through their mycorrhizalfungi, no mineralization is required to providethis nitrogen to plants. Although we know thatdirect uptake of organic nitrogen by plantsoccurs in many ecosystems (Read 1991,Kielland 1994, Näsholm et al. 1998, Lipson et al.1999, Raab et al. 1999), we know little about therelative importance of this pathway vs. the fluxof nitrogen through mineralization. In manycases, microbes probably have a competitiveadvantage over plant roots because the

microbes producing exoenzymes are closest tothe site of enzymatic activity. In this case, weexpect nitrogen mineralization to be the majorfate of DON in soils. In organic soils, whereDON concentrations are high, plants competewell for amino acids (Schimel and Chapin 1996)and meet a significant proportion (about 65%in one study) of their nitrogen requirementthrough direct uptake of DON (Lipson et al.2001) (see Chapter 8). Even in agriculturalecosystems, 20% of plant nitrogen uptake canbe met by DON.

DON is a chemically complex mixture ofcompounds, only a few percent of which con-sists of amino acids and other labile forms of nitrogen. The labile DON that is absorbed by microbial cells can be incorporated directlyinto microbial proteins and amino acids ortransformed to other organic compounds thatsupport microbial growth and respiration.DON can also be adsorbed onto the soilexchange complex or leached from the ecosys-tem in groundwater. Amino acids contain bothpositively and negatively charged groups (NH2

+ and COO-, respectively). Small neutrallycharged amino acids, such as glycine, are most mobile in soils and are most readilyabsorbed by both plants and microbes (Kielland 1994).

Production and Fate of Ammonium

The net absorption or release of ammonium bymicrobes depends on their carbon status. Whenmicrobial growth is carbon limited, microbesuse the carbon from DON to support growthand respiration and secrete NH4

+ as a wasteproduct into the soil solution. This process ofammonification is the mechanism by whichnitrogen is mineralized. Other nitrogen-limitedmicrobes may absorb, or immobilize, some ofthis ammonium and use it for growth. Mineral-ization and microbial uptake occur simultane-ously in soils so a given unit of nitrogen cancycle between microbial release and uptakemany times before it is used by plants or under-goes some other fate. Gross mineralization isthe total amount of nitrogen released via mineralization (regardless of whether it is subsequently immobilized or not). Net miner-


alization is the net accumulation of inorganicnitrogen in the soil solution over a given timeinterval. Net mineralization occurs when micro-bial growth is limited more strongly by carbonthan by nitrogen, whereas net immobilizationoccurs when microbial communities are nitro-gen limited.

Net nitrogen mineralization is an excellentmeasure of the nitrogen supply to plants inecosystems with rapid nitrogen turnover, wherethere is little competition for nitrogen betweenplants and microbes. The annual net mineral-ization in the deciduous forests of easternNorth America, for example, approximatelyequals nitrogen uptake by vegetation (Nadelhoffer et al. 1992). In less fertile ecosys-tems, such as arctic tundra, net nitrogen miner-alization rate substantially underestimates theamount of nitrogen that is annually acquired byplants (Nadelhoffer et al. 1992). There are atleast two explanations for this apparent dis-crepancy. (1) Plant roots and their mycorrhizalfungi, which are excluded in net mineralizationassays, may be good competitors with sapro-

phytic microbes for mineralized nitrogen in thereal world, so this assay may underestimate thequantity of nitrogen that would be mineralizedin the presence of roots (Stark 2000). (2) Plantsand their mycorrhizal fungi that absorb aminoacids may also not depend heavily on nitrogenmineralization to meet their nitrogen require-ments in infertile ecosystems due to absorptionof DON, so the low rates of net nitrogen min-eralization in these ecosystems may be an accurate reflection of small fluxes that naturallyoccur through this pathway.

Nitrogen mineralization rate is controlled bythe availability of DON and inorganic nitrogen,the activity of soil microbes, and their relativedemands for carbon and nitrogen. The quantityand quality of organic matter, including its relative amounts and forms of carbon andnitrogen, that enter the soil are the major determinants of the substrate available fordecomposition (see Chapter 7) and thereforethe substrate available for nitrogen mineraliza-tion (Fig. 9.3). Thus the state factors and inter-active controls that promote productivity and

LONG-TERMCONTROLS

SHORT-TERMCONTROLS

Plantfunctional

types

Soilresources

Microbial C:N

TemperatureH2O

NET NITROGENMINERALIZATION

(AMMONIFICATION)

BIOTA

PARENTMATERIAL

Climate

TIME

Dissolvedorganicnitrogen

Litterquantity

Carbon quality

Plantuptake of

DON

STATE FACTORS

Interactivecontrols

Directcontrols

Indirectcontrols

(-)

(-)

Figure 9.3. The major factors controlling ammonifi-cation (net nitrogen mineralization) in soils. Thesecontrols range from the proximate control overnitrogen mineralization (the concentration of DON,physical environment, and microbial carbon to nitrogen ratio) to the state factors and interactive

controls that ultimately determine the differencesamong ecosystems in mineralization rates. Thicknessof the arrows indicates the strength of effects. Theinfluence of one factor on another is positive unlessotherwise indicated (-).


high litter quality (see Fig. 7.14) also promotenitrogen mineralization.

Environmental conditions that promotemicrobial activity enhance both gross and netnitrogen mineralization. Net nitrogen mineral-ization rates are therefore generally higher intropical than in temperate forest soils, whereasarctic soils show net nitrogen immobilization (a net decrease in the concentrations of in-organic nitrogen) during the growing season(Nadelhoffer et al. 1992). Even within a biome,factors that improve the soil temperature andmoisture environment for microbial activityenhance net nitrogen mineralization. Acrossthe Great Plains of the United States, forexample, rates of mineralization are positivelyrelated to precipitation and temperature.Recently deforested areas also typically havehigher rates of net nitrogen mineralization than do undisturbed forests, at least in part dueto warmer, moister soils (Matson and Vitousek1981). Soil moisture that is high enough torestrict microbial activity also restricts netnitrogen mineralization.

Why do favorable litter quality, moisture, andtemperature lead to net nitrogen mineraliza-tion rather than immobilization of nitrogen ina growing microbial biomass? Several factorscontribute to this pattern. First, the limitationof microbial activity by carbon availability,particularly in environments where nitrogen isreadily available and litter nitrogen is relatively

high, causes microbes to use some DON tomeet their carbon requirements for growth and maintenance, secreting the ammonium as a waste product. Second, warm temperaturesincrease maintenance respiration and thereforethe carbon demands for microbial activity.Finally, increases in microbial productivitypromote predation by soil animals, causinggreater microbial turnover and release of nitro-gen to the soil.

Substrate quality influences nitrogen miner-alization rate, not only through its effects oncarbon quality, which governs decompositionrate (see Chapter 7) but also through its effectson the balance between carbon and nitrogenlimitation of microbial growth. The carbon tonitrogen (C :N) ratio in microbial biomass isabout 10 :1. As microbes break down organicmatter, they incorporate about 40% of thecarbon from their substrates into microbialbiomass and return the remaining 60% of thecarbon to the atmosphere as CO2 throughrespiration. With this 40% growth efficiency,microbes require substrates with a C :N ratio ofabout 25 :1 to meet their nitrogen requirement(Box 9.1). At higher C :N ratios, microbesimport nitrogen to meet their growth require-ments, and at lower C :N ratios nitrogenexceeds microbial growth requirements and issecreted into the litter and soil. In practice,microbes vary in their C :N ratio (5 to 10 in bac-teria and 8 to 15 in fungi) and in their growth

The critical C :N ratio that marks the divid-ing line between net nitrogen mineralizationand net nitrogen uptake by microbes can becalculated from the growth efficiency ofmicrobial populations and the C :N ratios ofthe microbial biomass and their substrate.Assume, for example, that the microbialbiomass has a growth efficiency of 40% and aC:N ratio of 10 :1. If the microbes breakdown 100 units of carbon, they will incorpo-rate 40 units of carbon into microbial

biomass and respire 60 units of carbon asCO2.The 40 units of microbial carbon require4 units of nitrogen to produce a microbial C :N ratio of 10 :1 (= 40 :4). If the 100 units oforiginal substrate is to supply all of this nitro-gen, its initial C :N ratio must have been 25 :1 (= 100 :4). At higher C :N ratios,microbes must absorb additional inorganicnitrogen from the soil to meet their growthdemands. At lower C :N ratios, microbesexcrete excess nitrogen into the soil.

Box 9.1. Estimation of Critical C :N Ratio for Net Nitrogen Mineralization


efficiency. Bacteria typically have a lowergrowth efficiency than do fungi. All microbesconvert substrates into biomass less efficientlywith less labile substrates or with greater envi-ronmental stress. Nonetheless, 25 :1 is oftenconsidered the critical C :N ratio above whichthere is no net nitrogen release from decom-posing organic matter. Note that there is a clearmechanistic effect of C :N ratio on the netimmobilization or mineralization of nitrogen,whereas its effect on decomposition (seeChapter 7) is much less certain.The growth effi-ciency of microbes (about 40%) is similar tothat of plants (net primary production averagesabout 47% of GPP; see Chapter 6), despitequite different mechanisms of acquiring carbonand nitrogen from the environment. This simi-larity may reflect a common underlying bio-chemistry of costs of synthesis and costs ofmaintenance.

The ammonium produced by nitrogen min-eralization has several potential fates. In addi-tion to being absorbed by plants or microbes,ammonium readily adsorbs to the negativelycharged surfaces of soil minerals and organicmatter (see Chapter 3), resulting in relativelylow concentrations of NH4

+ (often less than 1ppm) in the soil solution. When plant andmicrobial uptake depletes NH4

+ from the soilsolution, this shifts the equilibrium between dis-solved and exchangeable pools, and adsorbedions go back into solution from the exchangecomplex. The cation exchange complex thusserves as a storage reservoir of readily availableNH4

+ and other cations. NH4+ can also be fixed

in the interlayer portions of certain alumino-silicate clays or complexed with stabilized soilorganic matter, making it less available toplants or microbes. As long as the organiccomplex remains protected, the NH4

+ in thecomplex is unavailable to plants and microbes.Finally, NH4

+ can be volatilized to ammonia gas(NH3) or oxidized, mainly by bacteria, to NO2

-

and NO3-.

Production and Fate of Nitrate

Nitrification is the process by which NH4+ is

oxidized to NO2- and subsequently to NO3

-.Unlike ammonification, which is carried out by

a broad suite of decomposers, most nitrificationis carried out by a restricted group of nitrifyingbacteria. There are two general classes of nitri-fiers. Autotrophic nitrifiers use the energy yieldfrom NH4

+ oxidation to fix carbon used ingrowth and maintenance, analogous to the wayplants use solar energy to fix carbon via photo-synthesis. Heterotrophic nitrifiers gain theirenergy from breakdown of organic matter.

Autotrophic nitrifiers include two groups,one that converts ammonium to nitrite (e.g.,Nitrosolobus and other “nitroso-” genera) andanother that converts nitrite to nitrate (e.g.,Nitrobacter and other “nitro-” genera). Theseautotrophic nitrifiers are obligate aerobes thatsynthesize structural and metabolic carboncompounds by reducing CO2 and using energyfrom NH4

+ or NO2- oxidation to drive CO2 fix-

ation. In most systems, these two groups occurtogether, so NO2

- typically does not accumulatein soils. NO2

- is most likely to accumulate in dry forest and savanna ecosystems during thedry season, when the activity of Nitrobacter isrestricted, and in some fertilized ecosystems,where nitrogen inputs are high relative to plantand microbial demands.

Although autotrophic nitrification predomi-nates in many ecosystems, heterotrophic nitrifi-cation can be important in ecosystems with lownitrogen availability or acidic soils. Many het-erotrophic fungi and bacteria, including actino-mycetes, produce NO2

- or NO3- from NH4

+.Some also use organic nitrogen in the process.Because heterotrophs obtain their energy fromorganic materials, it is not clear what advantagethey gain from the NO3

- oxidation process.Nitrification has multiple effects on ecosys-

tem processes. The oxidation of NH4+ to NO2

-,which occurs in the first step of nitrification,produces 2 moles of H+ for each mole of NH4

+

consumed and therefore tends to acidify soils.The mono-oxygenase that catalyzes this stephas a broad substrate specificity and also oxi-dizes many chlorinated hydrocarbons, suggest-ing a role of nitrifiers in the breakdown ofpesticide residues. Finally, nitric oxide (NO)and nitrous oxide (N2O), which are producedduring nitrification (Fig. 9.4), are gases thathave important effects on atmospheric chemistry.


The availability of NH4+ is the most important

direct determinant of nitrification rate (Fig. 9.5)(Robertson 1989). The NH4

+ concentrationmust be high enough, at least in certain soilmicrosites, to allow nitrifiers to compete withother soil microbes. This is particularly impor-tant for autotrophic nitrifiers, which rely onNH4

+ as their sole energy source. NH4+ supply, in

turn, is regulated by the effects of substratequality and environment on ammonificationrate (Fig. 9.3). Fertilizer inputs and ammoniumdeposition are additional sources of ammoniumto many ecosystems. Conversely, plant rootsreduce NH4

+ concentration in the soil solution,thereby competing with nitrifiers for NH4

+.Many productive ecosystems have high nitrifi-cation rates despite low average NH4

+ concen-trations in the bulk soil, perhaps because ofspatial heterogeneity in NH4

+ concentration.Nitrification rates can also be substantial, evenwhen NO3

- concentrations in soils are low,because nitrate is relatively mobile and can beabsorbed by plants or microbes, leached fromsoils, and denitrified as rapidly as it is produced.

Nitrifier populations are often too small ininfertile soils to support significant nitrification.When ammonium substrate becomes available(e.g., through additions of nitrogen, or increasesin mineralization rates), nitrifier populationscan increase in size, and nitrification rates canincrease. The response can be rapid in somesoils but show a long delay in others (Vitouseket al. 1982). Secondary metabolites have beenhypothesized to inhibit nitrification in someecosystems, including those in late succession(Rice 1979), but the decline in nitrification inlate succession is generally best explained by a decline in ammonium supply rather than astoxicity to nitrifiers (Pastor et al. 1984, Schimelet al. 1996). Limitation of nitrifiers by otherresources is another possible cause of slow or delayed nitrification. In most cases, how-ever, the availability of ammonium ultimatelygoverns nitrification rate through its effects onboth the population density and activity ofnitrifying bacteria.

Oxygen is an important additional factorcontrolling nitrification because most nitrifiers

Process Reactions Processes Optimalenvironments

DON Ammonification(= N mineralization)

Warm,moist

NH4+

NO2-

NO2-

NO3-

N2O

N2O

N2

NO

NO

High O2

DenitrificationLow O2

High C

Assimilatorynitrate reduction(High C, low N)

NO

N2

N2O

Nitrification

Figure 9.4. Pathways of autotrophic nitrification and of denitrification and the nitrogen trace gases emittedby these pathways (Firestone and Davidson 1989).


require oxygen for oxidation of NH4. Oxygenavailability, in turn, is influenced by manyfactors, including soil moisture, soil texture, soilstructure, and respiration by microbes androots (Fig. 9.5).

Nitrifier activity is sensitive to temperature.It does, however, continue at low rates at lowtemperatures, so over a long winter season,substantial nitrification can occur, particularlyin nitrogen-rich agricultural soils. Nitrificationrates are slow in dry soils primarily because thinwater films restrict NH4

+ diffusion to nitrifiers(Stark and Firestone 1995). Under extremelydry conditions, low water potential furtherrestricts the activity of nitrifiers. The impor-tance of acidity in regulating nitrification ratesis uncertain. In laboratory cultures of agricul-tural soils, maximum nitrification rates occurbetween pH 6.6 and 8.0 and are negligiblebelow pH 4.5 (Paul and Clark 1996). Manynatural ecosystems with acidic soils, however,have substantial nitrification rates, even at pH4 (Stark and Hart 1997).

The fraction of mineralized nitrogen that isoxidized to nitrate varies widely among ecosys-

tems. In many unpolluted temperate coniferousand deciduous forests, nitrification is only asmall proportion of net mineralization (e.g., 0to 4%) but, as ecosystems receive increasingnitrogen deposition, the fraction of nitrificationcan increase to 25% (McNulty et al. 1990). Intropical forests, in contrast, net nitrification istypically nearly 100% of net mineralization,even in sites with low rates of net mineraliza-tion and without inputs of additional nitrogen(Vitousek and Matson 1988) (Fig. 9.6). In tropical ecosystems, plant and microbial growth are frequently limited by nutrientsother than nitrogen, and their demand fornitrogen is low, so nitrifiers have ready accessto NH4

+.The potential fates of nitrate are absorption

by plants and microbes, exchange on anionexchange sites, and loss from ecosystems viadenitrification or leaching. Because nitrate isrelatively mobile in soil solutions, it readilymoves to plant roots by mass flow or diffusion(see Chapter 8) or can be leached from the soil.Microbes also absorb nitrate and use it for synthesis of amino acids through assimilatory

LONG-TERMCONTROLS

SHORT-TERMCONTROLS

BIOTA

PARENTMATERIAL

CLIMATE

Plantfunctional

types

Soilresources

Litter quantity

Carbon quality

OxygenconcentrationTemperature

H2O

NITRIFICATION

TIME

Root / microbialrespiration

Soil texture

Ammoniumconcentration

Plant NH4+

uptake(-)

(-)

(-)

STATE FACTORS

Interactivecontrols

Directcontrols

Indirectcontrols

Figure 9.5. The major factors controlling nitrifica-tion in soils (Robertson 1989). These controls rangefrom concentrations of reactants that directly con-trol nitrification to the interactive controls, such asclimate and disturbance regime, that are the ultimate

determinants of nitrification rate. Thickness of thearrows indicates the strength of effects.The influenceof one factor on another is positive unless otherwiseindicated (-).


nitrate reduction (Fig. 9.4).This process is ener-getically expensive and occurs primarily whenmicrobes are nitrogen limited but have an ade-quate energy supply. The low nitrate concen-trations observed in many acidic conifer forestsoils reflect a combination of low nitrificationrates and nitrate absorption by soil microbesand plants (Stark and Hart 1997).

Although NO3- is more mobile than most

cations, it can be held on exchange sites of soilswith a high anion exchange capacity (seeChapter 3). All mineral soils have a variablecharge that depends on pH. In the layered-claysilicate minerals typical of the temperate zone,the zero point of charge (below which pH thecharge is positive and above which it is nega-tive) is typically near pH 2, well below the pHof most soils. In some soils, especially those inthe tropics, iron and aluminum oxide mineralshave a positive surface charge at their typicalpH. In these soils, there is sufficient anionexchange capacity to prevent leaching losses ofnitrate after disturbance (Matson et al. 1987).In most soils, the strength of the anion adsorp-

tion is PO43- > SO4

3- > Cl- > NO3-, so NO3

- isdesorbed relatively easily.

Temporal and Spatial Variability

Fine-scale ecological controls cause large tem-poral and spatial variability in nitrogen cycling.Nitrogen transformation rates in soils are noto-riously variable, with rates often differing by an order of magnitude between adjacent soilsamples or sampling dates (Robertson et al.1997). This variability reflects the fine temporaland spatial scales over which controlling factorsvary. Anaerobic conditions that support deni-trification (discussed later) in the interiors ofsoil aggregates, for example, can occur withinmillimeters of aerobic soil pores. Fine rootscreate zones of rhizosphere soils with highcarbon and low soluble nitrogen concentrationsadjacent to bulk soil, where carbon-limited soil microbes mineralize organic nitrogen tomeet their energy demands. In densely rootedmicrosites, plants deplete concentrations ofNH4

+ below levels that can sustain nitrification,whereas nitrification can be substantial in adja-cent non-root microsites. The effects of thisfine-scale spatial heterogeneity on nitrogencycling are difficult to study, so we know onlyqualitatively of their importance.

Temporal variability in environment andextreme events have a strong influence onnitrogen mineralization. Drying–wetting andfreeze–thaw events, for example, burst manymicrobial cells and release pulses of nutrients.For this reason, the first rains after a long dry-season often cause a pulse of nitrification andnitrate leaching (Davidson et al. 1993). Thespring runoff after snowmelt in northern ormountain ecosystems also frequently carrieswith it a pulse of nutrient loss to streamsbecause of both freeze–thaw events and theabsence of plant uptake of nitrogen duringwinter. For example, 90% of the annual nitro-gen input to Toolik Lake in arctic Alaska,occurs in the first 10 days of snowmelt (Whalenand Cornwell 1985).

The seasonality of nitrogen mineralizationoften differs from the seasonality of plant nitro-gen uptake. In ecosystems in which plants aredormant for part of the year, soil microbes

Net

nitr

ifica

tion

(µg

nitr

ogen

g-1

)

Net nitrogen mineralization (µg nitrogen g-1)

-10-10

0

0

10

10

20

20

30

30

40

40

50

50

60

60

70

70

Figure 9.6. The relationship between net nitrogenmineralization and net nitrification (per gram of dry soil for a 10-day incubation) across a range oftropical forest ecosystems (Vitousek and Matson1984). Nearly all nitrogen that is mineralized in thesesystems is immediately nitrified. In contrast, nitrifi-cation is frequently less than 25% of net mineraliza-tion in temperate ecosystems.

Pathways of Nitrogen Loss 211

continue to mineralize nitrogen during thedormant season; this leads to an accumulationof available nitrogen that plants use when they become active. In temperate forests, forexample, mineralization during winter (evenbeneath a snowpack) creates a substantial poolof available nitrogen that is not absorbed byplants until the following spring. This asyn-chrony between microbial activity and plantuptake is particularly important in low-nutrientenvironments, where microbes may immobilizenitrogen during the season of most active plantgrowth, effectively competing with plants fornitrogen (Jaeger et al. 1999b). In soils thatfreeze or dry, the death of microbial cells pro-vides additional labile substrates that supportnet mineralization by the remaining microbeswhen conditions again become suitable formicrobial activity. Plant storage of nutrients tosupport spring growth is particularly importantin low-fertility ecosystems, because nutrientsare not dependably available from the soil attimes when the environment favors plantgrowth (Chapin et al. 1990).

Pathways of Nitrogen Loss

Gaseous Losses of Nitrogen

Ammonia volatilization, nitrification, and den-itrification are the major avenues of gaseousnitrogen loss from ecosystems. These processesrelease nitrogen as ammonia gas, nitrous oxide,nitric oxide, and dinitrogen. Gas fluxes are con-trolled by the rates of soil processes and by soiland environmental characteristics that regulatediffusion rates through soils. Once in the atmos-phere, these gases can be chemically modifiedand deposited downwind.

Ecological Controls

Ammonia gas (NH3) can be emitted from soilsand scenescing leaves. In soils, it is emitted as aconsequence of the pH-dependent equilibriumbetween NH4

+ and NH3. At pH values greaterthan 7, a significant fraction of NH4

+ is con-verted to NH3 gas.

NH4+ + OH- ´ NH3 + H2O (9.1)

Ammonia then diffuses from the soil to theatmosphere. This diffusion is most rapid incoarse dry soils with large air spaces. In densecanopies, some of the NH3 emitted from soils isabsorbed by plant leaves and incorporated intoamino acids.

NH3 flux is low from most ecosystemsbecause NH4

+ is maintained at low concentra-tions by plant and microbial uptake and bybinding to the soil exchange complex. NH3

fluxes are substantial, however, in ecosystems inwhich NH4

+ accumulates due to large nitrogeninputs. In grazed ecosystems, for example, urinepatches dominate the aerial flux of NH3.Agricultural fields that are fertilized withammonium-based fertilizers or urea often lose20 to 30% of the added nitrogen as NH3, espe-cially if fertilizers are placed on the surface.Nitrogen-rich basic soils are particularly proneto NH3 volatilization because of the pH effecton the equilibrium between NH4

+ and NH3.Leaves also emit NH3 during senescence,when nitrogen-containing compounds arebroken down for transport to storage organs.Fertilization and domestic animal husbandryhave substantially increased the flux of NH3 tothe atmosphere (see Chapter 15).

The production of NO and N2O during nitri-fication depends primarily on the rate of nitri-fication. The conversion of NH4

+ to NO3- by

nitrification produces some NO and N2O as by-products (Fig. 9.4), typically at a NO to N2Oratio of 10 to 20. The quantities of NO and N2Oreleased during nitrification are correlated withthe total flux through the nitrification pathway,suggesting that nitrification acts like a leakypipe (Firestone and Davidson 1989), in which a small proportion (perhaps 0.1 to 10%) of the nitrogen “leaks out” as trace gases duringnitrification.

The reduction of nitrate or nitrite to gaseousnitrogen by denitrification occurs under condi-tions of high nitrate and low oxygen. Manytypes of bacteria contribute to biological deni-trification. They use NO3

- or NO2- as an elec-

tron acceptor to oxidize organic carbon forenergy when oxygen concentration is low. Mostdenitrifiers are facultative anaerobes and useoxygen rather than NO3

-, when oxygen is avail-able. In addition to biological denitrification,


chemodenitrification converts NO2- (nitrite)

abiotically to nitric oxide gas (NO) where NO2-

accumulates in the soil at low pH. Chemodeni-trification is typically much less important thanbiological denitrification.

The sequence of NO3- reduction is NO3

- ÆNO2

- Æ NO Æ N2O Æ N2. The last three prod-ucts, particularly N2O and N2, are released asgases to the atmosphere (Fig. 9.4). Most deni-trifiers have the enzymatic potential to carryout the entire reductive sequence but producevariable proportions of N2O and N2, dependingin part on the relative availability of oxidant(NO3

-) versus reductant (organic carbon).When NO3

- is relatively more abundant thanlabile organic carbon, more N2O than N2 is pro-duced. Other factors that favor N2O over N2

production include low pH, low temperature,and high oxygen. Although NO is oftenreleased during denitrification in laboratoryincubations, this seldom occurs in naturebecause its diffusion to the air is impeded by water-filled pore spaces. Some of the NOthat is produced serves as a substrate for

further reduction to N2O or N2 by denitrifying bacteria.

The three conditions required for significantdenitrification are low oxygen, high nitrate con-centration, and a supply of organic carbon (Fig.9.7) (Del Grosso et al. 2000). In most non-flooded soils, oxygen availability exerts thestrongest control over denitrification. Oxygensupply is reduced by high soil water content,which impedes the diffusion of oxygen throughsoil pores. Soil moisture, in turn, is controlled by other environmental factors such as slopeposition, soil texture, and the balance betweenprecipitation and evapotranspiration. Soiloxygen concentration is also sensitive to its rate of consumption by soil microbes and roots.It is consumed most quickly in warm, moistenvironments.

The second major control over denitrifica-tion is an adequate supply of the substrate NO3

-. Because nitrification is a primarily anaerobic process, the low-oxygen conditions thatare optimal for denitrification frequently limitNO3

- supply. Some wetlands, for example, have

STATE FACTORS

Interactivecontrols

DIRECTCONTROLS

LONG-TERMCONTROLS

SHORT-TERMCONTROLS

BIOTA

PARENTMATERIAL

CLIMATE

Plantfunctional

types

Soilresources

Litter quantity

Carbon quality

Oxygenconcentration

TemperatureH2O

DENITRIFICATION

Indirectcontrols

TIME

Root / microbialrespiration

Soil texture

Nitrateconcentration

Plant NO3-

uptake(-)

(-) (-)

(-)

Labilecarbon

Figure 9.7. The major factors controlling denitrifi-cation in soils. These controls range from concentra-tions of substrates that directly control nitrificationto the interactive controls such as climate and dis-turbance regime that are the ultimate determinants

of denitrification rate. Thickness of the arrows indicates the strength of effects. The influence of one factor on another is positive unless otherwiseindicated (-).

Pathways of Nitrogen Loss 213

low denitrification rates despite their saturatedsoils and large quantities of organic matter dueto low availability of nitrate. Wetlands supporthigh denitrification rates only if (1) they receiveNO3

- from outside the system (lateral transfer);(2) they have an aerobic zone above an anaer-obic zone (vertical transfer), as in partiallydrained wetlands; or (3) they go through cyclesof flooding and drainage (temporal separation),as in many rice paddies. At a finer scale, deni-trification can occur within soil aggregates orother anaerobic microsites (e.g., pieces of soilorganic matter) in moderately well drainedsoils due to fine-scale heterogeneity in soiloxygen concentration and nitrification rate.Finally, the availability of organic carbon sub-strates can limit denitrification because theprocess is carried out primarily by het-erotrophic bacteria. Long-term cultivation ofagricultural soils, for example, can reduce soilorganic matter concentrations sufficiently tolimit denitrification.

Fires also account for large gaseous losses ofnitrogen. The amount and forms of nitrogenvolatilized during fire depend on the tempera-ture of the fire. Fires with active flames produceconsiderable turbulence, are well supplied withoxygen, and release nitrogen primarily as NOx.Smoldering fires release nitrogen in morereduced forms, such as ammonia (Goode et al.2000). About a third of the nitrogen is emittedas N2. Fire is an important part of the nitrogencycle of many ecosystems. Fire suppression insome areas and biomass burning in others havealtered the natural patterns of nitrogen cyclingin many ecosystems.

Atmospheric Roles of Nitrogen Gases

The four nitrogen gases have different rolesand consequences for the atmosphere. NH3 thatenters the atmosphere reacts with acids andthus neutralizes atmospheric acidity.

NH3 + H2SO4 Æ (NH4)2SO4 (9.2)

In this reaction, NH3 is converted back to NH4

+, which can be deposited downwind on thesurface of dry particles or as NH4

+ dissolvedin precipitation. Ammonia volatilization and

deposition transfer nitrogen from one ecosys-tem to another.

In the atmosphere, the nitrogen oxides are inequilibrium with one another due to their rapidinterconversion. NOx is very reactive, and itsconcentration regulates several importantatmospheric chemical reactions. At high NOx

concentrations, for example, carbon monoxide(or methane and nonmethane hydrocarbons)are oxidized, producing tropospheric ozone(O3), an important component of photochemi-cal smog in urban, industrial, and agriculturalareas.

CO + 2O2 Æ CO2 + O3 (9.3)

When NOx concentrations are low, the oxida-tion of CO consumes O3.

CO + O3 Æ CO2 + O2 (9.4)

In addition to its role as a catalyst that altersatmospheric chemistry and generates pollution,NOx can be transported long distances and alterthe functioning of ecosystems downwind. In theform of nitric acid, it is a principle componentof acid deposition and adds both availablenitrogen and acidity to the soil. In its gaseousNO2 form, it can be absorbed through thestomata of leaves and be used in metabolism(see Chapter 5). It can also be deposited in par-ticulate form, another type of inadvertent fertilization.

In contrast to the highly reactive NOx, nitrousoxide (N2O) has an atmospheric lifetime of 150years and is not chemically reactive in the tro-posphere.The low reactivity of N2O contributesto a different environmental problem. N2O is agreenhouse gas that is 200 times more effectiveper molecule than is CO2 in absorbing infraredradiation (see Chapter 2). In addition, N2O inthe stratosphere reacts with excited oxygen inpresence of ultraviolet radiation to produceNO, which catalyzes the destruction of stratos-pheric ozone (O3).

Given that the atmosphere is already 78%N2, dinitrogen emissions to the atmosphere viadenitrification have no significant atmosphericeffects, although these losses may influenceecosystem nitrogen pools. Atmospheric N2 hasa turnover time of thousands of years.


Solution Losses

Nitrate and dissolved organic nitrogen accountfor most of the solution loss of nitrogen fromecosystems. Undisturbed ecosystems thatreceive low atmospheric inputs generally loserelatively little nitrogen, and these small lossesoccur primarily as dissolved organic nitrogen(Hedin et al. 1995). Although nitrate is alsohighly mobile, plants and microbes absorb mostnitrate before it leaches below the rooting zoneof intact ecosystems. Disturbance, however,often improves the environment for mineral-ization by increasing soil moisture and temper-ature and reduces the biomass of vegetationavailable to absorb nutrients (see Chapter 12).At the Hubbard Brook Forest in the north-eastern United States, for example, all vegeta-tion was removed from an experimentalwatershed to examine the consequences ofdevegetation.There were large losses of nitrate,calcium, and potassium to the groundwater andstreams when vegetation regrowth was pre-vented (Bormann and Likens 1979) (Fig. 9.8).

Once vegetation began to regrow, however, theaccumulating plant biomass absorbed most ofthe mineralized nutrients, and stream nutrientconcentrations returned to their preharvestlevels. Additions of fertilizer nitrogen or nitro-gen deposition that exceed plant and microbialnitrogen demands also increase nitrate leach-ing. Increased nitrate leaching is one of thecharacteristics of nitrogen saturation, thechanges that occur in ecosystem functioningwhen anthropogenic nitrogen additions relievenitrogen limitation to plants and microbes(Aber et al. 1998). Anthropogenic nitrogeninputs are generally correlated with nitrogenoutputs via leaching (Tietema and Beier 1995,Fenn et al. 1998) (Fig. 9.9).

Nitrate loss to groundwater can have impor-tant consequences for human health and for the ecological integrity of aquatic ecosystems.Nitrite, which forms from nitrate under reduc-ing conditions, can reduce the capacity ofhemoglobin in animals to transport oxygen,producing anemia, especially in infants.Groundwater in areas of intensive agriculture

0

10

20

30

40

0

1

2

3

6

9

40

60

1964 1966 1968 1970 1972 1974 1976

Nut

rient

loss

(g m

-2)

60 yr old forest

Reforestationprevented Recovery

Calcium

Potassium

Nitrate

Particulate matter

0

3

20

0

Deforested

Control

Figure 9.8. Losses of calcium, potassium,nitrate, and particulate organic matter instream water before and after deforestationof an experimental watershed at HubbardBrook Forest in the northeastern UnitedStates. The shaded area shows the time inter-val during which vegetation was absent due tocutting of trees and herbicide application.(Redrawn with permission from Springer-Verlag; Bormann and Likens 1979.)

Other Element Cycles 215

often has nitrate concentrations that exceedpublic health standards.

Nitrogen leached from terrestrial ecosystemsmoves in groundwater to lakes and rivers. Themovement of nitrate to the North AtlanticOcean from major rivers has increased 6- to 20-fold in the past century (Howarth et al. 1996),primarily due to increased inputs of fertilizer,atmospheric deposition, nitrogen fixation bycrops, and food imports (see Chapter 15).Nitrate in coastal marine systems frequentlyincreases productivity and detritus accumula-tion. The resulting stimulation of decomposi-tion can reduce oxygen concentrationssufficiently to kill fish, particularly in winter,when primary production is temperaturelimited (see Chapter 14). The nitrogen loadingfrom agricultural and urban systems in the Mississippi drainage, for example, has pro-duced a “dead zone” where this river enters theGulf of Mexico (Mitsch et al. 2001).

Solutions that move through the soil mustmaintain a balanced charge, with negativelycharged ions like nitrate balanced by cations.Therefore, every nitrate ion that leaches fromsoil carries with it a cation such as calcium,

potassium, and ammonium to maintain chargebalance. When cation loss by leaching exceedsthe rate of cation supply by weathering plusdeposition, the net loss of cations can lead tocation deficiency (Driscoll et al. 2001). Afterthese nutrient cations are depleted, nitratetakes with it H+ and/or Al3+, which are deleteri-ous to downstream ecosystems. Nitrificationalso generates acidity:

2NH4+ + 3O2 Æ 2NO2

- + H2O + H+ (9.5)

The hydrogen ion released in this reactionexchanges with other ions on cation exchangesites in the soil, making these cations more vul-nerable to leaching loss.

Erosional Losses

Erosion is a natural pathway of nitrogen lossthat often increases dramatically after land usechanges. As with leaching, erosional losses ofnitrogen include both organic and inorganicforms, although organic forms associated withsoil aggregates and particles are most impor-tant. In some ecosystems, especially those onunstable slopes or in areas exposed to highwinds, erosion is a dominant natural pathway ofnitrogen loss.

Other Element Cycles

Phosphorus

Weathering of primary minerals is the majorsource of new phosphorus to ecosystems. Incontrast to nitrogen, whose major source is theatmosphere, phosphorus enters ecosystems pri-marily by weathering of rocks (Fig. 9.10). Theweathering of phosphorus-containing apatiteby the carbonic acid generated from soil respi-ration, for example, releases phosphorus inavailable forms (Eq. 9.6) that can be taken updirectly by plants or microorganisms or beadsorbed or precipitated (see Chapter 3).

Ca5(PO4)3 + 4H2CO3 Æ5Ca2+ + 3HPO4

2- + 4HCO3 + H2O (9.6)

Phosphorus inputs from weathering dependon the mineralogy of the parent material, the

00 1

1

2

2

3

3

4

4

5

5

North America

Europe

Nitr

ogen

out

put (

g m

-2 y

r-1)

Nitrogen deposition (g m-2 yr-1)

Figure 9.9. Comparisons of inputs from nitrogendeposition and nitrogen outputs in solution fromforests of North America and Europe. There is astrong relationship between inputs and outputs innitrogen-saturated ecosystems. (Data from Tietemaand Beier 1995, Fenn et al. 1998.)


NH3

N2O

N2

E

N2NOx

NH4+Norg NO3

-

E

Rock

Sorg SO42-

H2S

SO42-

Rock

Caorg Ca2+E

EK+

Rock

Rock

PbE

E

Pbound

Porg PO43-

Rock

Figure 9.10. Comparison of natural element cycleswith respect to the relative importance of internalrecycling, inputs, and outputs. Inputs of nitrogencome primarily from the atmosphere, whereas inputs of phosphorus, potassium, calcium, and mostunessential elements such as lead (Pb) come pri-marily from rocks. Sulfur comes from both the

atmosphere and rocks. Over long time scales, atmos-pheric inputs of all elements can be important.Element losses occur through leaching, erosion (E),and, in the case of nitrogen and sulfur, gaseous emission. Subscripts indicate organic (org) or boundforms of the element.


climate, and the landscape age. Marine-derivedcalcareous rocks have relatively high phospho-rus content and weather readily (Eq. 9.6). Overmillions of years, phosphorus-containing min-erals become depleted, leading to increasingphosphorus limitation, as landscapes age. Theeffects of landscape age are most pronouncedin the tropics where weathering occurs mostrapidly. In areas downwind of deserts or agri-cultural areas, the deposition of phosphorus inrainfall or dry deposition often represents a sig-nificant input to ecosystems, particularly in oldlandscapes where inputs from weathering arelow.

As with nitrogen, the internal cycling ofphosphorus in ecosystems requires the cleavingof bonds with organic matter to produce a formthat is water soluble and can be absorbed bymicrobes and plants. Phosphorus turnover issomewhat less tightly linked to decompositionthan is nitrogen, because the ester linkages thatbind phosphorus to carbon (C—O—P) can becleaved without breaking down the carbonskeleton. Nitrogen, in contrast, is directlybonded to the carbon skeletons of organicmatter (C—N) and is generally released bybreakdown of the carbon skeleton into aminoacids and other forms of dissolved organicnitrogen.

The decomposition process that breaks downthe organic matter exposes the C—O—P bondsto enzymatic attack. Low soil phosphorus avail-ability induces plants and microbes to investnitrogen in enzymes to acquire phosphorus.Plant roots and their mycorrhizal associates—particularly arbuscular mycorrhizae, which areabundant in grasslands, tropical forests, andmany other systems—produce phosphatasesthat cleave ester bonds in organic matter torelease phosphate (PO4

3-). There is thereforetight cycling of phosphorus between organicmatter and plant roots in many ecosystems. Intropical forests, for example, mats of mycor-rhizal roots form in the litter layer and pro-duce phosphatases that cleave phosphate fromorganic matter. Mycorrhizal roots directlyabsorb much of this phosphate before it inter-acts with the mineral phase of the soil. Plantand microbial phosphatases are induced by lowsoil phosphate. This contrasts with protease,

whose activity correlates more strongly withmicrobial activity than with concentrations ofsoil organic nitrogen.

Microbial biomass frequently accounts for 20 to 30% of the organic phosphorus in soils(Smith and Paul 1990, Jonasson et al. 1999),much larger than the proportion of microbialcarbon (about 2%) or nitrogen (about 4%).Microbial biomass is therefore an importantreservoir of potentially available phosphorus,particularly in ecosystems with highly basic oracidic soils. Microbial phosphorus is potentiallymore available than inorganic phosphatebecause it is protected from reactions with themineral phase of soils, as described later.Although the biogeochemical literatureemphasizes the importance of C :N ratios, C :Pratios of dead organic matter can also be criti-cal. In ecosystems with low phosphorus avail-ability, the C :P ratio of dead organic mattercontrols the balance between phosphorus min-eralization and immobilization and thereforethe supply of phosphorus to plants.

Chemical reactions with soil minerals play akey role in controlling phosphorus availabilityin soils. Unlike nitrogen, phosphorus undergoesno oxidation–reduction reactions in soils andhas no important gas phases or atmosphericcomponents. In addition, many of the reactionsthat control phosphorus availability are geo-chemical rather than biological in nature. Phos-phate is the main form of available phosphorusin soils. Theoretically, soil pH determines themost common form of phosphate in the soilsolution:

H2PO4- ´ H+ + HPO4

2- ´ 2H+ + PO43- (9.7)

4 10 14 [pH range]

This is important because the less highlycharged forms of phosphate (H2PO4

-) are moremobile in soil and are therefore more availableto plants and microbes. The actual effects of pH depend, however, on the concentrations of other ions and minerals present in the soilmatrix (Fig. 9.11). At low pH, for example,where H2PO4

- should dominate, aluminum,iron, and manganese are also quite soluble and react with H2PO4

- to form insoluble compounds:


Al3+ + H2PO4- +

2H2O ´ 2H+ + Al(OH)2H2PO4 (9.8)

soluble insoluble

Very little of the phosphorus in soils issoluble at any time because many inorganic andorganic mechanisms retain phosphorus in insol-uble forms. Inorganic mechanisms are stronglypH dependent. At low pH, phosphorus can besorbed onto the surfaces of clays and oxides ofiron and aluminum. Phosphate is initially elec-trostatically attracted to positively charged siteson minerals through anion exchange. Oncethere, phosphate can become increasinglytightly bound (and correspondingly unavailableto plants) as it forms one or two covalent bondswith the metals on the mineral surface. Phos-phorus can also bind with soluble minerals(especially iron oxides) to form insoluble pre-cipitates. Chemical precipitates of phosphoruswith these oxides and phosphate sorption onoxide surfaces, explain why highly weatheredtropical soils (Oxisols and Ultisols) haveextremely low phosphorus availability and whythe growth of forests on those soils is typicallyphosphorus limited (see Chapter 3).The silicateclay minerals that dominate temperate soils fixphosphate to a lesser extent than do the oxidesof tropical oxisols.

In soils with high concentrations ofexchangeable calcium and calcium carbonate

(CaCO3), which typically occur at high pH,calcium phosphate precipitates, reducing phos-phate availability in solution:

Ca(H2PO4)2 + 2Ca2+ Æ Ca3(PO4)2 + 4H+ (9.9)

soluble insoluble

At high pH, phosphate combines with Ca toform (in order of decreasing availability)monocalcium, dicalcium, and tricalcium phos-phates. Precipitation of calcium phosphate isone of the main reasons that phosphate fertilizer immediately becomes unavailable incalcium-rich temperate agricultural ecosys-tems. Due to the precipitation reactions thatoccur at high and low pH, phosphorus is mostavailable in a narrow range around pH 6.5 (Fig.9.11). Organic compounds in the soil also regu-late, both directly and indirectly, phosphorusbinding and availability. Charged organic compounds, for example, can compete withphosphate ions for binding sites on the surfacesof oxides, thereby decreasing phosphorus fixa-tion. Organic compounds can also chelatemetals and prevent their reaction with phos-phate. On the other hand, organic compoundsform complexes with iron, aluminum, and phosphate that protect these compounds from enzymatic attack. In tropical allophanesoils, these complexes form a major sink forphosphorus.

Much of the phosphorus that precipitates asiron, aluminum, and calcium compounds isessentially unavailable to plants and is referredto as occluded phosphorus. During soil devel-opment, primary minerals gradually disappearas a result of weathering and erosional loss.Themass of phosphate in soils tends to shift frommineral, organic, and nonoccluded forms tooccluded and organically bound forms, causinga shift from nitrogen to phosphorus limitationin ecosystems over long time scales (see Fig.3.4) (Crews et al. 1995).

The tight binding of phosphate to organicmatter or to soil minerals in most soils causes90% of the phosphorus loss to occur throughsurface runoff and erosion of particulate phos-phorus rather than through leaching of solublephosphate to groundwater (Tiessen 1995). Twothirds of the dissolved phosphorus that entersgroundwater is organic and therefore less reac-

Soil pH

100

50

04.0 5.0 6.0 7.0 8.0

Perc

enta

ge d

istri

butio

n

Fixation by soluble Fe, Al, and Mn

Fixation by hydrousoxides of Fe, Al, and Mg

Relatively availablephosphates

Silicate reactions

Fixation mostly as calcium phosphate

Figure 9.11. Effect of pH on the major forms ofphosphorus present in soils. The low solubility ofphosphorus compounds at low and high pH result ina relatively narrow window of phosphate availabilitynear pH 6.5. (Redrawn from Nature and Propertiesof Soils, 13th edition by N.C. Brady and R.R. Weil ©2001, by permission of Pearson Education, Inc.,Upper Saddle River, NJ; Brady and Weil 2001.)


tive with soil minerals. The productivity ofaquatic ecosystems is so sensitive to phospho-rus additions that even small additions ingroundwater can cause large changes in theirfunctioning (see Chapter 13).

Sulfur

Sulfur cycling is particularly complex because itundergoes oxidation–reduction reactions, likenitrogen, and has both gaseous and mineralsources (Fig. 9.10). Rock weathering, which isthe primary natural source of sulfur in mostecosystems, is being increasingly supplementedby atmospheric inputs in the form of acid rain.Combustion of fossil fuels produces gaseoussulfur dioxide (SO2), which dissolves in clouddroplets to produce sulfuric acid (H2SO4), astrong acid that is responsible for much of thelake acidification downwind of industrial areas(see Chapter 15). Sulfur in plant tissues is bothcarbon and ester bonded, so microbial mineral-ization includes immobilization and releaseprocesses similar to those of nitrogen and phos-phorus. Like nitrogen, inorganic sulfur under-goes oxidation–reduction reactions and istherefore sensitive to oxygen availability in theenvironment. In anaerobic soils, sulfate acts asan electron acceptor that allows microbes tometabolize organic carbon for energy, withhydrogen sulfide being produced as a by-product. In aerobic environments, however,reduced sulfur can be an important energysource for bacteria. The productivity associatedwith deep-sea vents, for example, is basedentirely on the oxidation of hydrogen sulfide(H2S) from the vents. Sulfur is a component of most enzymes, including the nitrogenase ofnitrogen fixers, so low availability of sulfur inhighly weathered soils in unpolluted areas canlimit nitrogen inputs to ecosystems and there-fore plant production and nutrient turnover.Superphosphate fertilizer has a high sulfur con-centration, so vegetation responses to applica-tion of phosphate fertilizers in some ecosystemsmay include a response to sulfur (Eviner et al.2000). Sulfur compounds in the atmosphereplay critical roles as aerosols, which increase thealbedo of the atmosphere and therefore causeclimatic cooling (see Chapter 2).

Essential Cations

Rock weathering is the primary avenue forelement inputs of potassium, calcium, magne-sium, and manganese, the cations required inlargest amounts by plants. As with nitrogen,phosphorus, and sulfur, the quantities of thesecations cycling in ecosystems from soils toplants and back to soils are much larger thanare annual inputs to and losses from ecosys-tems. The availability of cations in the soil solu-tion is largely governed by exchange reactionsand depends on the cation exchange capacity ofthe soil and its base saturation (see Chapter 3),which, in turn, is influenced by parent materialand weathering characteristics. Calcium is animportant structural component of plant andfungal cell walls. Its release and cycling there-fore depends on decomposition in a way some-what similar to that of nitrogen and phosphorus(Fig. 9.11). Potassium, on the other hand, occursprimarily in cell cytoplasm and is releasedthrough the leaching action of water mov-ing through live and dead organic material.Magnesium and manganese are intermediatebetween calcium and potassium in their cyclingcharacteristics.

Potassium limits plant production in someecosystems, but calcium concentration in thesoil solution of most ecosystems is so high that it is actively excluded by plant cells duringthe uptake process (see Chapter 8). Availabil-ity of calcium and other cations may be lowenough to limit plant production on the old,highly weathered tropical soils or where acidrain has leached most available calcium fromsoils.

These cations have no gaseous phase, butatmospheric transfers of these elements (and of essential micronutrients) in dust can be animportant pathway of loss from deserts andagricultural areas that experience wind erosionand an important input to the open ocean andto ecosystems on highly weathered parentmaterials. Cations can also be lost via leaching.Nitrate or other anions that are leached fromecosystems must be accompanied by cations tomaintain electrical neutrality. Thus high nitrateleaching rates that occur in nitrogen-saturatedsites or as a result of excessive nitrogen fer-tilization are accompanied by high losses of


cations. The declines in forest productionobserved in Europe and the eastern UnitedStates in response to acid rain is at least partlya consequence of calcium and magnesium defi-ciencies induced by cation leaching (Schulze1989, Aber et al. 1998, Driscoll et al. 2001).

Nonessential Elements

The cycling of nonessential elements is domi-nated by the balance between inputs fromweathering, precipitation, and dust and outputsin leaching. Vegetation plays a relatively smallrole in the balance between inputs and outputsof elements such as chloride, mercury, and leadthat are not required by organisms (Fig. 9.11).Consequently, external cycling of elements(ecosystem inputs and outputs) dominates thecycling of nonessential elements, whereas inter-nal cycling through vegetation dominates thecycling of essential elements. The cycling ofnonessential elements is therefore not stronglyaffected by successional changes in vegetationactivity, whereas the losses of essential ele-ments decline dramatically during early suc-cession when organic matter and associatednutrients are accumulating in plant and micro-bial biomass (see Fig. 13.12) (Vitousek andReiners 1975).

Interactions Among Element Cycles

Across broad gradients in nutrient availability,the supply rate of the most strongly limitingnutrient determines the rate of cycling throughvegetation of all essential nutrients. Nutrientabsorption by vegetation depends on a dynamicbalance between the rate of supply of nutrientsin soil and nutrient demands by vegetation (seeChapter 8). Most plants exhibit a limited range of element ratios. The element that moststrongly limits plant growth has greatest impacton net primary production (NPP). Absorptionof other elements is then adjusted to maintainrelatively constant nutrient ratios in vegetation.

Despite this general synchrony of cyclingrates across broad gradients in nutrient supplyrates, there are important differences in factorsgoverning supply of different elements. Somephases of the cycles are controlled by the same

factors, such as uptake and release by vegeta-tion, but other components of the cycles, suchas input and losses, occur by separate pathwayswith different controlling factors. Conse-quently, essential elements do not cycle per-fectly in tandem in ecosystems. Differentialcycling leads to the opportunity for oneelement cycle to directly influence another. Thecycling of nitrogen, phosphorus, and sulfurthrough vegetation and dead organic matter,for example, is not perfectly coupled. Eachnutrient may be absorbed when it is most abun-dant and stored until it is required to supportgrowth (see Chapter 8). Availabilities of theseelements also differ in their amounts andtiming, due to their differential dependence oninputs, on decomposition, and on reactions withsoil minerals. Over long time scales, differentialavailabilities and plant requirements for thesenutrient elements set up the potential for inter-actions among them.

The importance of interactions among cyclesis illustrated by the changes in element cyclingduring long-term soil and ecosystem develop-ment. Newly exposed substrates such as freshglacial till, lava, or sand dune contain sub-stantial concentrations of phosphorus, majorcations (Ca, Mg, K), and metals (Fe, Cu, Zn,Mo) in primary minerals but low concentra-tions of nitrogen. When exposed to water andacidity, these minerals weather and release ele-ments into biologically available forms. Thebiological availability of these rock-derived ele-ments increases rapidly during early primarysuccession. Over time, a fraction of these ele-ments is lost via leaching and/or bound intoinsoluble or physically protected forms, andtheir availability declines.

Over the very long term (thousands to mil-lions of years, depending on the element, therock, and the climate), the primary minerals insoil are depleted, and most of the elements theysupplied are lost or irreversibly bound. Phos-phorus availability in particular decreases to thepoint that ecosystems on extremely old soils inhumid regions are strongly phosphorus limited(see Fig. 3.5) (Walker and Syers 1976), untilsome geological disturbance such as a landslideprovides a new source of unweathered minerals.This pattern of phosphorus decrease in old soils


has been observed in developmental sequencesin radically different climates, such as thosefound in New Zealand, Australia, and Hawaii(Vitousek and Farrington 1997).

In contrast to phosphorus, fixed nitrogen isnearly absent from most nonsedimentary rocks,so young soils must accumulate it from theatmosphere. Although sedimentary rocks con-tribute some nitrogen by weathering (0.01 to 2gm-2 yr-1) (Holloway et al. 1998), much of thismay be lost in groundwater rather than enter-ing ecosystems. The combination of substantialinputs of phosphorus and other elements with no nitrogen give nitrogen fixers a strongadvantage early in soil and ecosystem develop-ment. Indeed, many early successional systemsare dominated by symbiotic nitrogen fixers(Chapin et al. 1994). Where nitrogen fixersoccur, nitrogen accumulates relatively quickly.Where nitrogen fixers are sparse or absent,nitrogen enters from deposition and accumu-lates slowly. Nitrogen continues to accumulatein ecosystems until nitrogen availability comesinto approximate equilibrium with otherresources, including phosphorus (Walker andSyers 1976). Nitrogen limits forest growth onyoung substrates in Hawaii, for example; phos-phorus limits growth on old substrates; andnitrogen and phosphorus are both relativelyavailable in intermediate-aged sites (Vitousekand Farrington 1997).

Why does phosphorus rather than otherrock-derived elements limit biological pro-cesses in the long run? The major cations,especially calcium, are absorbed by organismsin much larger quantities than is phosphorusand are more readily leached from soils. On the Hawaiian sequence, rock-derived calcium,magnesium, and potassium virtually disappearwithin 100,000 years but do not limit forest pro-duction anywhere on the sequence (Vitousekand Farrington 1997). Atmospheric inputs ofcations prevent these elements from becominglimiting. Marine-derived aerosols containingcalcium, magnesium, and potassium aredeposited on forests in Hawaii through rain andcloud droplets. Phosphorus concentrations inmarine aerosols are low, however, because highphosphorus demands by marine organismsmaintain a low concentration in surface waters.The atmospheric inputs of calcium are 10-foldless than weathering inputs in young sites butare more than 1000-fold greater than weather-ing inputs in older sites (Chadwick et al. 1999).In continental interiors, dust from semiarid andother sparsely vegetated areas is a major sourceof cations. Even in Hawaii, dust from Asia,over 6000km away, is an important input ofphosphorus, especially during glacial times,when vegetation cover was sparse and windspeeds were high (Chadwick et al. 1999) (Box9.2). In situ weathering of parent material is

Geochemical tracers have been used to iden-tify dust and to determine its rate of input tothe Hawaiian Islands. Hawaiian rocks arederived from Earth’s mantle, whereas Asiandust comes from the crust.These two sourcesdiffer in the ratio of two isotopes of neo-dynium, in the ratio of europium to otherlanthanide elements, and in the ratio ofthorium to halfnium. All of these elementsare relatively immobile in soils, so changesover time in the isotopic or elemental ratioscan be used to calculate time-integratedinputs of Asian dust. Knowing the phospho-

rus content of the dust, it is then possible tocalculate phosphorus inputs by this pathway.Atmospheric inputs of phosphorus are muchlower than weathering for the first millionyears or more of soil development. However,by 4 million years, rock-derived phosphorushas nearly disappeared, and Asian dust pro-vides most of the phosphorus input to thesoil. The biological availability of phospho-rus is low in old sites, but it would be muchlower were it not for inputs of Asian dust,most of it transported more than 10,000years ago (Chadwick et al. 1999).

Box 9.2. Geochemical Tracers to Identify Source of Inputs to Ecosystems


therefore not always the dominant input ofminerals to ecosystems.

Summary

Nutrients enter ecosystems through the chemi-cal weathering of rocks, the biological fixationof atmospheric nitrogen, and the deposition ofnutrients from the atmosphere in rain, wind-blown particles, or gases. Human activities havegreatly increased these inputs, particularly ofnitrogen and sulfur, through combustion offossil fuels, addition of fertilizers, and plantingof nitrogen-fixing crops. Unlike carbon, theinternal recycling of essential plant nutrients ismuch larger than the annual inputs and lossesfrom the ecosystem, producing relatively closednutrient cycles.

Most nutrients that are essential to plant production become available to plants due tothe microbial release of elements from deadorganic matter during decomposition. Micro-bial exoenzymes break down the large poly-mers in particulate dead organic matter intosoluble compounds and ions that can beabsorbed by microbes or plant roots. The netmineralization of nutrients depends on thebalance between the microbial immobilizationof nutrients to support microbial growth andthe secretion of nutrients that exceed microbialrequirements for growth. The first product ofnitrogen mineralization is ammonium. Ammo-nium can be converted to nitrate by autotrophicnitrifiers that use ammonium as a source ofreducing power or by heterotrophic nitrifiers.Both plants and microbes use DON, ammo-nium, and nitrate in varying proportions asnitrogen sources, when their growth is nitrogenlimited. Soil minerals and organic matter alsoinfluence nutrient availability to plants andmicrobes through exchange reactions (primar-ily with soil cations, except in some tropicalsoils that have a substantial anion exchangecapacity), the precipitation of phosphorus withsoil minerals, and the incorporation of nitrogeninto humus.

Nutrients are lost from ecosystems throughthe leaching of elements out of the ecosystemin solution, emissions of gases, loss of nutrients

adsorbed on soil particles in wind or watererosion, and the removal of materials inharvest. Human activities, as with nutrientinputs, often increase nutrient losses from ter-restrial ecosystems.

Review Questions

1. What are the relative magnitudes of atmos-pheric inputs and mineralization from deadorganic matter in supplying the annualnitrogen uptake by vegetation?

2. If Earth is bathed in dinitrogen gas, why isthe productivity of so many ecosystemslimited by availability of nitrogen? What isbiological nitrogen fixation? What factorsinfluence when and where it occurs?

3. What are the mechanisms by which nitro-gen is deposited from the atmosphere intoterrestrial ecosystems?

4. What are the major steps in the mineral-ization of litter nitrogen to inorganicforms? What microbial processes mediateeach step and what are the products of eachstep? Which of these processes are extra-cellular and which are intracellular?

5. What ecological factors account for differ-ences among ecosystems in annual netnitrogen mineralization? How does each ofthese factors influence microbial activity?

6. What determines the balance betweennitrogen mineralization and nitrogen immo-bilization in soils?

7. What factors determine the balancebetween plant uptake and microbial uptakeof dissolved organic and inorganic nitrogenin soils?

8. How do ammonium and nitrate differ inmobility in the soil? Why? How does thisinfluence plant uptake and susceptibility toleaching loss?

9. What is denitrification and what regulatesit? What are the gases that can be produced,and what are their roles in the atmosphere?

10. What is the main mechanism by whichphosphorus enters ecosystems?

11. What factors control availability of phos-phorus for plant uptake? Why is phospho-rus availability low in many tropical soils?


12. Why are mycorrhizae so important forplant acquisition of phosphorus?

13. What is the main pathway of phosphorusloss from terrestrial ecosystems?

Additional Reading

Andreae, M.O., and D.S. Schimel. 1989. Exchange ofTrace Gases between Terrestrial Ecosystems andthe Atmosphere. John Wiley, New York.

Howarth, R.W., editor. 1996. Nitrogen Cycling in theNorth Atlantic Ocean and Its Watershed. Kluwer,Dordrecht.

Paul, E.A., and F.E. Clark. 1996. Soil Microbiologyand Biochemistry. 2nd Edition. Academic Press,San Diego, CA.

Sala, O.E., R.B. Jackson, H.A. Mooney, and R.W.Howarth, editors. 2000. Methods in EcosystemScience. Springer-Verlag, New York.

Schlesinger, W.H. 1997. Biogeochemistry. An Analy-sis of Global Change. Academic Press, San Diego,CA.

Tiessen, H., editor. 1997. Phosphorus in the GlobalEnvironment: Transfers, Cycles and Management[Scope Vol. 54]. John Wiley, New York.

Vitousek, P.M., J.D. Aber, R.W. Howarth, G.E.Likens, P.A. Matson, D.W. Schindler, W.H.Schlesinger, and G.D. Tilman. 1997. Human alter-ation of the global nitrogen cycle: Sources andconsequences. Ecological Applications 7:737–750.

Vitousek, P.M., and R.W. Howarth. 1991. Nitrogenlimitation on land and in the sea: How can itoccur? Biogeochemistry 13:87–115.

Introduction

The same general principles govern carbon and nutrient cycling in aquatic and terrestrialecosystems, but the physical differencesbetween water and air result in radically differ-ent ecological controls. Many of the basic prin-ciples of ecosystem ecology were developed inaquatic ecosystems, including the concepts oftrophic dynamics and energy flow (Lindeman1942), the interactions among biogeochemicalcycles (Redfield 1958), and species effects onecosystem processes (Redfield 1958, Carpenterand Kitchell 1993). These concepts are broadlyapplicable to terrestrial ecosystems. None-theless, many of the ecological patterns anddynamics of aquatic and terrestrial ecosystemsare quite different, largely due to the differ-ences between air and water as the basicsupport medium. In this chapter we explore theconsequences of this physical difference anddescribe the broad similarities and differencesin carbon and nutrient cycling between aquaticand terrestrial ecosystems.

Aquatic ecosystems are just as structurallyand functionally diverse as are terrestrialecosystems. We therefore initially focus ouraquatic–terrestrial comparison on pelagic(open water) marine ecosystems, which differ

most dramatically from terrestrial ecosystemsand then discuss in less detail the differencesbetween marine ecosystems and lakes andstreams.

Ecosystem Properties

The differences in physical properties betweenwater and air result in fundamental differencesin structure and functioning between aquaticand terrestrial ecosystems. Due to the greaterdensity of water than air, the physical supportfor photosynthetic organisms is greater in waterthan on land (Table 10.1). The primary produc-ers in pelagic ecosystems are therefore micro-scopic algae (phytoplankton) that float near thewater surface, where light availability is great-est. This contrasts with terrestrial plants, whichrequire elaborate support structures to raisetheir leaves above neighbors. Plants are themajor habitat-structuring feature in terrestrialecosystems. Their physical structure governsthe patterns of physical environment, organismactivity, and ecosystem processes. In the openocean, however, the environment is physicallystructured by vertical gradients in light, tem-perature, oxygen, and salinity. Ocean currentsand deep circulation frequently remove phyto-

10Aquatic Carbon and Nutrient Cycling

Aquatic ecosystems differ radically from their terrestrial counterparts in physicalenvironment and therefore in controls over ecosystem processes. This chapterdescribes the major differences in carbon and nutrient cycling between terrestrialand aquatic ecosystems.

224

Ecosystem Properties 225

plankton from the euphotic zone, the upper-most layer of water where there is enough lightto support photosynthesis. Coping with this fre-quent disturbance requires rapid cell division,small size, and, for larger organisms, the cap-acity to swim. The small size of phytoplanktonresults in a high surface to volume ratio thatmaximizes their effectiveness in absorbingnutrients. Many marine plankton are particu-larly small (nanoplankton are 2 to 20mm indiameter, and picoplankton are less than 2mmin diameter) and have a competitive advant-age where nutrients are extremely dilute.Picoplankton, for example, account for half ofthe plankton biomass of the highly nutrient-impoverished tropical oceans (Valiela 1995).The size and lifespan of marine organismsincreases going up the food chain (Fig. 10.1).

The size of aquatic organisms determinestheir feeding strategy.Water is a polar molecule

that sticks to the surface of organisms. Themovement of small organisms and particles isimpeded by these viscous forces. Large organ-isms, in contrast, can swim, and their speed islargely determined by inertia. The Reynoldsnumber (Re) is the ratio of inertial to viscousforces and is a measure of the ease with whichorganisms can move through a viscous fluid likewater.

(10.1)

The movement of organisms through water isnot strongly impeded for organisms with a largelength (l) and velocity (v) and under conditionsof low kinematic viscosity (Vk) (Fig. 10.2). Smallplanktonic organisms must deal with life at alow Reynolds number, where viscous forces are much stronger than inertial forces. At thesesmall sizes, swimming and filter feeding are

Re =lvVk

Property Water Air Ratio of water to air

Oxygen concentration (ml L-1) 7.0 209.0 1 :30Diffusion coefficient (mm s-1)

Oxygen 0.00025 1.98 1 :8000Carbon dioxide 0.00018 1.55 1 :9000

Density (kg L-1) 1.000 0.0013 800 :1Viscosity (cP) 1.0 0.02 50 :1Heat capacity (cal L-1 (°C)-1) 1000.0 0.31 3000 :1

Data from Moss (1998).

Table 10.1. Basic propertiesof water and air that influenceecosystem processes.

0.1 102

105

108

10

102

103

104

105

1

0.1

0.1 102

105

108

Length (µm) [log scale]

Gen

erat

ion

time

(d)

[log

scal

e]

PlantsHerbivores

Carnivores Herbivores and

carnivores

Plants

Marine Terrestrial

Figure 10.1. Body size and generation time fororganisms in the ocean (Steele 1991) and on land ofdominant plants, herbivores, and carnivores. In the ocean the dominant plants (picoplankton and

nanoplankton) are generally smaller than the herbi-vores that feed on them, whereas on land, the dom-inant plants are often as large or larger than theherbivores that eat them.

226 10. Aquatic Carbon and Nutrient Cycling

energetically expensive, so diffusion is themajor process that moves nutrients to the cellsurface, just as with fine roots growing in thesoil solution.

The small size and lack of nonphotosyntheticsupport structures in marine phytoplanktonmean that marine primary producers requirerelatively little biomass to support a given pho-tosynthetic capacity. The average primary pro-ducer biomass per unit area on land, forexample, is 660-fold greater than in the ocean,although the average net primary production(NPP) per unit area on land is only 5-fold greaterthan in the ocean (Table 10.2) (Cohen 1994).Phytoplankton biomass of oceans and lakesturns over 20 to 40 times per year, or even dailyunder conditions that are favorable for growth,whereas turnover for terrestrial plant biomassoften requires years to decades (Valiela 1995).

The air that surrounds terrestrial organismsdelivers oxygen and other gases orders of magnitude more rapidly than occurs in water.The surface ocean water, for example, has anoxygen concentration 30-fold lower than in air(Table 10.1), and aquatic sediments are muchmore likely to be anaerobic than are terrestrialsoils. Aquatic organisms therefore exhibit avariety of adaptations to acquire oxygen and

withstand anaerobic conditions, whereas onland the acquisition of water and the avoidanceor tolerance of desiccation are more commonevolutionary themes.

The buoyancy of water produces an environ-ment that is rich in small particles, includingalgal cells and suspended particles of detritus.Filter feeders are organisms that feed on sus-pended particles through use of a diverse arrayof tools, including hairy appendages on legs ormouth parts, sticky secretions, and silken nets.Filter feeders have no counterpart in terrestrial

103

103100

100

10-3

10-3

10-610-6

106

Spe

ed o

f tra

vel

(m s

-1)

Pure viscosityeffect

Effect of both factors

Pure inertialeffect

Reynolds number

Figure 10.2. Range of Reynolds numbers for organ-isms of different lengths and speeds. Small organismslike phytoplankton have small Reynolds numbersand derive their nutrition by diffusion. As size and

Reynolds number increase, nutrition based on movement (filter feeding and swimming) becomeprogressively more important. (Redrawn with per-mission from Halsted Press; Schwoerbel 1987.)

Table 10.2. Characteristics of oceans and continents.

Unit Oceans Continents

Surface area (% of Earth) 71 29Volume of life zone (% of Earth) 99.5 0.5Living biomass (1015 gC) 2 560Living biomass (gm-2) 5.6 3700Dead organic matter 5.5 10

(103 gm-2)Net primary production 69 330

(gC m-2 yr-1)Residence time of C in living 0.08 11.2

biomass (yr)

Data from Cohen (1994).

Ecosystem Properties 227

ecosystems (Gee 1991). Most filter feeders areabout 100-fold larger than the suspended parti-cles (seston) on which they feed. Filter feedersare often selective in the size of particles theyingest but process algae, bacteria, and sus-pended particles of organic and inorganic mate-rials relatively indiscriminately. Much of thisfood may therefore be of relatively low quality,and substantial quantities of water must be pro-cessed to acquire sufficient energy and nutri-ents to support growth.

There are many more species on land than inthe oceans but the broad phyletic diversity forcoping with life is greater in the ocean. About76% of all species occur on land, and most ofthese are insects. Of the 15% of the species thatare marine, most are benthic animals of theocean sediments. There are only about 20,000photosynthetic species in aquatic ecosystems incontrast to the 300,000 species of terrestrialplants, and there are very few aquatic insects(Falkowski et al. 1999). At higher taxonomiclevels, however, 80% the multicellular generaoccur in the sea, and only 20% are on land(Table 10.3) (May 1994). The greater diversityof genera and phyla in the ocean than on landcould reflect its longer evolutionary history,giving organisms more time to try out differentfundamental body plans and functional types(May 1994). The larger number of species onland than in the ocean could reflect the greaterheterogeneity and potential for spatial isolationin terrestrial habitats.

Life evolved in the sea. From there, bothplants and multicellular animals moved to freshwaters and then to land (Moss 1998). Severalgroups of plants and animals subsequently followed the reverse path from land to freshwater to oceans.These transitions have occurredmany times, indicating that the physiologicaladjustments required are not insurmountableon evolutionary time scales.

The marine environment is slightly moresaline (salty) than the internal body fluids ofmarine organisms, so organisms must minimizeloss of water and gain of salts to maintain ionicbalance. Movement of organisms from marineto fresh water reverses this osmotic gradientand intensifies the costs of osmoregulation.Terrestrial plants and animals confront twocontrasting problems. First, their source ofwater is usually fresh, requiring more energy tomaintain osmotic gradients. Second, they areexposed to an aerial environment that pro-motes water loss and dehydration. One greatadvantage to life on land is greater availabilityof oxygen and the greater energy provided by aerobic metabolism. Disadvantages includegreater dehydration and lower buoyancy of airthan water (Table 10.1).

The benthic environment of marine sedi-ments differs radically from terrestrial soils inits low oxygen availability. Oxygen concentra-tion in deep waters is much lower than in air,and oxygen diffusion into water-saturated sediments is much slower than through the air-filled pores of terrestrial soils. Mixing of sediments by benthic organisms plays a criticalrole in promoting oxygen flux into sedimentsand therefore benthic decomposition. Redoxreactions involving electron acceptors otherthan oxygen play a key role in the metabolismof benthic organisms and therefore in the patterns of carbon and nutrient cycling (seeChapter 3) (Valiela 1995).

The wide range in ecosystem structureamong aquatic ecosystems reflects variations inphysical environment. Most nonpelagic aquaticecosystems are intermediate in structural prop-erties between terrestrial and pelagic ecosys-tems. In the littoral zone, where the ocean andland meet, organisms can reduce the probabil-ity of being swept away by attaching to sub-strates. This gives rise to a diverse array of

Ocean OceanPhyla benthic pelagic Fresh-water Symbiotic Terrestrial

Total (33) 27 11 14 15 11Endemic 10 1 0 4 1

Data from May (1994).

Table 10.3. Generic meta-zoan diversity of land andoceans.


shallow-water ecosystems, including coral reefs,seagrass beds, and kelp forests. In ecosystemswhere physical attachment is possible, phyto-plankton, benthic algal mats, multicellular algaesuch as sea lettuce (e.g., Ulva) and kelp (e.g.,Laminaria), and vascular plants like eelgrass(Zostera) occur in various combinations. Therelative abundance of different types ofprimary producers depends on many factors,including nutrient availability, water depth, thestability of substrates, and disturbance regime.Nutrient addition, for example, favors rapidlygrowing phytoplankton, which reduce the lightavailable to multicellular algae at depth. Inter-mediate levels of disturbance and presence ofstable substrates favor multicellular primaryproducers (Sousa 1985).

Oceans

The large area and low productivity per unitarea of ocean cancel out, so the ocean con-tributes nearly half (about 40%) of global NPP.

Although oceans cover 70% of Earth’s surface,the average NPP per unit area is only 20% ofthat on land (Table 10.2). Aquatic productivity,however, is highly variable, just as on land.The most productive aquatic ecosystems, suchas coral reefs, kelp forests, and eutrophic lakes,can be at least as productive as the most pro-ductive terrestrial ecosystems (Fig. 10.3). NPPin the open ocean, which accounts for 90% ofthe ocean area, however, is similar to that ofterrestrial deserts and tundra. Because of itslarge area, the open ocean accounts for 60% ofmarine production, with picoplankton account-ing for about 90% of this production (Valiela1995).

Carbon and Light Availability

Photosynthesis is seldom carbon limited in the ocean. In marine pelagic ecosystems, forexample, only 1% of the carbon in a givenwater volume is involved in primary produc-tion, whereas the nitrogen in this water maycycle through primary production 10 times a

10 2 3 4

Production (kg C m-2 yr-1)

Corals

Kelp and rockweeds

Salt marsh grasses

Sea grasses

Mangroves

Benthic microalgae

Coastal phytoplankton

Open sea phytoplankton

FRESH-WATER PRODUCERS

Macrophytes

Phytoplankton (nutrient rich)

Phytoplankton (nutrient poor)

TERRESTRIAL PRODUCERS

Tropical wet forests

Temperate forests

Grasslands

Deserts and tundra

MARINE PRODUCERS

Figure 10.3. NPP of selected marine,fresh-water, and terrestrial ecosystems.Marine and fresh-water ecosystems exhibitthe same range of NPP that occurs on land,but unproductive marine ecosystems (theopen ocean) are much more widespread.(Redrawn with permission from Springer-Verlag; Valiela 1995.)

Oceans 229

year (Thurman 1991). One reason for the lackof carbon limitation in the ocean is that inor-ganic carbon is available in several forms,including CO2, bicarbonate (HCO3

-), carbonate(CO3

-), and carbonic acid (H2CO3). When CO2

dissolves in water, a small part is transformedto carbonic acid, which in turn dissociates tobicarbonate, carbonate, and H+ ions with a con-comitant drop in pH.

H2O + CO2 ´ H2CO3 ´H+ + HCO3

- ´ 2H+ + CO32- (10.2)

As expected from these equilibrium reactions,the predominant forms of inorganic carbon arefree CO2 and carbonic acid at low pH (theequation driven to the left), soluble bicar-bonate at about pH 8 (typical of ocean waters),and carbonates at high pH (equation driven to the right). Bicarbonate accounts for 90% of the inorganic carbon in most marinewaters. Phytoplankton in pelagic ecosystemsuse primarily CO2 as their carbon source.CO2 is then replenished from bicarbonate (Eq. 10.2). Some marine algae in the littoralzone, such as the macroalga Ulva, also usebicarbonate.

Water, algae, and suspended and dissolvedmaterials absorb light that enters aquaticecosystems, whereas light on land is absorbedprimarily by the plant canopy. Light energyfuels photosynthesis in the same way on landand in the water. On land this energy is modi-fied (absorbed or transmitted) primarily byvegetation as it moves down through thecanopy (see Chapter 5). In water, however, thesmaller biomass of plant cells and the signifi-cant absorption by water, dissolved organics,and suspended particles cause the mediumitself to contribute substantially to the expo-nential decrease in light from the surface todepth. Light is depleted rapidly in coastalwaters, where plankton, suspended sediments,and dissolved organic compounds are relativelyabundant. The euphotic zone, where there issufficient light to support phytoplanktongrowth, extends only to about 200m in the clearwaters of the open ocean (Fig. 10.4). Most of the ocean volume therefore cannot support thegrowth of primary producers whose carbon fix-ation depends on light.

In aquatic ecosystems blue light predomi-nates at depth whereas red light predominatesat depth in terrestrial canopies. In clear water there is an inverse relationship betweenwavelength and light transmission; blue light, which has a short wavelength and highenergy, penetrates to greatest depth (Fig. 10.5).Turbid waters both absorb and scatter light, causing the longer wavelengths to becomemore rapidly depleted with depth. Humic compounds, for example, absorb ultraviolet(UV) and blue wavelengths. In contrast toaquatic systems, the high density of chlorophyllin terrestrial plant canopies selectively depletesthe short-wavelength, high-energy blue light,so light at the bottom of a forest canopy isenriched in red compared to incoming solarradiation (Fig. 10.5). Green light, which is notabsorbed efficiently by chlorophyll, also pene-trates to the forest floor to a greater extent than does blue light. Eutrophic lakes with high chlorophyll concentrations have profiles of light quality more similar to those of foreststhan to the open ocean. Aquatic and terrestrialecosystems thus differ in both the quantity and the quality of light that penetrates to depth.

Marine phytoplankton are like terrestrialshade plants. Photosynthesis saturates at rela-

Clearestoce

an water

10-1

103

10-9

10-5

10-13

200

400

600

800

1000

1200

Lower limit of bioluminescence

Lower limit of phytoplankton

Lower limitof fish

Terrestrial forest

Irradiance (W m-2)

Dep

th (

m)

0

Coastal ocean water

Figure 10.4. Light availability at different depths inforests and in coastal and open oceans (Chazdon and Fetcher 1984a). (Modified with permission fromSpringer-Verlag; Valiela 1995.)


tively low light intensity, from 5 to 25% of full sun, depending on the algal group (Valiela1995). Photosynthesis declines at higher lightlevels, so maximum photosynthesis usuallyoccurs at about 10m depth on sunny days. Onereason that phytoplankton function as shadeplants is that they mix vertically through thewater column, so an individual cell spends rel-atively little time near the surface. Photosyn-thetic acclimation to high light requires about12h, which is probably longer than the time thatmost phytoplankton would be exposed to highlight. In the ocean and clear lakes, UV-B radi-ation may also contribute to low photosyntheticrates in surface waters, raising questions aboutthe consequences of ozone holes and increasedUV-B at high latitudes. Light appears to limitocean and lake production at the water surface,primarily beneath ice cover or during winter athigh latitudes when the sun angle is low. Atdepth, light limits production in all aquatichabitats.

Primary producers exhibit a greater diversityof pigment types in the ocean than on land,presumably related to the complex light environment of water. Red algae, which areabundant at depth in tropical oceans, and

brown algae like kelp, which are abundant atdepth in cool temperate oceans, have pig-ments that absorb the blue light that is available to them. Surface algae tend to have green pigments. These differences in algal pig-ments may also be adaptations to light quantityper se.

Reduced sulfur compounds provide theenergy for carbohydrate synthesis from CO2

in some anaerobic aquatic habitats. Althoughmost organisms depend on carbon fixedthrough light-dependent photosynthesis, somebacteria use the energy of reduced sulfur com-pounds to reduce CO2 to form organic com-pounds. Entire ecosystems are built on such abase near hydrothermal vents in zones of sea-floor spreading in midocean regions. Althoughhydrothermal vents account for only a tiny frac-tion of total ocean production, they supportunique communities and complex food websthat are completely independent of energyinput from the sun (Karl et al. 1980). Similarchemosynthesis occurs in anaerobic sediments,but these sulfur-dependent oxidation–reduc-tion reactions usually account for only a smallfraction of the total carbon budget of theseenvironments.

0.1 1 10 100 0.1 1 10 1000.1 1 10 100

GB

5

10

15

20

25

30W

0

RGB

0

10

20

Oceanic water Coastal water Forest

20

60

100

140

R

R

Y

G

B

Irradiance (% of incident light) [log scale]D

epth

(m

)

Figure 10.5. Light quality at different depths inocean and coastal waters and in forests (Chazdonand Fetcher 1984a). R, red; Y, yellow; G, green; B,

blue; W, white. (Modified with permission fromSpringer-Verlag; Valiela 1995.)

Oceans 231

Nutrient Availability

The euphotic zones of the ocean are frequentlynutrient poor. In pelagic ecosystems of theopen ocean, photosynthetic cells in the eup-hotic zone are spatially separated from thebenthic supply of nutrients. This contrasts withterrestrial ecosystems in which transport tissuescarry nutrients directly from the soil to photo-synthetic cells in the canopy. The small size ofphytoplankton causes diffusion to be the majorprocess that moves nutrients to the cell surface,as described earlier. Production in these pelagicecosystems is therefore generally nutrientlimited, and algal uptake maintains low nutri-ent concentrations in the water of the euphoticzone (Fig. 10.6). Some phytoplankton swim(flagellates or ciliates) or sink (through changesin buoyancy) to reduce nutrient limitation by diffusion. Swimming can increase nutrientuptake in microplankton by 50 to 200%, butpicoplankton cannot swim fast enough to over-

come diffusion (Valiela 1995). Only large-celledalgae can sink fast enough to overcome nutri-ent limitation by diffusion.

The nature of nutrient limitation in the openocean is a complex consequence of elementinteractions. The open ocean is a nutritionaldesert, remote from sources of nutrient input.In the open ocean, phosphorus appears to bethe master element that ultimately limits theproductive capacity of the oceans (Tyrrell 1999,Sigman and Boyle 2000). Its supply to the openocean depends on products of rock weatheringthat are transported to the ocean in rivers,deposited as dust from neighboring continents,or mixed upward from the deep ocean. When-ever phosphorus availability increases, nitrogenfixers such as cyanobacteria generally addnitrogen until phosphorus again limits theirproduction. The open ocean, however, seldombuilds up the high nitrate concentrations foundin lakes, and phytoplankton production fre-quently responds more strongly to nitrogenthan to phosphorus in short-term experiments(Fig. 10.7) (Valiela 1995, Tyrrell 1999). Oceanwater converges strongly on a relatively con-stant N :P ratio of 14 to 16, suggesting that bothnitrogen and phosphorus frequently limit pro-duction. This Redfield ratio reflects the relativerequirement of the two elements by phyto-plankton and most other organisms on Earth.Nitrogen limitation is widespread in coastaloceans, perhaps reflecting denitrification thatoccurs in anaerobic sediments (Falkowski et al.1999).

Trace elements—which are cofactors fornitrogenase, the nitrogen-fixing enzyme, andwhich are also required by other phytoplank-ton—often limit ocean productivity. In thesubequatorial gyres, the Subarctic Pacific, andthe Southern Ocean surrounding Antarctica,surface nitrogen and phosphorus concentra-tions are relatively high, and about half of theavailable nitrogen and phosphorus are mixed todepth without being used to support primaryproduction. In these regions, production fails torespond to addition of these nutrients, leadingto a syndrome known as high-nitrogen, low-chlorophyll (HNLC) syndrome (Valiela 1995,Falkowski et al. 1999). Large-scale iron-addition experiments in these regions have

Nitrate N (µg L-1)

00

20

40

60

5 10 15 20 25 30

Peru current

Californiacurrent

North CentralAtlantic

NorthPacific

80

Dep

th (

m)

120

100

140

160

Upwellingcurrents

Mid-oceangyres

Figure 10.6. Depth profiles of nitrate and phosphatein midocean gyres and upwelling zones of the ocean.(Redrawn with permission from Saunders; Dugdale1976.)


caused phytoplankton blooms large enough tobe seen from satellites, indicating that ironlimits the capacity of phytoplankton to usenitrogen and phosphorus. During glacialperiods there may have been 10-fold greaterinput of iron- and phosphorus-bearing dust tothe oceans, thus stimulating ocean productivityand in turn lowering atmospheric CO2 concen-trations (Falkowski et al. 1999). The key role ofiron in regulating production in some sectors ofthe open ocean has led to the suggestion that large-scale iron fertilization might stimulateocean production sufficiently to scavange largeamounts of CO2 from the atmosphere andsequester it in deep oceans in the form of deadorganic matter. The iron-addition experiments,however, show that this stimulation of pro-duction is relatively short-lived, presumablybecause other elements quickly become limit-ing to production, as soon as the iron demandsof phytoplankton are met. Grazing is anotherfactor that contributes to low phytoplanktonbiomass and productivity in portions of theopen ocean. In some HNLC zones of the oceanthere is simply not enough phytoplanktonbiomass to use the nutrients that are available(Valiela 1995).

The strong nutrient limitation and lack ofCO2 limitation of productivity in most of theworld’s oceans make it unlikely that marineproductivity will respond directly to increasingatmospheric CO2. Nutrient limitation of marineproduction, however, makes these ecosystemspotentially vulnerable to anthropogenic nutrient inputs. Estuaries, for example, havebeen substantially altered by the large nutrientinputs from agricultural lands. Runoff andsewage effluents have substantially alteredcoastal ecosystems near heavily populated or agricultural areas (Howarth et al. 1996).Oxygen depletion by decomposers in the watercolumn of the Mississippi Delta, for example,has created a large dead zone in the Gulf ofMexico. Although the impacts of nutrients onthe open ocean may be more subtle and diffi-cult to detect because most pollution sourcesare remote, they could, over the long term, beimportant because of the high degree of nutri-ent limitation of pelagic ecosystems and theirlarge aerial extent.

Ocean productivity is ultimately limited bythe rate of nutrient supply from the land ordeep ocean waters. For this reason, produc-tivity is greater in coastal waters than in the

Marine phytoplankton

Freshwater phytoplankton

N:P (by atoms) [log scale]

Freq

uenc

y

15

10

5

0

15

10

5

0

10 100

N limited

P limited

1 1000

No data available

Figure 10.7. Frequency distribu-tion of the N : P ratio in marine andfresh-water phytoplankton. Nutri-ent addition experiments (shadedbars) indicate that the high N :Pratios in lakes reflect phosphoruslimitation of algal growth, whereasthe generally low N:P ratios ofmarine phytoplankton commonlyreflect nitrogen limitation. (Modi-fied with permission from Springer-Verlag; Valiela 1995.)

Oceans 233

open ocean. Tidal mixing of sediment nutrientsinto the water column and oxygenation of thewater column contribute to the high productiv-ity of estuaries and intertidal and near-shoremarine ecosystems (Nixon 1988). The energyinput from waves, for example, exceeds thatfrom the sun in some intertidal areas. Coralreefs are among the most productive eco-systems on Earth (Fig. 10.3). Frequent tidalflushing supplies nutrients to algae that grow onthe surfaces of dead corals. These algae havehigh turnover rates because they are constantlygrazed by fish. The biomass of algae in thisecosystem is therefore small, just like thebiomass of phytoplankton in pelagic ecosystems.

In pelagic ecosystems, upwelling near thewest coasts of continents provides the greatestrate of nutrient supply. Upwelling supportssome of Earth’s major fisheries off Peru, north-west Africa, eastern India, southwest Africa,and the western United States (Valiela 1995)(Fig. 10.6). In these areas, Coriolis forces causewinds and surface waters to move offshore (seeChapter 2). These surface waters are replacedby nutrient-rich waters from depth. Upwellingalso occurs in the open ocean where majorocean currents diverge. This happens, forexample, in the equatorial Pacific, where oceancurrents diverge to the north and south and inthe Southern Ocean, the North Atlantic, andthe North Pacific (Valiela 1995). These regionshave relatively high nutrient availability andproductivity.

Vertical gradients in water density also influ-ence the effectiveness of nutrient transportfrom subsurface to surface waters. In thecentral subtropical oceans, where upwellingdoes not occur, the strong vertical temperaturegradient results in an extremely stable thermo-cline, in which low-density warm water isunderlain by high-density cold water (seeChapter 2). This stable stratification of waterminimizes the effectiveness of vertical mixingby waves and ocean currents, so nutrient avail-ability and productivity of the subtropicalocean is extremely low. As latitude increases,however, surface temperature declines. Thisweakens the vertical density gradient, so stormwaves and currents are more effective in mixing

deep nutrient-rich waters to the surface. Thestrong westerly winds and storm tracks associ-ated with the polar jet also contribute to effec-tive mixing in high-latitude oceans. Temperateand polar ocean waters are therefore morenutrient-rich and productive than are tropicaloceans.

The upward mixing of nutrients is greatestduring winter, when surface waters are coldest,and the vertical stratification is least stable.Winter is also the time of year when strongequator-to-pole heating gradients generate thestrongest winds (see Chapter 2). High-latitudeoceans typically experience a bloom of phyto-plankton in spring, after winter mixing hasoccurred and when light increases sufficientlyto support high photosynthetic and growthrates. The bloom ends when nutrients aredepleted by production, and most algae havebeen consumed by zooplankton grazers. Thehigh productivity of high-latitude oceans sup-ports rich fisheries, although many of thesehave been depleted by overfishing. The latitu-dinal variation in pelagic productivity alsoexplains several other interesting ecologicalpatterns, such as the annual migration of manywhales and sea birds between the Antarctic andthe Arctic Oceans to capitalize on springblooms of high-latitude productivity. In addi-tion, a high proportion of fish species at highlatitudes have an anadromous life history, inwhich they exploit the productive marine envi-ronment to support growth during the adultphase and use the relatively predator-freefreshwater environment to reproduce. Thisanadromous life history strategy is increasinglyfavored as latitude increases because marineproductivity increases with increasing latitude,whereas terrestrial productivity declines withincreasing latitude (Gross et al. 1988).

Carbon and Nutrient Cycling

Herbivory accounts for a threefold greater pro-portion of the carbon and nutrient transfer inpelagic than in terrestrial ecosystems (Fig. 10.8)(Cyr and Pace 1993). Although marine phyto-plankton exhibit a range of structural andchemical defenses against herbivores, just liketerrestrial plants, phytoplankton are relatively


digestible due to their lack of structural supporttissue. The resulting high rate of herbivory byzooplankton in pelagic ecosystems transfers alarge proportion of primary producer carbonfrom plants to animals. Herbivory is stronglycorrelated with NPP, so the secondary produc-tivity of marine fisheries and other componentsof secondary production depend strongly onNPP (see Chapter 11). Food webs in the three-dimensional pelagic environment are fre-quently longer and more complex than those in the two-dimensional benthic environment(Thurman 1991). Because predation is stronglysize dependent, the wide range of sizes ofpelagic plankton (0.1 to 2000mm) also con-tributes to long food chains and complex websin pelagic ecosystems.

Decomposition within the euphotic zonerecycles nutrients and contributes energy tohigher trophic levels. Phytoplankton releaseabout 10% (5 to 60%) of their production asexudates into the water column (Valiela 1995),a proportion of NPP similar to that which terrestrial plants transfer to the soil as root exudates and to support mycorrhizal fungi.Zooplankton spill phytoplankton cytoplasminto the water, as they eat, and excrete theirown waste products. Pelagic bacteria breakdown the resulting organic compounds andmineralize the associated nutrients, which are then available to primary producers.This decomposition occurs relatively quicklybecause the carbon substrates are mostly labile

organic compounds of low molecular weightwith a low C :N ratio (Fenchel 1994). This contrasts with the structurally complex,carbon-rich compounds (cellulose, lignin,phenols,tannins) that dominate terrestrial detritus.Viruses play an important role in planktonicfood webs, lysing bacteria and algae. Viral lysismay account for 5 to 25% of bacterial mortal-ity in pelagic ecosystems (Valiela 1995). Pelagicbacteria and viruses are grazed by small(nanoplankton) flagellate protozoans, which in turn are eaten by larger zooplankton. Thedetritus-based food web (see Chapter 11) istherefore tightly integrated with the plant-based trophic system in pelagic food webs andcontributes substantially to the energy andnutrients that support marine fisheries. Thismicrobial loop in pelagic ecosystems recyclesmost of the carbon and nutrients within the euphotic zone, so nutrients are recycledthrough food webs multiple times before beinglost to depth (Fig. 10.9).

Pelagic carbon cycling pumps carbon andnutrients from the ocean surface to depth (Fig.10.9). Although most of the planktonic carbonacquired through photosynthesis returns to theenvironment in respiration, just as in terrestrialecosystems, marine pelagic ecosystems alsotransport 5 to 20% of the carbon fixed in theeuphotic zone into the deeper ocean (Valiela1995). This process is called the biologicalpump. The carbon flux to depth correlatesclosely with primary production, so the envi-ronmental controls over NPP largely determinethe rate of carbon export to the deep ocean.This carbon export consists of particulate deadorganic matter (feces and dead cells) and thecarbonate exoskeletons that provide structuralrigidity to many marine organisms. Carbonateaccounts for about 25% of the biotically fixedcarbon that rains out of the euphotic zone(Houghton et al. 1996). The carbonates redis-solve under pressure as they sink to depth.Over decades to centuries, some of this carbonin deep waters recirculates to the surfacethrough upwelling and mixing. This long-termcirculation pattern will cause the effects of the current increase in atmospheric CO2 toinfluence marine biogeochemistry for centuriesafter its impacts are felt in terrestrial ecosys-

Her

bivo

ry (

g C

m-2

yr-1

)[lo

g sc

ale]

Net primary production (g C m-2 yr-1)[log scale]

Aquatic

Terrestrial

5 1010

-2

1

102

104

102

103

Figure 10.8. Comparative productivity and her-bivory rates between aquatic and terrestrial ecosys-tems. (Redrawn with permission from Nature; Cyrand Pace 1993.)

Oceans 235

tems. The net effect of the biological pump is tomove carbon from the atmosphere to the deepwaters and to ocean sediments. Carbon accu-mulation in midocean sediments is slow (about0.01% of NPP) because most decomposition

occurs in the water column before organicmatter reaches the sediments and because these well-oxygenated sediments support decom-position of much of the remaining carbon(Valiela 1995).

CO2

Navail

Navail

CO2

Carbon (C) fluxes

Nutrient (N) fluxesAtmosphere

Euphotic zone

Primaryproducers

Grazers,predators,

and viruses

POCPON

Decomposers

Water-columndecomposition

Biologicalpump

Deepocean

SedimentsBenthic

decomposition

Deep-ocean transport

Upwellingand mixing

DOCDON

CO2HCO3

-

Terrestrialinputs

Figure 10.9. Major pools and net fluxes of carbon(C) and nitrogen (N) in the ocean. Phosphorus andother essential nutrients cycle in patterns similar tothat shown for nitrogen. CO2 in the euphotic zoneequilibrates with bicarbonate in ocean water andwith CO2 in the atmosphere. CO2 is depleted by pho-tosynthesis by primary producers and is replenishedby respiration of organisms and by upwelling andmixing from depth. Grazers consume primary pro-ducers and bacteria and are eaten by other animalsand lysed by viruses. Each of these organismsreleases dissolved and particulate forms of carbonand nitrogen (DOC, DON; POC, PON).Animals anddecomposers also release available nitrogen (Navail).

DOC is consumed by bacteria, and available nutri-ents are absorbed by primary producers. Particulatecarbon and nutrients produced by feces and deadorganisms sink from the euphotic zone toward thesediments; as they sink, they decompose, releasingCO2 and available nutrients. Benthic decompositionalso releases CO2 and available nutrients. Bottomwaters, which are relatively rich in CO2 and availablenutrients, eventually return to the surface throughmixing and upwelling; this augments the supply ofavailable nutrients in the euphotic zone. DOC,dissolved organic carbon; DON, dissolved organicnitrogen; POC, particulate organic carbon; PON,particulate organic nitrogen.


The biological pump that transports carbonto depth carries with it the nutrients containedin dead organic matter. Decomposition contin-ues as particles sink, so much of the decompo-sition occurs in the water column rather than inthe sediments, particularly in the deep oceans.The rapid (about weekly) turnover of carbonand nutrients in phytoplankton in the euphoticzone (Falkowski et al. 1999) makes these nutri-ents vulnerable to loss from the ecosystem andcontributes to the relatively open nutrientcycles of pelagic ecosystems. The longer-livedand larger primary producers on land can storeand internally recycle nutrients for years. Thisreduces the proportion of nutrients that areannually cycled and contributes to the tightnessof terrestrial nutrient cycles.

Benthic decomposition is more important on continental shelves than in the deep oceanbecause the coastal pelagic system is more pro-ductive, generating more detritus. In addition,the dead organic matter has less time to decom-pose before it reaches the sediments. Hereoxygen consumption by decomposers depletesthe oxygen enough that decompositionbecomes oxygen limited, and organic matteraccumulates or becomes a carbon source formethanogens and denitrifiers.

Lakes

Lakes consist of a range of ecosystem types,from pelagic systems to wetlands dominated byvascular plants (Wetzel 2001). The centers ofdeep lakes are structurally similar to marineecosystems with discrete pelagic and benthicsystems; phytoplankton are the major primaryproducers, and zooplankton are the major her-bivores. The littoral zone of lakes generallyexperiences less disturbance from wave actionand currents than in marine systems.This allowsmats of algae to grow directly on lake sedi-ments, even where light is only 0.1% of thatpresent at the surface. The littoral zone of lakesoften has rooted vascular plants, whose leavesextend above the water surface and shade thewater, reducing the light available to phyto-plankton and benthic algae. Floating aquaticplants like water lilies cause a similar reduction

in light availability in the water column. Manylakeshore ecosystems and salt marshes, theirmarine equivalent, are structurally and func-tionally similar to terrestrial wetland ecosys-tems. There is therefore a continuum betweenthe structural and functional properties ofaquatic and terrestrial ecosystems.

The origin of lakes strongly affects theirstructure and functioning (Lodge 2001). Glaciallakes are abundant in young landscapes at highlatitudes and altitudes. They are frequentlyinterconnected with other lakes by short streamsegments and have a low degree of endemism.Rivers create lakes in several ways, includingisolation of former river channels (oxbowlakes) and periodically inundated swamps andfloodplains such as the Pantanal of Bolivia andBrazil. These lakes are generally shallow andoccasionally reconnect with adjacent riversduring floods. Tectonic lakes form along faults.They are often large, deep, and isolated fromone another, providing an environment for sub-stantial diversification, such as in the rift lakesin eastern Africa and Lake Baikal in Siberia.These large tectonically derived lakes harbormost of the endemic freshwater organisms.Over 80% of the open water animals of LakeBaikal, for example, are endemic (Burgis andMorris 1987). Other lakes form in volcaniccraters, by damming of rivers, and otherprocesses.

Controls over NPP

Photosynthesis in fresh-water ecosystems isseldom carbon limited, just as in the ocean.Groundwater entering fresh-water ecosystemsis supersaturated with CO2 derived from rootand microbial respiration in soil (Kling et al.1991). Most streams, rivers, and oligotrophiclakes are net sources of CO2 to the atmospherebecause the rates of water and CO2 input fromgroundwater generally exceed the capacity ofprimary producers to use the CO2 (Cole et al.1994, Hope et al. 1994). Eutrophic lakes withtheir high algal biomass have a greater demandfor CO2 to support photosynthesis than do olig-otrophic systems, but their organic accumula-tion and high decomposition rate in sedimentsprovide a large CO2 input to the water column.

Lakes 237

This creates a strong vertical gradient in CO2,with CO2 concentration being drawn down atthe surface, leading to CO2 absorption from theatmosphere during the day (Carpenter et al.2001). Some fresh-water vascular plants such as Isoetes use crassulacean acid metabolism(CAM) photosynthesis to acquire CO2 at nightand refix it by photosynthesis during the day(Keeley 1990). Other fresh-water vascularplants transport CO2 from the roots to thecanopy to supplement CO2 supplied from thewater column.

Vertical mixing is important in lakes, just as in marine pelagic ecosystems. Lake mixingoccurs not only by wave action, as in the ocean,but also by lake turnover in most temperate andhigh-latitude lakes (Wetzel 2001). In autumn,the surface waters cool to 4°C, the temperatureat which water is most dense. Once the surfacewaters cool to the point that water temperatureis similar from top to bottom, the water columnis readily mixed by wind. This causes surfacewaters to sink and brings nutrient-rich bottomwaters to the surface. Turnover also occurs inspring, when surface waters warm to 4°C,leading to a spring bloom in production. Whenlakes do not turn over, oxygen becomesdepleted at depth, leading to greater prevalenceof anaerobic conditions.Warm-climate lakes donot experience this seasonal lake turnover ifthe surface waters remain much warmer andless dense than deep water throughout the year.

Nutrients, rather than light, water, or CO2,are the resources that most consistently limitthe productivity of aquatic ecosystems. BothN:P ratios in algae and experimental nutrientadditions show that phosphorus limits algalproduction in the majority of unpolluted lakes,whereas nitrogen is the most common limitingelement in coastal marine and salt marshecosystems (Fig. 10.7) (Schindler 1977, Valiela1995). Why should nitrogen be the limitingelement in temperate terrestrial ecosystems butphosphorus the limiting element in lakes thatare embedded within this terrestrial matrix? Atleast two factors resolve this apparent paradox.The low mobility of phosphorus compared tonitrogen in soils retains phosphorus more effec-tively than nitrogen in terrestrial systems. Inaddition, lakes that receive large phosphorus

inputs from pollution or other sources gener-ally support the growth of nitrogen-fixing phytoplankton, such as cyanobacteria. Thesenitrogen fixers have a competitive advantageover nonfixers when nitrogen is scarce andphosphorus is available. Lakes therefore addtheir own nitrogen, whenever the phosphorus issufficient to support nitrogen fixation. Nitrogenfixation in the surface water of lakes is seldomlimited by light, as it may be in terrestrial andsome stream ecosystems. For these reasons,lakes are seldom nitrogen limited. Nitrate con-centrations are typically an order of magnitudehigher in lake than in ocean water (Valiela1995), again indicating the generally greateravailability of nitrogen than of phosphorus inlakes.

Nutrient inputs to lakes from streams,groundwater, and atmospheric depositionstrongly influence lake biogeochemistry. Lakesare generally small aquatic patches in a ter-restrial matrix; they are therefore strongly influenced by inputs of macronutrients andbase cations from groundwater and streams(Schindler 1978). The granitic bedrock of theCanadian Shield, from which soils wereremoved by continental glaciers during thePleistocene, for example, have low rates ofnutrient input from watersheds to lakes. Thestrong nutrient limitation of many of theselakes makes them vulnerable to change inresponse to nutrient inputs from agriculture oracid rain (Driscoll et al. 2001). Trout and othertop predators in oligotrophic lakes may requiredecades to reach a large size, whereas this mayoccur in a few months or years in eutrophiclakes.

Anthropogenic addition of nutrients to lakesfrequently causes eutrophication, a nutrient-induced increase in lake productivity. Eutroph-ication radically alters ecosystem structure andfunctioning. Increased algal biomass reduceswater clarity, thereby reducing the depth of theeuphotic zone. This in turn reduces the oxygenavailable at depth. The increased productivityalso increases the demand for oxygen tosupport the decomposition of the large detritalinputs. If mixing is insufficient to provideoxygen at depth, the deeper waters no longersupport fish and other oxygen-requiring het-


erotrophs. This situation is particularly severein winter, when low temperature limits oxygenproduction from photosynthesis. In ice-coveredlakes, ice and snow reduce light inputs thatdrive photosynthesis (providing oxygen) andprevent the surface mixing of oxygen into thelake. Lakes in which the entire water columnbecomes anaerobic during winter do notsupport fish. Even during summer, the accumu-lation of algal detritus at times of low surfacemixing can deplete oxygen from the watercolumn, leading to high fish mortality.


Carbon and nutrient cycling processes in lakesare similar to those described for the ocean.Phytoplankton account for most primary pro-duction in large lakes. As in the ocean, mostphytoplankton are grazed by zooplankton, sophytoplankton biomass is relatively low. Lakeswithout fish have large-bodied zooplanktonlike Daphnia that are efficient feeders onplankton and suspended organic matter. Whenfish are present, however, they reduce popula-tions of large-bodied zooplankton and proba-bly reduce the efficiency of energy transfer upthe food chain (Brooks and Dodson 1965). Asin the ocean, bacterial production is substantialin lakes, and the bacteria are an important com-ponent of the pelagic food web. Terrestrial dis-solved and particulate organic matter is animportant substrate for bacterial production insome lakes. Some of this terrestrial organicmatter may be more recalcitrant than the bac-terial substrates in the oceans and may be con-sumed more slowly by bacteria or accumulatein sediments. Decomposition in the sedimentstends to be more important in lakes than in the ocean because more detritus reaches thebottom before it decomposes, so there is oftena well-developed benthic food web similar tothat in coastal sediments. Some of the nutrientsreleased by decomposition return to the watercolumn and are mixed to the surface by waveaction and lake turnover.

The species composition of lakes stronglyinfluences their physical properties and biogeo-chemistry. Inadvertent experiments in whichfishermen or management agencies have intro-

duced fish or benthic organisms to lakes haveprovided a wealth of evidence that species traits strongly affect the functioning of aquaticecosystems (Spencer et al. 1991). In many lakes, the abundance of a top predator altersthe abundance of their prey and indirectly theabundance of phytoplankton (see Chapter 11).Changes in benthic fauna can have equallylarge impacts. Introduction of the zebra musselto the United States, for example, has displacednative mussels from many rivers and streams.The zebra mussel is a more effective filterfeeder than their native counterparts, filteringfrom 10 to 100% of the water column per day(Strayer et al. 1999). The resulting decrease indensity of phytoplankton and other edible particles reduced zooplankton abundance andshifted energy flow from the water column tothe sediments.

Streams and Rivers

The structure of stream and river ecosystemsdepends on stream width and flow rate. Thephysical environment and therefore the bioticstructure of stream ecosystems are dramaticallydifferent from those of lakes or the open ocean.Water is constantly moving downstream acrossthe riverbed, bringing in new material fromupstream and sweeping away anything that isnot attached to the substrate or able to swimvigorously. Phytoplankton are therefore unim-portant in streams, except in slow-moving orpolluted rivers.The major primary producers ofrapidly moving streams are periphyton, algaethat attach to stable surfaces such as rocks andvascular plants. The slippery surfaces of rocksin a riverbed consist of periphyton and associ-ated bacteria in a polysaccharide matrix.Submerged or emergent vascular plants andbenthic mats become relatively more importantin slow-moving stretches of the river. Within agiven stretch of river, alternating pools andriffles differ in flow rate and ecosystem struc-ture. Seasonal changes in discharge often radically alter the flow regime and thereforestructure of these ecosystems. Desert streams,for example, have flash floods after intenserains but may have no surface flow during dry

Streams and Rivers 239

times of year (Fisher et al. 1998). Other streamshave discharge peaks associated with snowmelt. Large rivers may overflow their banksonto a floodplain during periods of high flow.Many tropical rivers flood annually, so flood-plains alternate between being terrestrial andaquatic habitats.

The river continuum concept describes anidealized transition in ecosystem structure andfunctioning from narrow headwater streams tobroad rivers (Fig. 10.10) (Vannote et al. 1980).Headwater streams are often shaded by terrestrial vegetation. These plants reduce lightavailability to aquatic primary producers andprovide most of the organic matter input to the stream. Leaves and wood that fall into the

stream are colonized by aquatic fungi and to alesser extent by bacteria (Moss 1998). Theresulting leaf packs that accumulate behindrocks, logs, and other obstructions are con-sumed by invertebrate shredders that breakleaves and other detritus into pieces and digestthe microbial jam on the surface of these parti-cles, just as occurs in the soil (Wagener et al.1998). This creates fresh surfaces for microbialattack and produces feces and other fine mate-rial that is carried downstream. Some of the fineparticles are consumed in suspension by filterfeeders like black fly larvae or from benthicsediments by collectors like oligochaete worms.The abundance of algae and their grazers islimited in headwater streams by low light avail-

Figure 10.10. The river continuum concept(Vannote et al. 1980). Headwater streams have littlein-stream production (P), so respiration (R) bydecomposers and animals greatly exceeds produc-tion. Coarse particulate organic matter (CPOM)dominates the detrital pool. Shredders and collectorsare the dominant invertebrates. In middle sections ofrivers, more light is available, and in-stream produc-

tion exceeds respiration. Fine particulate organicmatter (FPOM) is the dominant form of organicmatter, and collectors and grazers are the dominantorganisms. Large rivers accumulate considerableorganic-rich sediments dominated by collectorsfeeding on FPOM from upstream. Respiration bydetrital organisms exceeds primary production.


ability. As headwater streams merge to formbroader streams, the greater light availabilitysupports more in-stream production, and theinput of terrestrial detritus contributes pro-portionately less to stream energetics. Thiscoincides with a change in the invertebratecommunity from one dominated by shreddersto one dominated by collectors and grazers(Fig. 10.3). These middle reaches of rivers aretypically less steep than headwaters and beginto accumulate sediments from upstreamerosion. These sediments support rooted vascu-lar plants and a benthic detrital community ofcollectors. The largest downstream reaches ofrivers typically have a sediment bed and aredominated by collectors that live in the sedi-ments. These large rivers may have submergedor emergent vascular plants, depending on thestability of the flow regime. There is a gradualincrease in fish diversity from headwaterstreams to large rivers, whereas the diversity ofbenthic invertebrates is generally greatest inmiddle reaches of rivers (Poff et al. 2001).

There is massive variation among streamsand rivers in their structure and functioning,just as in terrestrial and marine ecosystems.Theriver continuum concept provides a frameworkfor predicting patterns of variation within aregion but does not capture the large variationdue to substrate and climate. Nutrient-poorregions of the tropics and boreal peatlands, forexample, have large inputs of dissolved organiccarbon (DOC) leading to black-water rivers.White-water rivers result from inputs of silt and other mineral particulates from glacial andagricultural erosion. Clear-water rivers lackthese dissolved and suspended materials. Islandstreams are much shorter than the large riversthat drain the interiors of continents. Lakes andimpoundments on rivers create abrupt shifts inhabitats, food resources, and biota, punctuatingthe gradual changes of the river continuum(Ward and Stanford 1983).


Stream productivity is governed by its interfacewith terrestrial ecosystems (Hynes 1975).Terrestrial ecosystems influence stream pro-ductivity directly through the input of detritus

that fuels the detritus-based food chain andindirectly by determining the light environmentthat supports in-stream production. In forestheadwater streams, the dominant energy inputis terrestrial detritus that enters as coarseparticulate organic matter (CPOM) (Fig. 10.10).This includes leaves, wood, and other materiallarger than 1mm diameter. In forests, there isrelatively little algal production in headwaterstreams because of low light availability.Algal production becomes proportionately moreimportant in ecosystems with low canopy cover,such as in grasslands, tundras, and deserts. Fineparticulate organic matter (FPOM) comes pri-marily from within the stream through the pro-cessing of CPOM by shredders, the abrasion ofperiphyton from rocks, and other processes.About a third of the leaf material consumed by shredders, for example, is released into thestream as FPOM (Giller and Malmqvist 1998). The third major organic carbon input tostreams comes as dissolved organic carbonfrom terrestrial groundwater. DOC is thelargest pool of organic carbon in most streams.In tropical black-water rivers and boreal peat-lands, this carbon source to streams is particu-larly large and/or persistent. DOC inputs tostreams can be an important energy source ifthe compounds are readily assimilated andmetabolized by microbes. Tannins and otherrecalcitrant substances, however, are processedslowly in streams. Headwater streams are dom-inated by a detritus-based food chain, includingfungi, shredders, and their predators. Het-erotrophic respiration therefore considerablyexceeds photosynthesis. Downstream, whererivers are wide enough to allow substantial light input, photosynthesis may be similar to or exceed heterotrophic respiration (Vannote et al. 1980). In these middle sections of rivers,heterotrophic respiration is supported by amixture of FPOM imported from upstream,terrestrial inputs of CPOM (litter), DOC, andalgal production. In large rivers with large sedi-ment loads, water clarity may limit algal produc-tion, and detrital processing again dominates.

The frequency and magnitude of nutrientlimitation to algal production in streams andrivers are more variable than in lakes (Newbold1992). Many streams, particularly headwater

Streams and Rivers 241

streams, are not strongly nutrient limited, inpart because turbulence reduces diffusion limitation. In addition, uptake of nutrients bystream organisms does not influence the supplyof nutrients from upstream (Newbold 1992).The relative importance of nitrogen and phosphorus limitation varies among streams,depending on watershed parent material, land-scape age, and land use. Phosphorus limitationof stream production, for example, is morecommon in the eastern United States, wherethe parent material is relatively old and weath-ered, than on younger parent materials, wherephosphorus inputs are larger and nitrogen ismore likely to limit production (Home andGoldman 1994).

There is a strong interaction between top–down and bottom–up controls over primaryproduction in streams. Nutritional controls overthe energy available to support higher trophiclevels is generally the dominant control overstream productivity, but the types of predatorspresent strongly influence the pathway ofenergy flow, just as in lakes (see Chapter 12).

Carbon and nutrients spiral down streamsand rivers and the groundwater beneath them,rather than exchange vertically with the atmos-phere and groundwater. Streams are notpassive channels that carry materials from landto the ocean. The streams and their riparianzones process much of the material that entersthem. The strong directional flow of water instreams and rivers carries the resulting prod-ucts downstream, where they are repeatedlyreprocessed in successive stream sections.Energy and nutrients therefore spiral downstreams, rather than cycle vertically as they tendto do in most terrestrial ecosystems (Fisher etal. 1998). This leads to open patterns of nutri-ent cycling, in which the lateral transfers aremuch greater than the internal recycling (Gillerand Malmqvist 1998). Stream productivitytherefore depends highly on regular subsidiesfrom the surrounding terrestrial matrix and isquite sensitive to changes in these inputs due topollution or land use change. The spiralinglength of a stream is the average horizontal dis-tance between successive uptake events. Itdepends on the turnover length (the down-stream distance moved while an element is in

organisms) and the uptake length (the averagedistance that an atom moves from the time it isreleased until it is absorbed again). A repre-sentative spiraling length of a woodland streamis about 200m. Of this distance, about 10%occurs as microorganisms flow downstreamattached to CPOM and FPOM, 1% as con-sumers move downstream, and the remaining89% after release of the nutrient by mineral-ization (Giller and Malmqvist 1998). A unit ofnutrient therefore spends most of its time withrelatively little movement, but moves rapidlyonce it is mineralized and soluble in the water.Spiraling is therefore not a gradual process butoccurs in pulses. The patterns of drift of streaminvertebrates is consistent with these general-izations. Invertebrates drift downstream whenthey are dislodged from substrates or disperse.Drift is a an important food source for fish butrepresents only about 0.01% of the inverte-brate biomass of stream at any point in time. Inother words, stream invertebrates are so effec-tive in remaining attached to their substratesthat carbon and nutrients spiral downstreamprimarily in the dissolved phase.

Headwater streams less than 10m in widthare particularly important in nutrient process-ing because they are the immediate recipient ofmost terrestrial inputs and account for up to85% of the stream length within most drainagenetworks (Peterson et al. 2001). Small streamsare particularly effective in cycling nitrogen(have shorter uptake lengths) because theirshallow depths and high surface to volumeratios enhance nitrogen absorption by algaeand bacteria that are attached to rocks and sed-iments. Uptake lengths for ammonium rangefrom 10 to 1000m and increase exponentiallywith increases in stream discharge (Peterson etal. 2001). Streams generally have much highernitrate than ammonium concentrations, evenwhen they occur in ammonium-dominatedwatersheds, because of preferential uptake ofammonium over nitrate by stream organismsand because nitrification rates are frequentlyhigh in riparian zones and in streams. For thesereasons, the uptake length of nitrate is about10-fold greater than that of ammonium. Thusnitrate is much more mobile than ammoniumin streams, as on land, but for different reasons.


Because of high rates of nitrogen uptake andcycling by the streambed, most nitrogen thatenters streams from terrestrial ecosystems isabsorbed within minutes to hours and isprocessed multiple times before it reaches theocean (Peterson et al. 2001). Nitrification anddenitrification release to the atmosphere substantial proportions of the nitrogen thatenters streams. Consequently, nitrogen inputsto oceans are much less than the quantity thatenters the stream system.

The horizontal flow of carbon and nutrientsin streams is similar to the vertical movementof elements through the soil on land but occursover much larger distances. The basic steps indecomposition are identical on land and inaquatic ecosystems (Valiela 1995, Wagener etal. 1998). These steps include leaching ofsoluble materials from detritus, fragmentationof litter into small particles by invertebrates,and microbial decomposition of labile andrecalcitrant substrates (see Chapter 7). Onland, these processes begin at the soil surface,and small particles of organic matter movedownward in the soil profile due to mixing bysoil invertebrates, burial by new litter, and other processes. In stream ecosystems the sameprocesses occur, but materials move horizon-tally tens of meters in the process.

Rivers and streams have a belowgroundcomponent that is just as poorly understood asthe soils of terrestrial ecosystems. The hypor-rheic zone is the zone of groundwater thatmoves downstream within the streambed. Sub-stantial decomposition occurs in the hyporrheiczone, releasing nutrients that support in-streamalgal production. In intermittent streams, thehyporrheic zone is all that remains of thestream during dry periods. Water moves moreslowly and therefore has a shorter processinglength in the hyporrheic zone than in thestream channel, so the spiraling length is muchshorter (Fisher et al. 1998).

Summary

The major differences between aquatic and ter-restrial ecosystems result from the differencesin the surrounding medium.The greater density

of water than air supports photosyntheticorganisms with minimal investment in struc-tural support. Aquatic ecosystems thereforehave negligible plant biomass but account fornearly half of Earth’s NPP. Slow diffusion ofgases in water cause oxygen to be much morestrongly limiting in aquatic than terrestrialenvironments, particularly in sediments. Lightand nutrients are the resources that most fre-quently limit aquatic production. NPP is largelyrestricted to the uppermost part of the watercolumn, where there is sufficient light to drivephotosynthesis. Within this zone, nutrients gen-erally limit production. The magnitude andnature of nutrient limitation generally dependson nutrient inputs from terrestrial ecosystems.The open ocean, which receives least terrestrialinputs is strongly nutrient limited, especially insubtropical oceans where wind-driven mixingand upwelling are minimal.

Review Questions

1. How do water and air differ in density, vis-cosity, and rates of gas diffusion? How dothese physical differences give rise to thelarge differences in structure and function-ing that exist between terrestrial and aquaticecosystems?

2. How do the controls over carbon and lightavailability differ between terrestrial andaquatic ecosystems?

3. How do the controls over nutrient availabil-ity and nutrient cycling differ between ter-restrial and aquatic ecosystems?

4. What controls nutrient availability in theopen ocean? How does this differ betweenthe open ocean and the coastal zone?Between the open ocean, lakes and streams?

5. Describe the functioning of the microbialloop and biological pump in marine pelagicecosystems. How does this differ fromprocesses occurring in terrestrial soils?

6. How do ecosystem processes (communitycomposition, patterns of production anddecomposition, etc.) change from a head-water stream down to the mouth of a largeriver?


7. How do ecosystem structure and functioningchange, when a stream is blocked by the con-struction of a dam?

8. How does the terrestrial matrix influencecarbon and nutrient cycling in lakes andsteams?

9. What determines the spiraling length ofnutrients in a stream or river? How mighthuman activities influence this spiralinglength?

Additional Reading

Falkowski, P.G., R.T. Barber, and V. Smetacek.1999. Biogeochemical controls and feedbacks on ocean primary production. Science 281:200–206.

Giller, P.S., and B. Malmqvist. 1998. The Biology ofStreams and Rivers. Oxford University Press,Oxford, UK.

Moss, B. 1998. Ecology of Fresh Waters: Man andMedium, Past to Future. 3rd ed. Blackwell Scien-tific, Oxford, UK.

Schindler, D.W. 1977. Evolution of phosphorus limi-tation in lakes. Science 195:260–267.

Valiela, I. 1995. Marine Ecological Processes.Springer-Verlag, New York.

Vannote, R.I., G.W. Minshall, K.W. Cummings, J.R.Sedell, and C.E. Cushing. 1980. The river contin-uum concept. Canadian Journal of Fisheries andAquatic Sciences 37:120–137.

Wagener, S.M., M.W. Oswood, and J.P. Schimel. 1998.Rivers and soils: Parallels in carbon and nutrientprocessing. BioScience 48:104–108.

Wetzel,R.G.2001.Limnology:Lake and River Ecosys-tems. 3rd ed.Academic Press, San Diego, CA.

Introduction

Although terrestrial animals consume a rela-tively small proportion of net primary produc-tion (NPP), they strongly affect energy flowand nutrient cycling. In earlier chapters weemphasized the interactions between plantsand soil microbes, because these two groupsdirectly account for about 90% of the energytransfers in most terrestrial ecosystems. Plantsuse solar energy to reduce CO2 to organicmatter, most of which senesces, dies, anddirectly enters the soil, where it is broken downby bacteria and fungi. Similarly, most nutrienttransfers in ecosystems involve uptake byplants and return to the soil as dead organicmatter, where nutrients are released and madeavailable by microbial mineralization. In mostterrestrial ecosystems the uncertainties in ourestimates of primary production and decompo-sition exceed the total energy transfers fromplants to animals. It is perhaps for this reasonthat terrestrial ecosystem ecologists have fre-quently ignored animals in classical studies ofproduction and biogeochemical cycles. Aquaticecologists, however, have been unable to ignoreanimals because most of the energy and nutri-ents are transferred from plants to animalsrather than directly from plants to dead organicmatter (see Chapter 10). Perhaps for this

reason aquatic ecosystem ecologists have gen-erally led the theoretical developments in thisaspect of ecosystem ecology.

Understanding the factors governing energyand nutrient transfer to animals has societalimplications. Most human populations dependheavily on high-protein foods derived fromanimals. The exponentially increasing humanpopulation requires more food in a worldwhere many people already face an inadequatefood supply. An ecologically viable strategy forefficiently providing food to feed the growinghuman population requires a good understand-ing of the ecological principles regulating the efficiency with which plants and animalssupport their growth and maintenance.

Overview

Energy transfers define the trophic structure ofecosystems. The simplest way to visualize theenergetic interactions among organisms in anecosystem is to trace the fate of a packet ofenergy from the time it enters the ecosystemuntil it leaves—without worrying about theidentity of the organisms involved (Lindeman1942). Trophic transfers involve the feeding ofone organism on another or on dead organicmatter. Plants are called primary producers or

11Trophic Dynamics

Trophic dynamics govern the movement of carbon, nutrients, and energy amongorganisms in an ecosystem. This chapter describes the controls over trophic dynam-ics of ecosystems.

244

Overview 245

autotrophs because they convert CO2, water,and solar energy into biomass (see Chapter 5).Heterotrophs are organisms that derive theirenergy by eating live or dead organic matter.Heterotrophs function as part of two majortrophic pathways, one that is based on liveplants (the plant-based trophic system) andanother that is based on dead organic matter(the detritus-based trophic system). The secondtrophic system is less immediately obvious toaboveground animals like us and is often over-looked, even though it usually accounts formost of the energy transfers through animals.Consumers (also termed secondary producers)are organisms that eat other live organisms.These include herbivores, which eat plants,microbivores, which eat bacteria and fungi,and carnivores, which eat animals. A group oforganisms that are linked together by the lineartransfer of energy and nutrients from oneorganism to another are referred to as a foodchain. Grass, grasshoppers, and birds, for

example, form a food chain. Those organismsthat obtain their energy with the same numberof transfers from plants or detritus belong tothe same trophic level. Thus plants constitutethe first trophic level; herbivores, the second;primary carnivores, the third; secondary carni-vores that eat mainly primary carnivores, thefourth, etc., in a plant-based trophic system(Lindeman 1942, Odum 1959). Similarly, in thedetritus-based trophic system, bacteria andfungi directly break down dead soil organicmatter and absorb the breakdown products for their own growth and maintenance. Theseprimary detritivores are the first trophic level inthe detritus-based food chain and are fed on byanimals in a series of trophic levels analogousto those in the plant-based trophic system (Fig.11.1).

Although food chains are an easy way tovisualize the trophic dynamics of an ecosystem,they are a gross oversimplification for organ-isms that eat more than one kind of food.

Roots

Phytophagousnematodes

Saprophyticfungi

Collembolans

Detritus

Bacteria

Cryptostigmatidmites

Flagellates

Bacteriophagousmites

Amoebae

Enchytraeids

Fungivorousnematodes

Predaceousnematodes

Noncrypto-stigmatid mites

Nematode -feeding mites

Predaceouscollembolans

Predaceousmites

Earthworms

Bacteriophagousnematodes

Figure 11.1. Pattern of energy flow through below-ground portions of a grassland food web. Food websconsist of many interconnecting food chains, includ-

ing plant-based (heavy solid lines) and detritus-based (dashed lines). (Modified with permissionfrom Biology and Fertility of Soils; Hunt et al. 1987.)

246 11. Trophic Dynamics

People, for example, eat food from severaltrophic levels, including plants (first trophiclevel), cows (second trophic level), fish (secondand higher trophic levels), and mushrooms(detritivores). Similarly, many birds and somemammals consume both herbivorous insects(plant-based trophic system) and worms (detritus-based trophic system). The actualenergy transfers that occur in all ecosystems aretherefore complex food webs (Fig. 11.1).We cantrace the energy transfers through these foodwebs only by knowing the contribution of eachtrophic level to the diet of each animal in theecosystem.Although the structure of food webshas been partially described for many ecosys-tems (Pimm 1984), the quantitative patterns ofenergy flow through food webs are generallypoorly known, especially for detritus-basedfood webs.

The regulation of energy and nutrient flowthrough food webs is complicated and variesconsiderably among ecosystems. There are two idealized patterns, however, that bracketthe range of possible controls. The availabilityof food at the base of the food chain (eitherplants or detritus) limits the production ofupper trophic levels through bottom–up con-trols. Predators that regulate the abundance oftheir prey exert top–down control on foodwebs. Most trophic systems exhibit some com-bination of bottom–up and top–down controls,

and the relative importance of these controlsvaries both temporally and spatially (Polis1999).

Plant-Based Trophic Systems

Controls over Energy Flow Through Ecosystems

Plant production places an upper limit to theenergy flow through plant-based food webs.The energy consumed by animals in the plant-based trophic system, on average, cannotexceed the energy that initially enters theecosystem through primary production. Thisconstitutes a fundamental constraint to theanimal production that an ecosystem cansupport. When all terrestrial ecosystems arecompared, herbivore production tends toincrease with increasing primary production(Fig. 11.2). This relationship between primaryand secondary production is particularly strongwhen comparisons are made among similartypes of ecosystems. In the grasslands ofArgentina, for example, the biomass of mam-malian herbivores increases with increasingaboveground production along a gradient ofwater availability in both natural and managedgrasslands (Fig. 11.3) (Osterheld et al. 1992). Inthe Serengetti grasslands of Africa, the largeherds of ungulates also acquire most of their

D

D

D

D

T

T

T

T

T GG

GG

GGG

GG

S

SS

SS

S

S

S

S

SSS

S

F

F

F

F

0.001

0.01

0.1

1.0

10

100

1000

100 1,000 10,000 100,000

Sec

onda

ry p

rodu

ctio

n(k

J m

-2 y

r-1)

[log

scal

e]

D = DesertT = TundraG = GrasslandS = SavannaF = Forest

Aboveground NPP (kJ m-2 yr-1) [log scale]

Figure 11.2. Log–log relationshipbetween aboveground NPP and herbi-vore production. Note that 1 g of ash-free biomass is equivalent to 20 kJ ofenergy. Production of aboveground her-bivores correlates with abovegroundNPP across a wide range of ecosystems.(Redrawn with permission from Nature;McNaughton et al. 1989.)

Plant-Based Trophic Systems 247

food in the more productive grasslands (Sinclair 1979, McNaughton 1985). Similarly,productive forests generally have greater insectherbivory than do nonproductive forests. Whenforests are fertilized to increase their produc-tion, this usually increases the feeding by her-bivores (Niemelä et al. 2001). The correlationbetween primary production and animal pro-duction within an ecosystem type is the basis ofthe world’s large fisheries. Upwelling of nutri-ent-rich bottom waters supports a high pro-duction of algae, zooplankton, and fish (seeChapter 10). At the opposite extreme, olig-otrophic (nutrient-poor) lakes on the CanadianShield, an area whose soils were scraped awayby continental glaciers during the Pleistocene,have low production of algae, zooplankton,and fish.

Subsidies are an important exception to thegeneralization that NPP within an ecosystemconstrains secondary production. Most of the

energetic base for headwater streams in forests,for example, comes from inputs of terrestriallitter. This allochthonous input (i.e., an inputfrom outside the stream ecosystem) constitutesa subsidy that, together with autochthonousproduction (i.e., production occurring withinthe stream), provides the energy that supportsaquatic food webs (see Chapter 10). Terrestrialfood webs near oceans, rivers, and lakes are fre-quently subsidized by inputs of aquatic energy,for example, when birds or bears feed on fish orwhen spiders feed on marine detritus (Polis andHurd 1996). Agricultural production is gener-ally subsidized by inputs of nutrients, water,and/or fossil fuels (Schlesinger 1999).

Biome differences in herbivory reflect differ-ences in NPP and plant allocation to structure.The most dramatic differences in herbivoryamong ecosystem types are consequences ofvariation in plant allocation to physical support.Lakes, oceans, and many rivers and streams aredominated by algae that allocate most of theirenergy to cytoplasm rather than to celluloseand structural support. Most algal cells arereadily digested by zooplankton, so animals eata large proportion of primary production andconvert it into animal biomass. Even amongalgae, chlorophytes (naked green algae) aregenerally consumed more readily than algaethat produce a protective outer coating, such asdiatoms, dinoflagellates, and chrysophytes. Atthe opposite extreme, forests have a substantialproportion of production allocated to cellulose-and lignin-rich woody tissue that cannot bedirectly digested by animals. Some animals likeruminants (e.g., cows), caecal digesters (e.g.,rabbits), and some insects (e.g., termites)support symbiotic gut microbes capable of cellulose breakdown; these animals assimilatesome of the energy released by this microbialbreakdown.

Among terrestrial ecosystems, there is a 1000-fold variation in the quantity of plant biomass consumed by herbivores(McNaughton et al. 1989). Herbivores consumethe least biomass in unproductive ecosystemssuch as tundra (Fig. 11.4A). However, theenergy consumed by herbivores is quite vari-able within and among other biomes. Con-sumption by herbivores shows a much stronger

0

1

2

Her

bivo

re b

iom

ass

(log

kJ m

-2)

Aboveground NPP (log kJ m-2 yr-1)

3

-1

-2

-3

-42.6 2.8 3 3.2 3.4 3.6 3.8 4 4.2 4.4

Managed grassland

Natural grassland

Figure 11.3. Log–log relationship between mam-malian herbivore biomass and aboveground plantproduction in natural and managed grazing systemsof South America. Herbivore biomass increases withincreasing NPP. Animal biomass on the managedgrassland is 10-fold greater than on the natural grass-land at a given level of plant production, becausemanagers control predation, parasitism, and diseaseand provide supplemental drinking water and min-erals in managed systems. This difference in herbi-vore biomass between managed and unmanagedsystems indicates that NPP is not the only constrainton animal production. (Redrawn with permissionfrom Nature; Osterheld et al. 1992.)


relationship with production of edible tissue(e.g., leaves) (Fig. 11.4B) than with total aboveground NPP (Fig. 11.4A), because thewoody support structures produced by plants contribute relatively little to herbivore consumption.

Plant chemical and physical defenses againstherbivores reduce the proportion of energytransferred to herbivores in low-resource envi-ronments. It has been argued that predationrather than food availability must limit theabundance of herbivores because the world iscovered by green biomass that has not beeneaten by animals (Hairston et al. 1960). Not all green biomass, however, is sufficiently

digestible to serve as food. In low-resourcehabitats, plants have a low protein content, highconcentrations of chemical defenses againstherbivores, and (frequently) physical defensessuch as thorns. In African grasslands, forexample, fertile “sweet veldt” grasslandssupport a higher diversity and production ofherbivores than do the less fertile “sour veldt”grasslands. The same pattern is seen in tropicalforests, where there are higher levels of chemi-cal defense and lower levels of insect herbivoryin infertile than in fertile forests (McKey et al.1978).

Dry environments are also dominated byplants with low protein content and high levels

DT

TT

T

TG

G

GG G

GGG

G G

G G

G

GG

SS

SSSSS S

SSS

SSSS

SSSSS

SS SSS

FFFFF

F

F F F

F

F

F

F F

F

100,000

100,000

10,000

10,000

1,000

1,000

100

100

100,00010,0001,000100

10

1

D = DesertT = TundraG = GrasslandS = SavannaF = Forest

Con

sum

ptio

n by

her

bivo

res

(kJ

m-2

yr-1

) [lo

g sc

ale]

Foliage production (kJ m-2 yr-1) [log scale]

Aboveground NPP (kJ m-2 yr-1) [log scale]

G

GG

GG

GG

GG

GG

GG

G

D

D

DDT

TTT

TT

T

T

SSSSSS

SSS

SS

SS

S SSSSS

FF

F

FF

F

FF

FF

F

100,000

10,000

1000

100

10

1

Con

sum

ptio

n by

her

bivo

res

(kJ

m-2

yr-1

) [lo

g sc

ale]

A

B

Figure 11.4. Log–log relationshipbetween consumption by herbivoresand (A) aboveground NPP and (B)foliage production. Note that 1 g of ash-free biomass is equivalent to 20 kJof energy. Consumption by herbivores is more closely related to foliage pro-duction that to total aboveground NPP, because much of the abovegroundNPP is inedible to most herbivores.(Redrawn with permission from Nature;McNaughton et al. 1989.)


of plant defense. Most herbivory in deserts isconcentrated on seeds or on annual plants thatgerminate, grow, and reproduce within a fewweeks after the onset of rains rather than on thedominant perennial plants with their higherconcentrations of plant defenses. Thus, even indry habitats, herbivory is concentrated on thosesegments of the community that lack well-developed physical and chemical defenses.

Three factors govern the allocation todefense in plants: (1) genetic potential, (2) theenvironment in which a plant grows, and (3) the seasonal program of allocation. Ecosystem differences in plant defense are determinedmost strongly by species composition. Differentspecies in terrestrial and aquatic environmentsexhibit a wide range in both the type and quan-tity of defensive compounds produced. Terres-trial plants and marine kelps adapted tolow-resource environments generally producelong-lived tissues with high concentrations ofcarbon-based defense compounds (i.e., organiccompounds that contain no nitrogen, such astannins, resins, and essential oils) (see Chapter6). These compounds deter feeding by mostherbivores (Coley et al. 1985, Hay and Fenical1988). The tissue loss to herbivores is oftensimilar (1 to 10%) to the annual allocation toreproduction (i.e., the allocation that mostdirectly determines fitness). This suggests thatthere should be strong selection for effectivechemical defenses against herbivores. Whengenotypes of a species are compared, forexample, those individuals that allocate moststrongly to defense grow most slowly (Fig.11.5), suggesting a trade-off between alloca-tion to growth vs. defense (Coley 1986). Plantspecies typical of high-nitrogen environments,particularly nitrogen-fixing species, often pro-duce nitrogen-based defenses (i.e., organiccompounds containing nitrogen, such as alka-loids) that are toxic to generalist herbivores.Nitrogen-based defenses are well developed,for example, in terrestrial legumes and fresh-water cyanobacteria. Other species deter herbi-vores through production of sulfur-containingdefenses, accumulation of selenium or silica,etc.

Any given genotype is usually less palatablewhen grown in a low-resource instead of a high-

resource environment, due to a lower proteincontent and higher levels of plant defense.Under conditions of low nutrient or wateravailability, growth is constrained morestrongly than is photosynthesis, so carbon tendsto accumulate (Bryant et al. 1983) (see Chapter6). Under these circumstances carbon may beallocated to chemical defense with modest neg-ative impacts on growth rate. Finally, in a givenenvironment, plants vary seasonally in theirallocation to defense, with allocation to growthoccurring when conditions are favorable andallocation to tissue differentiation and defensewhen conditions deteriorate (Lorio 1986,Herms and Mattson 1992).

All three of these sources of variation in allo-cation to plant defense (genetics, environment,and seasonality) cause defensive measures tobe most strongly expressed under conditions of low resource supply. This is the situation inwhich carbon is unlikely to be the resource thatmost directly limits growth (see Chapter 6).Themost important effect of the plant secondarymetabolites on herbivores is their toxicity to theanimal, although toxicity to gut microbes andtheir tendency to bind with proteins (makingprotein less available to the animal) may alsobe important under some circumstances.

Herbivores magnify differences amongecosystems in production and energy flow

Num

ber

of le

aves

pro

duce

d

Tannin (mg g-1)

60

60

50

50

40

40

30

2010 30

Figure 11.5. Relationship between rate of leaf pro-duction (an index of growth rate) and leaf tanninconcentration in the tropical tree Cecropia peltata.Note the negative relationship between investmentin defense and growth rate. (Redrawn with permis-sion from Oecologia; Coley 1986.)


through food webs. Dominance by plants withwell-developed plant defenses in low-resourceenvironments tends to reduce the frequency ofherbivory in these ecosystems because herbi-vores select patches in the landscape wherefood quality is relatively high (see Chapter 14).In addition, many plant defenses are toxic tosoil microbes, which reduces decompositionrates (see Chapter 7) and further reduces soil fertility in low-resource environments(Northup et al. 1995) (Fig. 11.6).

In high-resource environments, however,where plants are more productive and morepalatable, herbivores are more abundant. Theirfeeding results in a large input of availablenutrients in feces and urine, which short-circuitsdecomposition and nitrogen mineralization andenhances the production of these ecosystems(Ruess and McNaughton 1987). Plants in theseenvironments are frequently well adapted toherbivory. Grasslands with an evolutionaryhistory of intensive grazing, for example, areoften more productive when moderately grazedthan in the absence of grazers. In the absenceof grazers, species composition shifts to speciesthat are less productive and have lower litterquality (McNaughton 1979, Milchunas andLauenroth 1993, Hobbs 1996). Thus the neteffect of herbivores in unmanaged terrestrialecosystems is to magnify the differences innutrient availability and production compared

to regional patterns that simply reflect directeffects of parent material on nutrient supplyand NPP (Chapin 1991a, Hobbs 1996). Inmanaged ecosystems, there is often moregrazing than would occur naturally (Fig. 11.3).Overgrazing can reduce production and plantcover and increase soil erosion, leading to adecline in soil resources and the productivepotential of an ecosystem (Milchunas andLauenroth 1993).

In contrast to terrestrial ecosystems, naturalrates of herbivory in lakes sometimes reducerates of nutrient cycling. Early in the season,high rates of herbivory often remove the small,edible, rapidly recycled phytoplankton in favorof more defended cyanobacteria. This reducesherbivory, causing the cycling of nutrients fromprimary producers to herbivores to decline.

Ecological Efficiencies

Energy losses at each trophic transfer limit theproduction of higher trophic levels. Not all ofthe biomass that is produced at one trophiclevel is consumed at the next level. Moreover,only some of the consumed biomass is digestedand assimilated, and only some of the assimi-lated energy is converted into animal produc-tion (Fig. 11.7). Consequently, a relatively smallfraction (generally less than 1 to 25%) of theenergy available as food at one trophic level is

Rapiddecomposition

Plantdefense

Plantdefense

Slow growth

Low soilfertility

Slowdecomposition

High soilfertility

Rapid growth

Highgrazing

Nutrientsreturned in

feces and urine

Soil fertility gradient

Figure 11.6. Feedbacks by which grazing and plantdefense magnify differences among sites in soil fer-tility (Chapin 1991a). In infertile soils, herbivoryselects for plant defenses, which reduce litter quality,

decomposition, and nutrient supply rate. In fertilesoils, herbivory speeds the return of available nutri-ents to the soil.


converted into production at the next link in afood chain.This has profound consequences forthe trophic structure of ecosystems becauseeach link in the food chain has less energy avail-able to it than did the preceding trophic link. Inany plant-based trophic system, plants processthe largest quantity of energy, with progres-sively less energy processed by herbivores,primary carnivores, secondary carnivores, etc.This leads to an inevitable energy pyramid(Elton 1927) and various ecological efficiencies(Lindeman 1942) that determine the quantityof energy transferred between successive tro-phic levels (Fig. 11.8). The production at eachtrophic link (Prodn) depends on the productionat the preceding trophic level (Prodn-1) and thetrophic efficiency (Etroph) with which the pro-duction of the prey (Prodn-1) is converted intoproduction of consumers (Prodn).

(11.1)

The trophic efficiency of each link in a foodchain can be broken down into three ecological

Pr Pr PrPr

Prod od E od

ododn n n

n

n

= ¥ = ¥ ÊË

ˆ¯- -

-1 1

1troph

efficiencies (Fig. 11.7) related to the efficienciesof consumption (Econsump), assimilation (Eassim),and production (Eprod) (Lindeman 1942, Odum1959, Kozlovsky 1968).

(11.2)

In terrestrial ecosystems, the distribution ofbiomass among trophic levels can be visualizedas a biomass pyramid that is similar in structureto the energy pyramid, with greatest biomass in primary producers and progressively lessbiomass in higher trophic levels (Fig. 11.8). Thisoccurs for at least two reasons: (1) As describedearlier, the energy pyramid results in lessenergy available at each successive trophic link.(2) The large proportion of structural tissue interrestrial plants minimizes the proportion ofplant production that can be converted to sec-ondary production. The decrease in biomasswith successive links is most pronounced inforests, where the dominant plants are longlived and produce a large proportion of inedi-ble biomass. Biomass pyramids are less broadin grasslands where plants have a lower alloca-tion to woody structures, and there is a rela-

E E E Etroph consump assim prod= ¥ ¥

Consumption efficiency (Econsump) =In

Prodn-1

Assimilation efficiency (Eassim) =An

In

Production efficiency (Eprod) =Prodn

An

Trophic efficiency (Etroph) =Prodn

Prodn-1(Econsump) ¥ (Eassim) ¥ (Eprod) =

Detritus

2o

Production

RespirationPrimaryproduction

Notconsumed

Prodn-1

Feces

ProdnAn

In

Figure 11.7. Components of trophic effi-ciency, which is the product of consump-tion efficiency, assimilation efficiency,and production efficiency. Production effi-ciency is the proportion of primary pro-duction that is ingested (In) by animals.Assimilation efficiency is the proportion ofingested food that is assimilated into theblood stream (An). Production efficiency isthe proportion of assimilated energy thatis converted to animal production. Mostprimary production is not consumed byanimals and passes directly to the soil asdetritus. Of the plant material consumedby herbivores, most is transferred to thesoils as feces. Of the material assimilatedby animals, most supports the energeticdemands of growth and maintenance (respiration), and the remainder is con-verted to new animal biomass (secondaryproduction).


tively large biomass of herbivores and highertrophic levels.

In contrast to terrestrial ecosystems, fresh-water and marine pelagic ecosystems have lessbiomass of primary producers than of highertrophic levels, leading to an inverted biomasspyramid (Fig. 11.8). This difference in trophicstructure between terrestrial and pelagic eco-systems reflects fundamental differences inallocation and life history.The high buoyancy ofwater provides physical support for photosyn-thetic organisms, so algae do not require elab-orate physical support structures and are muchsmaller and more edible than their terrestrialcounterparts. Consequently, they are rapidlygrazed, and their biomass does not accumulate.In summary, terrestrial ecosystems are charac-

terized by large, long-lived plants, leading to a large plant biomass and relatively smallbiomass of higher trophic levels. Aquaticecosystems, in contrast, are characterized byrapidly reproducing plants that are smaller andmore short lived than organisms at highertrophic levels (see Fig. 10.1).

Regardless of the biomass distributionamong trophic levels, there must always bemore energy flow through the base of a trophicchain than at higher trophic levels. It is theenergy pyramid rather than the biomasspyramid that describes the fundamental ener-getic relationships among trophic levels,because energy is lost at each trophic transfer;so there must always be a decline in energyavailable at each successive trophic level.

Biomass

Primary producers

Secondary carnivores

Primary carnivores

Herbivores

Energy flow

Primary producers

Secondary carnivores

Primary carnivores

Herbivores

Terrestrial ecosystem Aquatic ecosystem

Terrestrial ecosystem Aquatic ecosystem

Figure 11.8. Pyramids of biomass and energy in aterrestrial and an aquatic food chain. The width ofeach box is proportional to its biomass or energycontent. Pyramids of energy are structurally similarin terrestrial and aquatic food chains because energy

is lost at each trophic transfer. Biomass pyramidsdiffer between terrestrial and aquatic food chainsbecause most plant biomass is not eaten on land,whereas most plant biomass (phytoplankton) iseaten and is short lived in aquatic ecosystems.


Consumption Efficiency

Consumption efficiency is determined primar-ily by food quality and secondarily by preda-tion. Consumption efficiency is the proportionof the production at one trophic level that isingested by the next trophic level (In) (Fig.11.7).

(11.3)

Unconsumed material eventually enters thedetritus-based food chain as dead organicmatter. On average, the quantity of food con-sumed by a given trophic level must be less thanthe production of the preceding trophic level,or the prey will be driven to extinction. In otherwords, the average consumption efficiency of atrophic link must be less than 100%. There maybe times when the consumption by one trophiclevel exceeds that in the preceding level. Mostvertebrate herbivores, for example, consumeplants during winter, when there is no plantproduction. This is, however, typically offset by other times, such as summer, when plantsusually produce much more biomass thananimals can consume. Situations in which con-sumption efficiency is greater than 100% forprolonged periods lead to dramatic ecosystemchanges. If predator control, for example, leadsto a large deer population that consumes moreplant biomass than is produced, the plantbiomass will be reduced, altering plant speciescomposition in ways that profoundly affect allecosystem processes (see Chapter 12) (Pastoret al. 1988, Kielland and Bryant 1998, Paine2000). Sometimes this occurs naturally. Someherbivores, such as beavers, typically overex-ploit their local food supply and move to newareas when their food is depleted.

The proportion of NPP consumed by herbi-vores varies at least 100-fold among ecosys-tems, from less than 1% to greater than 40%(Table 11.1). The major factor accounting forthis wide range in herbivore consumption effi-ciency is differences in plant allocation to struc-ture. Herbivore consumption efficiency isgenerally lowest in forests (less than 1 to 5%),where there is a large plant allocation to wood.Herbivore consumption efficiencies are higher

EIod

n

nconsump =

-Pr 1

in grasslands (10 to 60%), where most above-ground material is nonwoody, and highest (generally greater than 40%) in pelagic aquaticecosystems, where most plant (i.e., algal) bio-mass is cell contents rather than cell walls. Inthese ecosystems, more algal biomass is oftenconsumed by herbivores than dies and decom-poses; this pattern contributes to invertedbiomass pyramids (Fig. 11.8). In grasslands,consumption efficiencies are generally greaterfor ecosystems dominated by large mammals(25 to 50%) than those dominated by insectsand small mammals (5 to 15%) (Detling 1988).The toxic nature of some plant tissues (due topresence of plant secondary metabolites) andinaccessibility of other tissues (e.g., roots toaboveground herbivores) constrain the her-bivore consumption efficiency of terrestrialecosystems. Nematodes, one of the majorbelowground herbivores, consume 5 to 15% of belowground NPP in grasslands (Detling1988). Consumption efficiencies for below-ground herbivores are not well documented,so whole-ecosystem estimates of consumptionefficiencies almost always emphasize above-ground consumption. The highest consumptionefficiencies in terrestrial ecosystems are ongrazing lawns, such as those found in someAfrican savannas (McNaughton 1985) andarctic wetlands (Jefferies 1988). These highlyproductive grasslands are maintained as a lawnby repeated herbivore grazing. Nutrient inputsin urine and feces from these herbivores

Table 11.1. Consumption efficiency of the herbi-vore trophic level in selected ecosystem types.

Consumption efficiencya

Ecosystem type (% of aboveground NPP)

Oceans 60–99Managed rangelands 30–45African grasslands 28–60Herbaceous old fields 5–15

(1–7yr)Herbaceous old fields (30yr) 1.1Mature deciduous forests 1.5–2.5

a Terrestrial estimates emphasize consumption by above-ground herbivores and may not accurately reflect the totalecosystem-scale consumption efficiency.Data from Wiegert and Owen (1971) and Detling (1988).


promote rapid recycling of nutrients and there-fore are a key factor supporting the grasslands’high production (Ruess et al. 1989) (Fig. 11.6).

Consumption efficiencies of carnivores areoften higher than those of herbivores, rangingfrom 5 to 100%. Vertebrate predators that feedon vertebrate prey, for example, often have aconsumption efficiency greater than 50%, indi-cating that more of their prey is eaten thanenters the soil pool as detritus. Invertebratecarnivores often have a lower consumption effi-ciency (5 to 25%) than vertebrate carnivores.Consumption efficiency of a trophic level at the ecosystem scale must integrate vertebrateand invertebrate consumption, including ani-mals that feed belowground, but these efficien-cies are not well documented at the ecosystemscale. More frequently, consumption efficiencyis documented for a single large herbivore foran ecosystem in which it is abundant.

The consumption efficiency of a trophic level depends on the biomass of consumers atthat trophic level and factors governing theirfood intake. Over the long term, the quantityand quality of available food constitute thebottom–up controls over the populationdynamics and biomass of consumers. In addi-tion, predators exert top–down controls overconsumer biomass. Bottom–up and top–downcontrols frequently interact. Insects feeding onlow-quality foliage, for example, must eat morefood over a longer time to meet their energeticand nutrient requirements for development.The longer development time required on low-quality food increases their vulnerability to predators and parasites. Rising atmosphericCO2 concentration, which reduces leaf quality,for example, often increases the quantity of leafmaterial eaten by a caterpiller, because it musteat more food to meet its energetic require-ments for development (Lindroth 1996). Theresulting increase in development time, how-ever, probably alters their interactions withhigher trophic levels. Bottom–up controlsrelated to NPP and food quality often explaindifferences among ecosystems in average con-sumer biomass and consumption, with greaterconsumer biomass in more productive ecosys-tems (Figs. 11.3 and 11.4). Predation, however,explains much of the interannual variation in

consumer biomass and the quantity of foodconsumed.

People have substantially altered the trophicdynamics of ecosystems through their effectson consumer biomass. Stocking of lakes withsalmonids, for example, increases predation onsmaller fish, such as exotic alewife. Overfishingcan have a variety of trophic effects, dependingon the trophic level of the target fish. Over-fishing of herbivorous fish in coral reefs, forexample, allows macroalgae to escape grazingpressure and overgrow the corals, killing themin places. On land, stocking of cattle at densi-ties higher than can be supported by primaryproduction causes overgrazing and a decreasein plant biomass; this has led to the loss of productive capacity in many arid lands(Schlesinger et al. 1990). The consequences ofhuman impacts on trophic systems are highlyvariable, but they often have profound effectson trophic levels up and down the food chainas well as on the target species (Pauly andChristensen 1995).

The bottom–up controls over consumptionefficiency can be described in terms of thefactors regulating food intake. Consumption byindividual animals depends on the time avail-able for eating, the time spent looking for food,the proportion of food that is eaten, and therate at which food is consumed and digested.Each of these four determinants of consump-tion has important ecological, physiological,morphological, and behavioral controls thatdiffer among animal species.

Animals do many things other than eating,including predator avoidance, digestion, repro-duction, and sleeping. In addition, unfavorableconditions often restrict the time available forforaging, especially for poikilothermic animalssuch as insects, amphibians, and reptiles, whosebody temperature depends on the environ-ment. Because of these constraints, deer con-centrate their feeding at dawn and dusk; desertrodents feed primarily at night; bears hibernatemost of the winter; and mosquitoes feed mostactively under conditions of low wind, moder-ate temperatures, and high humidity. Activitybudgets describe the proportion of the timethat an animal spends in various activities.Activity budgets differ among species, seasons,


and habitats, but many animals spend a rela-tively small proportion of their time consumingfood. Changes in climate or predator risk thatinfluence activity budgets of an animal can profoundly alter food intake and therefore the energy available for animal production andmaintenance.

Animals must find their food before they eat it. Most predators such as wolves spendmore time looking for food than ingesting it. Other animals, including most herbivores,search for favorable habitats within a land-scape, then spend most of their time ingestingfood. Animals generally consume food fasterthan they can digest it, so some of the timespent in other activities simultaneously con-tributes to digestion of food.

Once an animal finds its food, it generallyconsumes only some of it. Many herbivores,for example, select only the youngest leaves ofcertain plant species and avoid other plantspecies, older leaves, stems, and roots. Similarly,carnivores may eat only certain parts of ananimal and leave behind parts such as skin andbones. This selectivity places an upper limit onconsumption efficiency, because there are com-ponents of production from one trophic levelthat are not consumed at the next level. Manyanimals become more selective as food avail-ability increases. Lions, for example, eat less oftheir prey when food is abundant. Gypsy mothsand snowshoe hares also preferentially feed oncertain plant species, given the opportunity, butwill feed on almost any plant during populationoutbreaks, after most palatable species havebeen consumed.

Selectivity also depends on the nutritionaldemands of an animal. Caribou and reindeer,for example, have a gut flora that is adapted todigest lichens, which are avoided by most otherherbivores. These animals eat lichens in winterwhen low temperatures impose a high energydemand for homeothermy (maintenance of a constant body temperature). Lichens have a high energy content but little protein. Insummer, however, when there is a high proteinrequirement for growth and lactation, theseanimals increase the proportion of nitrogen-rich vascular plant species in their diet (Klein1982). Other herbivores may select plant

species to minimize the accumulation of planttoxins. Moose and snowshoe hares in the borealforest, for example, can consume only a certainamount of particular plant species before accu-mulation of plant toxins has detrimental phy-siological effects (Bryant and Kuropat 1980).They therefore tend to avoid plant species withhigh levels of toxic secondary metabolites.Selectivity by herbivores also depends on thecommunity context. Mammalian generalist her-bivores preferentially select plant species whenthey are uncommon because rare species areconsumed too infrequently to reach a thresholdof toxicity. Selectivity by these generalistbrowsers therefore tends to eliminate rareplant species and reduce plant diversity (Bryantand Chapin 1986).

Selectivity differs among animal species.Some grazers, like wildebeest in African savan-nas, are almost like lawn mowers. They followthe pulse of grass growth that occurs after rainsand consume most plants that they encounter.Other animals, like impala, select leaves of rel-atively high nitrogen and low fiber content,especially in the dry season. Among mammals,there is a continuum from large-bodied gener-alist herbivores, which are relatively nonselec-tive, to smaller-bodied specialist herbivores,which are highly specific in their food require-ments. Similar patterns are seen among fresh-water zooplankton; large-bodied cladoceranslike Daphnia are generalist filter feeders,whereas same-size or smaller copepods aremore selective (Thorp and Covich 2001). Spe-cialization is even more pronounced among terrestrial insects. Some tropical insects, forexample, eat only one part of a single plantspecies. The abundance of specialist insectscould contribute to the high diversity of tropi-cal forests, by preventing any one plant speciesfrom becoming extremely abundant.

Animals differ in their eating rate. This canbe quantified as the bite rate times bite size forvertebrate herbivores or the rate at which soilmoves through the gut of an earthworm. Thereis often an inverse relationship between feedingrate and selectivity, with selective herbivoresspending more time looking for food andselecting among species and plant parts thatthey consume. Because gut capacity is limited,


eating rate may be constrained by processingtime. Animals that are more selective and havea lower feeding rate generally eat food that ismore digestible, contributing to their shorterpassage time.

Assimilation Efficiency

Assimilation efficiency depends on both thequality of the food and the physiology of theconsumer. Assimilation efficiency is the pro-portion of ingested energy that is digested and assimilated (An) into the bloodstream (Fig. 11.7).

(11.4)

Unassimilated material returns to the soil asfeces, a component of the detrital input toecosystems.

Assimilation efficiencies are often higher (5to 80%) than consumption efficiencies (0.1 to50%). Carnivores feeding on vertebrates tendto have higher assimilation efficiencies (about80%) than do terrestrial herbivores (5 to 20%),because carnivores eat food that has less struc-tural material than is present in terrestrialplants. Carnivores that kill large prey can avoideating indigestible parts such as bones, whereasmost terrestrial herbivores consume the indi-gestible cell wall structure in combination withcell contents. Among herbivores, species thatfeed on seeds, which have high concentrationsof digestible, energy-rich storage reserves, havea higher assimilation efficiency than thosefeeding on leaves. Leaf-feeding herbivores,in turn, have higher assimilation efficienciesthan those feeding on wood, which has higherconcentrations of cellulose and lignin. Manyaquatic herbivores have a particularly highassimilation efficiency (up to 80%) because ofthe low allocation to structure in many algaeand other aquatic plants. Even in aquaticecosystems, however, herbivores that feed onwell-defended species have low assimilationefficiencies. Assimilation efficiencies of herbi-vores feeding on cyanobacteria, for example,can be as low as 20%.

The physiological properties of a consumerstrongly influence assimilation efficiency. Rumi-

EAI

n

nassim =

nants, which carry a vat of cellulose-digestingmicrobes (the rumen) have a higher assimila-tion efficiency (about 50%) than do most nonruminant herbivores. One reason for thehigh assimilation efficiency of ruminants is thegreater processing time than in nonruminantsof similar size, giving more time for microbialbreakdown of food. Homeotherms typicallyhave higher assimilation efficiencies than dopoikilotherms due to the warmer, more con-stant gut temperature, which promotes diges-tion and assimilation. Homeotherms thereforehave an advantage over poikilotherms in bothconsumption and assimilation efficiency.

Production Efficiency

Production efficiency is determined primarilyby animal metabolism. Production efficiencyis the proportion of assimilated energy that is converted to animal production (Fig. 11.7).Production efficiency includes both growth ofindividuals and reproduction to produce newindividuals.

(11.5)

Assimilated energy that is not incorporatedinto production is lost to the environment asrespiratory heat. Production efficiencies forindividual animals vary 50-fold from less than 1 to greater than 50% (Table 11.2) and differmost dramatically between homeotherms (Eprod

1 to 3%) and poikilotherms (Eprod 10 to 50%).Homeotherms expend most of their assimilatedenergy maintaining a relatively constant bodytemperature. This high constant body tempera-ture makes their activity less dependent onenvironmental temperature and increases theircapacity to catch prey and avoid predation butmakes homeotherms inefficient in producingnew animal biomass. Among homeotherms,production efficiency decreases with decreasingbody size because a small size results in a highsurface to volume ratio and therefore a highrate of heat loss from the warm animal to thecold environment. In contrast, the productionefficiency of poikilotherms is relatively high(about 25%) and tends to decrease withincreasing body size. Some large-bodied

Eod

An

nprod =

Pr


animals, such as tuna, that belong to groupsusually thought of as poikilotherms are par-tially homeothermic. Among poikilotherms,production efficiency is lowest in fish and socialinsects (about 10%), intermediate in noninsectinvertebrates (about 25%), and highest innonsocial insects (about 40%) (Table 11.2).Production efficiency often decreases withincreasing age, because of changes in allocationto growth and reproduction.

Note that belowground NPP—including exu-dates and transfers to mycorrhizae—is large,poorly quantified, and frequently ignored inestimating trophic efficiencies. Our views oftrophic efficiencies may change considerably as our understanding of belowground trophicdynamics improves. Fine roots, mycorrhizae,and exudates, for example, turn over quicklyand may support high belowground consump-tion and assimilation efficiencies for herbivoressuch as nematodes that specialize on thesecarbon sources (Detling 1988).

Food Chain Length and Trophic Cascades

Production interacts with other factors todetermine length of food chains and trophicstructure of communities. Both the NPP andthe inefficiencies of energy transfer at each

trophic link constrain the amount of energythat is available at successive trophic levels andthat therefore could influence the number oftrophic levels that an ecosystem can support.The least productive ecosystems, for example,may have only plants and herbivores, whereasmore productive habitats might also supportmultiple levels of carnivores (Fretwell 1977,Oksanen 1990). Detritus-based food chains alsotend to be longer in more productive ecosys-tems (Moore and de Ruiter 2000). In someaquatic ecosystems, however, the trend can goin the opposite direction. Oligotrophic habitatscan support inverted biomass pyramids inwhich large long-lived fish are more conspicu-ous than the algal and invertebrate populationsthat support them. Eutrophic lakes or rivers are often dominated by taxa at lower trophiclevels that are less edible (Power et al. 1996a).The death and decay of these organisms maydeplete dissolved oxygen, eventually makingthe habitat lethal for fish and other aquaticpredators, shortening the length of food chains(Carpenter et al. 1998). When ecosystems arecompared across broad productivity gradients,there is no simple relationship between NPPand the number of trophic levels (Pimm 1982,Post et al. 2000). Other factors such as environ-mental variability and the physical structure ofthe environment often have a greater effect on the number of trophic levels than does theenergy available at the base of the food chain(Post et al. 2000).

The number of trophic levels influences thestructure and dynamics of ecosystems throughthe action of trophic cascades, in which chan-ges in the abundance at one trophic level alterthe abundance of other trophic levels acrossmore than one link in a food web (Pace et al.1999). Trophic cascades result from strongpredator–prey interactions between particularspecies (Paine 1980). Predation by one organ-ism at one trophic level reduces the density oftheir prey, which releases its prey from con-sumer control (Carpenter et al. 1985, Pace et al.1999). This trophic cascade (a top–down effect)causes an alternation among trophic levels inbiomass of organisms (Power 1990). In streams,for example, if only algae are present, they growuntil their biomass becomes nutrient limited,

Table 11.2. Production efficiency of selectedanimals.

Production efficiencyAnimal type (% of assimilation)

HomeothermsBirds 1.3Small mammals 1.5Large mammals 3.1

PoikilothermsFish and social insects 9.8Nonsocial insects 40.7

Herbivores 38.8Carnivores 55.6Detritus-based insects 47.0

Noninsect invertebrates 25.0Herbivores 20.9Carnivores 27.6Detritus-based invertebrates 36.2

Data from Humphreys (1979).


producing a “green” surface (Fig. 11.9). If thereare two trophic levels (plants and herbivores),the herbivores graze the plants to a low biomasslevel, leaving a barren surface. With threetrophic levels, the secondary consumer reducesthe biomass and grazing pressure of herbivores,which again allows algae to achieve a highbiomass. Algal biomass is generally low whenthere is an even number (two, four, etc.) oftrophic levels. An odd number of trophic levelsin a trophic cascade reduces the biomass of her-bivores and releases the algae, producing a“green” world (Fretwell 1977).

Trophic cascades have been demonstrated in a wide range of ecosystems, from the openocean to the tropical rain forests and microbialfood webs (Pace et al. 1999). Trophic cascadesare best documented at the level of speciesrather than ecosystems, because they generallyresult from strong interactions between indi-vidual species (Paine 1980, Polis 1999). Trophiccascades are most important at the ecosystemscale, when a single species dominates a trophiclevel, for example when Daphnia is the domi-nant herbivore or a minnow-eating fish is thedominant carnivore in a lake (Polis 1999).Eutrophication often leads to strong species

dominance, thereby providing conditionswhere trophic cascades can play an importantrole (Pace et al. 1999). Trophic cascades haveimportant practical implications; introductionof minnow-eating fish, under the right circum-stances, can release populations of zooplanktongrazers, which graze down algal blooms andincrease water clarity. Manipulation of trophiccascades to address management issues re-quires a sophisticated understanding of theecology of the species involved and the factorsgoverning their interactions. Interactions thatwere not anticipated frequently become impor-tant when trophic dynamics are altered, leadingto unexpected responses to species introduc-tions (Kitchell 1992). In most ecosystems thereis a dynamic balance between bottom–up andtop–down controls that is governed by a widevariety of ecological feedbacks (Power 1992b).In only some situations are ecosystem-scaletrophic cascades a dominant feature of ecosys-tems (Polis 1999).

Seasonal Patterns

In terrestrial ecosystems, production by onetrophic level seldom coincides with consump-

1st

2nd

3rd

4th

A green world A barren world

1 trophic level 3 trophic levels 2 trophic levels 4 trophic levels

Figure 11.9. Effect of food chain length on primaryproducer biomass in situations in which trophic cascades operate. Plant biomass is abundant wherethere are odd numbers of trophic levels (1, 3, 5, etc.),

because these have a low biomass of herbivores;plant biomass is reduced where there are evennumbers of trophic levels (2, 4, 6, etc.), because thesehave a large biomass of herbivores.


tion by the next. The precise temporal rela-tionship between predator and prey is highlyvariable, but some common patterns emerge.Plants and their insect predators often usesimilar temperature and photoperiodic cues toinitiate spring growth. However, insects cannotafford to emerge before their food, so there isoften a brief window in spring when plants arerelatively free of invertebrate herbivory (Fig.11.10). After insect emergence, there is often abrief window before leaves become too toughor toxic for insects to feed (Feeny 1970, Ayresand MacLean 1987). In contrast to insects,homeotherm herbivores continue to consumefood during the cold season, when plants aredormant. These are, however, only two of manyhighly specific patterns of interactions betweenplants and their herbivores. Predation by highertrophic levels often focuses at times when preyare most vulnerable, such as when vertebratesare giving birth to young, when salmon aremigrating, or when insects are moving activelyin search of food. Again, the specific patternsare quite diverse and depend on the biology ofpredator and prey. The important point is thatproduction by one trophic level and consump-tion by the next are seldom equal at any pointin the annual cycle.

Nutrient Transfers

The pathway of nutrients through food chainsis usually similar to that of energy. Nitrogen,phosphorus, and other nutrients in plants andanimals are either organically bound or are dis-solved in the cell contents. Nutrients containedin biomass eaten by animals therefore generallyfollow the same path through food chains asdoes energy, from plants to herbivores, toprimary carnivores, to secondary carnivores,etc. At each link in the food chain, nutrients are digested and assimilated by animals, just asenergy is digested and assimilated, although the efficiencies may differ substantially.As withenergy, nutrient losses occur with each trophictransfer in the form of uneaten food, feces, andurine, so the quantity of nutrients transferredmust decline with each successive trophic link.The pyramids of nutrient transfers are there-fore similar in shape to those of energy flow.In terrestrial ecosystems, the pools of nutrientsin organisms also decline with each successivetrophic link, as is the case for pyramids ofbiomass. In pelagic aquatic ecosystems, how-ever, the high trophic efficiencies of zooplank-ton and high turnover of primary producersgenerally result in inverted pyramids of nutri-ent pools, just as for aquatic biomass pyramids.In summary, the general patterns of nutrienttransfer through ecosystems are similar tothose of energy, although the quantitativedynamics may differ.

An important exception to this rule issodium, which is required by animals for trans-mission of impulses in nerves and muscles. Mostplants actively exclude sodium from roots and from leaves, so tissue concentrations arelower in plants than would be expected basedon soil solution concentrations (see Chapter 8).Sodium is therefore more likely to be limiting toanimals than to plants. Many terrestrial herbi-vores supplement the sodium acquired fromfood by ingesting soil or salts from salt licks,which are mineral-rich springs or outcrops.Sodium may therefore show a different path-way of trophic transfer than do other minerals.

A larger proportion of the nutrients con-tained in plant production pass through terres-trial herbivores than is the case for energy. Most

Gro

wth

effi

cien

cy (

%) Growth efficiency

Leaf

tough

ness

Spec

ific

leaf

mas

s

Spe

cific

leaf

mas

s (m

g cm

-2)

Toug

hnes

s (P

a)

24

16

8 11 18 23 29 65

014

7

6

8

4

0

June July

Figure 11.10. Seasonal pattern of specific leaf massand leaf toughness of Finnish birch leaves and ofgrowth efficiency of fourth instar larvae of birchmoths. The herbivore grows at maximal efficiencyuntil leaves become tough and mature. After this 2-week window of leaf development, the herbivoregrows slowly. (Data from Ayres and MacLean 1987.)


terrestrial herbivores selectively feed on youngtissues with a high concentration of digestibleenergy and low concentrations of cellulose andlignin. These young tissues have high concen-trations of nitrogen and phosphorus. Because of selective herbivory on nutrient-rich tissues,a larger proportion of plant-derived nutrientscycle through plant-based trophic systems thanis the case for carbon.

Terrestrial herbivores not only select nutrient-rich tissues but cycle nutrients morerapidly than do plants. Plants resorb about halfthe nitrogen and phosphorus from leavesduring senescence, so plant litter generally hasonly half the nitrogen and phosphorus con-centrations as does the live tissue eaten by herbivores (see Chapter 8). For this reason,herbivory is at least twice as important anavenue for nitrogen and phosphorus cycling interrestrial ecosystems as it is for biomass andenergy. The turnover time for nutrients in ter-restrial herbivores is often shorter than in theplants on which they feed. Many terrestrialanimals, particularly carnivores and homeo-therms, eat more nutrients than they require forgrowth, due to the large energetic costs ofmovement, and in the case of homeotherms, fortemperature regulation.Animals excrete excessnutrients in inorganic form or as urea, which isquickly hydrolyzed in soils (see Chapter 9). In

summary, terrestrial herbivores speed nutrientcycling in at least two ways: (1) by removingplant tissues that are more nutrient-rich thanwould otherwise return to the soil in litterfalland (2) by returning nutrients to the soil informs that can be directly used by plants (Fig.11.6).

The ratio of elements required by plants andherbivores determines the nature of elementlimitation in organisms and the patterns ofnutrient cycling in ecosystems. Both aquaticand terrestrial plants require nitrogen andphosphorus in a ratio of about 15 :1 (the Redfield ratio; see Chapter 8) (Fig. 11.11). TheN:P ratio in herbivores is generally less than in plants, particularly in lakes (Elser et al. 2000).Lake herbivores must therefore concentratephosphorus more strongly than nitrogen tomeet their nutritional demands and tend toexcrete the excess nitrogen (Elser and Urabe1999). In this way, herbivory speeds the recy-cling of nitrogen, relative to phosphorus,making nitrogen more available to phytoplank-ton and reinforcing the phosphorus limitationthat characterizes many lakes. Differences in N :P ratios among grazers in lakes illustrate theimportance of this effect. Daphnia is a rapidlygrowing cladoceran grazer that has a higherphosphorus concentration (lower N :P ratio)than the more slowly growing copepods. Under

Freq

uenc

y (%

of o

bser

vatio

ns) 25

20

15

10

5

N:P ratio N:P ratio

Terrestrial

Freshwater

Terrestrial

Freshwater

0 0

10

20

30

40

2.3 5 10 15 20 25 2.3 5 10 15 20 25

Plants Invertebrate herbivores

A B

Figure 11.11. Frequency distribution of N : P ratiosin terrestrial and freshwater ecosystems in (A) plantsand (B) invertebrate herbivores. Ratios are mass ofnitrogen relative to mass of phosphorus. The N : Pratio is lower in herbivores than in the plants on

which they feed, particularly in fresh-water ecosys-tems. Herbivores therefore preferentially retainphosphorus and excrete nitrogen to the environ-ment. (Redrawn with permission from Nature; Elseret al. 2000.)

Integrated Food Webs 261

conditions of Daphnia dominance, grazers con-centrate more phosphorus and excrete morenitrogen than when copepods are the dominantgrazer; this leads to phosphorus limitation ofphytoplankton growth when Daphnia domi-nates and nitrogen limitation when copepodsdominate. N :P ratios differ less between ter-restrial plants and their herbivores than inaquatic ecosystems. Herbivory also accountsfor less of the total nutrient return from plantsto the environment in terrestrial than in aquaticecosystems, so the impact of trophic shifts in N :P ratios on nutrient cycling in terrestrialecosystems is uncertain. Nonetheless, the trendtoward lower N :P ratios in terrestrial herbi-vores than in plants (Fig. 11.11) suggests thatherbivores may be more phosphorus limitedthan the plants on which they feed and thatphosphorus could be a more important nutri-tional constraint for animals than is generallyrecognized.

Detritus-Based Trophic Systems

Detritus-based trophic systems convert a muchlarger proportion of available energy into pro-duction than do plant-based trophic systemsbecause they recycle unused organic matter.Plant and animal detritus is fed on by decom-poser organisms (primarily bacteria and fungi),just as herbivores feed on live plants. As in theplant-based trophic system, there is a foodchain of animals that feed on these decomposerorganisms (Fig. 11.12).The principles governingthis energy flow are similar to those in theplant-based food chain.

The quantity and quality of soil organicmatter is the major determinant of the quantityof energy that flows through the detritus-basedsystem. The detritus-based food chain exhibitslosses of energy to growth and maintenancerespiration and as feces, just as in plant-basedfood chains (Fig. 11.12). Moreover, each trophictransfer entails the excretion of inorganic N andP, which become available to plants, just as inthe plant-based trophic system.

The major structural distinction betweenplant- and detritus-based systems is that theplant-based system involves a one-way flow of

energy, as energy is either transferred up thefood chain or is lost from the food chain as respiration, unconsumed production, or feces.In the detritus-based food chain, however,uneaten food, feces, and dead organisms againbecome substrate for decomposers at the baseof the food chain (Fig. 11.12) (Heal andMacLean 1975). Energy flow in the detritus-based system therefore has a strong recyclingcomponent. Energy is conserved and is avail-able to support detritus-based production untilit is respired away or is converted to recalci-trant humic material. Due to the efficient use(and reuse) of energy that enters the base of the food chain, the detritus-based food webaccounts for most of the energy flow and supports the greatest animal diversity in ecosystems (Heal and MacLean 1975).

The trophic efficiencies of the detritus-basedtrophic system are generally higher than in theplant-based trophic system. The consumptionefficiency of detritus-based food chains is high(greater than 100%) because all of the poten-tial “food” is consumed several times until it is eventually respired away. Assimilation effi-ciency is also high in decomposers (bacteria andfungi) because their digestion is extracellular,so, by definition, all the material that is con-sumed by decomposers is assimilated. Herbi-vores, the first link in the plant-based trophicsystem, in contrast, have assimilation efficien-cies that are commonly 1 to 10%. Productionefficiencies of decomposers (40 to 60%; seeChapter 7) and animals in detritus-based foodchains (35 to 45%) are also higher than inplant-based trophic systems (Table 11.2).Together these high trophic efficiencies ex-plain why the detritus-based trophic systemaccounts for most of the secondary productionin ecosystems.

Integrated Food Webs

Mixing of Plant-Based and Detritus-Based Food Chains

Food webs blur the trophic position of eachspecies in an ecosystem. In the real world, mostanimals feed on prey from more than one


trophic level, and many animals feed from boththe plant-based and the detritus-based trophicsystems and at different trophic levels withineach system (Polis 1991). For this reason, it isdifficult to assign most organisms to a singletrophic level. Most fungivores, for example,feed on a mixture of mycorrhizal fungi thatderive their energy from plants and saprophyticfungi that decompose dead organic matter.Bacteria also derive energy from root exudates(a component of NPP) and from dead organicmatter. Soil animals that eat bacteria and fungiare therefore part of both the plant-based and the detritus-based trophic systems. Root-feeding mites and nematodes fall prey toanimals that also eat detritus-based animals(Fig. 11.1). All soil food webs therefore processa mixture of plant and detrital energy and

nutrients in ways that are difficult to unravel.Although food webs through abovegroundanimals have been studied more thoroughly,they also involve substantial detrital input fromanimals that feed on fungi or on soil animals.Robins, for example, feed on both earthwormsand herbivorous insects. Bears eat plant rootsand ants of terrestrial origin (plant- and largelydetritus-based food chains, respectively) andfish from aquatic food webs. Many insects aredetrital feeders at the larval stage but drinknectar or blood (plant-based trophic system), asadults. About 75% of food webs contain bothplant- and detritus-based components (Mooreand Hunt 1988), so mixed trophic systems arethe rule rather than the exception.

The ecosystem consequence of this blurringof food webs is that each food web subsidizes,

R

R

R

R

R

R

B + FH

CC2

C1 M

SOMNPP

Plant-basedtrophic system

Detritus-basedtrophic system

Figure 11.12. The two basic trophic systems inecosystems (Heal and MacLean 1975). In the plant-based trophic system, some energy is transferredfrom live plants to herbivores (H), primary carni-vores (C1), secondary carnivores (C2), etc. In thedetritus-based trophic system, energy is transferredfrom dead soil organic matter (SOM) to bacteria (B)and fungi (F), microbivores (M), carnivores (C), etc.In both trophic systems, energy that is not assimi-lated at each trophic transfer passes to the detritus

pool (as unconsumed organisms or as feces). Themajor difference between these two trophic systemsis that energy passes in a unidirectional flow throughthe plant-based trophic system to herbivores andcarnivores or to the detrital pool. In the detritus-based trophic system, however, material that is notconsumed returns to the base of the food chain andcan recycle multiple times through the food chainbefore it is respired (R) away or converted to recal-citrant humus.


and is subsidized by, other food webs. Subsidiesare therefore an important component of mostecosystems.

Scavengers such as vultures, hyenas, crabs,and many beetles are technically part of thedetritus-based food web, although their con-sumption, assimilation, and production effi-ciencies are similar to those of carnivores.Scavengers often kill weakened animals, andmany predators feed on prey that have beenrecently killed by other animals, further blur-ring the distinction between plant-based anddetritus-based food chains.

Food Web Complexities

Parasites, pathogens, and diseases are trophi-cally similar to predators. They derive theirenergy from host tissues and use the productsof these cells for their own growth and re-production, just like predators. It is difficult inpractice, however, to separate the biomass ofparasites, pathogens, and diseases from that oftheir hosts, so the concepts of consumption andassimilation efficiencies are seldom applied tothese organisms. Parasites, pathogens, and dis-eases are therefore often treated as agents ofmortality rather than as consumers.

Mutualists also confound the trophic picture.Mycorrhizal fungi can change from being mutu-alistic to parasitic, depending on environmentalconditions and the nutritional status of the hostplant (Koide 1991). Under mutualistic condi-tions, mycorrhizal fungi act as herbivores intransferring carbohydrates from plants to thefungus. However, nutrient transfer occurs in theopposite direction, so the trophic role of thesetwo organisms depends on the constituent ofinterest. In summary, although the broad out-lines of trophic dynamics have a clear concep-tual basis, the complexities of nature and ourpoor understanding of belowground processesoften make it difficult to describe these foodwebs quantitatively.

Summary

Resource supply and other factors controllingNPP constrain the energy that is available to

higher trophic levels in plant-based trophicsystems.These same factors govern the quantityand quality of litter input to the soil and there-fore the energy available to the detritus-basedtrophic system. These factors constitute thebottom–up controls over trophic dynamics. Thetrophic efficiency with which energy is trans-ferred from one trophic level to the nextdepends on the efficiencies of consumption,assimilation, and production. Consumption efficiency depends on the interaction of foodquantity and quality with predation by highertrophic levels. Consumption efficiency of her-bivores is lowest in unproductive habitats dominated by plants that are woody or welldefended. Carnivores generally have higherconsumption efficiency than do herbivores.

Assimilation efficiency is determined pri-marily by food quality. It is lower in unproduc-tive than in productive habitats and lower forherbivores than for carnivores. In contrast tothe other components of trophic efficiency, theproduction efficiency is determined primarilyby animal physiology; poikilotherms have ahigher production efficiency than do homeo-therms. Most secondary production in terres-trial ecosystems occurs in the detritus-basedtrophic system. In this system, material that is not consumed or assimilated returns to thebase of the food chain and continues to recyclethrough the food chain until it is respired orconverted to recalcitrant humus. Most foodwebs contain both plant- and detritus-basedcomponents. Impacts, including those resultingfrom human activities, on any link in food websfrequently propagate to other links in foodwebs.

Review Questions

1. Describe the pathways of carbon flow in anherbivore-based food chain. How does theefficiency of conversion of food into con-sumer biomass differ between herbivoresand carnivores? What determines the parti-tioning of assimilated energy between respi-ration and production?

2. What is the major structural differencebetween plant-based and detritus-based


food chains? Which food chain can supportthe greatest total production? Why?

3. What are the major structural differencesbetween terrestrial and aquatic food chains?Why do these differences occur?

4. What plant traits determine the amount of herbivory that occurs? What ecologicalfactors influence these plant traits?

5. What are the effects of herbivores on nitro-gen cycling?

6. What are the mechanisms by which toppredators influence abundance of primaryproducers in aquatic food chains? Whatdetermines the number of trophic links in anecosystem? How does this affect ecosystemstructure?

Additional Reading

Bryant, J.P., and P.J. Kuropat. 1980. Selection ofwinter forage by subarctic browsing vertebrates:The role of plant chemistry. Annual Review ofEcology and Systematics 11:261–285.

Carpenter, S.R., J.F. Kitchell, and J.R. Hodgson. 1985.Cascading trophic interactions and lake produc-tivity. BioScience 35:634–649.

Coley, P.D., J.P. Bryant, and F.S. Chapin III. 1985.Resource availability and plant anti-herbivoredefense. Science 230:895–899.

Heal, O.W., and J. MacLean, S.F. 1975. Comparativeproductivity in ecosystems-secondary productiv-ity. Pages 89–108 in W.H. van Dobben, and R.H.Lowe-McConnell, editors. Unifying Concepts inEcology. Junk, The Hague.

Herms, D.A., and W.J. Mattson. 1992. The dilemmaof plants: To grow or defend. Quarterly Review ofBiology 67:283–335.

Hobbs, N.T. 1996. Modification of ecosystems byungulates. Journal of Wildlife Management 60:695–713.

Lindeman, R.L. 1942.The trophic-dynamic aspects ofecology. Ecology 23:399–418.

Oksanen, L. 1990. Predation, herbivory, and plantstrategies along gradients of primary productivity.Pages 445–474 in J.B. Grace, and D.Tilman, editors.Perspectives on Plant Competition. AcademicPress, San Diego.

Paine, R.T. 2000. Phycology for the mammalogist:Marine rocky shores and mammal-dominatedcommunities. How different are the structuringprocesses? Journal of Mammalogy 81:637–648.

Pastor, J., R.J. Naiman, B. Dewey, and P. McInnes.1988. Moose, microbes, and the boreal forest.BioScience 38:770–777.

Polis, G.A. 1999. Why are parts of the world green?Multiple factors control productivity and the distribution of biomass. Oikos 86:3–15.

Power, M.E. 1992. Top-down and bottom-up forcesin food webs: do plants have primacy? Ecology73:733–746.

Introduction

Species traits interact with the physical envi-ronment to govern ecosystem processes. Up tothis point, we have emphasized only the mostgeneral properties of organisms. We discussedprimary producers, for example, as if they werea homogeneous group of organisms in anecosystem and indicated that primary produc-tion can be broadly predicted from climate andparent material. Species differences in traitssuch as photosynthesis, root allocation, andlitter quality, however, strongly affect the func-tioning of terrestrial ecosystems. Similarly, thephosphorus requirements and prey size prefer-ences of zooplankton govern patterns of nutri-ent cycling in lakes. Under what circumstancesmust we know the traits of individual organismswithin a trophic group to understand ecosystemprocesses? In this chapter, we explore this ques-tion through discussion of functional types—that is, groups of species that are ecologicallysimilar in their effects on ecosystem processes(Chapin 1993a, Smith et al. 1997). Termites,homeotherm herbivores, nitrifying bacteria,and evergreen shrubs are examples of func-tional types that have predictable generaleffects on ecosystem processes. Predators, such

as planktivorous fish, that feed on the sameprey also constitute a functional type or feedingguild. However, no two species or individualswithin a functional type are ecologically identi-cal, so, as our understanding improves or ourquestions become more refined, we expect torecognize situations in which species diversitywithin functional types or genetic diversitywithin species has detectable ecosystem consequences.

Natural ecosystems are currently experienc-ing major changes in species diversity and thetraits of dominant species. Earth is currently in the midst of the sixth major extinction eventin the history of life (Pimm et al. 1995).Although the causes of earlier extinction events(e.g., the extinction of dinosaurs) are uncertain,they probably resulted from sudden changes inphysical environment caused by factors such as asteroid impacts or pulses of volcanism.Current extinction rates are 100- to 1000-foldhigher than prehuman extinction rates andcould rise to 10,000-fold, if species that are cur-rently threatened become extinct (Fig. 12.1).The current extinction event is unique in thehistory of life in that it is biotically driven,specifically by the effect of the human specieson land use, species invasions, and atmospheric

12Community Effects on Ecosystem Processes

The traits and diversity or organisms and their interactions in communities stronglyaffect ecosystem processes. This chapter describes the patterns of community effectson ecosystem processes.

265

266 12. Community Effects on Ecosystem Processes

and environmental change. Although humanimpacts affect many processes at global scales(Vitousek 1994a) (see Chapter 15), the loss ofspecies diversity is of particular concernbecause it is irreversible. For this reason, it iscritical to understand the functional conse-quences of the current large losses in speciesdiversity (Chapin et al. 2000b).Although globalextinction of a species is a major conservationconcern, localized extinctions or large changesin abundances happen more frequently andhave the largest effects on functioning ofecosystems.

A second biotic change of global proportionsis the frequent introduction of exotic speciesinto ecosystems. People have intentionally andinadvertently moved thousands of speciesaround the globe, leading toward a homoge-nization of the global biota. Exotic speciesoften change the physical and biotic en-vironment enough to alter the dominance oreliminate native species from an ecosystem.Although extinction and immigration of speciesare natural ecological processes, the dramaticincrease in frequency of these events in recentdecades is rapidly changing the types andnumbers of species throughout the globe.Thereare many ethical, aesthetic, and economicreasons for concern about changes in biodiver-sity. The changes are occurring so quickly,however, that we must assess their functionalconsequences for population, community, and

ecosystem processes. Changes in species com-position often have a greater effect on eco-system processes than do the direct impacts ofglobal changes in atmospheric composition andclimate. Understanding the nature of bioticimpacts on ecosystem processes is thereforecritical to predictions of future changes inecosystems. In this chapter we focus on theecosystem consequences of changes in commu-nities and the associated changes in speciestraits.

Overview

The number, relative abundance, identity, andinteractions of species all affect ecosystemprocesses. Species in a given trophic levelalmost always differ in some traits that affectecosystem processes. Sun and shade species, forexample, differ in the conditions under whichthey contribute to carbon inputs; presence ofboth types of species therefore increases theefficiency with which light is converted to netprimary production (NPP) (see Chapter 5).Nitrogen-fixing cyanobacteria differ from non-fixing phytoplankton in their impacts on nitro-gen cycling; presence of both types of speciestherefore influences the response of nitrogencycling to variations in phosphorus inputs tolakes (see Chapter 10). No single species canperform all of the functional roles that areexhibited by a trophic level. A diversity ofspecies is functionally important because itincreases the range of organismic traits that are represented in an ecosystem and thereforethe range of conditions under which ecosystemproperties can be sustained. From an ecosystemperspective, species diversity is simply asummary variable that describes the total rangeof biological attributes of all the species in theecosystem. The functional consequences of achange in diversity depend on the number ofspecies present (species richness), their relativeabundances (species evenness), the identity ofspecies that are present (species composition),the interactions among species, and the tem-poral and spatial variation in these properties.Each of these components of diversity affectsthe functioning of ecosystems.

Birds Mammals Fish Plants

10

20

Ext

inct

ion

thre

aten

ed(%

of g

loba

l spe

cies

)

0

Figure 12.1. Percentage of major vertebrate andvascular plant species that are currently threatenedwith extinction. (Redrawn with permission fromNature; Chapin et al. 2000b.)

Overview 267

If all species were functionally different andcontributed in unique ways to a given process,rates of ecosystem processes might change lin-early as the number of species increased (Fig.12.2A) (Vitousek and Hooper 1993, Sala et al.

B. Effect of species abundance

Rare species

Dominantspecies

Eco

syst

em p

roce

ss

C. Effect of species type

Keystonespecies

Compensating species

Species richness (number of species)

similar to other speciesSome species are ecologically

A. Effect of species number

Eac

hsp

ecie

sis

uniq

ue

Figure 12.2. Expected relationship between eco-system processes and the number of species, theirrelative abundance, and the type of species in anecosystem (Vitousek and Hooper 1993, Sala et al.1996). A, Some processes (or stocks) may increaselinearly with increasing species number; others mayshow an assymptotic increase. B, Removal of domi-nant species from an ecosystem has greater impacton ecosystem processes than does removal of rarespecies. C, Similarly, the removal of keystone specieshas large ecosystem effects, whereas removal of onespecies of a functional type allows other species inthat functional type to increase in abundance; thiscompensation would cause only a moderate impacton ecosystem processes, until most species from thatfunctional type have been removed. The arrowsshow the expected change in ecosystem processes inresponse to species loss.

1996). Nitrogen retention, for example, mightincrease as species with different rootingdepths or preferred forms of nitrogen uptakeare added to the ecosystem. In practice,however, the relationship between speciesnumber and any given ecosystem process tendsto saturate with increasing numbers of species,because some species that are added are eco-logically similar to species already present inthe community (Tilman et al. 1996). Ecosystemprocesses probably differ in their sensitivity tothe number of species in a given trophic level.Nutrient retention may be particularly sensitiveto the number and functional diversity of plantspecies, whereas decomposition rate maydepend more on the diversity of the microbialcommunity.

If all else were equal, a change in the abun-dance of a dominant species is more likely tohave ecosystem effects than is a change inabundance of a rare species (Sala et al. 1996)(Fig. 12.2B), because dominant species accountfor most of the energy and nutrient flowthrough an ecosystem. Dominant species arealso most likely to have strong effects onmicroenvironment. Loss of dominant conifersdue to pathogen or insect outbreak, forexample, alters microclimate and plant biomassso strongly that most ecosystem processes areaffected (Matson and Waring 1984).

The functional attributes of species in a com-munity are at least as important as the numberof species present in determining the mecha-nisms by which species diversity influencesecosystem processes (Hooper et al., in press).Many species are ecologically more importantthan their abundance would suggest. Keystonespecies are an extreme example of species withstrong effects.A keystone species is a functionaltype represented by only one species. It is eco-logically distinct from other species in theecosystem and has a much greater impact onecosystem processes than would be expectedfrom its biomass (Power et al. 1996b) (Fig.12.2C). The tsetse fly in Africa, for example,has a large effect on ecosystem processes per unit of tsetse fly biomass, because it limitsthe density of many ecologically importantmammals, including people. Loss of a keystonespecies has greater ecological impact than does


the loss of one of a group of ecologically similarspecies because, in the latter case, the remain-ing species could continue to perform the rele-vant ecological functions of that functionaltype. The more species there are in a functionaltype, the less likely it is that a gain or loss of asingle species from that functional type willhave large ecosystem effects. Our challenge, asecologists, is to identify the traits of organismsthat have strong effects on ecosystems; specieswith these traits are likely to be strong inter-actors in ecosystems (Paine 2000).

Species interactions govern the traits that areexpressed most clearly in ecosystems. Theimpact of a species on ecosystem processesdepends on its interactions with other species.The impact of deer on terrestrial vegetation orthe impact of Daphnia on algal biomass oflakes, for example, depends on the density oftheir predators. The mechanisms by whichspecies diversity influences ecosystem pro-cesses often depend on species interactionssuch as competition, facilitation, and predation.

Species Effects on Ecosystem Processes

Species are most likely to have strong ecosys-tem effects when they alter interactive controls,which are the general factors that directly regulate ecosystem processes. These controlsinclude the supply of resources that are essen-tial for primary production, climate, functional

types of organisms, disturbance regime, andhuman activities (see Chapter 1).

Species Effects on Resources

Resource Supply

Species traits that influence the supply of limit-ing resources have major impacts. The supplyof resources required for growth of primaryproducers is one of the interactive controls towhich ecosystem processes are most sensitive(see Chapter 1). These resources include light,nutrients, and, on land, water. For this reason,species traits that alter the supply of limitingresources will substantially alter ecosystemprocesses.

The introduction of a strong nitrogen fixerinto a community that lacks such species cansubstantially alter nitrogen availability andcycling. Invasion by the exotic nitrogen-fixingtree Myrica faya in Hawaii, for example,increased nitrogen inputs, litter nitrogen con-centration, and nitrogen availability (Vitouseket al. 1987) (Fig. 12.3).A nitrogen-fixing invaderis most likely to be successful in ecosystems thatare nitrogen limited; have no strong nitrogenfixers; and have adequate phosphorus, micronu-trients, and light (Vitousek and Howarth 1991).Thus we expect large ecosystem changes frominvasion of nitrogen-fixing species primarily incombinations of the following circumstances:(1) low nitrogen supply (early primary suc-cession in the temperate zone and in other low-nitrogen environments), (2) distant from

N in

puts

(g

m-2

yr-1

)

Myricaabsent

Myricapresent

Myricaabsent

Myricapresent

Myricaabsent

Myricapresent

Net

N m

iner

aliz

atio

n ra

te (

g N

m-2

yr

-1)

Litte

r N

con

cent

ratio

n (%

)

0

1.2

2.4

0

3

6

0

0.7

1.4

Figure 12.3. Impact of the nitrogen-fixing tree Myrica faya on nitrogen inputs, litter nitrogen concentration,and nitrogen mineralization rate in a Hawaiian montane forest (Vitousek et al. 1987).

Species Effects on Ecosystem Processes 269

parent populations of nitrogen-fixing species(e.g., islands), (3) low competition for light orphosphorus (e.g., early in succession and inphosphorus-enriched lakes or soils), and (4)reduction of shading by grazing (e.g., pastures).Exotic nitrogen-fixing species have particularlylarge effects when they escape the suite ofspecies-specific herbivores or pathogens thatoften restrict nitrogen inputs from native nitrogen fixers.

Deep-rooted species can increase the volumeof soil tapped by an ecosystem and thereforethe pool of soil resources available to supportproduction. The perennial bunch grasses thatonce dominated California grasslands, forexample, have been largely replaced either byEuropean annual grasses or by forests of Aus-tralian Eucalyptus.The deep-rooted Eucalyptustrees access a deeper soil profile than do annualgrasses, so the forest absorbs more water andnutrients. In dry, nutrient-limited ecosystems,this substantially enhances ecosystem pro-ductivity and nutrient cycling (Fig. 12.4) butreduces species diversity. The introduction ofdeep-rooted phreatophytes in deserts alsoincreases the productivity in watercourses butreduces diversity, because litter accumulationon the soil surface inhibits the growth of desertannuals (Berry 1970). Deep-rooted species canalso tap nutrients that are available only atdepth. A deep-rooted tundra sedge, for

example, is the only species in arctic tussocktundra that accesses nutrients in the ground-water that flows over permafrost. By tappingnutrients at depth, the productivity of this sedgeincreases 10-fold in sites with abundant ground-water flow, whereas productivity of otherspecies is unaffected by deep resources (Chapinet al. 1988). In the absence of this species,ecosystem productivity and nutrient cyclingwould be greatly reduced. In general, we expectlarge ecosystem impacts from invasion of deep-rooted species where growth-limiting resourcesare available at depth. Ecosystem differences inmaximum rooting depth have implications forregional hydrology and climate.

Nitrification, denitrification and, conse-quently, gaseous nitrogen loss to the atmos-phere are controlled by relatively few speciesof microorganisms. Nitrification rate also influ-ences the susceptibility of nitrogen to leachingloss (see Chapter 7). Changes in the abundanceof microorganisms controlling these processesmight therefore alter nutrient availabilitythrough their effects on nutrient loss (Schimel2001). Mycorrhizal fungi also influence thequantity of nutrients that are available to vegetation (see Chapter 8).

Animals can influence the resource base ofthe ecosystem by foraging in one area anddepositing nutrients elsewhere in feces andurine (see Chapter 14). Sheep, for example,

Forest Grass

14

7

0

Litte

r tu

rnov

er ti

me

(yr)

Forest Grass

600

300

0

Abo

vegr

ound

NP

P (

g m

-2 y

r-1)

Forest Grass

7.0

3.5

0

Litte

r m

ass

(kg

m-2

)

Figure 12.4. Comparison of ecosystem processesbetween two exotic communities that differ inrooting depth: annual grassland and a Eucalyptus

forest in California (Robles and Chapin 1995). Dataare means ± SE.


enrich soils on hilltops where they bed down at night. Migrating salmon perform a similarnutrient-transport role in streams. They feedprimarily in the open ocean and then return tosmall streams where they spawn, die, anddecompose.The nutrients carried by the salmonfrom the ocean can sustain a substantial proportion of the algal and insect productivityof small streams. These nutrient subsidies can be transferred to adjoining terrestrial habitats by bears and otters that feed on salmonor by predators of insects that emerge fromstreams.

Nutrient Turnover

Species differences in litter quality magnify sitedifferences in soil fertility. Differences amongplant species in tissue quality strongly influencelitter decomposition rates (see Chapter 7).Litter from low-nutrient-adapted speciesdecomposes slowly because of the negativeeffects on soil microbes of low concentrationsof nitrogen and phosphorus and high concen-trations of lignin, tannins, waxes, and otherrecalcitrant or toxic compounds. This slowdecomposition of litter from species character-istic of nutrient-poor sites reinforces the lownutrient availability of these sites (Hobbie1992, Wilson and Agnew 1992). Species fromhigh-resource sites, in contrast, produce rapidlydecomposing litter due to its higher nitrogenand phosphorus content and fewer recalcitrantcompounds, enhancing rates of nutrientturnover in nutrient-rich sites.

Experimental planting of species on acommon soil shows that species differences inlitter quality can alter soil fertility quite quickly.Early successional prairie grasses, whose litterhas a low C :N ratio, for example, causes anincrease in the nitrogen mineralization rate ofsoil within 3 years, compared to the same soilplanted with late-successional species whoselitter has a high C :N ratio (Wedin and Tilman1990) (Fig. 12.5).

Seasonality of Resource Capture

Phenological specialization could increaseresource capture. Phenological specialization inthe timing of plant activity can increase the

total time available for plants to acquireresources from their environment. This is mostevident when coexisting species differ in thetiming of their maximal activity. In mixed grass-lands, for example, C4 species are generallymore active in the warmer, drier part of thegrowing season than are C3 species. Conse-quently C3 species account for most earlyseason production, and C4 species account formost late-season production. Similarly, in theSonoran desert, there is a different suite ofannuals that becomes active after winter thanafter summer rains. In both cases, phenological specialization probably enhances NPP andnitrogen cycling. In mixed-cropping agricul-tural ecosystems, phenological specialization ismore effective in enhancing production thanare species differences in rooting depth(Steiner 1982).

The ecosystem consequences of phenologicalspecialization to exploit the extremes of thegrowing season are less clear. Evergreenforests, for example, have a longer photosyn-thetic season than deciduous forests, but mostcarbon gain occurs in midseason in both foresttypes, when conditions are most favorable

14

7

0As Ar Pp Ss Ag

Net

N m

iner

aliz

atio

n ra

te(g

N m

-2 y

r-1)

Species

(44)

(56)(64)

(109)

(122)

C:N ratios

Figure 12.5. Effects of prairie grass species on nitro-gen mineralization when grown on soils with con-taining 100gNm-2 (Wedin and Tilman 1990). Grassesrange from early to late successional in the follow-ing order: Agrostis scabra (As), Agropyron repens(Ar), Poa pratensis (Pp), Schizochyrium scoparium(Ss), and Andropogon gerardi (Ag). Data are means±95% confidence interval (CI). Numbers are the C :N ratios of aboveground biomass.

Species Effects on Ecosystem Processes 271

(Schulze et al. 1977).The early spring growth ofspring ephemeral herbs in deciduous forestsalso has relatively little influence on nitrogencycling because most nitrogen turnover occursin midseason. Phenological specialization is anarea in which species effects on ecosystemprocesses could be important, but these effectshave been well documented primarily in agri-cultural ecosystems.

Species Effects on Climate

Species effects on microclimate influenceecosystem processes most strongly in extremeenvironments. This occurs because ecosystemprocesses are particularly sensitive to climate inextreme environments (Wilson and Agnew1992, Hobbie 1995). Boreal mosses, forexample, form thick mats that insulate the soilfrom warm summer air temperatures. Theresulting low soil temperature retards decom-position, contributing to the slow rates of nutrient cycling that characterize these eco-systems (Van Cleve et al. 1991). Some mossessuch as Sphagnum effectively retain water, aswell as insulating the soil, leading to cold anaer-obic soils that reduce decomposition rate andfavor peat accumulation. The accumulation ofnitrogen and phosphorus in undecomposedpeat reduces growth of vascular plants. Theshading of soil by plants is an important factorgoverning soil microclimate in hot environ-ments. Establishment of many desert cactuses,for example, occurs primarily beneath theshade of “nurse plants.”

Species effects on water and energyexchange can affect regional climate. Speciesdifferences in albedo or the partitioningbetween sensible and latent heat fluxes can

have strong effects on local and regionalclimate. The lower transpiration rate of pasturegrasses compared to deep-rooted tropical trees,for example, could lead to a significantlywarmer, drier climate following widespreadtropical deforestation because of the lowerevapotranspiration and greater sensible heatflux of pastures (see Chapter 2). Changes invegetation caused by overgrazing could alterregional climate. In the Middle East, forexample, overgrazing reduced the cover ofplant biomass. Model simulations suggest thatthe resulting increase in albedo reduced thetotal energy absorbed, the amount of sensibleheat released to the atmosphere, and conse-quently the amount of convective uplift of theoverlying air. Less moisture was thereforeadvected from the Mediterranean Sea,resulting in less precipitation and reinforcingthe vegetation changes (Charney et al. 1977).These vegetation-induced climate feedbackscould have contributed to the desertification ofthe Fertile Crescent. Vegetation changes asso-ciated with fire in the boreal forest can have acooling effect on climate. Late-successionalconifers, which dominate the landscape in theabsence of fire, have a low albedo and stomatalconductance and therefore transfer largeamounts of sensible heat to the atmosphere.Postfire deciduous forests, in contrast, absorbless energy, due to their high albedo, and trans-mit more of this energy to the atmosphere aslatent rather than sensible heat, resulting in lessimmediate warming of the atmosphere andmore moisture available to support precipita-tion (Chapin et al. 2000a) (Fig. 12.6). If thesevegetation changes were widespread, theycould have a negative feedback to high-latitudewarming and reduce the probability of fire.This

Alb

edo

(%)

Eva

potr

ansp

iratio

n(%

of R

n)

Sen

sibl

e he

at(%

of R

n)

Conifer Deciduous Conifer Deciduous Conifer Deciduous0

20 60

0 0

80

Figure 12.6. Sensible and latent heat fluxes from deciduous and conifer boreal forests (Baldocchi et al.2000).


is one of the few negative feedbacks to regionalwarming that has been identified.

Species Effects on Disturbance Regime

Organisms that alter disturbance regimechange the balance between equilibrium andnonequilibrium processes. Following distur-bance, there are substantial changes in mostecological processes, including increasedopportunities for colonization by new individu-als and often an imbalance between inputs toand outputs from ecosystems (see Chapter 13).For this reason, animals or plants that enhancedisturbance frequency or severity increase theimportance of processes, such as colonization,that are particularly important under non-equilibrium conditions. The identity of plantsthat colonize following disturbance, in turnaffects the capacity of the ecosystem to gaincarbon and retain nutrients.

One of the major avenues by which animalsaffect ecosystem processes is through physicaldisturbance (Lawton and Jones 1995, Hobbs1996). Gophers, pigs, and ants, for example,disturb the soil, creating sites for seedling estab-lishment and favoring early successionalspecies (Hobbs and Mooney 1991). Elephantshave a similar effect, trampling vegetation andremoving portions of trees (Owen-Smith 1988).By analogy, the Pleistocene megafauna mayhave promoted steppe grassland vegetation bytrampling mosses and stimulating nutrientcycling (Zimov et al. 1995). The shift towardearly successional or less woody vegetationgenerally leads to a lower biomass, a higherratio of production to biomass, and a litterquality and microenvironment that favordecomposition (see Chapter 13).The associatedenhancement of mineralization can either stimulate production (Zimov et al. 1995) orpromote ecosystem nitrogen loss (Singer et al.1984), depending on the magnitude of disturbance.

Beavers in North America are “ecosystemengineers” that modify the availability ofresources to other organisms by changing thephysical environment at a landscape scale(Jones et al. 1994, Lawton and Jones 1995). The

associated flooding of organic-rich ripariansoils produces anaerobic conditions thatpromote methanogenesis, so beaver pondsbecome hot spots of methane emissions (seeChapter 14) (Roulet et al. 1997). The recentrecovery of beaver populations in NorthAmerica after intensive trapping during thenineteenth and early twentieth centuries hassubstantially altered boreal landscapes, leadingto a fourfold increase in methane emissions inregions where beaver are abundant (Bridghamet al. 1995).

The major ecosystem engineers in soils areearthworms in the temperate zone and termitesin the tropics (Lavelle et al. 1997). Soil mixingby these animals alters soil development andmost soil processes by disrupting the formationof distinct soil horizons, reducing soil com-paction, and transporting organic matter todepth (see Chapter 7).

Plants also alter disturbance regime througheffects on flammability. The introduction ofgrasses into a forest or shrubland, for example,can increase fire frequency and cause thereplacement of forest by savanna (D’Antonioand Vitousek 1992). Similarly, boreal conifersare more flammable than deciduous treesbecause of their large leaf and twig surfacearea, canopies that extend to the groundsurface (acting as ladders for fire to move intothe canopy), low moisture content, and highresin content (Johnson 1992). For this reason,the invasion of the northern hardwood forestsby hemlock in the early Holocene caused anincrease in fire frequency (Davis et al. 1998).The resins in boreal conifers that promote fire also retard decomposition (Flanagan andVan Cleve 1983) and contribute to fuel accumulation.

Plants are often critical in stabilizing soils andreducing wind and soil erosion in early succes-sion. This allows successional development andretains the soil resources that determine thestructure and productivity of late-successionalstages. Introduced dune grasses, for example,have altered soil accumulation patterns anddune morphology in the western United States(D’Antonio and Vitousek 1992), and early suc-cessional alpine vegetation stabilizes soils andreduces probability of landslides.

Species Interactions and Ecosystem Processes 273

Species Interactions andEcosystem Processes

Species interactions modify the impacts of indi-vidual species on ecosystem processes. Mostecosystem processes respond in complex waysto changes in the presence or absence of certainspecies, because interactions among speciesgenerally govern the extent to which speciestraits are expressed at the ecosystem level.Species interactions, including mutualism,trophic interactions (predation, parasitism, andherbivory), facilitation, and competition, mayaffect ecosystem processes directly by modify-ing pathways of energy and material flow orindirectly by modifying the abundances ortraits of species with strong ecosystem effects(Wilson and Agnew 1992, Callaway 1995).

Mutualistic species interactions contributedirectly to many essential ecosystem processes.Nitrogen inputs to terrestrial ecosystems, forexample, are mediated primarily by mutualisticassociations between plants and nitrogen-fixingmicroorganisms. Mycorrhizal associations between plant roots and fungi greatly aid plantnutrient uptake from soil, increase primary pro-duction, and speed succession. Decompositionis accelerated by the presence of highly inte-grated communities (consortia) of soil microor-ganisms in which each species contributes adistinct set of enzymes (see Chapter 7). Many

mutualisms are highly specific, which increasesthe probability that loss of a single species will have cascading effects on the rest of thesystem.

Species that alter trophic dynamics can havelarge ecosystem impacts. When top predatorsare removed, prey populations sometimesexplode and deplete their food resources,leading to a cascade of ecological effects (seeChapter 11). These top–down controls are par-ticularly well developed in aquatic systems. Theremoval of sea otters by Russian fur traders, forexample, allowed a population explosion of seaurchins that overgrazed kelp (Fig. 12.7) (Estesand Palmisano 1974). Recent overfishing in theNorth Pacific may have triggered similar seaurchin outbreaks, as killer whales moved closerto shore in search of food and switched to seaotters as an alternate prey (Estes et al. 1998).In the absence of dense sea urchin populations,kelp provides the physical structure for diversesubtidal communities and attenuates wavesthat otherwise augment coastal erosion andstorm damage.The addition or removal of a fishspecies from lakes often has large keystoneeffects that cascade up or down the food chain(Carpenter et al. 1992, Power et al. 1996a).Many nonaquatic ecosystems also exhibitstrong responses to changes in predator abun-dance (Hairston et al. 1960, Strong 1992, Hobbs1996). Removal of wolves, for example, releasesmoose populations that graze down vegetation,

Naturaldensity

Otterhunted

Naturaldensity

Otterhunted

Naturaldensity

Otterhunted

Otte

r de

nsity

(#

km-2

)

Urc

hin

dens

ity (

# m

-2)

Alg

al c

over

(%

)

0 0

30

15

300

150

00 0

100

50

Figure 12.7. Density of sea otters and sea urchins,and percentage of macroalgal cover. The latter twoparameters were measured at 3 and 9m depth in the

Aleutian Islands of Alaska (Estes and Palmisano1974). Sites differed in otter density due to differen-tial hunting pressure 300yr earlier.


and removal of elephants or other keystonemammalian herbivores leads to encroachmentof woody plants into savannas (Owen-Smith1988). Disease organisms, such as rinderpest inAfrica, can also act as a keystone species bygreatly modifying competitive interactions andcommunity structure (Bond 1993). Plantspecies that are introduced without their host-specific insect herbivores or pathogensoften become aggressive invaders. The cactusOpuntia, for example, became surprisinglyabundant when introduced to Australia, in partdue to overgrazing; but it was reduced to man-ageable levels by a cactus-specific herbivoreCactoblastis. Other species that have becomeaggressive in the absence of their specialist her-bivores include goldenrod (Solidago spp.) inEurope, wild rose (Rosa spp.) in Argentina,and star thistle (Centaurea spp.) in California.

Often these top–down controls by predatorsor pathogens have much greater effect onbiomass and species composition of lowertrophic levels than on the flow of energy ornutrients through the ecosystem (Carpenter et al. 1985), because declines in producer biomassare compensated by increased productivity andnutrient cycling rates by other trophic levels.Intensely grazed grassland systems such as thesouthern and southeastern Serengeti, forexample, have a low plant biomass but rapidcycling of carbon and nutrients due to treadingand excretion by large mammals. Grazing pre-vents the accumulation of standing dead litter,which return nutrients to soil in plant-availableforms (McNaughton 1985, 1988). Keystonepredators or grazers thus alter the pathway ofenergy and nutrient flow, modifying the balancebetween herbivore-based and detritus-basedfood chains, but we know less about theireffects on total energy and nutrients cyclingthrough ecosystems.

Many species effects on ecosystems are indirect and not easily predicted. Species thatthemselves have small effects on ecosystemprocesses can have large indirect effects if theyinfluence the abundance of species with largedirect ecosystem effects, as described fortrophic interactions. Thus a seed disperser orpollinator that has little direct effect on eco-system processes may be essential for persis-

tence of a canopy species with a greater directecosystem impact. Stream predatory inverte-brates alter the behavior of their prey, makingthem more vulnerable to fish predation, whichleads to an increase in the weight gain of fish(Soluck and Richardson 1997).Thus all types oforganisms—plants, animals, and microorgan-isms—must be considered in understanding theeffects of biodiversity on ecosystem function-ing. Although each of these examples is uniqueto a particular ecosystem, the ubiquitous natureof species interactions with strong ecosystemeffects makes these interactions a generalfeature of ecosystem functioning (Chapin et al.2000b). In many cases, changes in these inter-actions alter the traits that are expressed byspecies and therefore the effects of species onecosystem processes. Consequently, simplyknowing that a species is present or absent isinsufficient to predict its impact on ecosystems.There is currently no clear theoretical frame-work to predict when these indirect effects aremost important. Consequently, the introductionor loss of a species, such as a popular sport fish, often generates unanticipated surprises (Carpenter and Kitchell 1993).

Diversity Effects on Ecosystem Processes

Diversity within a functional type may enhancethe efficiency of resource use and retention inecosystems. Many species in a communityappear functionally similar, for example, thenanoplankton in the ocean or the canopy treesin a tropical forest.What are the ecosystem con-sequences of changes in species diversity withina functional type? Evolutionary theory pro-vides some clues. Ecologically similar speciesco-exist in a community in part because ofniche partitioning. In other words, co-existingspecies differ slightly in their responses to environment, perhaps specializing to use dif-ferent soil horizons, canopy heights, or times ofseason. They may also differ in the range oftemperatures or water or nutrient availabilitiesthat they exploit effectively (Tilman 1988).These subtle differences in environmental

Diversity Effects on Ecosystem Processes 275

specialization might increase the efficiency ofresource use by the community if some speciesuse resources that would otherwise not betapped by other species.

In experimental grassland communities, forexample, plots that were planted with a largernumber of species had greater plant cover andlower concentrations of inorganic soil nitrogenthan did low-diversity plots (Fig. 12.8) (Tilmanet al. 1996). The more diverse plots might usemore resources because species have comple-mentary patterns of resource use; in otherwords, species might differ in the types ofresources, the location of their roots, or theirtiming of uptake. Alternatively, diverse plotsmight use resources more effectively becausethey are more likely to have a species that ishighly effective in capturing resources (sampling effect) or are more likely to includespecies with complementary patterns ofresource use (Hooper et al., in press). In othercases, low-diversity ecosystems are quite effi-cient in using soil resources. Crop or forestmonocultures, for example, are often just asproductive as mixed cropping systems (Vandermeer 1995) and mixed-species foreststands (Rodin and Bazilevich 1967). Althoughthere are many examples of a positive relation-ship between species number and productivityor efficiency of resource use, this does notalways occur. The effect of species richness

frequently saturates at a much lower number ofspecies (5 to 10) than characterize most naturalcommunities. Determining the circumstancesand mechanisms in which species number influ-ences ecosystem processes is an active area ofecosystem research (Hooper et al., in press).

Diversity of functionally similar species stabilizes ecosystem processes in the face oftemporal variation in environment. In eco-systems in which functionally similar speciesdiffer in environmental response, this canbuffer ecosystem processes from environmen-tal fluctuations (McNaughton 1977, Chapin andShaver 1985).Tropical tree species, for example,differ subtly in their growth response to nutri-ents (Fig. 12.9). Conditions that favor somespecies will likely reduce the competitiveadvantage of other functionally similar species,thus stabilizing the total biomass or activity bythe entire community. In other words, in com-pensation for the reduced growth by somespecies, other species grow more. For example,in one study, annual variation in weather causedat least a twofold variation production by eachof the major vascular plant species in arctictussock tundra. Years that were favorable forsome species, however, reduced the productiv-ity of others, so there was no significant differ-ence in productivity at the ecosystem scaleamong the 5 years examined (Chapin and Shaver1985). Directional changes in environment canalso cause less change in total biomass than inthe biomass of individual species for similarreasons; some species respond positively to thechange in environment, whereas other speciesrespond negatively. This stabilization of biomassand production by diversity has been observedin many (but not all) studies (Cottingham et al.2001), including grasslands, in response to theaddition of water and nutrients (Lauenroth et al. 1978) and to grazing (McNaughton 1977);in tundra, in response to changes in tempera-ture, light, and nutrients (Chapin and Shaver1985); and in lakes, in response to acidification(Frost et al. 1995). This stability of processesprovided by diversity has societal relevance.Many traditional farmers plant diverse crops,not to maximize productivity in a given year butto decrease the chances of crop failure in a badyear (Altieri 1990). Even the loss of rare species

0.4

0.3

0.2

0.15 10 15 20 25

Species richness treatmentNitr

ate

in r

ootin

g zo

ne (

mg

kg-1

)

0

Figure 12.8. Effect of the number of plant speciessown on a plot on the nitrate concentration in therooting zone. Measurements were made 3 years afterthe plots were sown. Data are means ± SE. (Redrawnwith permission from Nature; Tilman et al. 1996.)


may jeopardize the resilience of ecosystems.Rare species that are functionally similar toabundant ones in rangelands, for example,become more common when their abundantcounterparts are reduced by overgrazing. Thiscompensation in response to release from com-petition minimizes the changes in ecosystemproperties (Walker et al. 1999).

Species diversity also reduces the probabilityof outbreaks by pest species by diluting theavailability of their hosts. Often the spread ofoutbreak species depends strongly on hostdensity.

Diversity provides insurance against changein functioning under extreme or novel condi-tions. Species diversity not only stabilizesecosystem processes in the face of annual variation in environment but also providesinsurance against drastic change in ecosystemstructure or processes in response to extremeevents (Walker 1992, Chapin et al. 1997). Anychange in climate or climatic extremes that issevere enough to cause extinction of onespecies is unlikely to eliminate all membersfrom a functional type (Walker 1995).The morespecies there are in a functional type, the lesslikely it is that any extinction event or series ofsuch events will have serious ecosystem conse-quences (Holling 1986). In a study of temper-ate grassland, for example, an extreme drought

eliminated or reduced the growth of many plantspecies. The drought had least impact on pro-ductivity in those plots with highest speciesdiversity (Fig. 12.10). Results from this fieldexperiment are somewhat difficult to interpretbecause plots of low diversity were the result oflong-term addition of nitrogen fertilizer; thenitrogen addition caused competitive elimina-

20

40

60

80

100N

utrie

nt d

epen

denc

y (%

)

Slow-growing climax species

Rapidly growingpioneer species

Figure 12.9. Graded environmental response ofspecies within a functional type. Nutrient depen-dency is the growth rate of fertilized seedlings ofcanopy trees from the dry tropical forest relative tounfertilized seedlings. Groups of species that show acontinuum in environmental response are more

likely to compensate for the removal of an ecologi-cally similar species than are species that are radi-cally different in their ecological response. (Figureprovided by permission of P. Huante; Huante et al.1995.)

Dro

ught

res

ista

nce

Species richness before drought

1

1/2

1/4

1/8

1/160 5 10 15 20 25

Figure 12.10. Relationship between drought resis-tance of vegetation in a temperate grassland andplant species richness before drought. Drought resis-tance was the log of the ratio of plant biomass at theheight of the drought to plant biomass before thedrought. Data are means ± SE. (Modified with per-mission from Nature; Tilman and Downing 1994.)


tion of the more drought-tolerant species.In a laboratory experiment that manipulatedspecies diversity of mosses, communities withhigh species diversity maintained a higherbiomass when exposed to drought than did less-diverse communities by facilitating the survivalof tall dominant mosses (Mulder et al. 2001).Though theoretical studies also predict suchbuffering or insurance against extreme events,there is relatively little experimental evidenceto test for this effect.

Differences in environmental responseamong functionally different species mayaccentuate ecosystem change. In contrast to thebuffering provided by ecologically similarspecies, species that differ in their response tothe environment and in their effects on eco-system processes can make ecosystems vulner-able to change. Rising concentrations ofatmospheric CO2, for example, can reduce planttranspiration, resulting in increased magnitudeor duration of soil moisture (Owensby et al.1996). This, in turn, can shift the competitivebalance from grasses to shrubs, promotingshrub encroachment into grasslands and savan-nas and causing replacement of one biome byanother.

Global environmental change is causingmany ecosystems to experience novel condi-tions of nitrogen deposition and atmosphericCO2 concentrations. If the principles discussedin this chapter apply broadly, we expect that thediversity of natural ecosystems will be criticalin determining the biotic properties of ecosys-tems (i.e., the diversity of functional types) andtheir vulnerability to change (the buffering provided by diversity within functional types).

Summary

The species diversity of Earth is changingrapidly due to frequent species extinctions(both locally and globally), introductions, andchanges in abundance. We are, however, onlybeginning to understand the ecosystem conse-quences of these changes. Many species havetraits that strongly affect ecosystem processesthrough their effects on the supply or turnoverof limiting resources, microclimate, and distur-

bance regime. The impact of these species traitson ecosystem processes depends on the abun-dance of a species, its functional similarity toother species in the community, and speciesinteractions that influence the expression ofimportant traits at the ecosystem scale.

Diversity per se may be ecologically impor-tant if it leads to complementary use ofresources by different species or increases theprobability of including species with particularecological effects. Because species belonging tothe same functional type generally differ intheir response to environment, diversity withina functional type may stabilize ecosystemprocesses in the face of temporal variation ordirectional changes in environment. Introduc-tion of functionally different species to anecosystem, in contrast, may accelerate the rateof ecosystem change. The effects of speciestraits on ecosystem processes are generally sostrong that changes in the species compositionor diversity of ecosystems are likely to altertheir functioning, although the exact nature ofthese changes is frequently difficult to predict.

Review Questions

1. What are functional types? What is the use-fulness of the functional-type concept if allspecies are ecologically distinct?

2. How is the expected ecosystem impact of theloss of a species affected by (a) the numberof species in the ecosystem, (b) the abun-dance or dominance of the species that iseliminated, or (c) the type of species that is eliminated? Explain.

3. If a new species invades or is lost from anecosystem, which species traits are mostlikely to cause large changes in productivityand nutrient cycling? Give examples thatillustrate the mechanisms by which thesespecies effects occur.

4. Which species traits have greatest effects onregional processes such as climate andhydrology?

5. How do species interactions influence theeffect of a species on ecosystem processes?

6. How does the diversity of species within afunctional type affect ecosystem processes?


What is the mechanism by which this occurs?Why is it important to distinguish betweenthe effects of changes in species compositionwithin vs. between functional types?

7. What are the mechanisms by which speciesdiversity might affect nutrient uptake or lossin an ecosystem. Suggest an experiment todistinguish between these possible mecha-nisms. Design an agricultural ecosystem thatmaintains crop productivity but has tightnutrient cycles.

Additional Reading

Chapin, F.S. III, E.S. Zaveleta, V.T. Eviner, R.L.Naylor, P.M. Vitousek, S. Lavorel, H.L. Reynolds,D.U. Hooper, O.E. Sala, S.E. Hobbie, M.C. Mack,and S. Diaz. 2000. Consequences of changing bioticdiversity. Nature 405:234–242.

Frost,T.M., S.R. Carpenter,A.R. Ives, and T.K. Kratz.1995. Species compensation and complementarityin ecosystem function. Pages 224–239 in C.G.Jones, and J.H. Lawton, editors. Linking Speciesand Ecosystems. Chapman & Hall, New York.

Johnson, K.G., K.A. Vogt, H.J. Clark, O.J. Schmitz,and D.J. Vogt. 1996. Biodiversity and the pro-ductivity and stability of ecosystems. Trends inEcology and Evolution 11:372–377.

Lawton, J.H., and C.G. Jones. 1995. Linking speciesand ecosystems: Organisms as ecosystem engi-neers. Pages 141–150 in C.G. Jones, and J.H.Lawton, editors. Linking Species and Ecosystems.Chapman & Hall, New York.

Power, M.E., D. Tilman, J.A. Estes, B.A. Menge,W.J. Bond, L.S. Mills, G. Daily, J.C. Castilla,J. Lubchenco, and R.T. Paine. 1996. Challenges inthe quest for keystones. BioScience 46:609–620.

Vandermeer, J. 1995. The ecological basis of alterna-tive agriculture. Annual Review of Ecology andSystematics 26:201–224.

Vitousek, P.M. 1990. Biological invasions and ecosystem processes: Towards an integration ofpopulation biology and ecosystem studies. Oikos57:7–13.

Walker, B.H. 1995. Conserving biological divers-ity through ecosystem resilience. ConservationBiology 9:747–752.

Wilson, J.B., and D.Q. Agnew. 1992. Positive-feedback switches in plant communities. Advancesin Ecological Research 23:263–336.

Introduction

Ecosystems are always recovering from pastchanges. In earlier chapters we emphasizedecosystem responses to the current environ-ment. Ecosystems are, however, always re-sponding to past changes that have occurredover all time scales (Holling 1973, Wu andLoucks 1995). These changes include relativelypredictable daily and seasonal variations, lesspredictable changes in weather (e.g., passage of weather fronts, El Niño events, and glacialcycles), and occurrence of disturbances (e.g.,tree falls, herbivore outbreaks, fires, and vol-canic eruptions). Consequently, the behavior ofan ecosystem is always influenced by both thecurrent environment and many previous envi-ronmental fluctuations and disturbances.

The global environment is changing morerapidly than it has for millions of years. Thesechanges result from an exponentially risinghuman population that shows an every-increasing technological capacity to alterEarth’s environment and ecosystems. Perhapsthe most urgent need in ecosystem ecology is toimprove our understanding of factors govern-ing the stability and change in ecologicalsystems (Box 13.1). This understanding is criti-cal to managing ecosystems so they sustaintheir diversity and other important ecological

attributes and so ecosystems continue toproduce the goods and services required bysociety. This chapter addresses the basis of thetemporal dynamics of ecosystems.

Fluctuations in Ecosystem Processes

Interannual Variability

Ecosystem processes measured in 1 year areseldom representative of the long-term mean.Many ecosystem processes are sensitive tointerannual variations in weather and to fluctu-ations in the internal dynamics of ecosystems,such as outbreaks of herbivores or pathogens.The same ecosystem, for example, can changefrom being a carbon source in one year to acarbon sink in the next (Goulden et al. 1998).The production at a particular trophic level canchange from being limited by food to beinglimited by predation. Some of the interannualvariability in ecosystem dynamics reflectsprocesses that are potentially predictable, suchas the cyclic variation in climate. El Niño eventsare a consequence of large-scale oscillations inthe global ocean-atmosphere system that recurevery 2 to 10 years (see Chapter 2). These areassociated with relatively repeatable climatic

13Temporal Dynamics

Ecosystem processes constantly change in response to variation in environment overall time scales. This chapter describes the major patterns and controls over the tem-poral dynamics of ecosystems.

281

282 13. Temporal Dynamics

An emerging challenge in ecosystem ecologyis to improve our understanding of the prop-erties and processes that allow ecosystems topersist in the face of environmental change.Ecological stability is an issue that hasintrigued ecologists for decades. Discussionsof stability initially focused on changes innatural populations of plants and animalsbut have since been broadened to includeecosystem processes. Although we may havean intuitive feel for what constitutes a stableecosystem, stability is difficult to define.There are at least 160 different definitions ofstability in the literature (Grimm and Wissel1997). Despite the wide range of definitions,some fundamental generalizations emergethat provide a basis for discussing ecosystemresponses to change.

The response of any system to a perturba-tion (i.e., a external force that displaces thesystem from equilibrium) can be describedin terms of a few general properties (Holling1986, Berkes and Folke 1998) (Fig. 13.1).The response of a system to perturbationdescribes the direction and magnitude ofchange in the system after a perturbation.The resistance of a system describes its ten-dency to remain in its reference state in theface of a perturbation; in ecological terms,resistance is the capacity of a system tomaintain certain structures and functionsdespite disturbance. An ecosystem thatshows little change in structure, productivity,or rate of nutrient cycling in response to adrought or fire, for example, is resistant tothose perturbations. The resistance of anecosystem to perturbation depends on thenature, magnitude, and duration of the per-turbation as well as on the nature of thesystem. Ecosystems are often particularlyvulnerable to new types of disturbances thatthey have not previously experienced. Therecovery of a system describes the extent towhich it returns to its original state after per-turbation.An ecosystem recovers, if negativefeedbacks that resist changes in ecosystemproperties are stronger than positive feed

backs that push the ecosystem toward somenew state (see Chapter 1). The recovery ofan ecosystem depends on the magnitude of the response and the time since the perturbation.

Ecological resilience has been defined inmany ways. One use of the term denoteselasticity, or the rate at which a systemreturns to a reference state following per-turbation (May 1973, Pimm 1984). Systemswith low resilience may never recover totheir original state and are readily convertedto a new state (Berkes and Folke 1998). The

Box 13.1. Ecosystem Resilience and Change

Large disturbance or low resistance

Recovery

Recovery

Response

Response

Resilience

Low resilienceHigh resilience

Small disturbance or high resistance

Original state

Original state

Ecological state

Pro

babi

lity

of c

hang

e

A

B

Figure 13.1. Properties of a system that influenceits probability of changing state. The solid ballrepresents the state of the system after a pertur-bation.The open ball shows the most likely futurestates of the system. A, A system shows a smallresponse to a perturbation if the perturbation issmall or the system is resistant to change. B,Aftera perturbation, a system can assume many possi-ble states; if it is highly resilient, it may returnquickly to its original state; if it is less resilient, orif the perturbation is large, the system may moveto a new state.

Fluctuations in Ecosystem Processes 283

patterns, such as drought in Southeast Asia and the continental interior of North America and rains in southwestern North America.Glacial cycles occur over thousands of years inresponse to variations in solar input and feed-backs from the biosphere to climate (seeChapter 2). An emerging challenge of climateand ecosystem modeling is to understand theseclimate oscillations well enough to predict orexplain their effects on interannual variation inecosystem processes.

The internal dynamics of ecosystems alsogenerate large interannual fluctuations inecosystem processes. The population density ofherbivores, for example, can vary more than100-fold over a few years, causing large fluctu-ations in plant biomass, nutrient cycling, andother processes (Fig. 13.2). The causes of thesefluctuations and cycles are debated but proba-bly reflect interactions among plants, herbi-vores, predators, and parasites. One importantinteraction, for example, may occur betweenplants and their herbivores. Herbivore popula-tions often decline after a depletion of theirfood supply, due to insufficient food and/orbuildup of predators. Theoretical studies sug-

gest that, if the recovery of an herbivore popu-lation lags behind the vegetation recovery, thisgenerates a population cycle, with a period thatis four times the length of the time lag (May1973). When snowshoe hares heavily browsetheir food, for example, plants produce newshoots that remain toxic for 2 to 3 years. Thiscontributes to low densities of hare populationsfor 2 to 3 years after a population crash. Thistime lag may contribute to the 11-year periodin the snowshoe hare cycle (Bryant 1981). Animportant challenge for ecosystem ecologists isto improve understanding of the interactionsbetween external factors, such as climate, andinternal ecosystem dynamics in causing thenatural variability in ecosystem processes.

Long-Term Change

Today’s ecosystem processes depend on boththe current environment and past events. Lega-cies are the persistent effect of past events.Legacies affect ecosystem processes over awide range of time scales. Individual redwoodtrees (Sequoia sempervirens) in coastal California, for example, can live for thousands

close interdependence of resilience andresistance and the many alternative defini-tions of these terms that have been proposedhave contributed to the confusion in termi-nology related to resilience. The importantpoint is that systems that maintain theirproperties despite disturbance (i.e., are resis-tant to change) and that return rapidly totheir original state (i.e., are resilient) exhibitmore stable and predictable ecosystem prop-erties (Holling 1986).

These properties allow us to assess thesensitivity of ecosystem processes to change.They imply, however, that an ecosystem canbe characterized by a reference state that istypical of the system. All ecosystems arehighly resilient to regular daily and seasonalvariations. Their properties may change,however, in response to directional changesin environment or to disturbances that are

particularly novel or severe (Fig. 13.1). Thechanges that are most important to considerare those that have large effects and occuron time scales at least as long as those of theprocesses we are studying. The biomass of aforest stand, for example, is relatively resis-tant to rapid environmental variations, suchas interannual variations in climate; biomassis, however, influenced by more long-lastingeffects caused by wind storms, recent herbi-vore outbreaks, and successional develop-ment. Net primary production respondsmore rapidly to environment than doesbiomass and requires additional considera-tion of shorter-term changes, such as inter-annual variation in climate. Resilience musttherefore be defined with respect to particu-lar ecosystem properties and perturbations,rather than being considered an absoluteproperty of an ecosystem.


of years. During the Tertiary they occupied awarm moist environment throughout much ofwestern North America. Their range is nowrestricted to valleys where coastal fog mini-mizes summer drought stress. Aspen clones inthe Rocky Mountains range in age from a fewyears to as much as 10,000 years (Kempermanand Barnes 1976, Tuskan et al. 1996). Thecurrent distribution of redwoods and aspens is therefore a product of past population andcommunity processes and cannot be fullyunderstood with reference only to the presentenvironment. Many species are still migratingpoleward in response to the disappearance ofcontinental ice sheets 10,000 years ago. Thesoils beneath these recent arrivals may stillreflect the properties of earlier communitiesrather than being completely a function ofcurrent vegetation.

The current functioning of the East Siberiancoastal plain also reflects processes that oc-curred thousands of years ago. During the Pleistocene, this region was a steppe grasslandthat accumulated highly organic loess soils andice lenses that now occupy 50 to 70% of the soilvolume (Zimov et al. 1995). As the ice meltedin the warmer Holocene climate, the soils sub-sided, forming lakes with organic-rich sedi-ments. Pleistocene-age organic matter is nowthe major carbon source for methane produc-tion by these lakes (Zimov et al. 1997). In otherwords, the current processing of carbon in theselakes is strongly influenced by processes thatoccurred 10,000 to 100,000 years ago.

Current ecosystem processes also respond to changes that have occurred more recently.

Large areas of Europe and northeastern NorthAmerica were deforested for agriculture inrecent centuries and have more recentlyreverted to forests. Even forests older than 200years still exhibit composition and dynamicsthat reflect their earlier history (Foster andMotzkin 1998) (Fig. 13.3). A plow layer is stillevident in these forests, for example, resultingin a sharp vertical discontinuity in soil pro-cesses and nutrient supply. Net primary pro-duction (NPP) substantially exceeds the ratesof heterotrophic respiration in these ecosys-tems, so vegetation and soils are actively accu-mulating carbon (Goulden et al. 1996). Anystudy of carbon dynamics that ignored thesehistorical legacies would seriously misinterpretthe relationship between carbon balance of

1972 1975 1978 1981

Flow

ers

(num

ber m

-2)

Smallrodents

50

Sm

all r

oden

ts (i

ndex

)

160

Year

80

0

25

0

Flowers

Figure 13.2. Interannual variation in floweringdensity of an understory shrub (Vaccinium myrtillus)and of small rodents in northern Finland (Laine and

100

80

60

40

20

0

100

80

60

40

20

01650 1750 1850 1950

Cultivated

Pasture

Forest

Land

scap

e co

mpo

sitio

n (%

of l

and)

Popu

latio

n de

nsity

(peo

ple

km-2

)

Year

Figure 13.3. Changes in land use and populationdensity in central Massachusetts (northeasternUnited States) since European colonization. Mostforests in this region were previously croplands or pastures. (Redrawn with permission from Ecosystems; Foster and Motzkin 1998.)

Henttonen 1983). These herbivores and their foodplants show approximately 4-year cycles of abun-dance.

Disturbance 285

ecosystems and current climate. Similarly, thelegacies of the current environment will exertlarge effects on the future structure and func-tioning of ecosystems.

Disturbance

Conceptual Framework

Disturbance is a major cause of long-term fluc-tuations in the structure and functioning ofecosystems. We define disturbance as a rela-tively discrete event in time and space thatalters the structure of populations, communi-ties, and ecosystems and causes changes inresource availability or the physical environ-ment (Pickett and White 1985, Pickett et al.1999). Many natural disturbances, such as herbivore outbreaks, treefalls, fires, hurricanes,floods, glacial advances, and volcanic eruptions,exert these effects through reductions in liveplant biomass or sudden changes in the pool ofactively cycling soil organic matter. Distur-bance is difficult to define unambiguously.Events such as intensive grazing and subzerotemperatures that are normal features of someecosystems seriously disrupt the functioning ofothers. Disturbance must therefore be definedin the context of the normal range of environ-mental variation that an ecosystem experi-ences. The dividing line between disturbanceand normal function is somewhat arbitrary.Herbivory, for example, is often treated as partof the steady-state functioning of ecosystems,whereas stand-killing insect outbreaks aretreated as disturbances. The processes aresimilar, however, and there is a continuum insize, severity, and frequency between these twoextremes. Disturbance is clearly not an externalevent that “happens” to an ecosystem. Likeother interactive controls, disturbance is anintegral part of the functioning of all eco-systems that responds to and affects most ecosystem processes. Naturally occurring dis-turbances such as fires and hurricanes aretherefore not “bad”; they are normal propertiesof ecosystems.

Human activities have altered the frequencyand size of many natural disturbances, such as

fires and floods, and have produced new typesof disturbance, such as large-scale logging,mining, and wars. Many human disturbanceshave ecological effects that are similar to thoseof natural disturbances, so the study of eithernatural or human disturbances providesinsights into the regulation of ecosystemprocesses and human impacts on these pro-cesses. Natural and anthropogenic disturbancesinteract with environmental gradients to createmuch of the spatial patterning in landscapes(see Chapter 14).

After disturbance, ecosystems undergo succession, a directional change in ecosystemstructure and functioning resulting from bioti-cally driven changes in resource supply. Distur-bances that remove live or dead organic matter,for example, are colonized by plants that grad-ually reduce the availability of light at the soilsurface and alter the availability of water andnutrients (Tilman 1985). If there were nofurther disturbance, succession would proceedtoward a climax, the end point of succession(Clements 1916). At this point, the structureand rates of ecosystem processes approach a steady state in which resource demand byvegetation would be balanced by the rate ofresource supply. In practice, however, new dis-turbances usually occur before successionreaches a climax, so individual stands of anecosystem are seldom in steady state. None-theless, the concept of directional changes invegetation after disturbance provides a usefulframework for analyzing the role of distur-bance in ecosystem processes.

Succession occurs in response to bioticallydriven changes in resource supply, which typically occur over time scales of years to centuries. Succession does not thereforeinclude the seasonal fluctuations in ecosystemprocesses from summer to winter, which aredriven more directly by climate than by theinternal dynamics of ecosystems.

Disturbance Properties

The impact of disturbance on ecosystem pro-cesses depends on its severity, frequency, type,size, timing, and intensity. Together these attrib-utes of disturbance constitute the disturbance


regime (Heinselman 1973). Disturbance regimeis a filter that influences the types of organismspresent and therefore the structure and func-tioning of ecosystems.

Disturbance severity is the magnitude ofchange in resource supply or environmentcaused by disturbance. For those disturbancesthat remove vegetation and/or soils, distur-bance severity is the quantity of organic matterremoved from plants or soil by the disturbance.Primary succession occurs after severe distur-bances that remove or bury most products ofecosystem processes, leaving little or no organicmatter or organisms. Disturbances leading toprimary succession include volcanoes, glaciers,landslides, mining, flooding, coastal dune for-mation, and drainage of lakes. Secondary suc-cession occurs on previously vegetated sitesafter disturbances such as fire, hurricanes,logging, and agricultural plowing. These distur-bances remove or kill most live abovegroundbiomass but leave some soil organic matter andplants or plant propagules in place. Disturbanceseverity is probably the major factor determin-ing the rate and trajectory of vegetation devel-opment after disturbance. A severe fire thatkills all plants, for example, has a differenteffect on vegetation recovery than does a firethat burns only surface litter, allowing surviving

vegetation to resprout. There is also a contin-uum in disturbance severity between large-scale defoliation events and the removal of asingle leaf by a caterpillar or between land-slides and the burial of surface litter by anearthworm. In other words, there is a contin-uum in disturbance severity between the day-to-day functioning of ecosystems and eventsthat initiate primary succession (Fig. 13.4).

Many disturbances can be characterized byintensity, the energy released per unit area andtime. Intensity of a disturbance often influencesthe severity of its effect. Intense marine storms,for example, dislodge many intertidal organ-isms, and intense hurricanes blow down manytrees. Fire intensity is rate of heat production.It depends on the mass, consumption rate, andenergy content of the fuel (Johnson 1992) anddetermines the temperatures that organismsexperience and therefore their probability ofsurviving fire. Intense fires are not alwayssevere, however. In fact, slow smoldering firesof low intensity frequently consume more fuel(are more severe) than are intense fires thatmove rapidly through a stand.

Disturbance frequency varies dramaticallyamong ecosystems and among disturbancetypes. Herbivory occurs continuously in mostecosystems. At the opposite extreme, volcanoes

Steadystate

Secondarysuccession

Primarysuccession

Herbivory

Fire

Agricultural clearing

Flooding

Mining and wars

Glaciers andvolcanoes

Disturbance severity (% of organic material removed)0 100

Figure 13.4. Spectrum of disturbance severity associated with major types of disturbance, ranging fromnormal steady-state functioning of ecosystems to primary succession.

Disturbance 287

or floods may never have occurred in somelocations. Average fire frequency ranges fromonce per year in some grasslands to once everyseveral thousand years in some mesic forests.Ecosystems are usually most resilient to distur-bances that occur frequently. Ecosystems thatexperience frequent fire, for example, supportfire-adapted species that recover biomass morequickly than in ecosystems in which fire occursinfrequently (Fig. 13.5). Human activities oftenmodify disturbance frequency through initia-tion or suppression of disturbance. Damming ofstreams can eliminate spring floods that scoursediments and detritus from channels, resultingin large changes in stream food webs and capac-ity to support fish (Power 1992a). Fire suppres-sion in the giant sequoias (Sequoiadendrongigantea) of the Sierra Nevada mountains ofCalifornia made this ecosystem more vulnera-ble to fire as a result of the growth of under-story trees that formed a ladder for fire to reachfrom the ground to the canopy. Although thethick-barked sequoias are resistant to groundfires, they are vulnerable to fires that extendinto the canopy. In this way, fire suppression hasincreased the risk of catastrophic fires thatcould eliminate giant sequoias.

Disturbance type influences ecosystemprocesses independent of frequency and sever-

ity. Organisms adapt to disturbances that occurrelatively frequently in their current environ-ment or in their evolutionary past. They areoften vulnerable to novel disturbances. Benthiccommunities, for example, may recover slowlyfrom bottom trawling that scrapes surface sed-iments. Many upland species are intolerant of flooding, whereas many trees from wet envi-ronments, such as tropical wet forests, have thinbark and are killed by fire. Some traits, such asheat-induced germination of chaparral postfireannuals, enable species to respond appro-priately to specific types of disturbance. Othertraits enable species to colonize many types of disturbances. Weedy species, for example,produce abundant small seeds that disperselong distances or remain dormant in the soilfrom one disturbance to the next. Their germi-nation is often triggered by environmental con-ditions characteristic of most disturbed sites(Fenner 1985, Baskin and Baskin 1998), so theyare relatively insensitive to disturbance type.Novel disturbances are more likely to lead toslow recovery or to trigger a new successionaltrajectory than are disturbances to whichorganisms are well adapted.

Disturbance size is highly variable. Gap-phase succession, for example, occurs in smallgaps created by the death of one or a few plants.Many tropical wet forests or intertidal commu-nities, for example, are mosaics of gaps of different ages. Other ecosystems develop afterstand-replacing disturbances that can be hun-dreds of square kilometers in area. Disturbancesize influences ecosystems primarily through itseffects on landscape structure, which influenceslateral flow of materials, organisms, and distur-bance among patches in the landscape (seeChapter 14). Disturbance size, for example,affects the rate of seed input after fire. Smallfires are readily colonized by seeds that blow in from surrounding unburned patches or arecarried by mammals and birds, whereas regen-eration in the middle of large fires, fields, orclearcuts may be limited be seed availabilityand be colonized primarily by light-seededspecies that disperse long distances. Dis-turbance size also influences the spread of herbivores and pathogens that colonize early

0

5

10

15

20

25

0 10010

Fire frequency (yr) [log scale]

T50

(yr

)

Figure 13.5. Relationship of fire frequency to thetime (t50) required for an ecosystem to accumulate50% of its maximum biomass (Chapin and Van Cleve1981). Ecosystems with frequent disturbance recovermore quickly.


successional sites. Disturbance pattern on thelandscape influences the effective size of a dis-turbance event. Disturbances often leaveislands of undisturbed vegetation that act aspropagule sources, causing the effective size ofthe disturbance to be much smaller than its areawould suggest (Turner et al. 1997). A series ofdams on a river creates a chain of intercon-nected lakes that can be colonized much morereadily than isolated kettle lakes formed afterglacial retreat or farm ponds formed by restrict-ing groundwater flow.

The timing of disturbance often influences itsimpact. A strong freeze or fire that occursduring bud break has greater impact than onethat occurs 2 weeks before bud break. Similarly,anaerobic conditions associated with floodingof the Mississippi River during the 1993 grow-ing season caused more root and tree mortalitythan if the flood had occurred when roots wereinactive. Hydroelectric dams may eliminateseasonal flooding associated with rain orsnowmelt and alter flow based on electricitydemand. Human activities often change thetiming of disturbances such as grazing, fire, andflooding.

Disturbance is one of the key interactive controls that governs ecosystem processes (seeChapter 1) through its effects on other interac-tive controls (microclimate, soil resourcesupply, functional types of organisms, and prob-ability of future disturbance). Postfire stands,for example, often have warm soils that have ahigh water content; this occurs due to the lowalbedo of the charred surface and the decreasein leaf area that transpires water and shades thesoil. Fire both volatilizes nitrogen, which is lostfrom the site, and returns inorganic nitrogenand other nutrients to the soil in ash, thus alter-ing soil resource supply. The net effect of fire is usually to enhance nutrient availability,although the magnitude of this effect dependson fire severity and intensity (Wan et al. 2001).Fire affects the functional types of plants in anecosystem through its effects on differentialsurvival and competitive balance in the postfireenvironment. Because of its sensitivity to, andeffect on, other interactive controls, changes indisturbance regime alter the structure and func-tioning of the ecosystem.

Succession

Successional changes occurring over decades tocenturies explain much of the local variationamong ecosystems. Although climate, soils, andtopography explain most of the broad globaland regional patterns in ecosystem processes,disturbance regime and postdisturbance suc-cession account for many of the local pat-terns of spatial variability (see Chapter 14). Inthis section, we describe common patterns of successional change in major ecosystemprocesses. These successional changes are mostclearly delineated in primary succession, so webegin with a description of primary succes-sional processes and then describe how the pat-terns differ between primary and secondarysuccession.

Ecosystem Structure and Composition

Primary Succession

Succession involves a change from a commu-nity governed by the dynamics of colonizationto one governed by competition for resources.Vegetation development after disturbance isstrongly influenced by the initial colonizationevents, which in turn depend on environmentand the availability of propagules (Egler 1954,Connell and Slatyer 1977, Bazzaz 1996). Severedisturbances such as glaciers, volcanic erup-tions, and mining eliminate most traces of pre-vious vegetation and must be colonized fromoutside the disturbed site. Most initial colo-nizers of these primary successional sites havesmall seeds that can disperse long distances bywind. Fresh lava or land exposed by retreat ofglaciers, for example, is first colonized by wind-dispersed spores of algae, cyanobacteria, andlichens, which form crusts that stabilize soils(Worley 1973). These are followed by small-seeded wind-dispersed vascular plants (primar-ily woody species), whose arrival rates dependlargely on distance to seed source (Shiro anddel Moral 1995). Late successional species withheavier seeds generally arrive more slowly (Fig.13.6).

Many of the species that become abundantearly in primary succession are free living or

Succession 289

symbiotic nitrogen fixers.Vascular plant speciescapable of symbiotic nitrogen fixation occurfrequently (about 75% of sites studied) in earlyprimary succession, although they dominate thevegetation only about 25% of the time (Walker1993). These species are most common onglacial moraines and mudflows; intermediateon mine tailings, landslides, floodplains, anddunes; and least abundant on volcanoes androck outcrops (Walker 1993). When early suc-cessional colonizers fix abundant nitrogen, theirnet effect is generally to promote the establish-ment and growth of later successional species(Walker 1993) (Fig. 13.7).

The long-term successional trajectory of veg-etation is strongly influenced by the speciescomposition of the initial colonizers becausethe opportunities for colonization decline assuccession proceeds. In many forests, all treespecies colonize in early succession, and thesuccessional changes in dominance reflectspecies differences in size and growth rate(Egler 1954, Walker et al. 1986). In other cases,late successional species may establish moregradually. As succession proceeds, the soilbecomes covered by leaf litter, making a lessfavorable seed bed, and there is increasing competition for light and other resources. The

60

40

20

0

80

60

40

20

0

Primary successional colonizers

Secondary successional colonizers

Late successional species60

40

20

01 2 30

RGR (week-1)

Freq

uenc

y (%

of t

otal

spe

cies

)

Primary successional colonizers

Secondary successional colonizers

Late successional species30

15

0

30

15

0

Seed mass (mg seed -1)

30

15

0

Freq

uenc

y (%

of t

otal

spe

cies

)

0.030.1

0.31

310

30

Figure 13.6. Frequency distribution of log seed massand relative growth rate (RGR) for British speciesthat are primary successional colonizers, secondarysuccessional colonizers, and late successional species.

(Data from Grime and Hunt 1975; Grime et al. 1981.Redrawn with permission from Blackwell Scientific;Chapin 1993b.)


effects of initial colonizers on their physicalenvironment also influence the identity ofspecies that can subsequently establish andcompete effectively. After volcanic eruption in Hawaii, for example, there is usually a slowsuccession from short-statured vegetation dominated by algal crusts, herbaceous plants,and small shrubs to forests dominated by slowly growing tree-ferns and trees. An exotic nitrogen-fixing tree, Myrica faya, whose seedsare brought to early or mid-successional sitesby birds, can, however, add sufficient nitrogento alter substantially the nitrogen supply, pro-duction, and species composition of vegetation

and therefore the successional trajectory(Vitousek et al. 1987) (Fig. 13.3).

In some cases, one or a few successional path-ways predominate because there are only a fewpostdisturbance combinations of environmentand potential colonizers. In other cases, multi-ple successional pathways are possible. Afterglacial retreat in 1800 at Glacier Bay, Alaska,for example, Populus (poplar) and Picea(spruce) were the major initial colonizers.Further retreat of the glacier, however, broughtearly successional habitat within dispersal distance of nitrogen-fixing alders, which thenbecame an important early successional species

Stage

Facilitativeeffects

Pioneer Dryas Alder Spruce

SOM

N

Mycorrhizae

Growth (weak)

Growth

Germination

Inhibitoryeffects

Germination

Survivorship

Survivorship

Seed predationand mortality

N (weak)

Germination

Survivorship


Root competition

Light competition

Growth

Survivorship


Root competition

Light competition

N

Germination(weak)

Life history patterns

Dominance bylight-seeded

species

Dominance by tallshrubs of intermediate

longevity

Dominance bylong-lived

trees

Dominance byrapidly growing

species

Impacts ofherbivory

Minimal Reduce growthof early

successionalspecies

Eliminate earlysuccessional

species

Minimal

Figure 13.7. Interaction of life history patterns,facilitative and inhibitory effects, and herbivory incausing successional change after glacial retreat atGlacier Bay, Alaska. Life history patterns determinethe type of species that dominate at each succes-sional stage. The rate at which this dominancechanges is determined by facilitative or inhibitoryeffects of the dominant species and by patterns of

herbivory. In general, all four of these processes con-tribute simultaneously to successional change, withthe most important processes being life history patterns in the pioneer stage, herbivory in mid-successional stages, facilitation in the alder stage, andinhibition in late succession. SOM, soil organicmatter. (Modified with permission from EcologicalMonographs; Chapin et al. 1994.)

Succession 291

(Fastie 1995). Alders increased the nitrogeninputs and long-term productivity of later successional stages (Bormann and Sidle 1990).The late-successional forests on older sites atGlacier Bay therefore followed a different suc-cessional trajectory than forests on youngersites. Human activities strongly affect both thepostdisturbance environment and availabilityof propagules, so future trajectories of succes-sion will likely differ from those that currentlypredominate.

Secondary Succession

Secondary succession differs from primary suc-cession in that many of the initial colonizers arealready present on site immediately after dis-turbance. They may resprout from roots orstems that survived the disturbance or germi-nate from a soil seed bank—seeds producedafter previous disturbance events and thatremain dormant in the soil until postdistur-bance conditions (light, wide temperature fluc-tuations, and/or high soil nitrate) triggergermination (Fenner 1985, Baskin and Baskin1998). In many forests there is also a seedlingbank of large-seeded species that show negligi-ble growth beneath the dense shade of a forestcanopy but grow rapidly in treefall gaps tobecome the next generation of canopy domi-nants. Other colonizers of secondary successiondisperse into the disturbed site from adjacentareas, just as in primary succession. Those dispersing species include both small-seeded,wind-dispersed species and large-seeded,animal-dispersed species (Fig. 13.6). Initial col-onizers grow rapidly to exploit the resourcesmade available by disturbance. Gap-phase suc-cession is seldom limited by propagule avail-

ability (Shugart 1980), whereas the successionaltrajectory of large disturbed sites may dependon the species that disperse to the site (Fastie1995).

The changes in species composition thatoccur after the initial colonization of a siteresult from a combination of (1) the inherentlife history traits of colonizers, (2) facilitation,(3) competitive (inhibitory) interactions, (4)herbivory, and (5) stochastic variation in theenvironment (Connell and Slatyer 1977, Pickettet al. 1987, Walker 1999). Life history traitsinclude seed size and number, potential growthrate, maximum size, and longevity. These traitsdetermine how quickly a species can get to asite, how quickly it grows, how tall it gets,and how long it survives. Most early secondarysuccessional species arrive soon after a dis-turbance, grow quickly, are relatively shortstatured, and have a low maximum longevity,compared to late-successional species (Nobleand Slatyer 1980) (Fig. 13.6; Table 13.1). Evenif no species interactions occurred during suc-cession, life history patterns alone would causea shift in dominance from early to late succes-sional species because of differences in arrivalrate, size, and longevity.

Facilitation involves processes in which earlysuccessional species make the environmentmore favorable for the growth of later succes-sional species. Facilitation is particularly impor-tant in severe physical environments, such asprimary succession, where nitrogen fixation andaddition of soil organic matter by early succes-sional species ameliorates the environment andincreases the probability that seedlings of otherspecies will establish and grow (Callaway 1995,Brooker and Callaghan 1998). Competition isan interaction among two organisms or species

Table 13.1. Successional changes in life history patterns after glacial retreat in Glacier Bay, Alaska.

Successional Seed mass Maximum Age at first MaximumGenus stage (mg seed-1) height (m) reproduction (yr) longevity (yr)

Epilobium pioneer 72 0.3 1 20Dryas Dryas 97 0.1 7 50Alnus alder 494 4 8 100Picea spruce 2694 40 40 700

Data from Chapin et al. (1994).


that reduces the availability of resources toother individuals. Both competitive and facili-tative interactions are widespread in plantcommunities (Callaway 1995, Bazzaz 1996);their relative importance in causing changes in species composition during succession pro-bably depends on environmental severity(Connell and Slatyer 1977, Callaway 1995).Herbivores and pathogens account for much of the mortality of early successional plants.Selective browsing by mammals is particularlyimportant in eliminating early successionalspecies as succession proceeds (Bryant andChapin 1986, Paine 2000). In intertidal commu-nities, grazing by fish and invertebrates such aslimpets exerts a similar effect.

In general, life history traits generally de-termine the pattern of species change throughsuccession, whereas facilitation, competition,and herbivory determine the rate at which thisoccurs (Chapin et al. 1994). These processesinteract with other less predictable events, suchas storms or droughts, to cause the diversity of successional changes that occur in naturalecosystems (Pickett et al. 1987, Walker 1999)(Fig. 13.7).

Secondary succession can begin with soilsthat have either high or low nutrient availabil-ity. When initial nutrient availability is high,early successional species typically have highrelative growth rates, supported by high rates of photosynthesis and nutrient uptake. Thesespecies reproduce at an early age and allocatea large proportion of NPP to reproduction(Table 13.1). Their strategy is to grow quicklyunder conditions of high resource supply, thendisperse to new disturbed sites.These early suc-cessional species include many weeds that col-onize sites disturbed by people. As successionproceeds, there is a gradual shift in dominanceto species that have lower resource require-ments and grow more slowly. In ecosystemswith low initial availability of soil resources,succession proceeds more slowly and followspatterns similar to those described for primarysuccession. Because there is a continuum in dis-turbance characteristics between primary andsecondary succession, the patterns of establish-ment and succession differ among ecosystemtypes with different disturbance regimes and

even among different disturbance events in thesame ecosystem type.

Carbon Balance

Primary Succession

In primary succession productivity and de-composition rates are often greatest in mid-succession. Primary succession begins withlittle live or dead organic matter, so NPP anddecomposition are initially close to zero. NPPincreases slowly at first because of low plantdensity, small plant size, and strong nitrogenlimitation of growth. NPP and biomass gener-ally increase most dramatically after nitrogenfixers colonize the site. The planting of nitrogen-fixing lupines on English mine wastes(Bradshaw 1983) and the natural establish-ment of nitrogen-fixing alders after retreat ofAlaskan glaciers (Bormann and Sidle 1990), forexample, cause sharp increases in plant biomassand NPP. In primary successional sequencesthat lack a strong nitrogen fixer, successionalincreases in biomass and NPP depend on otherforms of nitrogen input, including atmosphericdeposition, plant and animal detritus, andfloods.

Long-term successional trajectories ofbiomass and NPP differ among ecosystems. Acommon pattern in forests is that NPP increasesfrom early to mid-succession and then declinesafter the forest reaches its maximum leaf areaindex (LAI) (Fig. 13.8) (Ryan et al. 1997).Several processes are thought to contribute tothese patterns. In some forests, hydraulic con-ductance declines in late succession, causingwater to limit the leaf area that can be sup-ported and therefore the NPP that the eco-system can sustain (see Chapter 6). In otherforests, nutrient supply declines in late succes-sion, leading to a corresponding reduction inNPP (Van Cleve et al. 1991). The mortality ofbranches and trees often increases in late succession, as trees age. The combination ofreduced NPP and increased mortality of plantsand plant parts in late succession slows the rateof biomass accumulation, so biomass appro-aches a relatively constant value (steady state)(Fig. 13.9). There is little support for the earliergeneralization (Odum 1969) that the decline in

Succession 293

production in late succession reflects increasedmaintenance respiration to support the increas-ing biomass. Most of the increased biomass offorests is wood, which consists mainly of deadcells that require no maintenance. In late suc-cession interannual variation in climate andbiotic processes such as pest outbreaks have astronger influence on biomass and NPP thandoes successional change. Biomass can eitherincrease through succession or decline in latesuccession, due to stand thinning.The long-termend points of successional trajectories in bio-mass and NPP are often uncertain because dis-turbance usually resets the successional clockbefore the ecosystem reaches steady state.

Over extremely long time scales, changes inrates of weathering and soil development leadto further changes in biomass and other ecosys-tem properties (see Chapter 3). Redwoods inCalifornia coastal forests, for example, arereplaced by a pygmy forest of evergreen treesand shrubs after hundreds of thousands ofyears due to the formation of a hardpan thatprevents drainage and creates anaerobic condi-tions that retard decomposition and rootgrowth (Westman 1978). The slow-growingplants capable of surviving under these low-

nutrient conditions produce litter with a highconcentration of phenolics, which furtherreduces decomposition rate, resulting in a pos-itive feedback that leads to progressively lowerbiomass, productivity, and nutrient turnover(Northup et al. 1995).

The decomposition rate at the start ofprimary succession is near zero, because thereis little or no soil organic matter. The loworganic content of these soils contributes totheir low moisture-holding capacity and cationexchange capacity (CEC) (Fig. 13.10) (seeChapter 3). The pattern of change in decompo-sition through primary succession is similar tothe pattern described for NPP. Decomposition,however, lags behind the changes in NPP,causing soil organic matter (SOM) to accumu-late (Fig. 13.11). Initially decomposition is slowin primary succession because it is limited bythe rate of litter input. Decomposition increasessubstantially in mid-succession in response toincreases in the quantity and quality of litter. Inforests, the late-successional decline in NPPreduces litter inputs to soils, causing decompo-sition (in grams per square meter) to decline.In ecosystems in which nutrient availabilitydeclines in late succession, this reduces litterquality and quantity, further reducing decom-position rate (Van Cleve et al. 1993).

Net ecosystem production (NEP) is the netrate of carbon accumulation by the ecosystem.It is determined primarily by the balancebetween NPP and carbon losses through

40 80 1601200

2

4

6

8

0

Stand age (yr)

NP

P (

kg m

-2)

Figure 13.8. Successional changes in abovegroundspruce production in eastern Russia. NPP declinesafter the forest reaches maximum LAI at about 60yr of age. (Redrawn with permission from Advancesin Ecological Research; Ryan 1997.)

Figure 13.9. Idealized patterns of successionalchange in plant biomass, NPP, plant respiration(Rplant), and plant mortality of a forest. NPP oftenreaches a peak in mid-succession, and both produc-tion and respiration decline in late succession. GPP,gross primary production.


heterotrophic respiration and leaching; het-erotrophic respiration is typically the largestavenue of carbon loss. Because of the lag of het-erotrophic respiration behind NPP, NEP is gen-erally positive during early and mid-succession,leading to a net accumulation of carbon inecosystems in both vegetation and soils (Fig.13.11). This explains why midlatitude, northtemperate forests that were established inabandoned agricultural lands one to two cen-turies ago are currently a net carbon sink(Goulden et al. 1996, Valentini et al. 2000).NEP should approach zero in late successionbecause NPP and heterotrophic respiration are approximately equal. At this point NEP isgoverned more by climate and pest outbreaksthan by successional dynamics. In carbon-accumulating ecosystems such as peatlands,boreal forests, and arctic tundra, however,decomposition declines more strongly in latesuccession than does NPP, so NEP remains pos-

Figure 13.11. Idealized patterns of change in carbonpools (plants and soils) and fluxes (NPP, Rheterotr,and NEP) in primary and secondary succession.In early primary succession, plant and soil carbonaccumulates slowly, because NPP is greater than heterotrophic respiration (Rheterotr)—that is, there is a positive net ecosystem production (NEP). In earlysecondary succession, soil carbon declines after dis-

turbance because carbon losses from heterotrophicrespiration exceed carbon gain from NPP, leading to a negative NEP. In late succession, plant and soilcarbon approach the steady state (in an idealized situation), and NEP approaches zero. In bothprimary and secondary succession, NPP and NEP aremaximal in mid-succession. The graphs assume neg-ligible carbon loss to groundwater.

Successional age (yr)

Soil C

CEC

Soi

l car

bon

(kg

m-2

)

CE

C (

meq

100

cm

-3)

3

3

9

6

6

00 100 200 250

Figure 13.10. Accumulation during succession ofsoil organic carbon (Crocker and Major 1955) andassociated change in cation exchange capacity (CEC)of mineral soil (Ugolini 1968) after deglaciation atGlacier Bay Alaska. Measurements were made to adepth of 45 cm in mineral soil. The accumulation ofsoil carbon contributes to the increased CEC, whichretains nutrients to support plant growth.

Succession 295

itive in late succession, and the ecosystem con-tinues to accumulate carbon. This explains whysome peatlands have larger carbon pools perunit area than most ecosystems, despite theirlow productivity. Because of ecosystem differ-ences in disturbance frequency and environ-mental effects on carbon cycling processes,there is a wide range of potential successionalpatterns in NPP, heterotrophic respiration, andNEP among ecosystems.


The initial carbon pools and fluxes are muchlarger in secondary than in primary succession.Carbon dynamics are dramatically different in secondary succession from those in primarysuccession, because secondary successionbegins with an initial stock of SOM. Immedi-ately after disturbance, NPP is low in secondarysuccession because of low plant biomass, just asin primary succession (Fig. 13.11). NPP recov-ers more quickly in secondary than in primarysuccession, however, due to the generally rapidcolonization and high growth rate of herbs,grasses, and resprouting perennial species.High availability of light, water, and nutrientssupports the high growth potential of early suc-cessional vegetation in many secondary succes-sional sequences. The herbaceous species thatdominate most early secondary successionalsites return most of their biomass to the soileach year. As perennial plants, particularlywoody species, increase in abundance, biomassand NPP increase more rapidly, because woodyspecies retain a larger proportion of theirbiomass than do herbs. Changes in biomass andNPP in middle and late secondary successionare similar to patterns described for primarysuccession (Fig. 13.11), because they are con-trolled by the same factors and processes—largely the soil resources available to supportproduction and the growth potential of thespecies typical of the ecosystem.

In contrast to primary succession, decompo-sition is often more rapid early in secondarysuccession than at any other time (Fig. 13.11)because many disturbances transfer largeamounts of labile carbon to soils and create anenvironment that is favorable for decomposi-tion. The size of the initial carbon pool depends

on the nature and severity of the disturbance.After a treefall, hurricane, or insect outbreakthere are large inputs of new labile carbon fromleaf and root death. Fire consumes some of thesurface SOM but also adds new carbon to thesoil through death of roots and unburnedaboveground plant material. Disturbance alsostimulates decomposition because the removalof vegetation allows more radiation to pene-trate to the soil surface and reduces transpira-tional water loss. The resulting increases in soiltemperature and soil water content generallyenhance decomposition. The large quantity andhigh quality of litter of early secondary succes-sional plants also promotes decomposition. Inmid-succession the regrowing vegetation usesan increasing proportion of the available waterand nutrients and reduces soil temperature byshading the soil surface. These changes in en-vironment cause a decline in decomposition.Decomposition continues to decline in late suc-cession because the decline in NPP reduceslitter input, litter quality often declines, and theenvironment becomes less favorable than inearly succession.

How do these contrasting patterns of NPPand decomposition affect NEP? In early sec-ondary succession, ecosystem carbon poolsdecline (i.e., NEP is negative) because decom-position causes large carbon losses, and there is little NPP (Fig. 13.11). In early to mid-succession, before the peak in NPP, ecosys-tems begin accumulating carbon again, as soonas NPP outpaces decomposition. In late succes-sion, ecosystems either approach a carbonbalance of zero or continue to accumulatecarbon at a slow rate, depending on the envi-ronmental limitations to NPP and decomposi-tion. Other avenues of carbon loss fromecosystems, such as leaching of dissolvedorganic carbon, also influence NEP, but theirpatterns of successional change are not welldocumented.

Although the successional patterns of NPP,decomposition, and carbon stocks in plants andsoils that we have described are frequentlyobserved, the details and timing of these pat-terns differ substantially among ecosystems,depending on factors such as initial ecosystemcarbon stocks, resource availability, disturbanceseverity, and successional pathway.


Nutrient Cycling

Primary SuccessionNutrient dynamics during succession are both a cause and a consequence of the dynamicinterplay between NPP and decomposition.The most dramatic change in nutrient cyclingduring early primary succession is the accumu-lation of nitrogen in vegetation and soils. Mostparent materials have extremely low nitrogencontents in the absence of biotic influences, sothe initial nitrogen pools in the ecosystem aresmall and depend on atmospheric inputs. Atthis initial stage of primary succession, nitrogenis the element that most strongly limits plantgrowth and therefore the rates of accumulationof plant biomass and SOM (Vitousek et al.1987, Chapin et al. 1994). The rate of nitrogeninput, which frequently is associated with the establishment of nitrogen-fixing plants(both free-living cyanobacteria and symbiotic nitrogen fixers), therefore governs the ini-tial dynamics of nutrient cycling in primary succession. Nitrogen typically accumulates atrates of 3 to 16gNm-2 yr-1 for 50 to 200 years,before approaching an asymptote of 200 to 500gNm-2 (Walker 1993). As leaves and roots of nitrogen-fixing plants senesce and are eaten by herbivores, the nitrogen is trans-ferred from plants to the soil, where it is mineralized and absorbed by both nitrogen-fixing and non-nitrogen-fixing plants. Litterfrom non-nitrogen-fixing plants becomes anincreasingly important source for nitrogen mineralization as primary succession proceeds.This causes the ecosystem to shift from an opennitrogen cycle, with substantial input fromnitrogen fixation (see Chapter 9), to a moreclosed nitrogen cycle, in which plant growthdepends on the mineralization of soil organicnitrogen. During mid-succession, plants and soil microbes are so efficient at accumulatingnutrients that losses of nitrogen and otheressential elements from ecosystems are oftennegligible (Fig. 13.12) (Vitousek and Reiners1975). In late succession, nitrogen inputs to theecosystem may largely balance nitrogen lossesfrom leaching and denitrification, causingecosystem nitrogen pools to approach a rela-tively stable size.

Most evidence for these successional changesin nitrogen cycling comes from studies ofchronosequences, series of sites that differ inage but are assumed to be similar with respectto other state factors (see Chapter 1). Since weseldom know whether sites in a chronose-quence began their successional developmentunder identical conditions, long-term studies of succession are critical in testing whether the patterns in nitrogen cycling observed inchronosequences truly reflect the actual suc-cessional changes that occur at a site.

The accumulation of nitrogen in the initialstages of primary succession governs the ratesof internal cycling of other essential elementsin ecosystems. Early in primary succession,

Net

bio

mas

s in

crem

ent

0

Non-essential

Successional timeE

lem

ent o

utpu

ts(g

m-2

yr-1

)0

Limiting

Dis

turb

ance

Essential

Figure 13.12. Changes through succession in netbiomass increment in vegetation and in the losses oflimiting, essential, and nonessential elements. Inearly succession, when there are large annual incre-ments in biomass, elements that are required for thisproduction (especially growth-limiting elements)accumulate in new plant and microbial biomass, sothey are not lost from the ecosystem by leaching. Inlate succession, when the element requirements fornew plant and microbial biomass are balanced byelement release from the breakdown of dead organicmatter, nutrient inputs to the ecosystem are approx-imately balanced by nutrient outputs, regardless ofwhether nutrients are required by vegetation or not. (Modified with permission from BioScience;Vitousek and Reiners 1975.)

Succession 297

inputs of most elements in precipitation andweathering may approximately equal outputs,because biological storage pools in vegetationand soils are small. Plants and soil microbeshave a limited range of element ratios. Plants innitrogen-limited ecosystems therefore accumu-late most essential elements approximately inproportion to their rate of nitrogen accumula-tion, preventing the loss of these elements fromthe ecosystem in mid-succession (Fig. 13.12).If biomass and element pools stabilize in latesuccession, element losses increase until theyapproximately equal element inputs. Over timescales of soil development, there are additionalchanges in nutrient cycling that occur when thesupply of weatherable minerals is depleted orbecomes bound in unavailable forms. Avail-ability of phosphorus and cations, for example,typically declines in old high-weathered sites, asthey leach or become bound in unavailableforms (see Chapters 3 and 9). Under these cir-cumstances, phosphorus or other elements maylimit plant production (Chadwick et al. 1999),and cycling rates of these limiting elements regulate rates of cycling of nitrogen and otherminerals.


Secondary succession after natural distur-bances differs from primary succession becauseit generally begins with higher nitrogen avail-ability. Natural disturbances that initiate secondary succession produce a pulse of nutrient availability because disturbance-induced changes in environment and litterinputs increase mineralization of dead organicmatter and reduce plant biomass and nutrientuptake. Fires, which may volatilize largeamounts of nitrogen, also return nutrients inash, as described earlier, leading to high nutrient availability after fire (Wan et al. 2001).Plant growth is therefore generally not stronglynutrient-limited early in secondary successionafter natural disturbances, and there is usuallyadequate nitrogen to support high rates of pho-tosynthesis and growth (Scatena et al. 1996).The pulse of nutrient availability and the reduc-tion in plant biomass and capacity for plantuptake after disturbance also increase the vul-

nerability of ecosystems to nutrient loss. Highrates of nitrogen mineralization produce NH4

+,which serves as a substrate for nitrification andits associated loss of nitrogen trace gases (seeChapter 9). The nitrate can then be denitrified,particularly if soils are wet, or leached belowthe rooting zone. The occurrence or extent ofthis nitrogen loss depends on the balancebetween nitrogen mineralization and uptake byplants and microbes. Rains that occur immedi-ately after a fire, for example, can leach nitrateinto groundwater and streams. The few studiesof nutrient losses associated with natural dis-turbances show surprisingly small nitrogenlosses to streams from wildfire (Stark andSteele 1977) or hurricanes (Schaefer et al.2000). We know little, however, about otheravenues of loss or whether these results are re-presentative of natural disturbances.

The vulnerability of ecosystems to nutrientlosses after disturbance has been illustrated inmany forest harvest experiments, such as thoseat Coweta and Hubbard Brook in the UnitedStates. After stream discharge and chemistryhad been monitored for several years, the forestwas cut on an entire watershed, and regene-rating vegetation was killed with herbicides(Bormann and Likens 1979). The combinationof high decomposition and mineralization ratesand absence of plant uptake after disturbancecaused large losses of essential plant nutrientsin stream water (see Fig. 9.7). When vegetationwas allowed to regrow, the increased plantuptake caused nutrient losses in stream waterto decline to preharvest levels. These studiesshow clearly that the dynamics of nutrient lossafter disturbance are highly variable, with theextent of nutrient loss often depending onnutrient availability at the time of disturbanceand the capacity of regenerating vegetation toabsorb nutrients.

Anthropogenic disturbances create a widerange of initial nutrient availabilities. Some dis-turbances, such as mining, can produce an initialenvironment that is even less favorable thanmost natural primary successional habitats forinitiation of succession. These habitats mayhave toxic by-products of mining or mineralmaterial with a low capacity for water andnutrient retention. Some agricultural lands


are abandoned to secondary succession aftererosion or (in the tropics) formation of lateritesoils (see Chapter 3), reducing the nutrient-supplying power of soils. Soils from somedegraded lands also have concentrations of alu-minum and other elements that are toxic tomany plants. Secondary succession in degradedlands may be quite slow. At the opposite ex-treme, abandonment of rich agricultural landsor the logging of productive forests may createconditions of unusually high nutrient availabil-ity, leading to the potential loss of nutrientsthrough leaching and denitrification. Thesenutrient losses are particularly dramatic in the tropics, where rapid mineralization andbiomass burning associated with forest clearingrelease large amounts of nitrogen as trace gases(nitric oxide, NOx, and nitrous oxide, N2O) andas nitrate in groundwater (Matson et al. 1987).The impact of agricultural nutrient additions isparticularly long lived for phosphorus becauseof its effective retention by soils. An under-standing of the successional controls over nutri-ent cycling provides the basis for managementstrategies that minimize undesirable environ-mental impacts (see Chapter 16). The return of topsoil or planting of nitrogen-fixing plants on mine wastes, for example, greatly speeds successional development on these sites (Bradshaw 1983). Retention of some organicdebris after logging may support microbialimmobilization of nutrients that would reduceleaching loss.

Trophic Dynamics

The proportion of primary production con-sumed by herbivores is maximal in early tomiddle succession. In early primary and sec-ondary succession, rates of herbivory may below because of low food density, insufficientcover to hide vertebrate herbivores from theirpredators, and insufficient canopy to create ahumid, nondesiccating environment for inver-tebrate herbivores. Herbivory is often greatestin early to middle secondary successionbecause the rapidly growing herbaceous andshrub species that dominates this stage havehigh nitrogen concentrations and a relativelylow allocation to plant defense (see Chapter

11). This explains why abandoned agriculturalfields, recent burn scars, or riparian areas arefocal points for browsing mammals, insect her-bivores, and their predators. In early succes-sional boreal floodplains, for example, mooseconsume about 30% of aboveground NPP andaccount for a similar proportion of the nitrogeninputs to soil (Kielland and Bryant 1998). Theabundant insect herbivores on these sitessupport a high diversity of neotropical migrantbirds. Similarly, in temperate and tropicalregions, early successional forests support largepopulations of deer and other browsers. Inecosystems in which nutrient availabilitydeclines from early to late succession, plantsshift allocation from growth to defense (seeChapter 11). The resulting decline in foragequality reduces levels of consumption by mostherbivores and higher trophic levels. Someinsect outbreak species are an important excep-tion to this successional pattern. They oftenattack late-successional trees that are weak-ened by environmental stress.

Vertebrate herbivores can either promote orretard succession, depending on their relativeimpact on early vs. late successional species.Vertebrate herbivores both respond to and contribute to successional change. The effectsof herbivores on succession differ among ecosystems depending on the nature and spe-cificity of herbivore–plant interactions. There are, however, several common patterns thatemerge.

In forested regions, birds, rodents, and other vertebrates often enhance the dispersalof early successional species such as black-berries, junipers, and grasses into abandonedagricultural fields and other disturbed sites.Birds and squirrels also disperse the large seedsof late-successional species such as oak andhickory into early successional sites. Theseanimal-mediated dispersal events are particu-larly important in secondary succession, wherethe rapid development of herbaceous vegeta-tion makes it difficult for small-seeded woodyspecies to compete and establish successfully.

The relatively low levels of plant defense inspecies that typically characterize early forestsuccession make these plants a nutritious targetfor generalist herbivores. Preferential feeding

Succession 299

on these species reduces their height growth.Browsed plants respond to aboveground her-bivory by reducing root allocation, makingthem less competitive for water and nutrients(Ruess et al. 1998). Many later successionalspecies produce chemical defenses that detergeneralist herbivores. Selective herbivory con-tributes to the competitive release of late suc-cessional species, enabling them to overtop andshade their early successional competitors. Inthis way, selective browsing by mammals fre-quently speeds successional change in forests(Pastor et al. 1988, Kielland and Bryant 1998,Paine 2000). In tropical rain forests mammalianherbivores maintain the diversity of understoryseedlings that become the next generation ofcanopy dominants, because they feed preferen-tially on the “weedy” tree seedlings that aremost common in the understory (Dirzo andMiranda 1991).

In contrast to forests, many grasslands andsavannas are maintained by mammalian herbi-vores that prevent succession to forests. Ele-phants, for example, browse and uproot trees inAfrican savannas. These savannas succeed toclosed forests in areas where elephant popula-tions have been reduced by overhunting. InNorth American prairies, grazers and firerestrict the invasion of trees. When thesesources of disturbance are reduced, trees fre-quently invade and convert the grassland toforest. Similarly, at the end of the Pleistocenethe decline in large mammals that occurred onmany continents, in part from human hunting,contributed to the vegetation changes thatoccurred at that time (Flannery 1994, Zimov et al. 1995).

Herbivores have multiple effects on nutrientcycling in early succession. In the short termthey enhance nutrient availability by returningavailable nutrients to the soil in feces and urine,which short-circuits the decomposition process(Kielland and Bryant 1998). Herbivory can alsoalter the temperature and moisture regime fordecomposition at the soil surface by reducingleaf and root biomass. The quality of litter thata given plant produces is also enhanced by her-bivory (Irons et al. 1991). Over the long term,however, herbivory accelerates plant succes-sion by removing early successional species.

This tends to reduce nutrient cycling rates andnutrient losses (Pastor et al. 1988, Kielland andBryant 1998) (see Fig. 11.4).

Water and Energy Exchange

Disturbances that eliminate plant biomassincrease runoff through a reduction in evapo-transpiration. One of the most dramatic conse-quences of forest cutting or overgrazing isincreased runoff to streams and rivers. Asforests recover from disturbance, evapotran-spiration increases, and runoff returns to predisturbance levels (Fig. 13.13). These watershed-scale observations provide a basisfor understanding successional controls overwater and energy exchange. The low evapo-transpiration in early succession results fromthe small biomass of roots to absorb water fromthe soil and of leaves to transfer that water to the atmosphere. This low evapotranspirationtherefore leads to high soil moisture (except atthe surface where soil evaporation occurs) andrunoff. In disturbances that remove the plantcanopy and litter layer, impaction by rain dropson mineral soil reduces infiltration and in-creases overland flow and runoff. As root and

Figure 13.13. Runoff from a watershed in a NorthCarolina forest in the southeastern United Statesunder natural conditions (the calibration period) andafter forest harvest. Water yield from the watershedgreatly increased in the absence of vegetation and approached preharvest levels within 20 years.(Hibbert 1967.)


leaf area increase through succession, there is acorresponding increase in evapotranspirationand decrease in runoff. The high nitrogen avail-ability, high photosynthetic rate, and high leafarea early in secondary succession contribute toa high canopy conductance so evapotranspira-tion increases more rapidly than plant biomassas succession proceeds (Bormann and Likens1979). As the canopy increases in height andcomplexity, solar energy is trapped more effec-tively, reducing albedo and increasing theenergy available to drive evapotranspiration.The high surface roughness of tall complexcanopies increases mechanical turbulence and mixing within the canopy. All of thesefactors contribute to high evapotranspiration in mid-succession.

Successional changes in albedo differ amongecosystems because of the wide range amongecosystems in albedo of bare soil (see Table4.2). Many recently disturbed sites have a lowalbedo because of the dark color of moistexposed soils or of charcoal. Albedo increaseswhen vegetation, with its generally higheralbedo, begins to cover the soil surface (Fig.13.14). Albedo probably declines again in latesuccession due to increased canopy complexity(see Chapter 4). In ecosystems that succeedfrom deciduous to conifer forest, this speciesshift causes a further reduction in albedo. Thewinter energy exchange of northern forests isinfluenced by snow, which has an albedo three-fold to fivefold higher than vegetation (Bettsand Ball 1997). Winter albedo of these forestsdeclines through succession, first as vegetationgrows above the snow, then as the canopybecomes more dense, and finally when (if) thereis a switch from deciduous to evergreen vege-tation. All of these changes increase the extentto which vegetation masks the snow fromincoming solar radiation.

High surface temperatures that contribute tohigh emission of longwave radiation dominateenergy budgets of early successional sites. Earlysuccessional sites often have a high surface tem-perature for several reasons. (1) The low albedoof recently disturbed sites maximizes radiationabsorption and therefore the quantity of energyavailable at the surface. (2) The low leaf area,small root biomass, and low hydraulic conduc-

tance of dry surface soils limits the proportionof energy dissipated by evapotranspiration. (3)The relatively smooth surface of unvegetatedor early successional sites minimizes mechani-cal turbulence that would otherwise transportthe heat away from the surface. The resultinghigh surface temperature promotes emission oflong-wave radiation (see Chapter 4).

The large longwave emission dissipates muchof the absorbed radiation after disturbance,so net radiation (the net energy absorbed by the surface) is not as great as we might expectfrom the low albedo of these sites. For example,net radiation actually declines after fire in theboreal forest despite a reduction in albedobecause of the large emission of longwave radi-ation (Chambers and Chapin, in press).The soilsurface of unvegetated sites is prone to dryingbetween rain events due to the combination ofhigh surface temperatures and the low resupplyof water from depth, due to the low hydraulicconductance of dry soils (see Chapter 4). Drysurface soils provide little moisture for surface

Alb

edo

(%)

Time (years since burn)

Pre-disturbance albedo

Post-disturbance albedo2

4

6

8

10

12

14

16

00

20 40 60 80 100

Figure 13.14. Successional changes in albedo afterfire in Alaskan boreal forests (Chambers and Chapin,in press). The black postfire surface causes a declinein albedo. Albedo increases during the herbaceousand deciduous forest phases of succession anddeclines in late succession due to a switch to conifervegetation. This successional change occurs morerapidly after moderate fires because of the morerapid replacement of deciduous species by conifers.

Temporal Scaling of Ecological Processes 301

evaporation and are good thermal insulators, soboth evapotranspiration and average groundheat flux are often relatively low on unvege-tated surfaces (Oke 1987). Consequently, sensi-ble heat flux accounts for the largest proportionof energy that is dissipated from these sites tothe atmosphere. The absolute magnitude ofsensible heat flux from early successional sitesdiffers among ecosystems and climate zonesand depends on both net radiation (the energyavailable to be dissipated) and the energy par-titioning among sensible, latent, and groundheat fluxes. As succession proceeds, latent heat fluxes become a more prominent compo-nent of energy transfer from the land to theatmosphere.

Temporal Scaling of Ecological Processes

Temporal extrapolation requires an under-standing of the typical time scales of importantecological processes. Measurements of ecolog-ical processes are generally made over shortertime periods than the time scales over which wewould like to make predictions. Few studies, forexample, provide detailed information aboutthe functioning of ecosystems over time scalesof decades to centuries—the time scale overwhich ecosystems are likely to respond toglobal environmental change. Temporal scalingis the extrapolation of measurements made atone time interval to longer (or occasionallyshorter) time intervals. Simply multiplying aninstantaneous flux rate by 24h to get a dailyrate or by 365 days to get an annual rate seldomgives a reasonable approximation because this ignores the temporal variation in drivingvariables and the time lags and thresholds inecosystem responses to these drivers. Rates ofphotosynthesis, for example, differ betweennight and day and between summer and winter.

One approach to temporal scaling is to selectmeasurements that are consistent with the timescale and question of interest. A secondapproach is to extrapolate results based onmodels that simulate processes accounting forimportant sources of variation over the time

scale of interest. The key to temporal scaling istherefore to focus clearly on the processes thatare important over the time scales of interest.Entire books have been written on temporalscaling based on isotopic measurements(Ehleringer et al. 1993), long-term measure-ments (Sala et al. 2000b), and modeling(Ehleringer and Field 1993, Waring andRunning 1998). Here we provide a brief over-view of these approaches.

Isotopic tracers provide an important tool for estimating long-term rates of net carbonexchange of plants and ecosystems becausethey integrate the net effect of carbon inputsand loses throughout the time period thatcarbon exchanges occur (see Boxes 5.1 and7.1). The 13C content of plants in dry environ-ments, for example, provides an integratedmeasure of water use efficiency (WUE) duringthe time interval during which the plant mate-rial was produced. The 13C content of soils inecosystems that have changed in dominant veg-etation from C3 to C4 plants provides an inte-grated measure of soil carbon turnover sincethe time that the vegetation change occurred.These measurements are appropriate for esti-mating long-term rates because they incorpo-rate effects of processes that occur slowly orintermittently that might not be captured inshort-term gas-exchange measurements. Sea-sonally integrated WUE measured with stableisotopes, for example, is affected by dry and wet periods that influence seasonal water andcarbon exchange; whereas instantaneous mea-surements of gas exchange are unlikely to berepresentative of the entire annual cycle. Otherexamples of measurements that integrate overlong time intervals include NPP, which inte-grates longer time periods than does photo-synthesis or respiration, and changes in soilcarbon stocks over succession, which integrateover longer time periods than do NPP anddecomposition.

Process-based models are an important toolfor temporal scaling because they make pro-jections of the state of the ecosystem overlonger time intervals (or at different times)than can be measured directly. The challenge indeveloping models for temporal extrapolationis the selection of the driving variables that


account for the most important sources of tem-poral variation over the time scale of interest.The diurnal pattern of net photosynthesis canoften be adequately simulated based on therelationship of net photosynthesis to light andtemperature. Annual estimates of photosyn-thetic flux (gross primary production, GPP),however, also require information on seasonalvariation in leaf biomass and photosyntheticcapacity. In annual simulations, the diurnalvariation in photosynthesis is less important tomodel explicitly because it is quite predictable,based on the empirical relationship betweendaily photosynthesis and mean daily tempera-ture and light. Slow variables, such as succes-sional changes in LAI or nitrogen availability,are often treated as constants in short-term eco-logical studies, but can become key controllingvariables over longer time scales (Carpenterand Turner 2000). We must therefore thinkcarefully about which critical driving variablesare likely to change over the time scale ofintended predictions and look for evidence ofthe relationship of ecological processes to theseslow variables. Models of carbon flux based onthe relationship of GPP and respiration to dailyor monthly climate, for example, can be vali-

dated by comparing model output to patternsof carbon flux observed over longer time scales(e.g., interannual variation in carbon flux)(Clein et al., in press). There is a wide range oftemporal scales over which important ecosys-tem controls vary (Fig. 13.15).

Spatial variation in driving variables some-times gives hints as to which slow variables areimportant to include in long-term extrapola-tions. The spatial relationship between the dis-tribution of biomes or plant functional typesand climate, for example, has been used as a basis for predicting how vegetation mightrespond to future climatic warming (Prentice etal. 1992, VEMAP Members 1995). Spatial rela-tionships with driving variables often reflectquasi-equilibrium relationships. Dry tropicalforests, for example, occur where the averageclimate is warm and has a distinct dry season.Temporal extrapolations should also considerextreme events and time lags that may not beevident from an examination of spatial pattern.Ice storms, a spring freeze, intense droughts,100-year floods, and other events with long-lasting effects strongly influence the structureand functioning of ecosystems long after theyoccur.

Figure 13.15. Variation in return time for variablesthat strongly affect ecosystem processes. For any par-ticular process, such as NPP, there are fast variables(e.g., stomatal closure) that can be ignored, slow vari-

ables (e.g., El Niño or stand-replacing disturbance)that strongly affect the process, and extremely slowvariables (e.g., glacial cycles) that can be treated asconstants.


Summary

Rates of all ecosystem processes are constantlyadjusting to past changes that have occurredover all time scales, ranging from sunflecks that last milliseconds to soil development thatoccurs over millions of years. Ecosystem pro-cesses that occur slowly, such as soil organicmatter development, deviate most stronglyfrom steady state and are most stronglyaffected by legacies of past events. Ecosystemprocesses are highly resistant and/or resilient to predictable changes in environment such asthose that occur diurnally and seasonally and inresponse to disturbances to which the organ-isms are well adapted.

Because disturbance is a natural componentof all ecosystems, the successional changes in ecosystem processes after disturbance areimportant for understanding regional patternsof ecosystem dynamics. These ecosystemchanges are particularly sensitive to the sever-ity, frequency, and type of disturbance. Carbonaccumulates in vegetation and soils, leading to positive NEP through primary successionbecause changes in decomposition lag behindchanges in NPP. NPP in forests is frequentlygreatest in mid-succession. Secondary succes-sion begins with a large negative NEP due tolow NPP and rapid decomposition, but carboncycling in middle and late succession are similarto the patterns in primary succession.

Nutrient cycling changes through earlyprimary succession as nitrogen fixers establishand add nitrogen to the ecosystem. Other ele-ments cycle in proportion to the cycling ofnitrogen. In secondary succession, however,nitrogen is generally most available in earlysuccession. At this time, nitrogen and other ele-ments are vulnerable to loss until the potentialof plants and microbes to absorb nutrientsexceeds the rate of net mineralization. Thistightens the nitrogen cycle. Recycling within theecosystem is strongest in mid-succession, whenrates of nutrient mineralization constrain therates of uptake by vegetation.

The role of herbivores in succession differsamong ecosystem types and with successionalstage. Mammals often accelerate the early successional changes in forests by eliminating

palatable early successional species. In grass-lands, however, herbivores prevent the establishment of woody species that might otherwise transform grasslands into shru-blands and forests. Some insects have theirgreatest impact in late succession, particularlyin forests, where they can be important agentsof mortality.

Stand-replacing disturbances greatly reduceevapotranspiration and increase runoff. Evapo-transpiration increases through successionmore rapidly than might be expected frombiomass recovery because early successionalvegetation has high transpiration rates. Sensibleheat flux tends to show the reverse successionalpattern with high sensible heat flux (and/orlongwave radiation) immediately after distur-bance and lower sensible heat flux as rapidlygrowing mid-successional vegetation estab-lishes, reflects radiation, and transfers water tothe atmosphere.

Review Questions

1. Provide examples of ways in which thecarbon and nitrogen cycling of an ecosystemmight be influenced by the legacy of eventsthat occurred 1 week ago, 5 years ago, 100years ago, 2000 years ago.

2. What properties of disturbance regimesdetermine the ecological consequences ofdisturbance? How do these properties differbetween treefalls in a tropical wet forest andfire in a dry conifer forest?

3. What are the major processes causing suc-cessional change in plant species? How dothe relative importance of these processesdiffer between primary and secondary succession?

4. How do NPP, decomposition, and the carbonpools in plants and soils change throughprimary succession? At what successionalstage does carbon accumulate most rapidly?Why? How do these patterns differ betweenprimary and secondary succession? Why dothese differences occur?

5. How does nitrogen cycling differ betweenprimary and secondary succession? At whatstages is this difference most pronounced?


6. How do trophic dynamics change throughsuccession? Why?

7. How do water and energy exchange changethrough succession? What explains thesepatterns?

8. What are the major issues to consider inextrapolating information from one tem-poral scale to another? Describe ways inwhich this temporal extrapolation might bedone.

Additional Reading

Bazzaz, F.A. 1996. Plants in Changing Environments.Linking Physiological, Population, and Commu-nity Ecology. Cambridge University Press, Cam-bridge, UK.

Bormann, F.H., and G.E. Likens. 1979. Pattern andProcess in a Forested Ecosystem. Springer-Verlag,New York.

Chapin, F.S. III, L.R. Walker, C.L. Fastie, and L.C.Sharman. 1994. Mechanisms of primary successionfollowing deglaciation at Glacier Bay, Alaska.Ecological Monographs 64:149–175.

Clements, F.E. 1916. Plant Succession: An Analysis ofthe Development of Vegetation. Publication 242.Carnegie Institution of Washington, Washington,DC.

Connell, J.H., and R.O. Slatyer. 1977. Mechanisms ofsuccession in natural communities and their rolein community stability and organization. AmericanNaturalist 111:1119–1114.

Crocker, R.L., and J. Major. 1955. Soil developmentin relation to vegetation and surface age at GlacierBay, Alaska. Journal of Ecology 43:427–448.

Fastie, C.L. 1995. Causes and ecosystem conse-quences of multiple pathways of primary succes-sion at Glacier Bay,Alaska. Ecology 76:1899–1916.

Vitousek, P.M., and W.A. Reiners. 1975. Ecosystemsuccession and nutrient retention: A hypothesis.BioScience 25:376–381.

Vitousek, P.M., L.R. Walker, L.D. Whiteaker, D.Mueller-Dombois, and P.A. Matson. 1987. Biolog-ical invasion by Myrica faya alters ecosystemdevelopment in Hawaii. Science 238:802–804.

Zimov, S.A., V.I. Chuprynin, A.P. Oreshko, F.S.Chapin III, J.F. Reynolds, and M.C. Chapin. 1995.Steppe-tundra transition: An herbivore-drivenbiome shift at the end of the Pleistocene.American Naturalist 146:765–794.

Introduction

Spatial heterogeneity within and among ecosys-tems is critical to the functioning of individualecosystems and of entire regions. In previouschapters we emphasized the controls over eco-system processes in relatively homogenousunits or patches of an ecosystem. The spatialpattern of ecosystems in a region, however, alsoinfluences ecosystem processes. Riparian eco-systems between upland agricultural systemsand streams or rivers, for example, may filternitrate and other pollutants that would other-wise move into streams. Spatial patterns withinecosystems also influence ecosystem processes.The most rapid rates of nutrient cycling andgreatest accumulations of organic matter inarid ecosystems, for example, occur beneath,rather than between, plants. The fragmentationof ecosystems into smaller units separated byother patch types influences the abundance anddiversity of animals. All of the processes andmechanisms that operate in ecosystems (seePart II) have important spatial dimensions. Inthis chapter, we first discuss the concepts andcharacteristics of landscapes that aid in under-standing and quantifying landscape interac-tions and then discuss sources of spatialheterogeneity within and among ecosystems

and the consequences of that heterogeneity forinteractions among ecosystems on a landscape.

Concepts of LandscapeHeterogeneity

Spatial pattern exerts a critical control overecological processes at all scales. Landscapesare mosaics of patches that differ in eco-logically important properties. Landscape eco-logy addresses the causes and consequences of spatial heterogeneity (Urban et al. 1987,Forman 1995, Turner et al. 2001). This fieldfocuses on both the interactions among patcheson the landscape and the behavior and func-tioning of the landscape as a whole. Landscapeprocesses can be studied at any scale, rangingfrom the mosaic of gopher mounds in a squaremeter of grassland to biomes that are patchilydistributed across the globe. Landscape pro-cesses are frequently studied at the scale ofstands of vegetation within a watershed orregion.

Some landscape patches are biogeochemicalhot spots with high process rates, causing themto be more important than their area wouldsuggest. Beaver ponds, for example, are bio-

14Landscape Heterogeneity and Ecosystem Dynamics

Landscape heterogeneity determines the regional consequences of processes occurring in individual ecosystems. In this chapter we describe the major causes andconsequences of landscape heterogeneity.

305

306 14. Landscape Heterogeneity and Ecosystem Dynamics

geochemical hot spots for methane emissions inboreal landscapes (Roulet et al. 1997), andrecently cleared pastures in the central AmazonBasin are hotspots for nitrous oxide emissions(Matson et al. 1987). Hot spots are defined withrespect to a particular process and occur at allspatial scales, from the rhizosphere surroundinga root to urine patches in a grazed pasture, towetlands in a watershed, to tropical forests onthe globe. The environmental controls over biogeochemical hot spots often differ radicallyfrom controls in the surrounding matrix—thatis, the predominant patch type in the landscape.Only by studying processes in these hot spotscan we understand these processes and extrap-olate their consequences to larger scales. Land-scape ecology therefore plays an essential rolein understanding the Earth System because ofthe importance of estimating fluxes (and theircontrols) of energy and materials at regionaland global scales.

The size, shape, and distribution of patches in the landscape govern interactions amongpatches. Patch size influences habitat hetero-geneity. Large forest fragments in an agricul-tural landscape, for example, contain greaterhabitat heterogeneity and support more speciesand bird pairs than do small patches (Freemarkand Merriam 1986, Wiens 1996). Patch size alsoinfluences the spread of propagules and distur-bance from one patch to another. Seeds musttravel farther to colonize large disturbed pa-tches, such as fire scars or abandoned agricul-tural fields, than to colonize small disturbedpatches. Patch size therefore affects recruit-ment and the capacity of regenerating ve-getation to use the pulse of nutrients thataccompanies disturbance (Rupp et al. 2000).Patch shape influences the effective size ofpatches by determining the average distance ofeach point in the patch to an edge. Patch sizeand shape together determine the ratio of edgeto area of the patch. The edge-to-area ratio oflakes and streams, for example, is critical indetermining the relative importance of aquaticand terrestrial production in supplying energyto aquatic food webs, which radically affectstheir functioning (see Chapter 10).

The population dynamics of many organismsdepend on movement between patches, which

is strongly influenced by their connectivity(Turner et al. 2001). Birds and small animals inan agricultural landscape, for example, usefence rows to travel among patches of suitablehabitat. In a patchy environment, local popu-lations may go extinct, and the dynamics ofmetapopulations—populations that consist ofpartially isolated subpopulations—depend onrelative rates of local extinctions in patches andcolonization from adjacent patches (Hanski1999). Species conservation plans often encour-age the use of corridors to facilitate movementamong suitable habitat patches (Fahrig andMerriam 1985), although the effectiveness ofcorridors is debated (Rosenberg et al. 1997,Turner et al. 2001). Connectivity may be par-ticularly critical at times of climatic change.Isolated nature reserves, for example, maycontain species that cannot adapt or migrate in response to rapid environmental change.The effectiveness of corridors among patchesdepends on the size and mobility of organismsor the nature of disturbances that move amongpatches (Wu and Loucks 1995).A fence row, forexample, may be a corridor for voles, a barrierfor cattle, and invisible to birds.

Ecological boundaries are critical to the interactions among neighboring landscape elements (Gosz 1991). Animals like deer, forexample, are edge specialists that forage in onepatch type and seek protection from preda-tion in another. The size of the patch and itsedge-to-area ratio determine the total habitatavailable to edge specialists. Edges often ex-perience a different physical environment thando the interiors of patches. Forest boundariesadjacent to clearcuts, for example, experiencemore wind and solar radiation and are drierthan are patch interiors (Chen et al. 1995). Intropical rain forests the trees within 400m of an edge experience more frequent blowdownsthan do trees farther from an edge (Lauranceand Bierregaard 1997). These differences inphysical environment affect rates of distur-bance and nutrient cycling, which translate intovariations in recruitment, productivity, andcompetitive balance among species. The depthsto which these edge effects penetrate differamong processes and ecosystems. Wind effects,for example, may penetrate more deeply from

Causes of Spatial Heterogeneity 307

an edge than would availability of mycorrhizalpropagules.

The abruptness of boundaries influencestheir role in the landscape (McCoy et al. 1986).Relatively broad gradients at the boundariesbetween biomes occur where there is a gradualshift in some controlling variable such as pre-cipitation or temperature. Sharper boundariestend to occur where there are steep gradientsin physical variables that control the distribu-tion of organisms and ecosystem processes orwhere an ecologically important functionaltype reaches its climatic limit. The boundarybetween a stream and its riparian zone reflectspresence or absence of water above the groundsurface. Similarly, the boundary between differ-ent parent materials can be quite sharp, withtwo sides of a boundary supporting differentecosystems. Climatically determined bound-aries, such as treeline or the savanna–forestborder, are useful places to study the effects of climatic change because species are at their physiologically determined range limits.Species in these situations may be sensitive tosmall changes in climate.

The configuration, or spatial arrangement, ofpatches in a landscape influences landscapeproperties because it determines which patchesinteract and the spatial extent of their interac-tions. Riparian areas are important becausethey are an interface between terrestrial andaquatic ecosystems. Their linear configurationand location make them much more importantthan their small aerial extent would suggest.An area of the same size that occurred else-where in the landscape would function quitedifferently.

Causes of Spatial Heterogeneity

Landscape heterogeneity stems from environ-mental variation, population and communityprocesses, and disturbance. Spatial variation instate factors (e.g., topography and parent mate-rial) and interactive controls (e.g., disturbanceand dominant plant species) determine thenatural matrix of spatial variability in ecosys-tems (Holling 1992). Human activities are an

increasing cause of changes in the spatial het-erogeneity of ecosystems.

State Factors and Interactive Controls

Differences in abiotic characteristics and asso-ciated biotic processes account for the basicmatrix of landscape variability. Temperature,precipitation, parent materials, and topographyvary independently across Earth’s surface.Some of these state factors, such as rock type,exhibit sharp boundaries and can therefore beclassified into distinct patches. Others, includingclimate variables, vary more continuously andgenerate gradients in ecosystem structure andfunctioning. Analysis of these landscape classesand gradients shows that different factorscontrol spatial patterns at different spatialscales. Regional-scale patterns of vegetation,net primary production (NPP), soil organicmatter (SOM), litter quality, and nutrient avail-ability in grasslands, for example, correlate withregional gradients in precipitation and temper-ature (Burke et al. 1989) (Fig. 14.1). In contrast,topography, soil texture, and land use historyexplain most variability at the scale of a fewkilometers, and microsite variation accounts for variability within patches (Burke et al.1999). Broad elevational patterns of ecosystemprocesses in tropical forests on the HawaiianIslands are also governed largely by climate,with local variation reflecting the type and ageof the parent material (Vitousek et al. 1997b,Chadwick et al. 1999). The resulting differencesin soils give rise to consistent differences innitrogen cycling (Pastor et al. 1984), phospho-rus cycling (Lajtha and Klein 1988) and nitrousoxide emissions (Matson and Vitousek 1987).These comparative studies provide a basis forextrapolating ecosystem processes to regionalscales based on the underlying spatial matrix ofabiotic factors.

Community Processes and Legacies

Historical legacies, stochastic dispersal events,and other community processes can modify theunderlying relationship between environmentand the distribution of a species. Ecosystemprocesses depend not only on the current envi-


Figure 14.1. Regional patterns of air temperature,precipitation, soil sand (a measure of the coarsenessof soil texture), and soil carbon content across theGreat Plains of the United States (Burke et al. 1989).Soil carbon content was modeled based on regional

databases of the environmental variables using theCentury model. Soil carbon content varies regionallyin ways that are predictable from climate and soiltexture. (Figure provided by I. Burke.)


ronment but also on past events that influencethe species present at a site (see Chapter 13).The patterns of local dominance resulting frompast events can then affect the spatial patternof processes, such as nutrient cycling (Frelichand Reich 1995, Pastor et al. 1998). Whitespruce and paper birch, for example, are com-mon boreal trees that differ in both populationparameters (e.g., age of first reproduction andseed dispersal range) and species effects (e.g.,tissue nutrient concentration and decay rate)(Van Cleve et al. 1991, Pastor et al. 1998). Thespatial dynamics of birch and spruce thereforetranslate into spatial patterns of ecosystemprocesses such as rates of nitrogen cycling. Insemiarid ecosystems, soil processes are stronglyinfluenced by the presence or absence of indi-vidual plants, resulting in “resources islands”beneath plant canopies (Burke and Lauenroth1995). Herbivory also affects spatial patterningin ecosystems through its effects on SOM,nutrient availability, and NPP (Ruess andMcNaughton 1987). The distribution of specieson a landscape results from a combination ofhabitat requirements of a species, historicallegacies (see Chapter 13), and stochastic events.Once these patterns are established, they canpersist for a long time, if the species effects arestrong. The fine-scale distribution of hemlockand sugar maple that developed in Minnesotaseveral thousand years ago, for example, hasbeen maintained because these two tree specieseach produce soil conditions that favor theirown persistence (Davis et al. 1998).

Disturbance

Natural disturbances are ubiquitous in ecosys-tems and cause spatial patterning at manyscales. The literature on patch dynamics viewsa landscape as a mosaic of patches of differentages generated by cycles of disturbance andpostdisturbance succession (see Chapter 13)(Pickett and White 1985). In ecosystems char-acterized by gap-phase succession, the vegeta-tion at any point in the landscape is alwayschanging; but, averaged over a large enougharea, the proportion of the landscape in each successional stage is relatively constant,

forming a shifting steady state mosaic. Al-though every point in the landscape may be at different successional stages, the landscape asa whole may be close to steady state (Turner etal. 1993). This pattern is observed (1) in envi-ronmentally uniform areas, where disturbanceis the main source of landscape variability; (2)when disturbances are small relative to the sizeof the landscape; and (3) when the rate ofrecovery is faster than the return time of thedisturbance (Fig. 14.2). When disturbances aresmall and recovery is rapid, most of the land-scape will be in middle to late successionalstages. The formation of treefall gaps in gap-phase succession, for example, is the primarysource of canopy turnover in ecosystems suchas tropical rain forests, where large-scale disturbances are rare (Rollet 1983). Over time, gap-phase disturbance contributes to the maintenance of the productivity and nutrientdynamics of the entire forest. In the primaryrain forests of Costa Rica, for example,the regular occurrence of treefalls results inmaximum tree age of only 80 to 140 years(Hartshorn 1980). Light, and sometimes nutrient availability, increase in treefall gaps,providing resources that allow species withhigher resource requirements to grow quicklyand maintain themselves in the forest mosaic(Chazdon and Fetcher 1984b, Brokaw 1985).Disturbances by animals in grassland andshrublands can also generate a shifting steadystate mosaic. Gophers, for example, disturbpatches of California serpentine grasslands,causing patches to turn over every 3 to 5 years(Hobbs and Mooney 1991).

Large-scale infrequent disturbances alter thestructure and processes of some ecosystemsover large parts of a landscape. These distur-bances result in large expanses of the landscapein the same successional stage and are termednon–steady state mosaics. After Puerto Rico’shurricane Hugo in 1989, for example, most ofthe trees in the hurricane path were broken offor blown over or lost a large proportion of theirleaves, resulting in a massive transfer of carbonand nutrients from vegetation to the soils.The large pulse of high-quality litter increaseddecomposition rates substantially over largeareas (Scatena et al. 1996).


Fire can also create large patches of a singlesuccessional stage on the landscape (Johnson1992). The 1988 fires in Yellowstone NationalPark that burned 3200km2 of old pine forestaltered large areas of the landscape. Fires of this magnitude and intensity recur every fewcenturies (Romme and Knight 1982). Long-term human fire suppression has increased theproportion of late-successional communities inmany areas. This results in a more homoge-neous and spatially continuous, fuel-rich envi-ronment in which fires can burn large areas.Even large disturbed areas, however, are ofteninternally quite patchy. Fires, for example,generate islands of unburned vegetation and patches of varying burn severity. Theseunburned islands act as seed sources for post-fire succession and protective cover for wild-life, greatly reducing the effective size of the disturbance (Turner et al. 1997). In many cases, these patches become less distinct as

succession proceeds, so spatial heterogeneitymay decline with time in non–steady statemosaics.

Human-induced disturbances alter thenatural patterns and magnitude of landscapeheterogeneity. Half of the ice-free terrestrialsurface has been transformed by human activi-ties (Turner et al. 1990). We have cleared orselectively harvested forests, converted grass-lands and savannas to pastures or agriculturalsystems, drained wetlands, flooded uplands, andirrigated dry lands. Isolated land use changesmay augment landscape heterogeneity by cre-ating small patches within a matrix of lar-gely natural vegetation. As land use changebecomes more extensive, however, the human-dominated patches become the matrix in whichisolated fragments of natural ecosystems areembedded, causing a reduction in landscapeheterogeneity. These contrasting impacts ofhuman actions on landscape heterogeneity

10.0

5.0

1.0

0.5

0.1

0.05

0.01

Stable low

variance

Disturbance extent / Landscape extent [log scale]

Dis

turb

ance

inte

rval

/ R

ecov

ery

inte

rval

[log

sca

le]

0.750.500.25

Equilibriumor steady

state

Unstablesystem,

malfunctionor crash

Stable, low variance Stable,

very highvariance

Stable high

variance

Figure 14.2. Effect of disturbance size (relative tothe size of the landscape) and disturbance frequency(relative to the time required for the ecosystem torecover) on the stability of landscape processes.Landscapes are close to equilibrium when distur-bances are small (relative to the size of the studyarea) and when they are infrequent (relative to thetime required for ecosystem recovery). As distur-

bances become more frequent or larger, the land-scape becomes more heterogeneous and there isincreasing probability that the individual patchesmay undergo a different successional trajectory.(Redrawn with permission from Landscape Ecology,Vol. 8 © 1993 Kluwer Academic Publishers; Turneret al. 1993.)


are illustrated by the practice of shifting agriculture.

Shifting agriculture is a source of landscapeheterogeneity at low population densities butreduces landscape heterogeneity as humanpopulation increases. Shifting agriculture,also known as slash-and-burn agriculture orswidden agriculture, involves the clearing offorest for crops followed by a fallow periodduring which forests regrow, after which thecycle repeats. Shifting agriculture is practicedextensively in the tropics and in the past playedan important role in clearing the forests ofEurope and eastern North America. Smallareas of forest are typically cleared of mosttrees and burned to release organically boundnutrients. The soil is left untilled, causing littleloss of SOM. Crops are planted in species mix-tures, with multiple plantings and harvests(Vandermeer 1990). As soil fertility drops, andinsect and plant pests encroach, often within 3to 5 years, the agricultural plots are abandonedand allowed to regrow to forest. The regrowingforests provide fuel and other products for 20to 40 years until the cycle repeats. Shifting agriculture generates landscape heterogeneityat many scales, ranging from different agedpatches within a forest to different crop specieswithin a field.

With moderate human population densitiesthat allowed sufficient fallow periods and judi-cious selection of land for cultivation, shiftingagriculture existed sustainably for thousands of years and caused no directional change inbiogeochemical cycles (Ramakrishnan 1992).As population density increases, land becomesscarcer, and the fallow periods are shortened oreliminated, leading to a more homogeneousagricultural landscape. Under these conditions,nutrient and organic matter losses during theagricultural phase cannot be recouped, and the system degrades, requiring larger areas to provide sufficient food. As the landscapebecomes dominated by active cropland or earlysuccessional weedy species, the seed sources ofmid-successional species are eliminated, pre-venting forest regrowth and further reduc-ing the potential for landscape heterogeneity.In northeast India, for example, this shift-ing agriculture appears unsustainable when

the rotation cycle declines below 10 years(Ramakrishnan 1992).

Interactions Among Sources of Heterogeneity

Landscape heterogeneity and disturbancehistory interact to influence further distur-bance. Disturbance is more than a simpleoverlay on the spatial patterns governed byenvironment, because even slight variations intopography or edaphic factors can influence thefrequency, type, or severity of natural dis-turbance or the probability that land will becleared by people. Slope and aspect of a hill-side affect solar irradiance, soil moisture,soil temperature, and evapotranspiration rate.These factors, in turn, contribute to variation inbiomass accumulation, species composition,and fuel characteristics. Different parts of thelandscape may therefore differ in susceptibilityto fire. The resulting mosaic of patch types withdifferent flammabilities can prevent a small,locally contained fire from moving across largeareas. Slope and aspect can also directly influ-ence the exposure of ecosystems to fire spreadbecause fire generally moves uphill and tendsto halt at ridgetops. Elevation and topographicposition also influence the susceptibility offorest trees to wind-throw (Foster 1988).

Patchiness created by disturbance and otherlegacies influences the probability and spreadof disturbance, thereby maintaining the mosaicstructure of landscapes. The spread of fire, forexample, creates patches of early successionalvegetation in fire-prone ecosystems that areless flammable than late-successional vegeta-tion (Starfield and Chapin 1996). In this way,past disturbances create a legacy that governsthe probability and patch size of future distur-bances. The effectiveness of these disturbance-generated early successional firebreaksdepends on climate. At times of extreme fireweather, almost any vegetation will burn.

The past history of insect or pathogen out-breaks also generates a spatial pattern thatdetermines the pattern of future outbreaks. Inmountain hemlock ecosystems of the north-western United States, low light and nutrientavailability in old-growth stands makes trees


vulnerable to a root pathogen. The resultingtree death increases light, nitrogen mineraliza-tion, and nutrient availability, making theregrowing forest resistant to further attack(Matson and Boone 1984) (Fig. 14.3).The infec-tions tend to move through stands in a wave-like pattern, attacking susceptible patches and creating resistant patches in their wake(Sprugel 1976), just as described for fire. Simi-larly, hurricanes that blow down large patchesof trees generate early successional patches ofshort-statured trees that are less vulnerable towind-throw. Even the fine-grained steady statemosaics that characterize gap-phase successionare self-sustaining because young trees thatgrow in a gap created by treefall are less likely

to die than are older trees. In summary, distur-bances that reduce the probability of future dis-turbance generate a negative feedback thattends to stabilize the disturbance regime of anecosystem, resulting in a shifting steady statemosaic with a characteristic patch size andreturn interval. Any long-term trend in climateor soil resources that alters disturbance regimewill probably alter the characteristic distribu-tion of patch sizes on the landscape.

Disturbances that increase the probability ofother disturbances complicate predictions oflandscape pattern. Insect outbreaks that killtrees in a fine-scale mosaic, for example, canincrease the overall flammability of the for-est, increasing the probability of large fires(Holsten et al. 1995). The public concern aboutlarge fires after insect outbreaks then createspublic pressure for salvage logging of insect-killed stands. This logging creates patches ofclearcuts that are intermediate in size betweenthose created by insects and those that mighthave been produced by a catastrophic fire. It isdifficult to predict in advance which of thesethree alternative patch structures, or combina-tion of them, will occur, because disturbancesthat increase the probability of other distur-bances create a positive feedback that destabi-lizes the existing pattern of disturbance regimeand landscape heterogeneity. Rule-basedmodels that define conditions under which par-ticular scenarios are likely to occur provide aframework for predictions in the face of multi-ple potential outcomes (Starfield 1991).

Human activities create positive and negativefeedbacks to disturbances that alter the patchstructure and functioning of landscapes. Inprinciple, the effect of human-induced distur-bances, such as land clearing, on landscapestructure is no different from that of any otherdisturbance. However, the novel nature and theincreasingly extensive occurrence of humandisturbances are rapidly altering the structureof many landscapes. The construction of a roadthrough the tropical wet forests of Rondonia,Brazil, for example, created a simple linear dis-turbance of negligible size.The sudden increasein human access, however, led to rapid clearingof forest patches that were much larger than thenatural patches created by treefalls or the hand-

00

50

1 2 3 4 5

0 1 2 3 4 5

100

Rel

ativ

e he

ight

Time (sample position)

0

5

10

20

25

15

6

Am

mon

ium

-N

(µg

g-1

)

Matureforest

Rootpathogen

Forestrecovery

Figure 14.3. Transect of tree height and net nitro-gen mineralization rate across a disturbance causedby root pathogens in a hemlock stand in the north-western United States (Matson and Boone 1984).Nitrogen mineralization increases dramaticallybeneath trees recently killed by the pathogen (posi-tion 2). As trees recover (positions 3, 4, and 5), netnitrogen mineralization declines toward rates typicalof undisturbed forest (position 1).


cleared patches created by shifting agriculture.Similarly, road access is the major factor deter-mining the distribution of fire ignitions in theboreal forest of interior Alaska (Fig. 14.4). Ingeneral, road access is one of the best predic-tors of the spread of human-induced distur-bances in relatively natural landscapes (Dale et al. 2000).

Socioeconomic factors, such as farmerincome, interact with site characteristics toinfluence human effects on the landscapepattern. Heterogeneous landscapes are oftenconverted to fine-scale mosaics of agriculturaland natural vegetation, whereas large areassuitable for mechanized agriculture are morelikely to be deforested in large blocks. In north-ern Argentina, for example, patches of drydeciduous forests on the eastern slopes of theAndes were converted to small patches of crop-land or modified by grazing into thorn-scrubgrazing lands and secondary forests (Cabidoand Zak 1999) (Fig. 14.5). On the adjacentplains, however, larger parcels were initiallydeforested for grazing and more recently con-verted to mechanized agriculture. Large hold-ings on the plains are owned by companies thatmake land use decisions based on the globaleconomy. Small family producers in the moun-tains maintain a more traditional lifestyle thatinvolves smaller, less frequent changes in landuse.

Disturbance is increasingly used as a man-agement tool to generate more natural standand landscape structures. Forest harvest varies

from 0 to 100% tree removal, and the sizes andshapes of clearcuts can be altered from thestandard checkerboard pattern to mimic morenatural disturbances (Franklin et al. 1997).Forest harvest regimes can also be designed toretain some of the functional attributes of latesuccessional forests, such as the filtering func-tion of riparian vegetation, the presence oflarge woody debris, and the retention of a fewlarge trees as seed source and nesting habitat.Protection of these features can significantlyreduce the ecological impact of forest harvest.Prescribed fire is increasingly used as a man-agement tool, particularly in areas where acentury of Smoky the Bear policy of completefire suppression has led to unnaturally largefuel accumulations. Prescribed fires are typi-cally lit when weather conditions are such thatfire intensity and severity are low, so the fire canbe readily controlled. In populated mediter-ranean regions, vegetation may be physicallyremoved as a substitute for fire, because pre-scribed fires are considered unsafe. Natural fire,prescribed fire, and physical removal of vege-tation are likely to differ in their impacts onecosystem processes due to differences in the quantity of organic matter and nutrientsremoved; these differences affect subsequentregrowth. Ecologists are only beginning tounderstand the long-term consequences of different disturbance regimes for the structureand functioning of ecosystems and landscapes.As this understanding improves, more in-formed decisions can be made in using distur-

LightningFires

AnthropogenicFires

MajorRoads

Figure 14.4. Locations of naturally and human-caused fires in Alaska (Gabriel and Tande 1983).Thehuman-caused fires mirror the road and river trans-

portation corridors, indicating the importance ofhuman access in altering the regional fire regime.


bance as a tool in ecosystem management (seeChapter 16). The goal of manipulating distur-bance regime as a management tool is to mimicthe ecological effects of disturbance under con-ditions in which the natural disturbance patternhas unacceptable societal consequences, forexample in the rapidly expanding exurban-wildland interface where homes are being builtin fuel-rich habitats that are the product of firesuppression.

Patch Interactions on the Landscape

Interactions among patches on the landscapeinfluence the functioning of individual patchesand the landscape as a whole. Landscapepatches interact when things move acrossboundaries from one patch to another. Thisoccurs through topographically controlled in-teractions, transfers through the atmosphere,biotic transfers, and the spread of disturbance.These transfers are critically important to the

long-term sustainability of ecosystems becausethey represent losses from donor ecosystemsand subsidies to recipient ecosystems. Largechanges in these transfers constitute changes in inputs and/or outputs of resources and therefore substantially alter the functioning ofecosystems.

Topographic and Land-Water Interactions

Topographically controlled redistribution ofmaterials is the predominant physical pathwayby which materials move between ecosystems(Fig. 14.6). Gravity is a potent force for land-scape interactions. It causes water to movedownhill, carrying dissolved and particulatematerials. Gravity is also the driving force forlandslides, soil creep, and other forms of soilmovement. These topographically controlledprocesses transfer materials from uplands tolowlands, from terrestrial to aquatic systems,and from fresh-water ecosystems to estuariesand oceans (Naiman 1996).

Figure 14.5. Seminatural vegetation (black), landsthat have been modified by grazing (gray) and crop-lands (white) of the Cordoba region of northernArgentina in 1999 (Satellite based) (Cabido and Zak1999). The plains to the east are more suitable formechanized agriculture and are large land holdingswith substantial areas converted to croplands. Lands

to the west are more mountainous and less suitablefor mechanized agriculture; they are owned by smallfarmers, each of whom maintains a heterogeneousmosaic of land use.The proportion of area convertedto cropland is greater in large land holdings suitablefor intensive agriculture. (Figure provided by M.Cabido and M. Zak.)

Patch Interactions on the Landscape 315

The nature of donor ecosystems and theirmanagement govern the transfer of dissolvedmaterials. Regions with intensive agricultureand those receiving substantial nitrogen depo-sition transfer substantial quantities of nitrateand phosphorus to rivers, lakes, and ground-water (Carpenter et al. 1998). There is, forexample, a strong relationship between thetotal nitrogen input to the major watersheds of the world and nitrate loading in rivers (Fig.14.7) (Howarth et al. 1996). At more localscales, the patterns of land use and urbanizationinfluence the input of nutrients to lakes andstreams. These increased fluxes of dissolvednitrogen have multiple environmental conse-quences, including health hazards, acidification,eutrophication, and reduced biodiversity ofdownstream fresh-water and marine ecosys-tems (Howarth et al. 1996, Nixon et al. 1996).

Erosion moves particulate material contain-ing nutrients and organic matter from oneecosystem and deposits it in another. Erosionranges in scale from silt suspended in flowingwater to movement of whole mountainsides in landslides. The quantity of material moveddepends on many physical factors, includingslope position, slope gradient, the types of rocks and unconsolidated material underlying

soils, and the types of erosional agents (e.g.,amount and intensity of rainfall events) (seeChapter 3). The biological characteristics ofecosystems are also critical. Vegetation type,root strength, disturbance, management, andhuman development can be as important as thevertical gradient or parent material. Forestharvest on steep slopes in the northwesternUnited States, for example, has increased the

Uplands

Subsurfaceflow

Crop

Overlandflow

Ripariancommunity

Uptake

Denitrification

Groundwaterflow

Ground surface

Figure 14.6. Topographically controlled interac-tions among ecosystems in a landscape via erosionand solution transfers in subsurface flow or ground-water. Riparian forest trees absorb nutrients primar-ily from well-aerated soils, whereas denitrification

requires anoxic conditions, which generally occurbelow the water table. Nitrogen uptake and deni-trification are the most important mechanisms bywhich riparian zones filter nitrogen from ground-water between upland ecosystems and streams.

N SeaNW European coast

SE United States

SW Europeancoast

Mississippi

St. LawrenceBaltic

NE United States

Anthropogenic nitrogen inputs (kg km-2

yr-1

)

Nitr

ogen

out

puts

(kg

km

-2 y

r-1)

1600

1200

800

400

00 40002000 6000

N Canada

Figure 14.7. Relationship between the total anthro-pogenic nitrogen input to the major watersheds ofthe world and nitrate loading in rivers. (Redrawnwith permission from Biogeochemistry, Vol. 35 ©1996 Kluwer Academic Publishers; Howarth et al.1996.)


frequency of landslides. Similarly, upland agri-culture often increases sedimentation and theassociated transfer of nutrients and contami-nants (Comeleo et al. 1996). Proper manage-ment of upslope systems through use of covercrops, reduced tillage, and other management practices can reduce erosional transfers ofmaterials.

Landscape pattern influences the transfer of materials among ecosystems. Even in un-managed landscapes, ecosystems interact with one another along topographic sequences, withnutrients leached from uplands providing anutrient subsidy to midslope or lowland ecosys-tems (Shaver et al. 1991). The configuration ofthese ecosystems in the landscape determinesthe pattern of nutrient redistribution and theiroutputs to groundwater and streams. Riparianvegetation zones, including wetlands and flood-plain forests, act as filters and sediment trapsfor the water and materials moving fromuplands to streams (Fig. 14.6). Dominance ofriparian zones by disturbance-adapted plantsthat tolerate soil deposition and have rapidgrowth rates contributes to the effectiveness ofthese landscape filters. Riparian zones play aparticularly crucial role in agricultural water-sheds, where they remove fertilizer-derivednitrogen as well as phosphorus and erodingsediments. The fine-textured, organic-rich soilsand moist conditions characteristic of mostriparian areas also promote denitrification ofincoming nitrate. Plant uptake and denitrifica-tion together account for the decline in nitrateconcentration as groundwater flows from agricultural fields through riparian forests tostreams. Phosphorus is retained in riparianareas primarily by plant and microbial uptakeand physical adsorption to soils because phos-phorus has no pathway of gaseous loss.

The high productivity and nutrient status ofriparian vegetation and the presence of watercause riparian areas to be intensively used byanimals, including livestock in managed ecosys-tems. People also use riparian areas intensivelyfor water, gravel, transportation corridors, andrecreation. Long-term elevated inputs fromheavily fertilized agricultural areas or fromwetlands used for tertiary sewage treatment(i.e., to remove the products of microbial

decomposition) can saturate their capacity tofilter nutrients from groundwater. Any mode ofoverexploitation of riparian areas increasessediment and nutrient loading to streams andreduces shading, making fresh-water ecosys-tems more vulnerable to changes in land usewithin the watershed (Correll 1997, Lowranceet al. 1997, Naiman and Decamps 1997).

The properties of recipient ecosystems influ-ence their sensitivity to landscape interactions.The vulnerability of ecosystems to inputs fromother patches in the landscape depends largelyon their capacity to sequester or transfer theinputs. Riparian areas, for example, may have ahigher capacity to retain a pulse of nutrients ortransfer them to the atmosphere by denitrifica-tion than do upland late-successional forests.Streams characterized by frequent floods areless likely to accumulate sediment inputs thanare slow-moving streams and rivers, becausefloods flush sediments from river channels ofsteep stream reaches. Lakes on calcareous sub-strates or those that receive abundant ground-water input due to a location low in a watershedare better buffered against inputs of acidity andnutrients than are oligotrophic lakes on graniticsubstrates or lakes high in a watershed thatreceive less groundwater input (Webster et al.1996).

Estuaries, the coastal ecosystems locatedwhere rivers mix with seawater, are a strikingexample of the way in which ecosystem prop-erties influence their sensitivity to inputs fromthe landscape. They are among the most pro-ductive ecosystems on Earth (Howarth et al.1996, Nixon et al. 1996). Their high productiv-ity stems in part from the inputs they receivefrom land and from the physical structure of theecosystem, which is stabilized by the presenceof seagrasses and other rooted plants. Thistends to dampen wave and tidal energy, reduc-ing resuspension and increasing sedimentation.Salinity and other geochemical changes thatoccur as the waters mix lead to flocculation andsettling of suspended particulates. Nutrientuptake by the rooted vegetation and phyto-plankton, burial by sedimentation, and denitri-fication in anoxic sediments function as sinksfor nutrients flowing from upstream water-sheds, just as in riparian zones. Estuaries differ


from one another in their capacity to act assinks for incoming materials, due to variation inbasin geometry, sediment input, and tidal inter-actions. The stability of the landscape on theMississippi River Delta, for example, dependson regular delivery of sediments from upstreamto replace soils removed by tidal erosion. Chan-nels, levees, and other engineering solutions to flood control and water management mayreduce the probability of flooding but greatlyaugment the land loss to coastal erosion(Costanza et al. 1990). Many estuaries, includ-ing the Gulf of Mexico near the entrance of theMississippi River, are becoming saturated bynutrient enrichment within their watersheds,resulting in harmful algal blooms, loss of sea-grass, and increasing frequency of anoxia orhypoxia and related fish kills (Mitsch et al.2001).

Atmospheric Transfers

Atmospheric transport of gases and particleslinks ecosystems over large distances andcoarse spatial scales. Gases emitted frommanaged or natural ecosystems are processed

in the atmosphere and can be transported fordistances ranging from kilometers to the globe.Particulates from biomass burning, wind-blowndust, sea spray, and anthropogenic sources canalso be carried through the atmosphere fromone ecosystem to another. Once deposited, theycan alter the functioning of the recipientecosystems (Fig. 14.8), just as with topographi-cally controlled transfers.

In areas downwind of agriculture, ammoniagas (NH3) and nitric oxides (NOx) can repre-sent a significant fraction of nitrogen deposi-tion. Dutch heathlands, for example, receive10-fold more nitrogen deposition than wouldoccur naturally. The magnitude of these inputsis similar to the quantity of nitrogen that annu-ally cycles through vegetation, greatly increas-ing the openness of the nitrogen cycle. Areasdownwind of industry and fossil fuel combus-tion receive nitrogen largely as NOx. Sulfurgases, including sulfur dioxide (SO2), are alsoproduced by fossil fuel combustion, althoughimproved regulations have reduced these emis-sions and deposition relative to NOx.

The large nitrogen inputs to ecosystems haveimportant consequences for NPP, nutrient

Fertilizer N

Fossil fuel andbiomass combustion

Tracegases

Tracegases

Deposition

Leaching Leaching

Crop

Source region Sink region

Figure 14.8. Atmospheric transfers of gases, solutions, and particulates among ecosystems. Inputs come fromfossil fuel and biomass combustion and from trace gases originating from natural and managed ecosystems.


cycling, trace gas fluxes, and carbon storage.Chronic nitrogen deposition initially reducesnitrogen limitation by increasing nitrogencycling rates, foliar nitrogen concentrations,and NPP. Above some threshold, however, theecosystem becomes saturated with nitrogen(Fig. 14.9) (Aber et al. 1998). As excess nitrateand sulfate leach from the soil, they carry withthem cations to maintain charge balance, in-ducing calcium and magnesium deficiency invegetation (Driscoll et al. 2001). In southernSweden, for example, over half of the plant-available cations have been lost from the upper70cm of soil in the past half century, probablydue to chronic exposure to acid precipitation(Hallbacken 1992). The exchange complexbecomes more dominated by aluminum andhydrogen ions, increasing soil acidity and thelikelihood of aluminum toxicity. Together thissuite of soil changes often enhances frost sus-ceptibility, impairs root development, and pro-motes herbivory, leading to forest decline inmany areas of Europe and the northeasternUnited States (Schulze 1989, Aber et al. 1998).The major surprise, however, has been howresilient many forests have been to acid rain,often retaining most of the nitrogen inputswithin the ecosystem for as much as twodecades. The resilience of ecosystems dependsin part on the magnitude of inputs (related todistance from pollution sources and amount ofprecipitation received) and initial soil acidity,which in turn depends on parent material andspecies composition. Many ecosystems, how-

ever, now show clear signs of nitrogen satura-tion, resulting in forest decline, loss of acid-neutralizing capacity in lakes, and increasingnitrogen inputs to streams (Aber et al. 1998,Carpenter et al. 1998, Driscoll et al. 2001).

Nearly all research on the transport, deposi-tion, and ecosystem consequences of anthro-pogenic nitrogen has been conducted in thetemperate zone. Further increases in nitrogendeposition will, however, likely occur primarilyin the tropics and subtropics (Galloway et al.1995), where plant and microbial growth arefrequently limited by elements other thannitrogen. These ecosystems might thereforeshow more immediate nitrogen loss in tracegases or leaching in response to nitrogen depo-sition (Matson et al. 1998). On the other hand,soil properties such as high clay content orcation exchange capacity may allow tropicalsoils to sequester substantial quantities of nitro-gen before they become leaky.

Biomass burning transfers nutrients directlyfrom terrestrial pools to the atmosphere andthen to down-wind ecosystems. Biomass com-bustion releases a suite of gases that reflect theelemental concentrations in vegetation and fireintensity. About half of dry biomass consists ofcarbon, so the predominant gases released arecarbon compounds in various stages of oxida-tion, including carbon dioxide (CO2), methane(CH4), carbon monoxide (CO), and smallerquantities of nonmethane hydrocarbons. Theatmospheric role of these gases varies. CO2 andCH4 are greenhouse gases, whereas carbon

Foliar NN mineralization

NPP

Ca:Al andMg:N ratios

Nitrification

Leaching

Stage of N saturation1 2 30

0

50

100

150

200

Rel

ativ

e un

its

Figure 14.9. Changes hypothesizedto occur as forests undergo long-term nitrogen deposition and ni-trogen saturation. (Redrawn withpermission from BioScience; Aberet al. 1998.)


monoxide and nonmethane hydrocarbons reactin the troposphere to produce ozone and otheratmospheric pollutants that can affect down-wind ecosystems (see Chapter 2). Nitrogen isalso released in various oxidation states, includ-ing nitrogen oxides (NO and NO2, togetherknown as NOx) and ammonia. The propor-tional release of these forms also depends onthe intensity of the burn, with NOx typicallyaccounting for most of the emissions. Sulfur-containing gases; organic soot and other aerosolparticles; elemental carbon; and many tracespecies of carbon, nitrogen, and sulfur also haveimportant regional and global effects. Satelliteand aircraft data show that these gases andaerosols in biomass-burning plumes can betransported long distances.

Windblown particles of natural and anthro-pogenic origins link ecosystems on a landscape.The role of the atmosphere as a transportpathway among ecosystems differs among ele-ments. For some base cations (Ca2+, Mg2+, Na+,and K+) and for phosphorus, dust transport isthe major atmospheric link among ecosystems.At the local to regional scales, dust from roadsor rivers can alter soil pH and other soil

properties that account for regional zonation of vegetation and land–atmosphere exchange(Walker and Everett 1991, Walker et al. 1998).At the global scale, Saharan dust is transportedacross the Atlantic Ocean and deposited on theAmazon by tropical easterlies. Although theannual input of dust is small, it contributes sub-stantially to soil development over the longterm (Graustein and Armstrong 1983). Simi-larly, dust from the Gobi desert is deposited inthe Hawaiian Islands at the rate of 0.1gm-2 percentury. In old soils (those more than 2 millionyears old), dust input can be the largest sourceof base cations (Chadwick et al. 1999).

Land–atmosphere exchange of water andenergy in one location influences downwindclimate. Oceans and large lakes moderate theclimate of adjacent land areas by reducing tem-perature extremes and increasing precipitation(see Chapter 4). Human alteration of the landsurface is now occurring so extensively that italso has significant effects on downwind ecosys-tems. Conversion of Australian heathlands toagriculture has, for example, increased precipi-tation over heathlands and reduced it by 30%over agricultural areas (Fig. 14.10) (see Chapter

Crop Natural vegetation

Moistair

rises

Dry air subsides

Low albedo

High roughness

High sensible heat

High air temperature

High albedo

High RH

+10% Precipitation-30% Precipitation

Air

m ovement

Figure 14.10. Effects on regional climate of conver-sion from heathland to barley croplands in south-western Australia (Chambers 1998). The heathlandabsorbs more radiation (low albedo) and transmits alarger proportion of this energy to the atmosphereas sensible heat than does adjacent croplands. Thiscauses air to rise over the heathland and draws in

moist air laterally from the irrigated cropland; thiscauses subsidence of air over the cropland, just aswith land–sea breezes. Rising moist air increases pre-cipitation by 10% over heathland, whereas subsidingdry air reduces precipitation by 30% over the crop-land. RH, relative humidity.


4). At a global scale, the clearing of land foragriculture has reduced regional albedo andevapotranspiration, leading to greater sensibleheat flux (Chase et al. 2000). At all spatialscales, the atmospheric transfer of heat andwater vapor from one ecosystem to anotherstrongly affects ecosystem processes in down-wind ecosystems. The climatic impacts ondownwind ecosystems of reservoirs, irrigationof arid lands, and land use change are seldomincluded in assessments of the potential effectsof these management projects.

Movement of Plants and Animals onthe Landscape

The movement and dispersal of plants andanimals link ecosystems on a landscape. Largeanimals typically consume forage from high-quality patches and deposit it where they restor sleep. Sheep in New Zealand, for example,often camp on ridges at night, moving nutrientsupward and counteracting the downward nutri-ent transport by gravity. Marine birds transferso much phosphorus from marine foods to theland that the guano deposited in their tradi-tional nesting areas has served as a majorsource of phosphorus for fertilizer. Anadro-mous fish—that is, marine fish that enter freshwater to breed, also transport marine-derivednutrients to terrestrial ecosystems. These fishcarry the nutrients up rivers and streams, wherethey become an important food item for ter-restrial predators, which transport the marine-derived nutrients to riparian and uplandterrestrial ecosystems (Willson et al. 1998,Helfield and Naiman 2001).These nutrient sub-sidies by animals contribute to the spatial pat-terning in ecosystem processes. Insects thatfeed on seaweed and other marine detritus arean important food source for spiders on islands,merging marine and terrestrial food webs (Polisand Hurd 1996).

Animals also transfer plants, especially asseeds, on fur and in feces. Many plants haveevolved life history strategies to take advantageof this efficient form of dispersal. This dispersalmechanism has contributed to the spread ofinvasive plants. Feral pigs, a nonnative herbi-vore in Hawaiian rain forests, for example,

transfer seeds of invasive plants such as thepassion vine, which alters patterns of nutrientcycling. Similarly, the alien bird white eyespreads the alien nitrogen fixer Myrica faya(Woodward et al. 1990), which alters the nitro-gen status of native ecosystems (Vitousek et al.1987). Thus invasions of both plants and ani-mals from one ecosystem to another can con-tribute to a variety of ecosystem changes.

Animals that move among patches can haveeffects that differ among patch types. Edge specialists such as deer, for example, may con-centrate their browsing in one habitat type butseek protection from predators and depositnutrients in another. At a larger scale, migra-tory birds move seasonally among differentecosystem types. Lesser snow geese, for ex-ample, overwinter in the southern UnitedStates and breed in the Canadian Arctic. Popu-lations of this species have increased by morethan an order of magnitude as a result ofincreased use of agricultural crops (rice, corn,and wheat) on the wintering grounds andreduced hunting pressure. This species nowexceeds the carrying capacity of its summerbreeding grounds and has converted productivearctic salt marshes into unvegetated barrens(Jefferies and Bryant 1995).

People are an increasing cause of lateraltransfers of materials among ecosystems,through addition of fertilizers, pesticides, etc.,and removal of crops and forest products, anddiversion of water. The nutrient transfers infood from rural to urban areas are substantial.The resulting nutrient inputs to aquatic systemsoccur in locations where riparian zones andother ecological filters are often degraded orabsent. Water diversion by people has substan-tially altered rates and patterns of land usechange in arid areas at the expense of rivers andwetlands (see Chapter 4). As water becomesincreasingly scarce in the coming decades, pres-sures for water diversion are likely to increase.

Disturbance Spread

Patch size and arrangement determine thespread of disturbance across a landscape.Disturbance is a critical interactive control over ecosystem processes that is strongly influ-

Human Land Use Change and Landscape Heterogeneity 321

enced by horizontal spread from one patch toanother. Fire and many pests and pathogensmove most readily across continuous stretchesof disturbance-prone vegetation. Fire breaks ofnonflammable vegetation, for example, are aneffective mechanism of reducing fire risk at theurban-wildland interface. Fires create their ownfire breaks because postfire vegetation is gen-erally less flammable than that which precedesa fire. Theoretical models suggest that, whenless than half of the landscape is disturbanceprone, frequency is less important than severityin determining the impacts of disturbance.When large proportions of the landscape aresusceptible to disturbance, however, the fre-quency of disturbance becomes increasinglyimportant (Gardner et al. 1987, Turner et al.1989). The size of patches also influences thespread of disturbance. Landscapes dominatedby large patches tend to have a low frequencyof large fires. Landscapes with small patches,have greater edge-to-area ratio, so fires tend tospread more frequently into less flammablevegetation (Rupp et al. 2000).

Patchy agricultural landscapes are less proneto spread of pests and pathogens than are largecontinuous monocultures. Intensive agriculturehas reduced landscape patchiness in severalrespects. The average size of individual fieldsand the proportion of the total area devoted to agriculture has generally increased, as hasthe use of genetically uniform varieties. This can lead to rapid spread of pests across thelandscape.

Human Land Use Change andLandscape Heterogeneity

Human modification of landscapes has funda-mentally altered the role of ecosystems inregional and global processes. Much of the landuse change has occurred within the last two tothree centuries, a relatively short time in thecontext of evolution or landscape development.Since 1700, for example, the land area devotedto crop production has increased 466% to acurrent 15 million km2 worldwide, an areaalmost twice the size of the conterminousUnited States. Many areas of the world are

therefore dominated by a patchwork of agri-cultural fields, pastures, and remnant unman-aged ecosystems. Similar patchworks of cut andregrowing forest interspersed with small areasof old-growth forest are common on every con-tinent. Human-dominated landscapes supplylarge amounts of food, fiber, and other ecosys-tem services to society. Two general patterns ofland use change emerge: (1) extensification, orthe increase in area affected by human activi-ties, and (2) intensification, or the increasedinputs applied to a given area of land or water.

Extensification

Land use changes include both conversions and modifications (Meyer and Turner 1992).Land use conversion involves a human-inducedchange in ecosystem type to one dominated bydifferent physical environment or plant func-tional type, for example, the change from forestto pasture or from stream to reservoir. Landuse modification is the human alteration of an ecosystem in ways that significantly affectecosystem processes, community structure, andpopulation dynamics without radically chang-ing the physical environment or dominant plantfunctional type. Examples include alteration ofnatural forest to managed forest, savanna man-agement as grazing lands, and alteration of tra-ditional low-input agriculture to high-intensityagriculture. In aquatic ecosystems this includesthe alteration of flood frequency by dams andlevees or the stocking of lakes for sport fishing.Both types of land use change alter the func-tioning of ecosystems, the interaction of patcheson the landscape, and the functioning of land-scapes as a whole.

Deforestation is an important conversion in terms of spatial extent and ecosystem andglobal consequences. Forests cover about 30%of the terrestrial surface, about three times thetotal agricultural land area. Globally, forestarea has decreased about 15% (i.e., by 9 millionkm2) since preagricultural times. Much of the European and the Indian subcontinents,for example, were prehistorically blanketed byforests but over the last 5 to 10 centuries havesupported extensive areas of agriculture. Simi-larly, North America was once contiguously


wooded from the Atlantic seaboard to the Mississippi River, but large areas of this forestwere cleared by European settlers at ratessimilar to those that now characterize tropicalforests (Dale et al. 2000).

Today, conversion of forests to pasture oragriculture is one of the dominant land usechanges in the humid tropics. The magnitude ofthis land use change is uncertain, but one inter-mediate estimate suggests that 75,000km2 ofprimary tropical forest—that is, forest that hadnever been cleared—are now cleared annually,either permanently or through the expansion of shifting cultivation (Melillo et al. 1996).Another 37,000km2 of primary forest arelogged annually for commercial use. Finally,approximately 145,000km2 of secondary for-est—that is, forests that have regrown afterearlier clearing—are cleared each year. At thisrate, most primary tropical forests may disap-pear in the next several decades. The rates andpatterns of forest clearing vary regionally, soprimary forests are likely to persist longer insome regions such as Amazonia and CentralAfrica than elsewhere.

The trajectory of landscape change caused by deforestation depends on both the nature of the original forests and the land use thatfollows. The permanent or long-term conver-sion of forests to managed ecosystems involvesburning or removal of most of the biomass andoften leads to large losses of carbon, nitrogen,and other nutrient elements from the system.Logging, in contrast, removes only the com-mercially valuable trees and may cause lesscarbon and nutrient loss from soils. The nutri-ent losses that accompany deforestation cancause nutrient limitations to plant and micro-bial growth. They can also alter adjacent eco-systems, particularly aquatic ecosystems, andinfluence the atmosphere and climate throughchanges in trace gas fluxes (see Chapter 9) andwater and energy exchange (see Chapter 4).

Reforestation of abandoned agriculturalland through natural succession or active treeplanting is also changing landscapes, particu-larly in the eastern United States, Europe,China, and Russia (see Fig. 13.2). In the easternUnited States, for example, much of the land

that was originally cleared reverted to forestdominated by native species. In Chile, however,primary forests are being replaced by planta-tions of rapidly growing exotic trees such asPinus radiata (Armesto et al. 2001).These plan-tations have low diversity and a quite differentlitter chemistry and pattern of nutrient cyclingthan do the primary forests that they replace.The characteristics of the regrowing forests also depend on the previous types of land use(Foster et al. 1996, Foster and Motzkin 1998).Long-term and intensive agricultural practicescan compact the soil, alter soil structure anddrainage capability, deplete the soil organicmatter, reduce soil water-holding capacity,reduce nutrient availability, deplete the seedbank of native species, and introduce newweedy species. The forests that regenerate onsuch land may therefore differ substantiallyfrom the original forest and from those regrowon less-intensively managed lands (Motzkin et al. 1996). Grazing intensity and accompany-ing land management practices also influencepotential revegetation, with more intensivelygrazed systems often taking longer to regainforest biomass. Natural reforestation underthese conditions may proceed slowly or not at all.

Use of native grasslands, savannas, andshrublands for cattle grazing is the most ex-tensive modification of natural ecosystemsoccurring today. Globally, thousands of squarekilometers of savanna are burned annually to maintain productivity for cattle grazing.Although both fire and grazing are naturalcomponents of most grasslands, changes in the frequency and/or severity of burning andgrazing alters ecosystem processes. Burningreleases nutrients and stimulates the produc-tion of new leaves that have a higher proteincontent and are more palatable to grazers.Conversely, grazers reduce fire probability byreducing the accumulation of grass biomass andleaf litter. Fire and grazers both prevent estab-lishment of most trees, which might other-wise convert savannas to woodlands or forests.When fire frequency increases substantially,however, the loss of carbon and nitrogen fromthe system can reduce soil fertility and water

Human Land Use Change and Landscape Heterogeneity 323

retention (and therefore productivity). It canalso affect regional trace gas budgets and deposition in downwind ecosystems and thetransfer of nutrients and sediments to aquaticecosystems.

Vegetation in many grassland ecosystems co-evolved with large herbivores. The steppeecosystems of the central United States, easternAfrica, and Argentina, for example, supportedhigh densities of animals, so grazing by do-mestic cattle does not necessarily degrade them (Milchunas et al. 1988, Owen-Smith 1988,Osterheld et al. 1992). Grazing is, in fact, anessential component of nutrient cycling and themaintenance of productivity of these grasslands(McNaughton 1979, 1985). In these ecosystems,grazers maintain the competitive advantage ofnative grazing-tolerant grasses. Overgrazing ismore likely to occur in environments with lowwater or nutrient availability, where intensivegrazing may exceed the productive capacity ofthe ecosystem. Under these circumstances, thecover of palatable grasses may decline, makingthe soil more prone to erosion, which feedsback to further loss of productivity (Schlesingeret al. 1990). Overgrazing can contribute to the encroachment of shrubs or succulents intograsslands. In other cases, cattle contribute toshrub encroachment through dispersal of seedsin dung (Brown and Archer 1999). Some of themost productive grasslands, such as the pampasof Argentina or the tallgrass prairies of Ukraineor the midwestern United States have beenlargely converted to mechanized agriculture.

Expansion of marine fishing has alteredmarine food webs globally, with cascadingeffects on most ecosystem processes. The areaof the world’s oceans that are actively fishedhas increased substantially, in part becausetechnological advances allow fish and benthicinvertebrates to be harvested more efficientlyand stored for longer times before returning tomarkets. Most of Earth’s continental shelves,the most productive marine ecosystems, arenow actively fished, as are productive high-latitude open oceans. Removal of fish has cas-cading effects on pelagic ecosystems becausefish predation has large top–down effects onthe biomass and on species composition of zoo-

plankton, which in turn affect primary produc-tivity by phytoplankton and the recycling ofnutrients within the water column (see Chapter10). Harvesting of benthic invertebrates, such asclams, crabs, and oysters, also has large ecosys-tem effects because of direct habitat distur-bance and the effects of these organisms ondetrital food webs and benthic decomposition.The globalization of marine fisheries has abroader impact than we might expect, becausemany large fish are highly mobile and migratefor thousands of kilometers. Large changes in their populations therefore have ecologicaleffects that diffuse widely thoughout the oceanand even into fresh-water ecosystems in thecase of anadromous fish.

Intensification

Intensification of agriculture frequently re-duces landscape heterogeneity and increasesthe transfer of nutrients and other pollutants toadjacent ecosystems. Agricultural intensifica-tion involves the intensive use of high-yieldcrop varieties combined with tillage; irrigation;and industrially produced fertilizers, pesticides,and herbicides. Intensification has allowed foodproduction to keep pace with the rapid humanpopulation growth (see Fig. 8.1) (Evans 1980).Although this practice has minimized the aerialextent of land required for agriculture, it hasnearly eliminated some ecosystem types thatwould naturally occupy areas of high soil fertil-ity. Intensive agriculture is most developed onrelatively flat areas such as floodplains andprairies that are suitable for irrigation and useof large farm machinery. The high cost of thisequipment requires that large areas be culti-vated, largely eliminating natural patterns oflandscape heterogeneity.

Agricultural intensification generates bio-geochemical hot spots that alter ecosystemprocesses in ways that impact the local, re-gional, and global environment (Matson et al.1997). Tillage increases decomposition rate byreducing the physical protection of SOM andaltering soil microclimate (see Chapters 3 and7). In this way 30 to 50% of the original soilcarbon is lost from permanently cultivated agri-


cultural systems (Fig. 14.11). These soil carbonlosses often occur within 20 years in temperateecosystems or within 5 years in tropical systems,depending on soil temperature and moisture.The large regular inputs of nutrients requiredto sustain intensive agriculture (see Fig. 8.1)increases the emissions of nitrogen trace gasesthat play a significant role in the global nitro-gen cycle and link these ecosystems with down-wind ecosystems (see Chapters 9 and 15).

Nutrient loading on land increases nonpointsources of pollution for neighboring aquaticecosystems (Carpenter et al. 1998). Phosphorusadditions on land have particular large effectsfor at least two reasons. First, primary produc-tion of most lakes is phosphorus limited andtherefore responds sensitively to even smallphosphorus additions. Second, much of thephosphorus added to agricultural fields ischemically fixed, so orders of magnitude morephosphorus are generally added to fields thanis absorbed by crops. These large additions rep-

resent a massive reservoir of phosphorus thatwill continue to enter aquatic ecosystems longafter farmers stop adding fertilizer. Phosphorusinputs from human and livestock sewage havesimilarly long-lasting effects.

Intensive agriculture for rice production inflooded ecosystems produces a different type ofbiogeochemical hot spot. Rice feeds half of theworld’s population, with 90% of the productionoccurring in Asia. The most intensive rice culti-vation takes place in periodically flooded fields,where reduced oxygen supply and decomposi-tion lead to greater accumulation of SOM thanin upland agricultural systems. Flooding createsan ideal environment for methanogens, whichproduce methane during decomposition oforganic matter under anaerobic conditions (see Chapter 7). Paddy agricultural systems aretherefore important sources of atmosphericmethane and now produce about half as muchmethane as all natural wetlands. Together withcattle production, rice cultivation accounts formuch of the increase in atmospheric methane(see Chapter 15).

Land use change caused greater ecologicalimpact during the twentieth century than anyother global change. Understanding and pro-jecting future changes in land use are there-fore critical to predicting and managing futurechanges in the Earth System. Development ofplausible scenarios for the future requires closecollaboration among climatologists, ecologists,agronomists, and social scientists. Optimisticscenarios that assume that the growing humanpopulation will be fed rather than die fromfamines, wars, or disease epidemics project con-tinued large changes in land use, particularly in developing countries (Alcamo 1994). Whatactually occurs in the future is, of course, uncer-tain, but these and other scenarios suggest thatland use change will continue to be the majorcause of global environmental change in thecoming decades. Ecologists working togetherwith policy makers, planners, and managershave the opportunity to develop approachesthat will minimize the impact of future land-scape changes (see Chapter 16). This visionmust recognize the large effects of land usechange on landscape processes and their con-sequences on local to global scales.

Year1900 1920 19801940 1960

5

4

3

6

Soi

l C (

kg m

-2)

Conversion to agriculturefrom native vegetation

Historical management

Conventionaltillage

Reducedtillage

61% of 1907

53% of 1907

7

Figure 14.11. Simulation of loss of SOM after con-version of grassland to agriculture, followed by asmall increase with conversion to low-till agriculture.Losses of soil carbon reduce the productive poten-tial of the soil and transfer carbon in the form of CO2

from the soil to the atmosphere. New techniques oflow-till and no-till agriculture can reduce the mag-nitude of soil carbon loss or, under some circum-stances, lead to a small net carbon accumulation.(Redrawn with permission from Science, Vol. 277 ©1997 American Association for the Advancement ofScience; Matson et al. 1997.)

Spatial Heterogeneity and Scaling 325

Spatial Heterogeneity and Scaling

Extrapolation of ecosystem processes to largespatial scales requires an understanding of the role of spatial heterogeneity in ecosystemprocesses. Efforts to estimate the cumulativeeffect of ecosystem processes at regional andglobal scales has contributed to the increasedrecognition of the importance of landscapeprocesses in ecosystem dynamics. Estimates of global productivity or annual carbon se-questration, for example, require that ratesmeasured (or modeled) in a few locations be extrapolated over large areas. Many appro-aches to spatial extrapolation have been used,each with its advantages and disadvantages. Apaint-by-numbers approach estimates the fluxor pool for a large area by multiplying theaverage value for each patch type (e.g., theyield of major types of crops or the carbonstocks of different forest types) times the aerialextent of that patch type. This provides a roughapproximation that can guide process-basedresearch. This approach requires the selectionof representative values of processes and accu-

rate estimates of the area of each patch type.Satellite imagery now provides improved esti-mates of the aerial extent of many patch types,but spatial and temporal variation in processesmakes it difficult to find good representativesites from which data can be extrapolated.This extrapolation approach can be combinedwith empirical regression relationships (ratherthan a single representative value) to estimateprocess rates for each patch type. Carbon poolsin forests, for example, might be estimated as a function of temperature or normalized dif-ference vegetation index (NDVI) rather thanassuming that a single value could represent thecarbon stocks of all forests.

Process-based models make up anotherapproach to estimating fluxes or pools overlarge areas. These estimates are based on maps of input variables for an area (e.g., mapsof climate, elevation, soils, and satellite-basedindices of leaf area) and a model that relatesinput variables to the ecosystem property thatis simulated by the model (Potter et al. 1993,VEMAP Members 1995) (Box 14.1). Regio-nal evapotranspiration, for example, can be estimated from satellite data on vegetationcomposition and maps of temperature and pre-

Box 14.1. Spatial Scaling Through Ecological Modeling

The complexity of ecological controls overall the processes that influence ecosystemcarbon balance makes long-term projectionsof terrestrial carbon storage a daunting task.Making these projections is, however, criti-cal to improving our understanding of therelative role of terrestrial ecosystems in theglobal carbon balance. Experiments that testthe multiple combinations of environmen-tal conditions influencing terrestrial carbonstorage are difficult to design. Modelingallows a limited amount of empirical in-formation to be greatly extended throughsimulation of complex combinations of environmental–biotic interactions. One im-portant use of ecosystem models has been to identify the key controls that govern

long-term changes in terrestrial carbon storage (net ecosystem production, NEP).This application of ecosystem understand-ing is central to societal issues and policy formulation.

Many of key processes regulating NEPinvolve changes that occur over decades tocenturies. The temporal resolution of themodels must therefore be coarse, with timesteps (the shortest unit of time simulated bythe model) of a day, month, or year. Use ofrelatively long time steps such as weeks ormonths reduces the level of detail that canbe considered. Temperature and moisturecontrols over decomposition, for example,can still be observed with an annual timestep.The short-term pulses of decomposition


associated with drying and wetting cycles orgrazing by soil fauna, however, are subsumedin the shape of the annual temperature andmoisture response curves of decompositionand in the decomposition coefficients. Theprocesses that are unique to the rhizosphereand bulk soil must also be ignored, whenthese environments are lumped into a singlesoil pool in a model. In this case, only themore general controls such as temperature,moisture, and chemistry can be included.

The basic structure of a model of NEPmust include the pools of carbon in the soils and vegetation. It must also include thefluxes of carbon from the atmosphere toplants (gross primary production [GPP] orNPP), from plants to the atmosphere (plantrespiration, harvest, and combustion), fromplants to soil (litterfall), and from soil to the atmosphere (decomposition and dis-turbance). Models differ in the detail withwhich these and other pools and fluxes arerepresented. Plants, for example, might beconsidered a single pool or might be sepa-rated into different plant parts (leaves,stems, and roots), functional types of plants(e.g., trees and grasses in a savanna), orchemical fractions such as cell wall and cellcontents. Under some circumstances, certainfluxes (e.g., fire and leaching) are ignored.There is no single “best” model of NEP. Eachmodel has a unique set of objectives, and themodel structure must be designed to meetthese objectives. We briefly describe howthree models incorporate information aboutcontrols over NEP, emphasizing how the differences in model structure make eachmodel appropriate to particular questions orecosystems.

Perhaps the biggest challenge in modeldevelopment is deciding which processes toinclude. One approach is to use a hierar-chical series of models to address differentquestions at different scales (Reynolds et al.1993). Models of leaf-level photosynthesisand of microclimate within a canopy havebeen developed and extensively tested foragricultural crops, based on the basic prin-ciples of leaf biochemistry and the physics of radiation transfer within canopies. One

output of these models is a regression rela-tionship between environment at the top ofthe canopy and net photosynthesis by thecanopy. This environment–photosynthesisregression relationship can then be incorpo-rated into models operating at larger tem-poral and spatial scales to simulate NPP,without explicitly including all the details of biochemistry and radiation transfer. Thishierarchical approach to modeling providesan opportunity to validate the model out-put (i.e., compare the model predictions with data obtained from field observationsor experimental manipulations) at severalscales of temporal and spatial resolution,providing confidence that the model cap-tures the important underlying processes ateach level of resolution.

The Terrestrial Ecosystem Model (TEM)(Fig. 14.12) was designed to simulate thecarbon budget of ecosystems for all locationson Earth at 0.5° longitude by 0.5° latituderesolution (60,000 grid cells) for time periodsof a century or more (McGuire et al. 2001).TEM has a relatively simple structure and a monthly time step, so it can run efficientlyin large numbers of grid cells for longperiods of time. Soil, for example, consists of a single carbon pool. The model assumessimple universal relationships between en-vironment and ecosystem processes based on general principles that have been es-tablished in ecosystem studies. The modelassumes, for example, that decompositionrate of the soil carbon pool depends on the size of this pool and is influenced by the temperature, moisture, and C :N ratio of the soil. TEM incorporates feedbacks thatconstrain the possible model outcomes.The nitrogen released by decomposition, forexample, determines the nitrogen availablefor NPP, which in turn governs carbon in-puts to the soil and therefore the pool of soil carbon available for decomposition.Thissimplified representation of ecosystem car-bon dynamics is sufficient to capture globalpatterns of carbon storage (McGuire et al.2001), making the model useful in simulatingregional and global patterns of soil carbonstorage.


Figure 14.12. The decomposition portion ofthree terrestrial ecosystem models: TEM(McGuire et al. 1995), LINKAGES (Pastor andPost 1986), and CENTURY (Parton et al. 1987).Inputs from the vegetation component of thesemodels is shown as plant litter. Arrows indicatethe fluxes of carbon from litter to other pools and

eventually to CO2. The bow ties indicate controlsover these fluxes (or the partitioning of the fluxto two pools) as functions ( f ) of C : N ratio (C : N), lignin (L), lignin : N ratio (L : N), tempera-ture (T), and moisture (M). In CENTURY weshow representative residence times of differentcarbon pools in grassland soils.

The CENTURY model was originallydeveloped to simulate changes in soil carbonstorage in grasslands in response to variationin climate, soils, and tillage (Parton et al.1987, 1993) (Fig. 14.12). It has since beenadapted to most global ecosystem types.In CENTURY, the soil is subdivided intothree compartments (active, slow, and pas-sive soil carbon pools) that are defined em-pirically by turnover rates observed in soils.The active pool represents microbial bio-mass and labile carbon in the soil that has aturnover time of days to years.The slow poolconsists of more recalcitrant materials, witha turnover time of years to decades. The pas-sive pool is humified carbon that is stabilizedon mineral surfaces. It has turnover times ofhundreds to thousands of years. The detailedrepresentation of soil pools in CENTURYenables it to estimate changes in decomposi-tion under situations in which a change indisturbance regime or climate alters the de-

composition of some soil pools more thanother pools. A change in climate, for ex-ample, primarily affects the active and slowpools, with the passive pool remaining pro-tected by clay minerals; tillage, however,enhances the decomposition of all soil pools.

The litter layer is much better developedin forests than in grasslands, so much of theforest decomposition occurs in the forestfloor above mineral soil. Soil texture there-fore has less influence on decomposition inforests than in grasslands. The LINKAGESmodel follows the decomposition of eachlitter cohort (i.e., each year’s litterfall) sepa-rately for 8 years, based on the temperature,moisture, and lignin :N ratio of that littercohort (Pastor and Post 1986) (Fig. 14.12).After 8 years, the remaining organic matteris transferred to an SOM pool, which decom-poses as a function of the size of the pool,temperature, and moisture, as in the otherecosystem carbon models.

Plant litter

SOM

TEM

Plant litter

Litter 1

Litter 2

Litter 8

LINKAGES

SOM

CO2

f (C:N, T, M)

f (L:N, T, M)

f (L:N, T, M)

f (L:N, T, M)

CO2

CO2

CO2

Plant litter

Active soil C

Slow soil C

CENTURY

Passive soil C

(3 yr)Metabolic CStructural C

(0.5 yr)

(25 yr)

(1000 yr)

(1.5 yr)

f (L:N)

CO2

CO2

CO2

CO2

CO2

f (L)0.45

0.55

f (T, M,Tex)

f (T,M)

f (T,M)


cipitation that are used as inputs to an ecosys-tem model (Running et al. 1989). Estimatesfrom ecosystem models are sensitive to thequality of the input data (which will always beinadequate at large spatial scales) and to thedegree of generality of the relationships simu-lated by the models. Rapid improvements in thetechnology and availability of remotely senseddata from satellites are providing new sourcesof spatially explicit data on ecosystem variablessuch as leaf area index (LAI), soil moisture,and surface temperature. Field research canthen relate these remote-sensing signatures to properties and processes that are important in ecosystems. The relationship of ecosystemcarbon exchange to light, temperature, and LAIcan be determined in field studies and incor-porated into the model structure (Clein et al.,in press).The generality of relationships used inecosystem models can then be tested throughcomparisons of model output with field dataand through intercomparisons of models thatdiffer in their structure but use the same inputdata (VEMAP Members 1995, Cramer et al. 2000).

Any extrapolation exercise requires consid-eration of biogeochemical hot spots with high

How do we know whether the patterns of NEP estimated by global-scale models are realistic? A comparison of model re-sults with field data for the few locationswhere NEP has been measured provides onereality check. At these sites, measurementsof NEP over several years spanning a rangeof weather conditions provides a measure of how that ecosystem responds to variationin climate. This allows a test of the model’sability to capture the effects of ecosystemstructure and climate on NEP.

The seasonal and interannual patterns ofatmospheric CO2 provide a second realitycheck for global models of NEP. The tempo-ral and spatial patterns of atmospheric CO2

are the direct consequence of net ecosys-tem exchange by the terrestrial biosphere(including human activities) and the oceans.Atmospheric transport models describe the

patterns of redistribution of water, energy,and CO2 through Earth’s atmosphere. Thesetransport models can be run in inverse modeto estimate the spatial and temporal patternsof CO2 uptake and release from the land andoceans that are required to produce theobserved patterns of CO2 concentration inthe atmosphere (Fung et al. 1987, Tans et al.1990). The global patterns of CO2 sourcesand sinks estimated from the atmospherictransport models can then be compared with the patterns estimated from ecosys-tem models. Any large discrepancy betweenthese two modeling approaches providehints about processes and/or locations whereeither the ecosystem or the atmospherictransport models may have not adequatelycaptured the important controls over carbonexchange and transport.

process rates. Regional extrapolation ofmethane flux at high latitudes, for example,should consider beaver ponds (Roulet et al.1997) and thermokarst lakes (Zimov et al.1997) because they have high fluxes relative totheir area. Estimates of NEP, however, requiredifferentiation between young and old forests,because forest age is the major determinant ofNEP (see Chapter 13).

Extrapolation requires a careful considera-tion of characteristic length scales of key pro-cesses (Fig. 14.13). General circulation modelsthat simulate climate at the global scale, forexample, often have a spatial resolution of 2°by 2° of longitude and latitude, which is suffi-cient for incorporating the differential heatingby land and ocean but inadequately representsfine-scale processes, such as cloud formation.Successional models of ecosystems character-ized by gap dynamics might simulate processesat the scale of individual trees, whereas eco-systems characterized by stand-replacing firesmight be adequately represented by patchesseveral hectares or square kilometers in size.Results from models that simulate processes at large spatial and temporal scales often focuson slow variables that strongly influence long-


term processes in ecosystems (Carpenter andTurner 2000). The availability of data and thecomputing power required for model simula-tions provide additional pragmatic constraintsto the spatial resolution of regional models.Satellite data are readily available at 1-km resolution for the globe, but data with finer res-olution are more expensive and less continu-ously available.

Some processes can be extrapolated to largescales without explicitly considering landscapeinteractions. The extrapolation of carbon flux,for example, may be adequately represented inthe short term from an understanding of itsresponse to climate, vegetation, and stand age.Over the long term, erosion, leaching, and firetransfer significant amounts of carbon amongpatches. Other processes, such as vegetationchange in response to changing climate, are sensitive to rates of species migration and dis-turbance spread. Spatially explicit models thatincorporate the spread of disturbance amongpatches on a landscape are critical for projec-

tions of long-term changes in vegetation anddisturbance regime (Gardner et al. 1987, Ruppet al. 2000).

Development of new measurement tech-niques that quantify processes at coarse spatialscales provides an additional basis for scaling.Measurements of carbon and water exchangethat were traditionally measured on individualleaves or patches of soil can now be measuredby eddy correlation techniques over entireecosystems, using either micrometeorologicaltowers or aircraft. These approaches measureecosystem fluxes at scales of tens of meters tokilometers and therefore integrate fluxes acrossboth hot and cold spots in the landscape.

A global monitoring network that measuresatmospheric CO2 and CH4 concentrations pro-vides an additional scaling tool. Atmospherictransport models can be run in inverse mode toestimate the regional and global patterns ofCO2 or CH4 fluxes. Inverse modeling asks whatseasonal and temporal patterns of fluxes wouldbe required to produce the observed patterns

Figure 14.13. Temporal and spatial scales at which selected ecosystem processes occur. The studyof any ecosystem process requires understanding at least one level below (to provide mechanistic

understanding) and one level above (to providecontext with respect to patterns of temporal andspatial variability).

Herbivory

Succession

Treereplacement

Migration

∆ Thermohalinecirculation

Allocation

Resourceuptake

Metaboliteturnover

(cent)

(yr)

(mo)

(d)

(hr)

(min)

Length (m2) [log scale]

Tim

e (y

r) [l

og s

cale

]

(mm)(µm) (m) (km)

106

104

102

10-2

10-4

10-6

10-6

10-4

10-2

102

104

106

108

1

1


of CO2 concentration in Earth’s atmosphere.Each of the currently available approaches toscaling has limitations, so the development of new approaches and comparisons of resultsfrom different approaches are active areas ofresearch on global change.

Summary

Spatial heterogeneity within and among ecosys-tems is critical to the functioning of individualecosystems and of entire regions. Landscapesare mosaics of patches that differ in ecologi-cally important properties. Some patches, forexample, are biogeochemical hot spots that aremuch more important than their area wouldsuggest. The size, shape, connectivity, and con-figuration of patches on a landscape influencetheir interactions. Large patches, for example,may contain greater heterogeneity of resourcesand environment and have a smaller propor-tion of edge habitat. The shape and connectiv-ity of patches influences their effective size andheterogeneity in ways that differ among organ-isms and processes. The distribution of patcheson a landscape is important because it deter-mines the nature of transfers of materials and disturbance among adjacent patches. Theboundaries between patches have unique prop-erties that are important to edge specialists.Boundaries also have physical and biotic prop-erties that differ from the centers of patches,so differences among patches in edge-to-arearatios, due to patch size and shape, influence theaverage rates of processes in a patch.

State factors, such as topography and parentmaterial, govern the underlying matrix of spa-tial variability in landscapes. This physicallydetermined pattern of variability is modified by biotic processes and legacies in situationswhere species strongly affect their environ-ment. These landscape patterns and processesin turn influence disturbance regime, whichfurther modifies the landscape pattern. Hu-mans are exerting increasing control over land-scape patterns and change. Land use decisionsthat convert one land-surface type to anther(e.g., deforestation, reforestation, shifting agri-culture) or that modify its functioning (e.g.,

cattle grazing on rangelands) influence both thesites where those activities are being carried outand the functioning of neighboring ecosystemsand the landscape as a whole. Human effects onecosystems are becoming both more extensive(i.e., affecting more area) and more intensive(i.e., having greater impact per unit area).

Ecosystems do not exist as isolated units onthe landscape. They interact through the move-ment of water, air, materials, organisms, and disturbance from one patch to another. Topo-graphically controlled movement of water andmaterials to downslope patches depends on thearrangement of patches on the landscape andthe properties of those patches. Riparian areas,for example, are critical filters that reduce thetransfer of nutrients and sediments fromupland ecosystems to streams, lakes, estuaries,and oceans.Aerial transport of nutrients, water,and heat strongly influences the nutrient inputsand climate of downwind ecosystems. Theseaerial transfers among ecosystems are now solarge and pervasive as to have strong effects on the functioning of the entire biosphere.Animals transport nutrients and plants at amore local scale and influence patterns of col-onization and ecosystem change. The spread ofdisturbance among patches influences both thetemporal dynamics and the average propertiesof patches on a landscape. The connectivity ofecosystems on the landscape is rarely incorpo-rated into management and planning activities.The increasing human impacts on landscapeinteractions must be considered in any long-term planning for the sustainability of managedand natural ecosystems.

Review Questions

1. What is a landscape? What properties ofpatches determine their interactions in alandscape?

2. How do fragmentation and connectivityinfluence the functioning of a landscape?

3. Give examples of spatial heterogeneity inecosystem structure at scales of 1m, 10m, 1km, 100km, and 1000km. How does spatialheterogeneity at each of these scales affectthe way in which these ecosystems func-


tion? In other words, if heterogeneity ateach scale disappeared, what would be the differences in the way in which theseecosystems function?

4. What are the major natural and anthro-pogenic sources of spatial heterogeneity ina landscape? How do these sources of het-erogeneity influence the way in which theselandscapes function? How do interactionsamong these sources of heterogeneityaffect landscape dynamics?

5. What is the difference between a shiftingsteady state mosaic and a non–steady statemosaic? Give examples of each.

6. What is the difference between intensifica-tion and extensification? What has been the role of each in ecosystem and globalprocesses?

7. Which ecosystem processes are moststrongly affected by landscape pattern?Why?

8. What properties of boundaries influencethe types of interactions that occur betweenpatches within a landscape?

9. Describe how patches within a landscapeinteract through (1) the flow of water, (2)transfers of materials through the atmos-phere, (3) movement of animals, and (4) themovement of disturbance. What propertiesof landscapes and patches influence the relative importance of these mechanisms of patch interaction?

10. What issues must be considered in extrap-olating processes measured at one scale tolarger areas? How does the occurrence ofhot spots influence approaches to spatialscaling?

Additional Reading

Forman, R.T.T. 1995. Land Mosaics: The Ecology ofLandscapes and Regions. Cambridge UniversityPress, Cambridge, UK.

Foster, D.R., and G. Motzkin. 1998. Ecology and conservation in the cultural landscape of NewEngland: Regional forest dynamics in central NewEngland. Ecosystems 1:96–119.

Holling, C.S. 1992. Cross-scale morphology, geome-try, and dynamics of ecosystems. Ecological Mono-graphs 62:447–502.

Matson, P.A.,W.J. Parton,A.G. Power, and M.J. Swift.1997. Agricultural intensification and ecosystemproperties. Science 227:504–509.

Meyer, W.B., and B.L. Turner III. 1992. Human pop-ulation growth and global land-use/cover change.Annual Review of Ecology and Systematics23:39–61.

Naiman, R.J., and H. Decamps. 1997. The ecology ofinterfaces: Riparian zones. Annual Review ofEcology and Systematics 28:621–658.

O’Neill, R.V., D.L. DeAngelis, J.B. Waide, and T.F.H. Allen. 1986. A Hierarchical Concept ofEcosystems. Princeton University Press, Princeton,NJ.

Pickett, S.T.A., and M.L. Cadenasso. 1995. Landscapeecology: Spatial heterogeneity in ecologicalsystems. Science 269:331–334.

Turner, M.G. 1989. Landscape ecology: The effect ofpattern on process. Annual Review of Ecology andSystematics 20:171–198.

Turner, M.G., R.H. Gardner, and R.V. O’Neill. 2001.Landscape Ecology in Theory and Practice: Patternand Process. Springer-Verlag, New York.

Urban, D.L., R.V. O’Neill, and H.H. Shugart. 1987.Landscape ecology. BioScience 37:119–127.


IntroductionA thorough understanding of the cycling ofwater, carbon, nitrogen, phosphorus, and sulfuris key to understanding ecosystems, the bios-phere, and the entire Earth System. Humanactivities have dramatically altered elementcycles since the beginning of the Industrial Revolution. Burning of fossil fuels in particularhas increased emissions of CO2, nitric oxides,and several sulfur gases. Mining and agriculturehave also altered the availability and mobilityof carbon, nitrogen, phosphorus, and sulfur.Changes in these biogeochemical cycles havealtered Earth’s climate, speeding up the globalhydrologic cycle, which in turn feeds back toother biogeochemical cycles. These changesaffect ecosystems at all scales, ranging fromindividual organisms to the entire biosphere. Inthis chapter, we focus on the global cycles ofcarbon, nitrogen, phosphorus, sulfur, and water,summarizing at the global scale the naturalpools and fluxes in the cycles and the factorsresponsible for change.

The Global Carbon Cycle Photosynthetic uptake of carbon from theatmosphere and oceans provides the fuel formost biotic processes. This reduced carbon

makes up about half of the mass of Earth’sorganic matter. Biological systems, in turn,respire CO2 when they use organic carbon forgrowth and metabolism. The controls over thecarbon cycle depend on time scale, rangingfrom millions of years, by which cycling is controlled by movements of Earth’s crust, toseconds, by which cycling is controlled by pho-tosynthetic rate and surface–air exchange (seeChapters 5 to 7).

Carbon is distributed among four majorpools: the atmosphere, oceans, land (soils andvegetation), and sediments and rocks (Fig. 15.1)(Reeburgh 1997, McCarthy et al. 2001). Atmos-pheric carbon, which consists primarily of CO2,is the smallest but most dynamic of these pools.It turns over—that is, is completely replen-ished—every 3 to 4 years, primarily through itsremoval by photosynthesis and return by re-spiration.The metabolic processes of organismstherefore constitute the engine that drives theglobal carbon cycle on time scales of seconds tocenturies.

Carbon is present in the oceans as dissolvedorganic carbon (DOC), dissolved inorganiccarbon (DIC), and particulate organic carbon(POC), which consists of both live organismsand dead materials. Most (98%) of this carbonis in inorganic form, primarily as bicarbonate(90%), with most of the rest as carbonate. Free

15Global Biogeochemical Cycles

The magnitude of biotic and human impacts on ecosystem processes becomes clearwhen summed at the global scale. This chapter describes the global cycles of waterand several biotically important elements.

335

336 15. Global Biogeochemical Cycles

CO2, the form that is directly used by mostmarine primary producers, accounts for lessthan 1% of this inorganic pool. These threeforms of DIC are in a pH-dependent equilib-rium (see Chapter 10). The marine biotaaccount for only 2Pg (2 ¥ 1015 g) carbon,although they cycle as much carbon annually as does terrestrial vegetation. The carbon inmarine biota turns over every 2 to 3 weeks.

The ocean’s surface waters that interact withthe atmosphere contain about 1000Pg carbon,similar to the quantity in the atmosphere (Fig.15.1). The capacity of the ocean to take upcarbon is constrained by three categories ofprocesses that operate at different time scales(Schlesinger 1997). In the short term, thesurface exchange rate depends on wind speed,surface temperature, and the CO2 concentra-

Ocean

mixin

gMarine biota

2

Biol

ogica

l

pu

mp

Intermediate and deepocean waters 38,000

Surface sediments150

Ocea

n flu

x

Rocks60 x 106

Cem

ent 0

.1Fo

ssil

fuel

s 5.

3

Atmosphere 760

Fire

s 4

Land

use

cha

nge

NPP

60

90 1.7

1.9

The GlobalCarbon Cycle

4233

11

45

3460

Het

erot

roph

ic re

spira

tion

55

0.2

0.8

88

Surface oceanwater 1,000

Vegetation650

Soils1,500

Figure 15.1. The global carbon cycle showingapproximate magnitudes of the major pools (boxes)and fluxes (arrows) in units of pedagrams per year(1 pedagram = 1015 g). The carbon pools that con-tribute to carbon cycling over decades to centuriesare the atmosphere, land (vegetation and soils), andoceans. On land, the carbon gain by vegetation isslightly greater than the carbon loss in respiration,leading to net carbon storage on land.The net carboninput to the oceans is also slightly greater than thenet carbon return to the atmosphere. The terrestrialbiosphere accounts for 50 to 60% of global netprimary production (NPP). Most (80%) of themarine NPP is released to the environment by het-

erotrophic respiration, and the remaining 20% goesto the deep oceans by the biological pump. Oceanupwelling returns most of this carbon to the surfaceocean waters. Human activities cause a net carbonflux to the atmosphere through combustion of fossilfuels, cement production, and land use change. Thecarbon content of the atmosphere was calculatedfrom the 1999 CO2 concentration of 367ppmv (Prentice et al. 2001). Data for pools of carbon invegetation are from Saugier et al. (2001); for thesurface ocean from Houghton et al. (1996); and forthe deep ocean, sediments, and rocks from Reeburgh1997). Remaining data are from Prentice et al.(2001).

The Global Carbon Cycle 337

tion of surface waters. On daily to monthly timescales, the CO2 concentration in the surfacewater depends on photosynthesis and pH-dependent buffering reactions. Finally, thesurface waters are a relatively small pool (only75 to 200m deep) of water that exchanges relatively slowly with deeper layers because the warm, low-salinity surface water is lessdense than deeper layers (see Fig. 2.8). Carbonthat enters surface waters is transported slowlyto depth by two major mechanisms. Organicdetritus and its calcium carbonate (CaCO3)skeletal content, which form in the euphoticzone, sink to deeper waters, a process termedthe biological pump (see Chapter 10). Bottom-water formation in the polar seas transports dis-solved carbon to depth, a process termed thesolubility pump (see Chapter 2). Once carbonreaches intermediate and deep waters, it isstored for hundreds to thousands of yearsbefore returning to the surface throughupwelling. Most (97%) of the ocean carbon isin the intermediate and deep waters (Fig. 15.1).

The terrestrial biosphere contains the largestbiological reservoir of carbon. There is nearlyas much carbon in terrestrial vegetation as inthe atmosphere, and there is at least twice asmuch carbon in soils as in the atmosphere (Fig.15.1) (Jobbágy and Jackson 2000). Terrestrialnet primary production (NPP) is slightlygreater than that in the ocean, but due to themuch larger plant biomass on land, terrestrialplant carbon has a turnover time of about 11years, compared to 2 to 3 weeks in the ocean.NPP is about half of gross primary production(GPP; i.e., photosynthetic carbon gain) on land(60Pgyr-1 out of 120Pgyr-1) and in the ocean(45Pgyr-1 out of 103Pgyr-1) (Prentice et al.2001). Soil carbon turns over on average every25 years. These average turnover times masklarge differences in turnover time among com-ponents of the terrestrial carbon cycle. Photo-synthetically fixed carbon in chloroplasts turnsover on time scales of seconds through photo-respiration (see Chapter 5). Leaves and rootsare replaced over weeks to years, and wood isreplaced over decades to centuries. Compo-nents of soil organic matter (SOM) also havequite different turnover times, with labile formsturning over in minutes and humus having

turnover times of decades to thousands of years(see Chapter 7).

Carbon in rocks and surface sedimentsaccounts for well over 99% of Earth’s carbon(107 Pg) (Reeburgh 1997, Schlesinger 1997).This carbon pool cycles extremely slowly, withturnover times of millions of years. Factors governing the turnover of these pools are geologic processes associated with the rockcycle, including the movement of continentalplates, volcanism, uplift, and weathering (seeChapter 3).

Human activities make a significant contri-bution to the global carbon cycle. Combustionof fossil fuels releases CO2 from petroleum pro-ducts. Cement production releases CO2 fromcarbonate rocks. Land use conversion releasescarbon by biomass burning and enhanceddecomposition. Together these fluxes are about15% of the carbon cycled by terrestrial or bymarine production, making human activitiesthe third largest biotically controlled flux ofcarbon to the atmosphere.

Long-Term Change in Atmospheric CO2

Critical processes in the carbon cycle occur onall time scales. The important processes arephotosynthesis and respiration on time scalesof seconds to years; NPP, SOM turnover, anddisturbance on time scales of decades to cen-turies; and uplift, volcanism, weathering, andocean sedimentation over thousands to millionsof years. Atmospheric CO2 concentration haschanged dramatically through Earth’s history,with concentrations 10-fold higher than today(greater than 3000ppmv) likely to haveoccurred several times in the last hundred mil-lions years. CO2 concentrations have also beenbelow 300ppmv, first about 20 million yearsago, in part due to reduced volcanism (Paganiet al. 1999, Pearson and Palmer 2000) and mostrecently just before the beginning of the Industrial Revolution 150 years ago (Webb andBartlein 1992). Geochemical processes deter-mine variation in CO2 on geological time scales.These include the weathering of silicate rocks(which consumes CO2 and releases bicar-


bonate), burial of organic carbon in sediments,and volcanism (which release CO2) (Berner1997). Biological processes influence geochem-ical cycling in many ways, for example byincreasing weathering rates (see Chapter 3).Although critical on long time scales, the ratesof these geochemical processes are so slowcompared to anthropogenic changes that theydo not influence current trajectories of changein atmospheric CO2.

Over the last 400,000 years, changes in solarinput associated with variations in Earth’s orbit(see Fig. 2.15) have caused cyclic variation inatmospheric CO2 concentrations associatedwith glacial–interglacial cycles (Fig. 15.2) (Petitet al. 1999, Sigman and Boyle 2000). CO2 con-centration declined during glacial periods andincreased during interglacials. These changes inCO2 concentration are much larger than can beexplained simply by changes in light intensityand temperature in response to altered solarinput. The large biospheric changes must resultfrom amplification by biogeochemical feed-backs in the Earth System. Several feedbacks

could contribute to these atmospheric changes(Sigman and Boyle 2000). (1) Increased trans-port of dust off the less-vegetated continentsduring glacial periods may have increased iron,phosphorus, and silica transport and enhancedNPP in high-latitude oceans, leading toincreased CO2 uptake and transport to depthvia the biological pump (see Chapter 10). (2)Extensive winter sea ice around Antarctica mayhave reduced out-gassing of CO2 in locations ofupwelling of CO2-rich deep waters (Stephensand Keeling 2000). (3) Some additional carbonmay have been stored on land during glacialperiods on continental shelves exposed by thedrop in sea level or in permafrost at high lati-tudes (Zimov et al. 1999). However, terrestrialsystems were probably a net carbon sourceduring glacial periods, due to the replacementof forests by grasslands, deserts, tundra, and icesheets. Marine foraminiferan fossils acquired a more terrestrial 13C signature during glacialperiods; this could reflect the movement of ter-restrial carbon from the land to the atmosphereand subsequently to the ocean (Bird et al. 1994,

400 200 100 0

Age (thousands of years ago)

CO

2co

ncen

trat

ion

(ppm

v)

300

Tem

pera

ture

rel

ativ

e to

pre

sent

clim

ate

(o C)

275

325

350

250

225

200

175

-8

4

0

-4

Temperature

CO2

Figure 15.2. Variations in temperature and atmos-pheric CO2 concentrations derived from air trappedin Antarctic ice cores (Petit et al. 1999). (Adapted

from IPCC Assessment Report 2001; Folland et al.2001.)


Crowley 1995). An improved understanding ofcontrols over these potential feedback mecha-nisms is important because it could indicatehow the Earth System will respond to currenttrends of increasing temperature and atmos-pheric CO2.

Atmospheric CO2 concentration has beenrelatively stable over the last 12,000 years,ranging from about 260 to 280ppmv in pre-industrial times. Several high-resolution ice-core records demonstrate that human-inducedincreases in CO2 over the past century are anorder of magnitude more rapid than any thatoccurred over the previous 20,000 years andprobably the last 400,000 years (Petit et al.1999). This sudden change in the global cyclesof carbon and other elements induced byhuman activities are so large as to suggest thatEarth has entered a new geologic epoch, theAnthropocene (see Fig. 2.12).

Anthropogenic Changes in theCarbon Cycle

The burning of fossil fuels is the primary causeof current increases in atmospheric CO2. About5.3Pgyr-1 of carbon are emitted from fossil fuel combustion and 0.1Pgyr-1 from cementproduction (Prentice et al. 2001). We are lesscertain about the estimated 1.7 ± 0.8Pgyr-1 ofcarbon emitted due to deforestation, becausethe aerial extent and carbon dynamics associ-ated with this land use change are less well documented. CO2 flux from land use changemust include both the net sources of carbon tothe atmosphere (e.g., deforestation and agricul-tural conversion) and the net sinks (e.g., forestregrowth, forest plantations, fire suppression,woody encroachment, and changes in agricul-tural management) (Houghton et al. 1999).

Anthropogenic emissions have causedatmospheric CO2 concentration to increaseexponentially since the beginning of the Indus-trial Revolution (Fig. 15.3). A comparison ofthe annual increment in carbon content of theatmosphere with known emissions shows thatonly about half the anthropogenic carbon thatis emitted to the atmosphere remains there.Theremainder is taken up on land or in the oceans.Measurements of d13C in atmospheric CO2 and

measurements of atmospheric O2 (Keeling etal. 1993, Bender et al. 1996) help identify moreprecisely the location of these missing sinks ofCO2 (Box 15.1).

Estimates of carbon sources and sinks canalso be developed using biomass inventories(Dixon et al. 1994), inverse atmospheric modeling (Tans et al. 1990), and carbon cyclemodeling. Inventory studies measure changesin regional or global carbon pools over time by repeatedly measuring the same forest plots.These records suggest that 0.5Pgyr-1 of carbonis accumulating in northern forests due toincreasing rates of tree growth (FCCC 2000).Data from the tropics are less certain butatmospheric sampling suggests that tropicalregions are approximately in balance with theatmosphere, so there must be a net sink in un-managed forests that approximately balancesthe carbon losses from deforestation andbiomass burning (Schimel et al. 2001).The largequantity and spatial variability of soil carbonpools makes it difficult to use an inventoryapproach to detect changes in soil carbon pools.Even the quantity of carbon in terrestrial veg-etation is uncertain; two recent global synthe-ses differed by 25% (500Pg vs. 650Pg) in theirestimate of the size of this pool (Prentice et al.2001, Saugier et al. 2001).

Inverse modeling can be used to estimate theglobal pattern of net sources and sinks of CO2

that are required to produce the observed lati-tudinal and seasonal patterns of atmosphericCO2 concentration. Atmospheric transportmodels and global networks of observations ofCO2 concentrations suggest that the gradient inCO2 concentration from the Northern Hemi-sphere to the Southern Hemisphere is not aslarge as might be expected from the geograph-ical distribution of anthropogenic CO2 emis-sion, which is largely a Northern Hemispheresource. These models therefore support theexistence of a Northern Hemisphere sink forCO2. Latitudinal gradients in the 13C and O2

suggest that about half of the sink is on land;about 1.9Pgyr-1 of anthropogenic carbon isabsorbed by both land and ocean (Prentice et al. 2001). Much of the terrestrial sink appearsto be in north temperate forests; NorthAmerica and Eurasia contribute about equally


to this sink per unit area of land, despite sub-stantial differences in climate and humanimpact (Schimel et al. 2001). Boreal forests may now be a source of carbon, in part due toincreased wildfire. Tropical forests appear to bein approximate balance with the atmosphere,due to similar magnitudes of net carbon uptakefrom unmanaged forests and carbon loss fromdeforestation.

Terrestrial Sinks for CO2

Four potentially important mechanisms con-tribute to the Northern Hemisphere terrestrialsink for CO2: land use change, CO2 fertilization,inadvertent nitrogen fertilization, and climateeffects (Schimel 1995). The conversion offorests to agricultural lands dominated land use change in the middle and high latitudesuntil the mid-twentieth century. Today, forest

regrowth in previously harvested areas or inabandoned agricultural lands has enhancedcarbon storage. The widespread suppression of wildfire also enhances the mid-latitudecarbon sink, because it reduces fire emissionsand allows woody plants to encroach into grasslands (Houghton et al. 2000). These areprobably the most important causes of the north-temperate terrestrial carbon sink(Schimel et al. 2001).

CO2 fertilization contributes to carbonstorage. Photosynthesis typically increases 20 to40% under doubled CO2 in short-term studies.Carbon storage by ecosystems is, however,much less responsive to CO2 than is the short-term response of individual plants becauseplant growth becomes nutrient limited as nutri-ents become sequestered in live and deadorganic matter (Shaver et al. 1992, Schimel1995) (see Chapter 5). Nutrient availability

2000

1750

1500

1250

1000

750

5001000 1200 1400 1600 1800 2000

CH

4 co

ncen

trat

ion

(ppb

v)

1000 1200 1400 1600 1800 2000

320

310

300

290

280

270

260

250

N2O

con

cent

ratio

n (p

pbv)

220

260

300

340

380

CO

2co

ncen

trat

ion

(ppm

v)

1000 1200 1400 1600 1800 2000

B

C

A

Time

Time

Figure 15.3. Changes since 1750 in the atmosphericconcentrations of three radiatively active gases thatare influenced by human activities (Prather et al.2001, Prentice et al. 2001). Data shown are a com-

posite of time series from air trapped in Antarctic icecores and from direct atmospheric measurements.(Adapted from IPCC Assessement Report 2001;Prentice et al. 2001.)


therefore limits the long-term rate of terrestrialcarbon sequestration, and the overall effect ofCO2 fertilization on terrestrial carbon storageappears to be much smaller than that due toreforestation.

Nitrogen typically limits the growth of plantsin nontropical terrestrial ecosystems (seeChapter 8). Nitrogen additions through fertil-ization or atmospheric deposition of air pollu-tants like NOx and nitric acid from fossil-fuelburning might therefore augment carbonstorage, although this effect may be relativelysmall (Nadelhoffer et al. 1999). Increases incarbon storage due to nitrogen fertilizationmay be limited in scope because nitrogen depo-sition causes forest degradation above somethreshold (Schulze 1989, Aber et al. 1998).Further nitrogen deposition might causecarbon loss from forests. The net effect ofanthropogenic nitrogen deposition on carboncycling and sinks is regionally variable andhighly uncertain (Holland et al. 1997).

Finally, climate changes (including changes in temperature, moisture, and radiation) affectcarbon storage through their effects on carboninputs (photosynthesis) and outputs (respira-tion). In the short term, warming is expected to reduce carbon storage in soils by increasingheterotrophic respiration. The associatedincrease in rates of mineralization of nitrogenand other nutrients from organic matter could,however, increase nutrient uptake and produc-tion by vegetation (Shaver et al. 2000). Vegeta-tion generally has a much higher C :N ratio (160 :1) than does soil organic matter (15 :1)(Schlesinger 1997), so the transfer of a givenquantity of nitrogen from the soil to plantsenhances carbon storage (Vukicevic et al.2001). In addition, ecosystem respiration accli-matizes to temperature, so it increases less inresponse to warming than might be expectedfrom short-term measurements (Luo et al.2001). The response of Alaskan arctic tundra toregional warming is consistent with these ideas.

Box 15.1. Partitioning of Carbon Uptake Between the Land and the Oceans

Only about half of the anthropogenic CO2

that enters the atmosphere remains there.The land or oceans take up the remainder.Changes in the oxygen content of the atmos-phere provide a measure of relative im-portance of land and ocean uptake. Netterrestrial uptake of CO2 is accompanied bya net release of oxygen, with a 1 :1 ratio of moles of CO2 absorbed to moles of O2

released. When CO2 dissolves in oceanwater, however, this causes no net release of oxygen. This difference in exchangeprocesses can be used to partition the totalCO2 uptake between terrestrial and oceancomponents (Keeling et al. 1996b).

The relative abundance of the two stableisotopes of carbon (13C and 12C) in theatmosphere provides a measure of the rela-tive activity of the terrestrial and oceaniccomponents of the global carbon cycle (Ciaiset al. 1995). Fractionation during photosyn-thesis by C3 plants discriminates against 13C,

causing biospheric carbon to be depleted in13C by about 18‰ relative to the atmosphere.Exchanges with the ocean, however, involverelatively small fractionation effects.Changes in the 13C : 12C ratio of atmosphericCO2 therefore indicate the relative magni-tude of terrestrial and oceanic CO2 uptake.

Measurement of the global pattern andtemporal changes in oxygen concentrationand the 13C : 12C ratio of atmospheric CO2

suggest that the land and ocean contributeabout equally to the removal of anthro-pogenic CO2 from the atmosphere (Prenticeet al. 2001). There are, however, manyassumptions and complications in usingeither of these approaches to estimate therelative magnitudes of terrestrial andoceanic carbon uptake. The advantage ofatmospheric measurements is that they give an integrated estimate of all uptakeprocesses on Earth, because of the relativelyrapid rate at which the atmosphere mixes.


These ecosystems initially responded towarming with increased carbon loss but subse-quently returned to near-zero carbon balancewith the atmosphere (Oechel et al. 2000).Increases in precipitation can also cause vari-able responses in ecosystem carbon balance.Improved soil moisture can increase plantgrowth, but in some ecosystems this carbonstorage term may be offset by increased decom-position. Although climate effects can be sub-stantial, their net effect on terrestrial carbonstorage is uncertain.

Terrestrial carbon exchange with the atmos-phere varies substantially among years, due to interannual variation in climate. Terrestrialecosystems tend to be a net carbon source in warm years and a net carbon sink in coolyears, because fires and respiration increasemore strongly with temperature than does NPP(Vukicevic et al. 2001, Schimel et al. 2001).Some of this interannual variation may reflectEl Niño southern oscillation (ENSO) events,perhaps explaining why the terrestrial bios-phere was a particularly strong sink for carbonin the early 1990s.

The relative importance of the various mech-anisms of enhanced carbon storage in the terrestrial biosphere is uncertain (Schimel1995, Schimel et al. 2001). Nonetheless, the multiple and potentially interacting sinksappear sufficient to account for the observedmovement of anthropogenic CO2 from theatmosphere to land (Table 15.1). These sinkmechanisms (forest regrowth, CO2 fertilization,and land use change) are likely to saturate andbecome less effective in the future (Schimel et al. 2001). The most effective mechanism ofstabilizing atmospheric CO2 concentration istherefore to reduce emissions.

Even if anthropogenic carbon emissions tothe atmosphere were stopped immediately,the elevated atmospheric CO2 concentrationcaused by past emissions will persist fordecades to centuries. The lifetime of anthro-pogenic effects on atmospheric CO2 dependsprimarily on the turnover time of terrestrial and oceanic carbon pools that interact with theatmosphere (Braswell and Moore 1994). Manyof these key pools turn over slowly. Soil, forexample, has an average turnover time of 25

years (Fig. 15.1), with some soil pools turningover even more slowly (see Chapter 7). Thereequilibration time of the atmosphere isusually 1.5 to 3 times longer than the turnovertime, explaining the persistence of anthro-pogenic effects on the atmosphere. Delays inefforts to reduce fossil-fuel emissions maytherefore have unexpectedly long-lastingeffects.

The Global Methane Budget

Human activities are responsible for increasingmethane concentrations in the atmosphere.Although the methane (CH4) concentration ofthe atmosphere (1.8ppmv) is much less thanthat of CO2 (370ppmv), CH4 is about 20 timesmore effective per molecule as a greenhousegas than is CO2. Like CO2, the CH4 concentra-tion of the atmosphere has increased exponen-tially since the beginning of the IndustrialRevolution (Fig. 15.3). The CH4 increaseaccounts for 20% of the increased greenhousewarming potential of the atmosphere (Ciceroneand Oremland 1988, Khalil and Rasmussen1990). Documenting the major global sourcesand sinks of atmospheric CH4 is therefore

Table 15.1. Average (1980–1989) annual emissionsand fate of anthropogenic carbon.a

Sources and sinks of Annual net fluxanthropogenic carbon (PgCyr-1)

Anthropogenic carbon sources 7.1 ± 1.1Fossil fuel and cement production 5.5 ± 0.5Net emissions from tropical 1.6 ± 1.0

land use changeCarbon sinks 7.1

Storage in the atmosphere 3.2 ± 0.2Oceanic uptake 1.6 ± 1.0Terrestrial uptake 2.1

CO2 fertilization 1.0 ± 0.5Forest regrowth in the 0.5 ± 0.5

Northern HemisphereNitrogen deposition 0.6 ± 0.3

Other 0.2 ± 2.0

a About half of the anthropogenic carbon emissions remainin the atmosphere. The oceans and terrestrial biosphereabsorb the remaining anthropogenic emissions.Data from Schimel (1995).

The Global Nitrogen Cycle 343

important for understanding the recentincreases in global temperature and the poten-tial for future climate warming.

Methane is produced only under anaerobicconditions (see Chapter 7). Wetlands accountfor 70% of the naturally produced CH4, withthe remainder coming primarily from fresh-water sediments, fermentation in the guts ofanimals (e.g., termites and ruminants), andvarious geological sources (Table 15.2).Anthropogenic methane sources are abouttwice as large as the natural sources, whichexplains why CH4 accumulates in the atmos-phere, despite its high reactivity and rapidturnover (9 years). Fossil-fuel extraction andrefining; waste management (landfills, animalwastes, and domestic sewage treatment); andagricultural sources (rice paddies, biomassburning, and fermentation in guts of domesticruminants like cattle) are each an importantCH4 source.

Estimates of the magnitude of most of thenatural and anthropogenic CH4 sources arehighly uncertain (Table 15.2) (McCarthy et al.2001). Important new sources are still being

identified, including high-latitude lakes andreservoirs on organic-rich substrates (Zimov et al. 1997, St. Louis et al. 2000).

CH4 reacts readily with OH radicals in theatmosphere in the presence of sunlight. Thisphotochemical process is the major sink foratmospheric CH4, accounting for 85% of theCH4 consumption (Table 15.2). Additional CH4

mixes into the stratosphere, where it reacts with ozone (see Chapter 2) or is removed by methanotrophs in soils (see Chapter 7). Theannual atmospheric accumulation of CH4 isabout 10% of the annual anthropogenic flux.

The Global Nitrogen Cycle

The productivity of many unmanaged ecosys-tems on both land and sea and of mostmanaged agricultural and forestry ecosystemsis limited by the supply of available nitrogen. Incontrast to carbon, almost all of the nitrogenthat is relevant to biogeochemistry is in a singlepool (the atmosphere) with comparativelysmall quantities in the oceans, rocks, and sedi-ments (Reeburgh 1997) (Fig. 15.4). Organicnitrogen pools are minuscule relative to theatmospheric pool and occur primarily in soilsand terrestrial vegetation. Although nitrogenmakes up 78% of the atmosphere, it is nearlyall N2 and is unavailable to most organisms.Themajor pathway by which N2 is transformed tobiologically available forms is via nitrogen fixation by bacteria in soils and aquatic systemsor living in association with plants. The globalquantity of nitrogen fixed annually by naturalecosystems is quite uncertain, ranging between90 and 190Tgyr-1 for terrestrial ecosystems and between 40 and 200Tgyr-1 for marineecosystems. Lightning strikes probably add anadditional 3 to 10Tgyr-1 of nitrogen to theavailable pool. Before human alteration, theamount of nitrogen coming into the biospherevia nitrogen fixation was approximately bal-anced by return to the unavailable pools viadenitrification and burial in sediments). In con-trast to carbon, nitrogen cycles quite tightlywithin terrestrial ecosystems, with the annualthroughput being about 10-fold greater thaninputs and losses.

Table 15.2. Global sources and sinks of methane.

Methane sources Annual fluxand sinks (TgCH4 yr-1)

Natural sources 160Wetlands 115Termites and ruminants 20Ocean sediments 10Fresh-water sediments 5Geological sources 10

Anthropogenic sourcesa 375Fossil-fuel use 100Waste management 90Fermentation by cattle 85Biomass burning 40Rice paddies 60

Total sources 535Sinks 515

Reaction with OH 445Removal in stratosphere 40Removal by soils 30

Atmospheric increase 30

a Reservoirs are estimated as an additional 70Tgyr-1

anthropogenic source (St. Louis et al. 2000).Data from Schlesinger (1997).


Anthropogenic Changes in theNitrogen Cycle

In the past century human activities haveapproximately doubled the quantity of nitrogencycling between terrestrial ecosystems and theatmosphere. Globally, human activities nowconvert N2 to reactive forms at about the samerate as natural processes, through industrial fix-ation of nitrogen and the planting of nitrogen-fixing crops. The Haber process, which uses

energy from fossil fuels to convert N2 toammonia gas (NH3) to produce fertilizers, fixesmore nitrogen than any other anthropogenicprocess. Industrial fixation of nitrogen by theHaber process increased substantially in the1940s, reaching 30Tgyr-1 by 1970 and 80Tgyr-1

by 2000 (Fig. 15.5); it is projected to increase to 120Tgyr-1 by 2025 (Galloway et al. 1995).Initially, most nitrogen fertilizer was applied in developed nations, but by 2000 almost twothirds of the fertilizer nitrogen was applied in

Ocean

mixin

g

Soils100,000

Atmospheric N2 3,900,000,000

Vegetation4,000

Oce

an N

17

Marine biota300

Surface oceanwater 60,000

pu

mp


8000

8000

Atmospheric N2O, NOx, NH3

1200

1200Biom

ass

burn

ing

N 1

3

Rocks and sediments400,000,000

Soil

biog

enic

N 3

5

Den

itrifi

catio

n 11

0

Anim

al N

34

Foss

il fu

el N

34

Indu

stria

l N fi

xatio

n 80

Biol

ogic

al N

fixa

tion

100

Bio

logi

cal N

fixa

tion

140

Den

itrifi

catio

n <

200

2515

250

850 600

Biolo

gical

10

36

The GlobalNitrogen Cycle

Dep

ositi

on 4

0

Figure 15.4. The global nitrogen cycle showingapproximate magnitudes of the major pools (boxes)and fluxes (arrows) in units of teragrams per year (1teragram = 1012 g). The atmosphere contains the vastmajority of Earth’s nitrogen.The amount of nitrogenthat annually cycles through terrestrial vegetation is9-fold greater than inputs by nitrogen fixation. In theocean the annual cycling of nitrogen through thebiota is 80-fold greater than inputs by nitrogen fixa-tion. Denitrification is the major output of nitrogento the atmosphere. Human activities increase nitro-gen inputs through fertilizer production, planting of

nitrogen-fixing crops, and combustion of fossil fuels.Nitrogen fluxes in the biological pump and oceanmixing were calculated from Fig. 15.6, assuming anN:P ratio of 15. Data for marine nitrogen fixationare from Karl et al. (in press) for nitrogen depositionfrom Holland et al. (1997), for nitrogen in vegetationand soils (calculated from Fig. 15.1, assuming a C :Nratio of 160 for vegetation and 15 for soils) fromSchlesinger (1997), and for nitrogen in marine biotafrom Galloway (1996) and Reeburgh (1997).Remaining data are from Table 15.3 and Schlesinger(1997).

The Global Nitrogen Cycle 345

developing nations, with 40% of this total beingapplied in the tropics and subtropics. Much of the projected increase in fertilizer use isexpected to occur in the less-developed nationsof the world.

Cultivation of nitrogen-fixing crops such assoybeans, alfalfa, and peas adds fixed nitrogenover and above that which is added via biolog-ical fixation in natural ecosystems. Some nitro-gen fixation is also carried out by free-livingand associative nitrogen fixers like Azolla that

commonly occur in rice paddies. Annual fixa-tion rates in crop systems are about 32 to 53Tgyr-1, 20 to 40% of total biotic fixation thatoccurs on land.

Human activities account for most of thenitrogen trace gases transferred from Earth tothe atmosphere (Table 15.3). In addition to the large pool of relatively unreactive N2, theatmosphere contains several nitrogen tracegases, including nitric oxides (NO and NO2),nitrous oxide (N2O), and NH3. Although thepools and fluxes of these nitrogen trace gasesare much smaller than those of N2 (Fig. 15.4),they play a much more active role in atmos-pheric chemistry and have been more stronglyaffected by human activities (see Chapter 9).

N2O, which is increasing at the rate of 0.2 to0.3% yr-1 (Fig. 15.3), is an inert gas that is 200-fold more effective than CO2 as a greenhousegas and contributes about 6% of the green-house warming (Ramaswamy et al. 2001). Nitri-fication and denitrification in the oceans and in tropical soils are the major natural sources of N2O (Schlesinger 1997). Human activitieshave nearly doubled N2O flux from Earth to the atmosphere, primarily through agricultural fertilization (Fig. 15.3). Other anthropogenicN2O sources include cattle and feedlots,biomass burning, and various industrial sources.

150

100

50

1960 1970 1980 1990

Time

Terr

estr

ial N

fixa

tion

(Tg

yr-1

) Total anthropogenicN fixation

Range of estimates ofnatural N fixation N fertilizer

Fossil fuel

Legume crops

Figure 15.5. Anthropogenic fixation of nitrogen interrestrial ecosystems over time compared with therange of estimates of natural biological nitrogen fixation on land. (Redrawn with permission fromGlobol Biogeochemical Cycles; Galloway et al. 1995).

Sources and sinks NOx-Na N2O-Na NH3-Nb Total N

Natural biogenic sources 7.8 10 23 40.8Oceans 0 4c 13 17Soils 2.8d 6 10 18.8Lightning 5 0 0 5

Anthropogenic sources 44.2 8.1 52.2 104.5Cultivated soils 2.8d 4.2 9 16Biomass burning 7.1 0.5 5 12.6Domestic animals 0 2.1 32 34.1Fossil fuels/industrial 33 1.3 2.2 36.5Other 0.7 0 4 4.7

SinksAtmospheric destruction ?? 12.3 1Deposition on land 40e?? 0 57

Annual accumulation 0 3.9 0

a Data from Prather et al. (2001).b Data from Schlesinger (1997).c Data from Karl et al. (in press).d Soil NOx flux uncertain; assumed to be 50% natural, 50% anthropogenic.e Data from Holland et al. (1997).

Table 15.3. Global sources toand sinks from the atmosphereof nitrogen trace gases.


N2O is broken down in the stratosphere, whereit catalyzes the destruction of stratosphericozone.

Human activities have tripled the flux of NH3 from land to the atmosphere (Table 15.3).Domestic animals are now the single largestglobal source of ammonia; agricultural fertil-ization, biomass burning, and human sewageare important additional sources. Cultivatedsoils, which account for only 10% of the ice-freeland area (see Table 6.6), account for about halfof the ammonia flux from soils to the atmos-phere. In summary, activities associated withagriculture (animal husbandry, fertilizer addi-tion, and biomass burning) are the major causefor increased ammonia transport to the atmos-phere and account for 60% of the global flux.Ammonia is a reactant in many atmosphericreactions that form aerosols and generate airpollution. Ammonia is also the main acid-neutralizing agent in the atmosphere, raisingthe pH of rainfall, cloud water, and aerosols.Most of the ammonia emitted to the atmos-phere returns to Earth in precipitation.

Human activities have increased NOx flux tothe atmosphere sixfold to sevenfold, primarilythrough the combustion of fossil fuels (Table15.3). Nitrification is the largest natural terres-trial source of NO (see Chapter 9). Fertilizeraddition has increased the magnitude of thissource, with additional NO coming frombiomass burning. Preindustrial NOx fluxes weregreater in tropical than temperate ecosystems,due to frequent burning of tropical savannas,soil emissions, and production by lightning(Holland et al. 1999). Most NOx deposition nowoccurs in the temperate zone, where depositionrates have increased fourfold since preindus-trial times. Unlike N2O, NO is highly reactiveand alters atmospheric chemistry rather thanaccumulating in the atmosphere. Nitric oxide isa precursor to the photochemical production oftropospheric ozone (O3), a major component ofsmog, and is often the rate-limiting reactant inozone formation. When its concentrations arehigh, O3 can be produced via the oxidation ofcarbon monoxide, nonmethane hydrocarbons,and methane. It also affects the concentrationof the hydroxyl radical, the main oxidizing (or

cleansing) chemical in the atmosphere and thusindirectly affects the concentrations of manyother gases. The oxidation of NO leads to theformation of nitric acid, a component of acidprecipitation and an increasingly large sourceof nitrogen inputs to ecosystems.

Nitrogen deposition affects many ecosystemprocesses. The widespread nitrogen limitationof plant production in nontropical ecosystemsresults in an effective retention of a large proportion of anthropogenic nitrogen that isdeposited in ecosystems, particularly in young,actively growing forests that are accumulatingnutrients in vegetation (see Fig. 13.11). In somecases nitrogen deposition may stimulate carbonuptake and storage, although the global magni-tude of carbon storage accounted for by nitro-gen deposition is uncertain (Table 15.1)(Holland et al. 1997).

Nitrogen accumulation in production andorganic matter storage cannot increase indefi-nitely. After long-term chronic nitrogen inputs,nitrogen supply may exceed plant and micro-bial demands, resulting in nitrogen saturation(Agren and Bosata 1988, Aber et al. 1998).When ecosystems become nitrogen saturated,nitrogen losses to stream water, groundwater,and the atmosphere should increase and even-tually approach nitrogen inputs. Nitrogen satu-ration is often associated with declines in forestproductivity and increased tree mortality inconiferous forests in Europe (Schulze 1989)and the United States (Aber et al. 1995).

Temperate forests vary regionally in the rateat which they approach nitrogen saturation,depending on rates of nitrogen inputs and thecapacity of soils to buffer these inputs(Berendse et al. 1993, Vitousek et al. 1997a,Aber et al. 1998). In tropical forests, wherenitrogen availability is typically high relative toplant and microbial demands, anthropogenicnitrogen deposition may lead to immediatenitrogen losses (Hall and Matson 1998); thiscould have potentially negative effects on plant and soil processes (Matson et al. 1999).In general, the capacity of a forest ecosystem to retain nitrogen is linked to its productivepotential and to its current degree of nitrogenlimitation (Aber et al. 1995, Magill et al. 1997).

The Global Phosphorus Cycle 347

The addition of limiting nutrients can alterspecies dominance and reduce the diversity ofecosystems. Nitrogen addition to grasslands or heathlands, for example, increases the dom-inance of nitrogen-demanding grasses, whichthen suppress other plant species (Berendse et al. 1993). These species changes can convertnutrient-poor, diverse heathlands to species-poor forests and grasslands (Aerts andBerendse 1988). Loss of species diversity withnitrogen addition therefore occurs at bothpatch and landscape scales.

Human activities increase the nitrogen lossesfrom terrestrial ecosystems and nitrogen trans-fer to aquatic ecosystems. The massive nitrogenadditions to terrestrial ecosystems, in the formof deposition, fertilization, food imports, andgrowth of nitrogen-fixing crops, have led to adramatic increase in nitrogen concentrations in surface and groundwaters over the pastcentury. Nitrate concentrations in the Missis-sippi River have more than doubled since the1960s (Turner and Rabalais 1991), and nitrateconcentrations in other major rivers of theUnited States have increased 3- to 10-fold inthe past century (see Fig. 14.10) (Howarth et al.1996). Nitrate concentrations in many lakes,streams, and rivers of Europe have likewiseincreased, as have concentrations in mostaquifers (Vitousek et al. 1997a).

The Global Phosphorus Cycle

Phosphorus is unique among the major bio-geochemical cycles because it has only a tinygaseous component and has no biotic pathwaythat brings new phosphorus into ecosystems.Therefore, until recently, ecosystems derivedmost available phosphorus from organic forms,and phosphorus cycled quite tightly within ter-restrial ecosystems. Like nitrogen, phosphorusis an essential nutrient that is frequently inshort supply. Marine and fresh-water sedimentsand terrestrial soils account for most phos-phorus on Earth’s surface (Fig. 15.6). Most of this store is not directly accessible to thebiota. Most phosphorus in soils, for example,occurs primarily in insoluble forms such as

calcium or iron phosphate. Most organic phos-phorus is in plant or microbial biomass, and therecycling of that organic matter when it dies isthe major source of available phosphorus toorganisms.

The physical transfers of phosphorus aroundthe global system are constrained by the lack of a major atmospheric gaseous component.Leaching losses in natural ecosystems are alsolow due to the low solubility of phosphorus.Instead, phosphorus moves around the globalsystem primarily through wind erosion and run-off of particulates in rivers and streams to theoceans. The major flux in the global phosphoruscycle (excluding human activities) is via hydro-logic transport from land to the oceans. In theoceans, some of those phosphorus-containingparticulates are recycled by marine biota, but amuch larger portion (90%) is buried in sedi-ments. Because there is no atmospheric linkfrom oceans to land, the flow is one-way onshort time scales (Smil 2000). On geologicaltime scales (tens to hundreds of millions ofyears), phosphorus-containing sedimentaryrocks are exposed and weathered, resupplyingphosphorus to the terrestrial biosphere.

Anthropogenic Changes in thePhosphorus Cycle

Human activities have enhanced the mobility ofphosphorus and altered its natural cycling byaccelerating erosion and wind- and water-bornetransport. Inorganic phosphorus fertilizershave been produced since the mid-1800s,but the amount produced and applied hasincreased dramatically since the mid-twentiethcentury (Fig. 15.7), coincident with the intensi-fication of agriculture that accompanied theGreen Revolution (Smil 2000). Between 1850and 2000, agricultural systems received about550Tg of new phosphorus. The annual applica-tion of phosphorus to agricultural ecosystems(10 to 15Tgyr-1) is 20 to 30% of that whichcycles naturally through all terrestrial ecosys-tems (Fig. 15.6).

Human land use change has also increasedphosphorus losses from ecosystems. Water andwind erosion cause a 15Tgyr-1 phosphorus loss


from the world’s croplands, an amount similarto the annual fertilizer inputs. Overgrazing hasalso increased erosional losses, mobilizingabout 13Tgyr-1 of phosphorus from grazinglands (Smil 2000). The production of humanand animal wastes have led to point and non-point sources of phosphorus. The total phos-phorus losses from terrestrial ecosystems dueto human activities are about twice the annualfertilizer inputs.

Together, these changes have increased thetransport of phosphorus around the world(Howarth et al. 1995). Because phosphoruscommonly limits production in lakes, the inad-

vertent phosphorus fertilization of fresh-waterecosystems can lead to eutrophication andassociated negative consequences for aquaticorganisms and society (see Chapter 14). Phos-phorus transport by wind-blown dust can alsoaffect down-wind ecosystems, such as theSouthern Ocean.

The Global Sulfur Cycle

The global cycle of sulfur shares characteristicswith the global cycles of nitrogen and phos-phorus. The sulfur cycle, like the nitrogen cycle,

Ocean

mixin

gSoils

200,000

Vegetation500

Marine biota70

Biol

ogica

l

pu

mp


1000

1000

Surface sediments4 x 109

Atmosphere 0.028

Mineable P10,000

Fertilizer P12

Crops

60

60

The GlobalPhosphorus

Cycle

Dus

t 1

Dus

t 1

18

135

58

21

42

Surface oceanwater 8000

Figure 15.6. The global phosphorus cycle showingapproximate magnitudes of the major pools (boxes)and fluxes (arrows) in units of teragrams per year.Most phosphorus that participates in biogeochemi-cal cycles over decades to centuries is present in soils,sediments, and the ocean. Phosphorus cycles tightlybetween vegetation and soils on land and betweenmarine biota and surface ocean water in the ocean.The major human effects on the global phosphorus

cycle have been application of fertilizers (about 20%of that which naturally cycles through vegetation)and the erosional loss from crop and grazing lands.Data for phosphorus pools in vegetation, marinebiota, and ocean water are from Smil (2000), for theatmospheric pool and for ocean mixing and the bio-logical pump from Reeburgh (1997). Remaining dataare from Schlesinger (1997).

The Global Sulfur Cycle 349

has a significant atmospheric component. Thegaseous forms in the atmosphere have low con-centrations but play important roles. Like phos-phorus, sulfur is primarily rock derived. Seawater, sediments, and rocks are the largestreservoirs of sulfur (Fig. 15.8). The atmospherecontains little sulfur. Before human activities ofthe past several centuries, sulfur became avail-able to the biosphere primarily through theweathering of sedimentary pyrite. Once weath-ered, sulfur moves through the global system byhydrologic transport or emission to the atmos-phere as a reduced sulfur gas or sulfur-containing particles. About 100Tgyr-1 of sulfur,moving mostly as dissolved sulfate, was trans-ported through rivers to the coastal margins oropen oceans in the preindustrial world (Gal-loway 1996).

Sulfur can be reduced to sulfide or to othertrace sulfur gases in anaerobic environmentssuch as wetlands and coastal sediments. Theemission of sulfate from sea water and sulfurtrace gases from oceans (160Tgyr-1) is about10-fold greater than that from continents (Fig. 15.8). Marine biogenic emission includedimethylsulfide (DMS), one of the primarysources of atmospheric sulfate; emissions ofsulfur dioxide (SO2) from volcanic eruptionsare the other major source.

Sulfur emitted to the atmosphere typicallyhas a short residence time. It is oxidized tosulfate by reaction with OH radicals. Sulfaterains out downwind within a few days, gener-ally as sulfuric acid. Sulfuric acid has low equi-librium vapor pressure, so it quickly condensesto form sulfate in cloud droplets, which readilyevaporate to form sulfate aerosols. Theseaerosols have both direct and indirect effects onEarth’s energy budget. Their direct effect is tobackscatter (reflect) incoming shortwave radia-tion, thus reducing solar inputs and tending to reduce global temperature. Their indirecteffects are more complicated and difficult topredict. As particulates, they act as cloud condensation nuclei by providing a surface onwhich water can condense, thereby influenc-ing cloud formation, cloud lifetimes, and clouddroplet size. The density of cloud condensationnuclei and droplet size in turn govern cloudalbedo. The uncertainty of the direction andmagnitude of these effects of sulfate aerosolson cloud albedo is a key reason for concernabout the anthropogenic changes in the globalsulfur cycle.

Human activities now transfer about 100Tgyr-1 of sulfur to the atmosphere and oceans,increasing the natural cycling rate by about50% (Fig. 15.8). Half of this sulfur arises fromfossil-fuel combustion and ore refining; and therest comes from mobilization of sulfur in dustfrom farming, animal husbandry, erosion ofexposed sediments, and other sources. Much ofthe anthropogenic sulfur moves through theatmosphere and is deposited on land, where it can accumulate in soils or biota, or is discharged to the oceans in solution.

Reconstruction of global temperaturerecords from ice cores shows that sulfur dioxidefrom volcanic emissions is a major cause ofinterannual climate variation over long timescales. Consequently, the dramatic increase insulfur aerosols due to anthropogenic emissionswill undoubtedly play an important role infuture climate changes. The negative forcingdue to sulfur emissions and their associateddirect and indirect effects could range from 0 to 1.5Wm2, a negative forcing that partially offsets the warming due to greenhouse gases(Penner et al. 2001).

15

20

15

5

01900 1920 1940 1960 1980 2000

Time

Pho

spho

rus

fert

ilize

r us

ed (

Tg

yr-1

)

Figure 15.7. Changes in the global use of inorganicphosphorus fertilizers during the twentieth century.(Redrawn with permission from Annual Review ofEnergy in the Environment, Volume 25 ©2000 byAnnual Reviews www.AnnualReviews.org; Smil2000.)


The Global Water Cycle

Only a tiny fraction of Earth’s water (0.01%) is in soils, where it is accessible to plants andavailable to support the activities of terrestrialorganisms. Most of Earth’s water is in theoceans (96.5%), ice caps and glaciers (2.4%),and groundwater (1%) (Fig. 15.9) (Schlesinger1997). About 91% of the water that evaporatesfrom oceans returns there as precipitation.Another 9% of ocean evaporation (40,000km3)moves over the land, where it falls as precipita-tion and returns to the oceans as river runoff.The total evaporation from land (71,000km3)

is about 15% of total global evaporation,although land occupies about 30% of Earth’ssurface; this indicates that average evapotran-spiration rates are about half as great on landas over the oceans. In regions of adequate soil moisture, vegetation enhances evaporationfrom land compared to a free water surface (seeChapter 4). There are large regional variationsin evaporation rate over both land and oceanrelated to climate and, in the case of land,to water availability and transpiration rates of vegetation. Of the terrestrial precipitation(111,000km3), about one third comes from the oceans (40,000km3), and two thirds (71,000km3) is evaporated from land and recy-

Ocean

mixin

gSoils

300,000

Vegetation8,500

Marine biota30

Surface oceanwater 100 x 106

Biolo

gical

pu

mp

Intermediate and deepocean waters 1,200 x 106

Sediments2,400 x 106

Depo

sition

180

Sedimentary rocks7,800 x 106

Volc

anoe

s 5

Bio

geni

c ga

ses

16

Atmosphere 3

150Mining

Sea

sal

t 144

Emis

sion

s 75

Dep

ositi

on 9

0

Biog

enic

4

Vol

cano

es 5

The GlobalSulfur Cycle

135

130

Figure 15.8. The global sulfur cycle showing ap-proximate magnitudes of the major pools (boxes)and fluxes (arrows) in units of teragrams per year.Most sulfur is in rocks, sediments, and ocean waters.The major fluxes in the sulfur cycle are through thebiota and various trace gas fluxes. Human activities

have doubled the global fluxes of sulfur throughmining and increased gas emissions. Data for anthro-pogenic emissions are from Penner et al. (2001) and for the sulfur pools in marine biota and soilsfrom Reeburgh (1997). Remaining data are fromGalloway (1996) and Schlesinger (1997).

The Global Water Cycle 351

cled. Evaporation and precipitation are bothhighly variable, both regionally and seasonally.

The quantity of water in the atmosphere is only 2.6% of that which annually cyclesthrough the atmosphere in evaporation andtranspiration, giving a turnover time of about10 days. Precipitation is therefore tightly linkedto evapotranspiration from upwind ecosystemsover time scales of hours to weeks. Soil water,in contrast, has a turnover time of about a year,with inputs from precipitation and outputs toevapotranspiration and runoff. This turnovertime makes soil moisture sensitive to seasonaland interannual variations in precipitation andevapotranspiration.

Anthropogenic Changes in the Water Cycle

Human activities have altered the global hydro-logic cycle primarily through changes in climateand Earth’s energy budget. The twentieth-century rise in global mean air temperature of 0.6°C has increased rates of evaporation,which in turn increased precipitation. Precipi-tation in the contiguous United States, forexample, increased by 10% between 1910 and2000, with projections of a 3 to 15% increase inresponse to the projected temperature increaseof 1.5 to 3.5°C during the twenty-first century.Past changes in precipitation have varied spa-

Soil water120,000

Vegetation10,000

Oceans1.350 x 106

40,000

Groundwaters15.3 x 106

Atmosphere 13,000

The GlobalWater Cycle

Tran

spira

tion

Ice33 x106

71,0

00

Lakes230,000

Surface evaporation

Pre

cipi

tatio

n 11

0,00

0

Evap

orat

ion 4

25,0

00

Prec

ipita

tion

385,

000

Figure 15.9. The global water cycle showing approx-imate magnitudes of the major pools (boxes) andfluxes (arrows) in units of cubic kilometers per year(Schlesinger 1997). Most water is in the oceans, ice,

and groundwater, where it is not directly accessibleto terrestrial organisms. The major water fluxes areevapotranspiration, precipitation, and runoff.

PTE15 10/9/2002 7:36 PM Page 351


tially, as will certainly be the case in the future(Fig. 15.10). The ecological impact of a givenchange in precipitation also varies amongecosystems. A small change in precipitation inarid regions, for example, could have muchgreater ecological impact than larger changes inareas that currently receive abundant rainfall.Projected future changes in precipitation willaffect river flows, groundwater recharge, thewater relations of natural ecosystems, and thewater available to managed ecosystems.

Land use changes alter the hydrologic cycleby altering (1) the quantity of energy absorbed,(2) the pathway of energy loss, and (3) the mois-ture content and temperature of the atmos-phere. Conversion from tropical rain forest topasture, for example, leads to less energyabsorption because of increased albedo and alarger proportion of energy dissipated to theatmosphere as sensible rather than latent heat(Gash and Nobre 1997). The warmer, drieratmosphere allows less precipitation, favoringthe persistence of pastures rather than returnto rain forests (see Fig. 2.11). Conversion ofheathland to agriculture in western Australiaalso reduced precipitation by 30% (Chambers1998) (see Fig. 14.10). When land use changesare extensive, they can have continental-scaleeffects on temperature and precipitation, oftenat locations remote from the region of land-cover change, as a result of large-scale adjust-

ments in atmospheric circulation (Chase et al.2000). Land-cover changes in Southeast Asia,for example, have particularly large effects onglobal-scale climate through atmospheric teleconnections.

Ecosystems are generally more sensitive tosoil moisture (terrestrial ecosystems) or runoff(aquatic ecosystems) than to precipitation. Theprojected increases in both evaporation andprecipitation complicate projections of changesin water availability. Decreases in precipitationwill probably reduce soil moisture, but regionswith increased precipitation may still experi-ence reduced soil moisture, if evaporationincreases more than precipitation. Models gen-erally project increased soil moisture at highlatitudes, large-scale soil drying in continentalportions of northern mid-latitudes in summerdue to higher temperatures and insufficientincreases (or reductions) of rainfall. Many areasthat are currently important for agriculture,such as Ukraine or the midwestern UnitedStates, may be particularly prone to futuredrought; and grain-producing areas maymigrate northward to areas that are currentlytoo cold to support intensive agriculture. Thesechanges in location of soil moisture suitable for agriculture will have major regional andnational economic and social impacts.

Consequences of Changes in theWater Cycle

Society depends most directly on some of the smallest and most vulnerable pools in the global hydrologic cycle. Agriculture, forexample, relies on soil water derived from pre-cipitation, a relatively small pool that would bealtered quickly in response to major changes inthe balance between precipitation and evapo-transpiration. In some areas soil moisturederived from precipitation is supplemented byirrigation, which withdraws water from lakes,rivers, and groundwater. Lands under irrigationhave increased fivefold during the twentiethcentury (Fig. 15.11) (Gleick 1998). During thisperiod there was an eightfold increase in thewater used to support human activities, whichparalleled a fourfold increase in human popu-lation and a 50% increase in per capita water

0

5

-5

10

-10

US

A

Nor

way

Japa

n

S. C

hina

Tha

iland

Bra

zil

Eth

iopi

a

E. R

ussi

a

Cha

nges

in a

nnua

l pre

cipi

tatio

n (

% d

ecad

e-1)

Temperate Tropical

Figure 15.10. Regional changes in precipitationbetween 1950 and 2000 (Folland et al. 2001). Thereis substantial regional variation in the direction ofprecipitation changes.

The Global Water Cycle 353

Figure 15.11. Trends in (A) world population andglobal land area under irrigation; (B) water with-drawals to support human activities, expressed as aglobal total and on a per capita basis; and (C) water

withdrawal in the United States, separated by eco-nomic sector. (Redrawn with permission from IslandPress; Gleick 1998.)


consumption. Humans now use 54% of theaccessible runoff (see Chapter 4). Most of thiswater is used for hydroelectric power and irri-gation (Fig. 15.11).

The scarcity of water is only part of thehydrologic problems facing society. Approxi-mately 40% of the world’s population had noaccess to adequate sanitation in 1990 (Gleick2000), and 20% had no clean drinking water.The shortage of clean water is particularlysevere in the developing nations, where futurepopulation growth and water requirements arelikely to be greatest. The projected increases inhuman demands for fresh water will certainlyhave strong impacts on aquatic ecosystems,through eutrophication and pollution, diversionof fresh water for irrigation, and modificationof flow regimes by dams and reservoirs (seeChapter 10).

Summary

Ecological processes and human activities playmajor roles in most biogeochemical cycles. Themagnitude of biotic and human impacts onecosystem processes becomes clear whensummed at the global scale. Biotic processes(photosynthesis and respiration) constitute theengine that drives the global carbon cycle. Thefour major carbon pools that contribute tocarbon cycling over decades to centuries are theatmosphere, land, oceans, and surface sedi-ments. On land the carbon gain by vegetationis slightly greater than the carbon loss in respi-ration, leading to net carbon storage on land.The net carbon input to the oceans is alsoslightly greater than the net carbon return tothe atmosphere. Marine primary production isabout the same as that on land. Most (80%) ofthis marine NPP is released to the environmentby respiration, with the remaining 20% going tothe deep oceans by the biological pump. Oceanupwelling returns most of this carbon to thesurface ocean waters; only small quantities aredeposited in sediments. Human activities causea net carbon flux to the atmosphere throughcombustion of fossil fuels, cement production,and land use change. This flux is equivalent to14% of terrestrial heterotrophic respiration.

The atmosphere contains the vast majority ofEarth’s nitrogen. The amount of nitrogen thatannually cycles through terrestrial vegetation is9-fold greater than inputs by nitrogen fixation.In the ocean the annual cycling of nitrogenthrough the biota is 500-fold greater thaninputs by nitrogen fixation. Denitrification isthe major output of nitrogen to the atmosphere.Human activities have doubled the quantity ofnitrogen fixed by the terrestrial biospherethrough fertilizer production, planting of nitro-gen-fixing crops, and combustion of fossil fuels.

Most phosphorus that participates in biogeo-chemical cycles over decades to centuries ispresent in soils, sediments, and the ocean. Phos-phorus cycles tightly between vegetation andsoils on land and between marine biota andsurface ocean water in the ocean. The majorhuman impact on the global phosphorus cyclehas been application of fertilizers (about 20%of that which naturally cycles through vegeta-tion) and erosional loss from crop and grazinglands (about half of that which annually cyclesthrough vegetation). Most sulfur is in rocks,sediments, and ocean waters. The major fluxesin the sulfur cycle are through the biota andvarious trace gas fluxes. Human activities havesubstantially increased global fluxes of sulfurthrough mining and increased gas emissions.

Most water is in the oceans, ice, and ground-water, where it is not directly accessible to ter-restrial organisms. The major water fluxes areevapotranspiration, precipitation, and runoff.Human activities have speeded up the globalhydrologic cycle by increasing global tempera-ture, which enhances evapotranspiration andtherefore precipitation, and by diverting morethan half of the accessible fresh water forhuman use.Availability of adequate fresh waterwill be an increasingly scarce resource forsociety, if current human population trendscontinue.

Review Questions

1. How do the major global cycles (carbon,nitrogen, phosphorus, sulfur, and water)differ from one another in terms of (a) themajor pools and (b) the major fluxes? In


which cycles are soil pools and fluxeslargest? In which cycles are atmosphericpools and fluxes largest?

2. How do the controls over the global carboncycle differ between time scales of months,decades, and millennia? How has atmos-pheric CO2 varied on each of these timescales, and what has caused this variation?

3. How have human activities altered theglobal carbon cycle? What are the mecha-nisms that explain why some of the CO2

generated by human activities becomessequestered on land?

4. What are the major causes and the climaticconsequences of increased atmosphericconcentrations of CO2, CH4, and N2O?What changes in human activities would berequired to reduce the rate of increase ofthese gases? What policies would be mosteffective in reducing atmospheric concen-trations of these gases, and what would bethe societal consequences of these policychanges?

5. What are the major natural sources andsinks of atmospheric methane? How mightthese be changed by recent changes inclimate and atmospheric composition?

6. What are the major natural sources andsinks of atmospheric N2O? How mightthese be changed by recent changes inclimate and land use?

7. How have human activities changed theglobal nitrogen cycle? How have thesechanges affected the nitrogen cycle inunmanaged ecosystems?

8. How do changes in the nitrogen cycle affectthe global carbon cycle? In what types ofecosystems would you expect these nitro-gen effects on the carbon cycle to bestrongest? Why?

9. How have human activities changed theglobal phosphorus and sulfur cycles? Howdo changes in these cycles affect the globalcycles of other elements?

10. How have human activities changed theglobal water cycle? If the world has so

much water, and this water is replenishedso frequently by precipitation, why arepeople concerned about changes in theglobal water cycle? In what regions of theworld will changes in the quantity andquality of water have greatest societalimpact? Why?

Additional Reading

Aber, J., W. McDowell, K. Nadelhoffer, A. Magill, G.Bernstson, M. Kamakea, S. McNulty, W. Currie, L.Rustad, and I. Fernandez. 1998. Nitrogen satura-tion in temperate forest ecosystems. BioScience48:921–934.

Cicerone, R.J., and R.S. Oremland. 1988. Biogeo-chemical aspects of atmospheric methane. GlobalBiogeochemical Cycles 2:299–327.

Galloway, J.N. 1996. Anthropogenic mobilization of sulfur and nitrogen: Immediate and delayedconsequences. Annual Review of Energy in theEnvironment 21:261–292.

Houghton, J.T., Y. Ding, D.J. Griggs, M. Noguer, P.J.van der Linden, X. Dai, K. Maskell, and C.A.Johnson, editors. 2001. Climate Change 2001: TheScientific Basis. Cambridge University Press, Cam-bridge, UK.

Matson, P.A., W.H. McDowell, A.R. Townsend, andP.M. Vitousek. 1999. The globalization of N depo-sition: Ecosystem consequences in tropical envi-ronments. Biogeochemistry 46:67–83.

Reeburgh,W.S. 1997. Figures summarizing the globalcycles of biogeochemically important elements.Bulletin of the Ecological Society of America 78:260–267.

Schimel, D.S., et al. 2001. Recent patterns and mech-anisms of carbon exchange by terrestrial ecosys-tems. Nature 414:169–172.

Schlesinger, W.H. 1997. Biogeochemistry: An Analy-sis of Global Change. Academic Press, San Diego,CA.

Smil, V. 2000. Phosphorus in the environment:Natural flows and human interferences. AnnualReview of Energy in the Environment 25:53–88.

Vitousek, P.M., and P.A. Matson. 1993. Agriculture,the global nitrogen cycle, and trace gas flux. Pages193–208 in R.S. Oremland, editor. The Biogeo-chemistry of Global Change: Radiative TraceGases. Chapman & Hall, New York.

Introduction

Humans depend on Earth’s ecosystems forfood, shelter, and other essential goods and services. Ecosystems provide well-recognizedgoods, including timber, forage, fuels, medi-cines, and precursors to industrial products.Ecosystems also provide underrecognized ser-vices, such as recycling of water and chemicals,mitigation of floods, pollination of crops, andcleansing of the atmosphere (Daily 1997). Theharvest and management of these resources area major component of the global economy. Ourpurposeful use and misuse of these resourceshave endangered them, and many apparentlyunrelated activities have had indirect and unin-tended negative effects on them.

The overuse or misuse of resources can alterthe functioning of ecosystems and the servicesthey provide (see Chapter 1). Land use change,for example, can degrade the capacity of water-sheds to purify water, leading to large treat-ment costs to cities. Degradation and loss ofwetlands can expose communities to increaseddamage from floods and storm surges. Decima-tion of populations of insect pollinators hasreduced yields of many crops (Daily 1997).Introductions and invasions of nonnativespecies such as killer bees, fire ants, and zebra

mussels, through the actions of humans, causeenormous damage to living resources andthreaten human health. Human activities alsoindirectly affect ecosystem goods and servicesthrough changes in the atmosphere, hydrologicsystems and climate (see Chapter 15).

The growing scale of human activities sug-gests that all ecosystems are influenced, directlyor indirectly, by our activities. No ecosystemfunctions in isolation, and all are influenced by human activities taking place in adjacentcommunities and around the world. Humanactivities are leading to global changes in mostmajor interactive controls over ecosystemprocesses: climate (global warming), soil andwater resources (nitrogen deposition, erosion,diversions), disturbance regime (land usechange, fire control), and functional types oforganisms (species introductions and extinc-tions). In many cases, at the scale of regions,these global changes interact with each otherand with local changes. All ecosystems aretherefore experiencing directional changes inecosystem controls, creating novel conditionsand, in many cases, positive feedbacks that leadto new types of ecosystems. These changes ininteractive controls will inevitably change theproperties of ecosystems; some of thesechanges are detrimental to society.

16Managing and Sustaining Ecosystems

Human activities influence all of Earth’s ecosystems. This chapter summarizes thenature of these impacts, the principles by which important ecological properties canbe sustained, and the management approaches that have been developed to maxi-mize sustainability.

356

Ecosystem Concepts in Management 357

Our use, mismanagement, and unintendedeffects on ecosystems have put some of them atrisk, imperiling our own well-being. In thischapter, we describe some general ecologicalprinciples that may contribute to formulat-ing management approaches and minimizingimpacts. We conclude that maintaining Earth’secosystems, even the “wild” ones, in the face ofanthropogenic changes will require new man-agement approaches. We review some of thesethat draw on ecosystem ecology and other sci-ences to manage and sustain ecosystems andthe benefits we derive from them.

Ecosystem Concepts inManagement

Given the strong directional changes in mostinteractive controls, all ecosystems on Earthmust be managed to sustain their importantproperties. Ecosystems that are managed forfood and fiber must be managed to maintaintheir productive potential. Ecosystems thathave been degraded by human activities shouldbe managed to restore their original propertiesand ecosystem services. In ecosystems that havebeen less influenced by human activities, themanagement challenge is to protect ecosystemfunctions and conserve biological diversity, byboth reducing rates of land conversion andplanning for conservation in the face of con-tinued human development, climate change,and other global changes. Management of all ofthese conditions requires an application of solidscientific information and principles. In thissection, we discuss several of the emerging con-cepts and principles of ecosystem ecology thatare important for sustainable management.

Natural Variability

Observations about natural temporal andspatial variability in ecosystem functioningprovide hints about the potential responses ofecosystems to human-caused changes. Naturalvariability has allowed us to understand some-thing about the driving forces that determinethe structure and functioning of ecosystems,

and how changes in those forces lead to eco-logical change. Hypotheses about the mecha-nisms of ecosystem response to change can bedeveloped and tested with spatial and temporalhistorical data (Swetnam et al. 1999).

Historical and regional information on variability can be useful in many managementdecisions. Records of hydrologic flows andecosystem characteristics, for example, allowedpredictions of the impacts of altered hydrologicflows in the Everglades (Harwell 1997). Firerecords and information on associated ecosys-tem processes provide insights into conse-quences of fire management. In cases in whichecosystem conditions desired by managers orstakeholders are not within the bounds ofnatural ecosystem variability, the mismatchindicates that the desired conditions requirereassessment (Landres et al. 1999).

Past patterns of ecosystem variability are notalways a good predictor of current and futurechanges. The multiple and interacting nature of current changes (see Chapter 1) differ fromthose that ecosystems experienced in the past.Moreover, time lags and nonlinearities inresponse of different parts of ecosystems tochange make the responses of whole ecosys-tems difficult to predict and manage based onpast experience. Management therefore alsobenefits from experimentation and hypothesistesting.

Resilience and Stability

Understanding the basis for ecosystemresilience provides the basis for sustainingecosystem properties in the face of human-induced change. Ecosystems that maintain theirproperties despite disturbance (i.e., are resis-tant to change) and that return rapidly to theiroriginal state (i.e., are resilient) exhibit morestable and predictable ecosystem properties(Holling 1986) (see Box 13.1). Lakes providesome of the most useful tests of hypothesesabout resistance and resilience. Phosphorusinputs to lakes can push a system from onestable state (clear-water, oligotrophic system)to another (turbid-water, eutrophic system).The relationship between phosphorus turnoverrate and food chain length provides an index of

358 16. Managing and Sustaining Ecosystems

resilience that is useful for comparing theresponse of different lakes to a given distur-bance (Cottingham and Carpenter 1994). Lakeswith high phosphorus turnover remove phosphorus from actively cycling pools morequickly than do lakes with slow turnover. Short(planktivore-dominated) food chains processphosphorus more rapidly than do long (pisci-vore) food chains, except under high phospho-rus levels. This type of study provides the basisfor predictions about vulnerability of lakes topulses of phosphorus inputs from agriculture orlivestock in the watershed. Lakes also differ intheir resistance to changes in pH resulting fromacid rain. Lakes on granitic parent materialwith low acid-neutralizing capacity are morelikely to acidify in response to acid rain than are lakes on limestone (Driscoll et al. 2001).The resulting changes in acidity alter trophicstructure and abundance of fish.

Sustainable management of ecosystems benefits from a landscape perspective that con-siders interactions among ecosystems. A lakecannot be managed sustainably, for example,without considering the nutrient inputs fromthe surrounding landscape, and forest produc-tion can be managed most sustainably as a landscape mosaic by taking account of distur-bances such as hurricanes, fire, and logging. Theresilience and sustainability of lakes dependson a range of control processes that function at different scales to mitigate the effects of disturbance (Carpenter and Cottingham 1997).These control processes include the filtrationeffects of riparian vegetation and wetlands, therole of game fish in trophic dynamics, and theabsorption of nutrients by macrophytes. Whenthese components are intact, landscapes containing lakes can withstand perturbationssuch as droughts, floods, forest fires, and someland use change. Management of landscapes atcoarse spatial scales requires different infor-mation from management of individual lakes,fields, or forest stands. At coarse spatial scales,monitoring of food webs in lakes is not feasible,so land use records, remote sensing, and surveysof fishing activity and success provide moreuseful input to models. An important implica-tion of a landscape focus is that it requires the

recognition of ecosystem response to multipleforcings.

State Factors and Interactive Controls

Directional changes in state factors or interac-tive controls limit the sustainability of ecosys-tems. State factors provide a useful frameworkwithin which to examine regulation of ecosys-tem processes (see Chapter 1). State factors andinteractive controls exert such strong controlover ecosystem processes that changes in thesecontrolling factors as a result of human activi-ties inevitably alter ecosystems and reduce theextent to which their current properties can besustained. Management practices can, however,strongly influence the degree of sustainability.In particular, management that focuses on negative feedbacks, which tend to maintain theecosystem in its current state, is the key toecosystem sustainability (see Chapter 1). Thesenegative feedbacks counterbalance positivefeedbacks that tend to push the ecosystemtoward some new state. Ecosystems will be sustainable when the net effect of negativefeedbacks exceeds the effects of positive feedbacks.

At least three considerations influence thedegree to which ecosystems can be managedsustainably. First, the sustainability of pro-ductivity and other ecosystem characteristicsrequires that state factors and interactive con-trols be conserved as much as possible. Theinteractive controls most readily managed areresources (e.g., through fertilizer and irriga-tion), disturbance regime, and functional typesof organisms. Maintenance of long-term agri-cultural productivity requires managementpractices that retain soil organic matter (SOM),which in turn provides a buffered supply ofwater and nutrients (see Chapter 7).

Second, negative feedbacks among impor-tant interactive controls increase the sustain-ability of ecosystem processes. Biocontrol inagriculture uses negative feedbacks betweenpredators and prey to limit the impact of pestinsects on crops (Huffaker 1957). When thesenegative feedbacks are weakened, managementmust be intensified. Positive feedbacks such as

Application of Ecosystem Knowledge in Management 359

mycorrhizal or nitrogen-fixing mutualisms maypush an ecosystem toward some new state.These can also be constructive managementtools in promoting the recovery of degradedecosystems (Perry et al. 1989).

Finally, linkages among ecosystems in a landscape are most likely to enhance sustain-ability when they generate negative feedbacksamong processes in these ecosystems. Aquaticecosystems, for example, are recipients of nutrient runoff from land. They are vulnerableto ecological changes that occur on land buthave only modest direct effects on their donorecosystems. Land use practices that consideronly terrestrial ecosystems therefore unavoid-ably affect aquatic ecosystems, unless laws or regulations constrain fertilizer inputs to land. Laws and regulations can create eitherpositive or negative feedbacks betweenunmanaged ecosystems and human society.Hunting regulations, for example, that limithuman harvest when game populations arereduced (a negative feedback) provide a morestable population regulation than regulationsthat provide subsidies or price supports whengame or fish populations decline (a positivefeedback).

Application of EcosystemKnowledge in Management

Forest Management

The challenge for sustainable forestry is todefine the attributes of forested ecosystemsthat are ecologically and societally importantand to maximize these ecosystem services in theface of change. Management for sustainabletimber production is one of many possibleobjectives for forest ecosystem managementand provides a good example of the need forecosystem ecology in management. Severalissues are addressed by sustainable forestry.Nutrient supply rates, for example, must be suf-ficient to support rapid growth yet not so highthat they lead to large nutrient losses. The rateat which stands are harvested must be balancedwith their rate of regeneration following

logging. Species diversity typical of naturalmosaics of forest stands should be maintained.The sizes and arrangement of logged patchesshould provide a seminatural landscape mosaicwith dependable seed sources and patterns offorest edges that allow natural use and move-ment of animal populations.

Addressing these issues requires attention toand management of interactive controls; under-standing disturbance regime, plant functionaltypes, and soil resources is critical. In the north-western United States, for example, old-growthDouglas fir forests attain an age of more than500 years (Wills and Stuart 1994). Fire andwind-throw of individual trees are the majornatural disturbances, creating mosaics of treeages at multiple spatial scales. Logging is nowthe most widespread disturbance in this region.Logging differs from the natural disturbanceregime by affecting larger areas, occurring morefrequently, removing the nitrogen bound inbiomass, and increasing the probability of soilerosion. On some sites, planting of nitrogen-fixing alders in association with regeneratingDouglas fir can compensate for nitrogen lossesduring logging (Binkley et al. 1992) and couldalso reduce erosion. On the other hand, man-agement with alder could have undesirableeffects in nitrogen-rich sites, where nitrogen is not limiting, and competition from aldersduring early succession could reduce the productivity of tree seedlings and potentiallylead to higher nutrient losses. Strategies forforest management that embrace ecologicalprinciples will recognize inherent variability inecosystem state factors and interactive controlsand will select management practices in a broadenvironmental context.

Fisheries Management

Formulation of management options for fish-eries requires an understanding of ecosystemresilience. Management options include marinereserves, quota systems, new approaches forsetting fishing limits based on population sizes,and economic incentives for long-term popula-tion maintenance. Unrestricted fish harvest canreduce sustainability by replacing the natural


negative feedbacks to population changes withpositive feedback responses that drive har-vested populations to low levels. Supply-and-demand economics and government subsidies,for example, often maintain or increase fish-ing intensity when fish populations decline(Ludwig et al. 1993, Hilborn et al. 1995, Paulyand Christensen 1995). This contrasts with the decreasing predation pressure that wouldaccompany a decline in prey population in anunmanaged ecosystem (Francis 1990).

Management of the North Pacific salmonfishery has instituted a negative feedback onfishing pressure through tight regulation offishing activity. Commercial and subsistencefishing are allowed only after enough fish havemoved into spawning streams to ensure ade-quate recruitment. This negative feedback tofishing pressure may contribute to the record-high salmon catches from this fishery after 30 years of management (Ludwig et al. 1993).Sustaining the fishery also requires protectionof spawning streams from changes in otherinteractive controls. These include dams thatprevent winter floods (disturbance regime),warming of streams by removal of riparian veg-etation of logged sites (microclimate), speciesintroductions (functional types), and inputs ofsilt and nutrients in runoff or municipal sewage(nutrient resources).

A common approach to sustainable manage-ment is to harvest only the production in excessof that which would occur when density-dependent mortality limits fisheries stocks(termed surplus production) (Rosenberg et al.1993, Hilborn et al. 1995). The existence andmagnitude of surplus production may dependon the stability of interactive controls (e.g.,physical environment, nutrients, and predationpressure) and the extent to which these inter-active controls respond to changes in fisheriesstocks. The major challenge in fisheries man-agement is to estimate surplus production inthe face of fluctuating interactive controls anduncertainty in the relationship between thesecontrols and the fish population size. It has beensharply debated whether any ecosystem is sus-tainable when subjected to continuous humanharvest (Ludwig et al. 1993, Rosenberg et al.1993).

Ecosystem Restoration

Ecosystem restoration often benefits fromintroduction of positive feedbacks that push theecosystem to a new, more desirable state. Manyecosystems become degraded through a combi-nation of human impacts, including soil loss, airand water pollution, habitat fragmentation,water diversion, fire suppression, and introduc-tion of exotic species. In degraded agriculturalsystems and grazing lands, the challenge is torestore them to a sufficiently productive stateto provide goods and services to humans. Inother cases, the goal is to restore the naturalcomposition, structure, processes, and dynamicsof the original ecosystem (Christensen et al.1996). Advances in restoration practicesinvolve identifying the impediments to recov-ery of ecosystem structure and function andovercoming these impediments with artificialinterventions that often use or mimic naturalprocesses and interactive controls.

Interventions can be applied to any compo-nent of ecosystems, but hydrology and soil andplant community characteristics are commonlythe focus of effort (Dobson et al. 1997) (Box16.1). Soil organic matter loss and low soil fer-tility are common problems in heavily managedagricultural and pasture systems and in forestsor grasslands reestablishing on mine wastes.Fertilizers and nitrogen-fixing trees can restoresoil nutrients and organic inputs (Bradshaw1983). Reduced tillage often slows or eliminateslosses of SOM. Once soil characteristics areappropriate, plant species can be reintroducedby seeding, planting, or natural immigration(Dobson et al. 1997). The scientific basis forrestoration ecology is actively developing.Remaining questions include the steps requiredto regain the full range of species in restoredsites, the importance of the initial compositionof species in determining long-term character-istics, and the importance of soil organisms inecosystem recovery.

Management for Endangered Species

Management for endangered species requires a landscape perspective. The focus of endan-gered-species protection has generally been

Box 16.1. Everglades Restoration Study

of Everglades National Park and providingsufficient water to meet the demands of a large urban and agricultural economy.Planned construction projects include thecreation of storm water treatment areas toremove phosphorus from the water and toallow increased water diversion into theEverglades (DeAngelis et al. 1998). Addi-tional land is being purchased to provideareas of water storage and a buffer zonebetween natural areas and the expandingurban zone.

An ecosystem model was developed toevaluate the potential effectiveness ofvarious rehabilitation and managementoptions. This spatially explicit landscapemodel was linked with individual-basedmodeling of 10 higher trophic-level indicatorspecies to provide quantitative predictionsrelevant to the goals of the EvergladesRestoration (DeAngelis et al. 1998). Theseindicator species—including the Floridapanther, white ibis, and American croco-dile—all differ in their use of the landscapeand resources and span a range of habitatneeds and trophic interactions (Davis andOgden 1994).The simultaneous success of allof these species in a restored Evergladeswould imply health of the overall ecosystem(DeAngelis et al. 1998). In this ecosystemapproach, models of higher trophic-levelindicator species use information frommodels at intermediate trophic levels (fish,aquatic macroinvertebrates such as crayfish,and several reptile and amphibian functionaltypes), simulated as size-structured popula-tions, and lower trophic levels (periphyton,aggregated mesofauna, and macrophytes),using process models. These species-specificmodels are then layered on a landscape geo-graphic information system (GIS) modelthat includes hydrological and abiotic factorssuch as surface elevations, vegetation types,soil types, road locations, and water levels(DeAngelis et al. 1998). South Florida pro-vides an example of the incorporation of scientific knowledge of ecosystem processesinto long-term state and national ecosystemmanagement efforts.

Major human impacts on the natural hydrol-ogy of the Everglades ecosystem in thesoutheastern United States began in theearly twentieth century. In response to hurricanes, flooding, and the resulting loss of human life and property, the U.S. ArmyCorps of Engineers built levees, canals,pumping stations, and water-control struc-tures that separated the remaining Ever-glades from growing urban areas anddivided them into basins (Davis and Ogden1994). The northern Everglades were parti-tioned into a series of water conservationareas (WCAs) and the Everglades Agricul-tural Area south of Lake Okeechobee, whichincluded 1900km2 of sugar cane plantations(DeAngelis et al. 1998). The water flow tothe remaining “natural” Everglades declinedsharply and occurred as pulsed releases bywater-control structures. These hydrologicchanges caused pronounced fluctuations inwater levels and increased the frequency ofmajor drying events (DeAngelis et al. 1998).The survival of many species, including birds,alligators, and crocodiles, depends on reasonable regularity in the rise and fall ofwater level throughout the year. Since the1940s, the nesting populations of wadingbirds declined by 90% (Davis and Ogden1994). Land use change such as agriculturaldrainage destroyed many high-elevation,short-hydroperiod wetlands. Other human-induced effects on the system include stormwater runoff from the phosphorus-enrichedEverglades Agricultural Area, mercury pollution, and invasive species (DeAngeliset al. 1998).

The large-scale changes in the FloridaEverglades and associated ecosystemsrequired a landscape approach to restore thenatural hydrology on which so much of the flora and fauna depend. The goals of thesouth Florida ecosystem restoration includethe maintenance of ecological processes(e.g., disturbance regimes, hydrologicprocesses, and nutrient cycles) and mainte-nance of viable populations of all nativespecies. The U.S. Army Corps of Engineerswas charged with both improving protection

361


the establishment of protected areas containingpopulations of the target species and vegetationassociated with those species. Establishment of parks is, however, insufficient protection forspecies when humans continue to influenceimportant state factors and interactive controls,such as climate, fire regime, water flows, orspecies introductions (Jensen et al. 1993). If theclimate changes, for example, animals may betrapped inside a park that no longer has a suit-able climate or vegetation. Selection of parksthat have a range of elevations provides anopportunity for organisms to migrate verticallyto higher elevations in response to climatewarming. Habitat fragmentation and land usechange also alter the natural linkages amongecosystems inside and outside of parks. Nearlyall parks therefore require management tocompensate for human impact. The boundariesof Yellowstone National Park, for example,block migration of elk to traditional winteringareas, so winter food supplements must be provided. These winter food supplements incombination with the extirpation of naturalpredators release the elk population from theirnatural population controls. Managers musttherefore allow hunting or relocation of elk as an alternative mechanism of population regulation. Using management to replace interactive controls, rather than to protect theinteractive controls, is a complex and difficulttask, especially when management has multi-ple, often conflicting goals.

Integrative Approaches toEcosystem Management

Managing and sustaining ecosystems in arapidly changing world requires new manage-ment approaches that consider ecosystems asinteracting components of social and biophysi-cal landscapes and a broader ecological per-spective than management focusing on a singlespecies or product. Ecological sustainabilitycannot be divorced from economic and culturalsustainability. A policy that promotes ecologi-cal sustainability at the expense of its humanresidents cannot be effectively implemented.

Conversely, programs of economic develop-ment that sacrifice long-term ecological or cul-tural sustainability cannot be sustained over thelong term. An emerging challenge is to addressregional sustainability in ways that simultane-ously consider the ecological, economic, andcultural costs and benefits of particular policies.Development of a scientific basis for ecosystemsustainability requires close collaborationamong ecologists, resource managers, econo-mists, sociologists, anthropologists, and others.Development and implementation of theresulting policies requires involvement of sci-entists, resource managers, policy makers, landowners, industrial and recreational users, andother stakeholders. This comprehensive ecosys-tem approach (Bengston 1994) uses both theecological principles outlined in this book andthe principles and understanding developed inmany fields of social science. Its objectives,scale, and roles for science and managementdiffer significantly from more traditional man-agement approaches (Table 16.1).

In general, an ecosystem approach considersthe range of goods and services provided by anecosystem and manages them in light of theirinteractions and trade-offs. Management mustfocus on the scale and pattern of the system,rather than being constrained by jurisdictionalboundaries. An ecosystem approach is placebased, designed around the traits of the ecosys-tem and its political and social landscape. Itsgoal is to sustain or increase the capacity of asystem to provide desired goods and servicesover the long term. Finally, the ecosystemapproach integrates social and economic information with biophysical information andexplicitly considers the provision of humanneeds.

The use of an ecosystem approach is evidentin the concepts and implementation of ecosys-tem management plans, integrated conserva-tion and development plans, and approachesthat assign value to ecosystem functions.

Ecosystem Management

Ecosystem management is the application ofecological science to resource management to

Integrative Approaches to Ecosystem Management 363

promote long-term sustainability of ecosystemsand the delivery of essential ecosystem goodsand services to society. The concept wasadopted by the U.S. Forest Service in 1992 andhas since been developing in theory and appli-cation. Although there are virtually hundreds

of definitions of ecosystem management (Table16.2), the concept includes a set of commonprinciples: (1) long-term sustainability as a fundamental value; (2) clear, operational goals;(3) sound ecological understanding; (4) under-standing of connectedness and complexity; (5)

Table 16.1. Differences between traditional forest management and an ecosystem approach to forest management.

Characteristic Traditional forest management Forest ecosystem management

Objectives • maximizes commodity production • maintains ecosystem integrity, whileallowing for sustainable commodityproduction

• maximizes net present value • maintains future options• maintains forest harvest at levels less than or • aims to sustain ecosystem productivity over

equal to their growth or renewal time, with short-term considerations of factors such as aesthetics and socialacceptability of harvest practices

Scale • stand scale within political or ownership • ecosystem and landscape scaleboundaries

Role of science • views forest management as an applied • views forest management as combiningscience science and social factors

Role of • focuses on outputs demanded by people (e.g., • focuses on inputs and processes (e.g.,management timber, recreation, wildlife) diversity and ecological processes) that give

rise to outputs• strives for management that fits industrial • strives for management that mimics natural

production processes and productivity• considers timber the primary output • considers all species important and considers

that services (protecting watersheds,recreation, etc.) are of equal importance with goods timber

• strives to avoid impending timber shortages • strives to avoid biodiversity loss and soildegradation

• views forests as a crop production system • views forests as a natural ecosystem• values economic efficiency • values cost effectiveness and social

acceptability

Adapted from WRI (2000), after Bengston (1994).

Table 16.2. Selected definitions of ecosystem management.

• Regulating internal ecosystem structure and function, plus inputs and outputs to achieve socially desirable conditions(Agee and Johnson 1987).

• The careful and skillful use of ecological, economic, social, and managerial principles in managing ecosystems toproduce, restore, or sustain ecosystem integrity and desired conditions, uses, product, values, and services over thelong term (Overbay 1992).

• The strategy by which, in aggregate, the full array of forest values and functions is maintained at the landscape level.Coordinated management at the landscape level, including across ownerships, is an essential component (Society ofAmerican Foresters 1993).

• Integration of ecological, economic, and social principles to manage biological and physical systems in a manner thatsafeguards the ecological sustainability, natural diversity, and productivity of the landscape (Wood 1994).

Modified from Christensen et al. (1996).


recognition of the dynamic character of ecosys-tems; (6) attention to scale and context; (7)inclusion of humans as a component of ecosystems; and (8) incorporation of adaptiveapproaches (Christensen et al. 1996). Most integrated ecosystem management programsexplicitly facilitate public participation and collaborative decision making.

Long-term sustainability is the fundamentalobjective of ecosystem management. It isachievable in part via the inclusion of soundecological models and understanding thatincorporate the complex and dynamic charac-ter of ecosystems and acknowledge humans asinherent components of ecosystems. Changeand uncertainty are intrinsic characteristics ofmost ecosystems, and ecosystem managementis an approach that acknowledges the occur-rence of stochastic events as well as predictablevariability (Holling 1993). Ecosystem manage-ment must therefore be flexible enough to learnfrom scientific analysis and advances and to adapt to institutional and environmentalchange.

An ecosystem management approach isespecially critical for the management ofcomplex systems such as watersheds andmarine fisheries, where management must con-sider multiple changes and the linkage amongecosystems through the movement of water, air,animals, and plants. The ocean ecosystem thatis relevant to salmon fisheries combines fresh-water rivers and streams, coastal ecosystems,intermediate continental shelf waters, and thedeep ocean, all of which are characterized bycomplex dynamics that vary in space and timein ways that are poorly understood. The com-plexity and scales of change in marine ecosys-tems are only partially understood, includingseasonal variations in productivity, regional-scale El Niño climatic events, and long-termchanges in salinity and ocean temperature.Large natural fluctuations in the abundances ofmarine fish are the rule more than the excep-tion. Our limited powers of direct observationalso result in a fragmented knowledge of thediversity, abundance, and interactions of marineorganisms. One challenge of ecosystem man-agement is to reconcile the disparity betweenthe spatial and temporal scales at which

humans make resource management decisionsand those at which ecosystem propertiesoperate (Christensen et al. 1996).

Ecosystem management goes beyond asingle focus on commodity resources and har-vesting limits. Instead, it embraces sustainabil-ity as the criterion for commodity provisionand/or other uses. Ecosystem management istherefore concerned with multiple functions,thresholds in processes, and trade-offs amongdifferent management consequences. It fre-quently considers, for example, both productiv-ity and biodiversity (Johnson et al. 1996). Allecosystem management projects strive for anintegrated understanding and management ofthe ecological, social, economic, and politicalaspects of resource use to maximize long-termsustainability (Box 16.2).

Adaptive management, involving experimen-tation in the design and implementation of policies, is central to effective management ofecosystems. An adaptive policy is one that isdesigned from the outset to test hypothesesabout the ways in which ecosystem processesrespond to human actions. In this way, if thepolicy fails, learning occurs, so better policiescan be applied in the future. Perhaps as a resultof frequent management failures and gaps inscientific knowledge, the concept of adaptivemanagement has become central to the imple-mentation of ecosystem management. Oneadvantage of adaptive management stems fromthe high degree of uncertainty in real-lifecomplex systems. Instead of delaying timelyaction due to the lack of certainty, adaptivemanagement promotes the opportunity to learnfrom management experience. The lack ofaction in the face of uncertainty can haveecosystem and societal consequences that areat least as great as actions based on reasonablehypotheses about how the ecosystem functions.Hypotheses that underlie adaptive manage-ment may consider the probabilities of bothdesired outcomes and ecological disasters(Starfield et al. 1995). The optimal policy,for example, may be one that has a moderateprobability of desirable outcomes and a lowprobability of causing an ecological disaster.

While hundreds of projects have embracedan ecosystem management perspective, most


still fall short in implementation. Impedimentsto success include jurisdictional debates; lack of political will or foresight; conflicting desiresfor, and trade-offs among, goods and services;and the limitations of scientific information.Nonetheless, adoption of this approach hasimproved the extent to which managementmeets its goals and gives optimism that futuredevelopment of the theory and practice ofecosystem management will be ecologically andsocietally beneficial (Yaffee et al. 1996, Peine1999, WRI 2000).

Integrated Conservation andDevelopment Projects

Integrated conservation and development projects (ICDPs) represent a new approach toconservation in the developing world. ICDPsfocus equally on biological conservation andhuman development, typically through exter-nally funded, locally based projects (Wells andBrandon 1993, Kremen et al. 1994). In the past,conservation and development projects typi-cally were considered separately, by different

organizations, sometimes with conflicting goals and conflicting consequences (Sutherland2000). It is now widely accepted that the twodirectives are more likely to be successful ifconsidered together; the main goal of ICDPs isto link these previously opposing goals. Inresponse to the failure of conservation anddevelopment projects to succeed separately,ICDPs emerged in the 1980s and establishedformal partnerships between conservationorganizations and development agencies in an effort to create environmentally sound,economically sustainable alternatives todestructive land use change (Kremen et al.1994, Alpert 1996).

An important objective of ICDPs is todetermine the types, intensities, and distribu-tion of resource use that are compatible withthe conservation of biodiversity and the main-tenance of ecological processes (Alpert 1996).Most ICDPs therefore have the following char-acteristics: (1) They link conservation of naturalhabitats with the improvement of living condi-tions in the local communities. (2) They are sitebased and tailored to specific problems such as

Box 16.2. Great Barrier Reef, Australia

zones, which permit various uses, includingrecreational and commercial fishing underguidelines established to maintain ecosystemintegrity. The preservation zones provide avaluable baseline for understanding andevaluating patterns of change in overallecosystem behavior. Additional special-usezones protect critical breeding or nestingsites and provide important protection fornatural areas or research. An ecotourismstrategy has been developed that allowsfloating structures in special areas forviewing the natural reef free from fishingimpacts. Public involvement and educationare an essential component to this ecosys-tem management plan that incorporatespublic accountability, operational efficiency,minimal regulation, adaptability to changingcircumstances, and scientific credibility.

Coral reef ecosystems have recently comeunder increasing consumptive pressure anddamage from polluting activities on land.The Great Barrier Reef Marine ParkAuthority (GBRMPA) in Australia devel-oped an ecosystem management plan toprotect this valuable and diverse coral reefbased on the principles of ecosystem man-agement (Christensen et al. 1996). The parkis a protected area managed for multipleuses with oversight by the GBRMPA. Pro-hibited activities include oil and gas explo-ration, mining, littering, spear fishing, andharvesting of large fish. The park is dividedinto three zones of use intensity: preserva-tion zones, which allow only strictly con-trolled scientific research; marine nationalpark zones, which allow scientific, educa-tional, and recreational use; and general-use


impending loss of exceptional habitat. (3) Theyattract international expertise, local support,and external sources of income, and (4) theyadapt to conditions in the developing worldsuch as heavy dependence on natural resources,high population growth, and high opportunitycosts of protected areas (Alpert 1996). ICDPsoften team a nongovernmental organization; aforeign donor agency; a national agency incharge of forestry, wildlife, or parks; and, if possible, local traditional and official leaders.Projects incorporate biological information and scientific knowledge of ecosystemprocesses, as well as the interests of managersand local communities in their design andimplementation.

One major challenge of successful ICDPs isto develop an appropriate research mechanismto collect the scientific data needed to guide thedual objective of conservation and develop-ment (Kremen et al. 1994). It is critical tomonitor biodiversity and ecosystem processesacross space and time and at multiple levels ofecological organization (species, communities,ecosystems, and landscape) and their responsesto management (Noss 1990, Kremen et al.1994). Ecological and socioeconomic indicatorscan identify the causes and consequences ofhabitat loss, monitor changes in resource useand harvesting impacts, and evaluate thesuccess of various management programs(Kremen et al. 1994, 1998). A successful moni-toring program is essential to test the hypo-thesis that economic development linked toconservation promotes conservation.

In the 1990s, more than 100 ICDP projectswere initiated, including over 50 in at least 20countries in sub-Saharan Africa (Alpert 1996).A review of African projects concluded thatICDPs do not provide a definitive solution tohabitat loss, but they can offer medium-termsolutions to local conflicts between biologicalconservation and natural resource use in eco-nomically poor, remote areas of exceptionalecological importance (Alpert 1996).An earlierreview of 36 ICDPs worldwide concluded thatonly 5 of the projects had contributed positivelyto wildlife conservation (Kremen et al. 1994).Limited tourist revenue potential, lack of local

management capacity, political unrest, largehuman populations, customary rights to land orresources enclosed by reserves, or the absenceof an official protected area can pose significantimpediments to the success of a project (Alpert1996). The ICDPs most successful in promotingconservation contain significant communityparticipation, which fosters improved commu-nity attitudes toward conservation. As withother kinds of ecosystem management, gettingthe science right is an essential, but insufficient,step. Over the long run, as we learn from suc-cesses and failures, the approaches employed in ICDPs will evolve to address remainingimpediments and challenges.

Valuation of Ecosystem Goods and Services

The concept of ecosystem goods and servicestakes a pragmatic view in asking how changesto ecosystems will affect the products and services that humans derive from ecologicalsystems. Recognizing and estimating the eco-nomic value of the broad range of services pro-vided by ecosystems is an important aspect ofthe ecosystem approach to management(Costanza et al. 1997, Daily 1997). The conceptof goods and services has been particularlyimportant in reminding the public and policymakers that they depend on ecosystems forfood, fiber, fuel, pharmaceuticals, and industrialproducts as well as many services like waterpurification, mitigation of floods, pollination ofcrop and wild plants, pest control, compositionof the atmosphere, and aesthetic beauty.

The recognition that the world’s ecosystemsare capital assets, and that they can yieldecosystem goods and services under propermanagement, has led to a need and an oppor-tunity for ecosystem valuation. Human soci-eties must often choose between alternativeuses of the environment. Should a wetland be preserved, used for sewage treatment, ordrained and converted to agriculture? Whichservices should fresh-water systems be man-aged for (Table 16.3)? Individuals and societiesare constantly making decisions about how touse ecosystem goods and services, but these


decisions often ignore the value of the resourceand assume that the ecosystem service was“free” (Daily et al. 2000).

Valuation of ecosystem services is a way toorganize information to help make such deci-sions (Daily et al. 2000).Valuation of ecosystemservices requires sound ecological informationand a clear understanding of alternatives andimpacts. Ecological understanding is critical, forexample, to characterize the services providedby ecosystems and the processes by which they

are generated. This information is frequentlysite specific, so local ecological knowledge isneeded. Ecological and economic informationmust then be integrated to make sound deci-sions. Several approaches have been used toassign economic value to ecosystem goods and services. These include direct valuationapproaches, by which economic worth isassigned according to avoided costs or tomarket value, and indirect approaches, whichestimate the worth of a good or service throughsurveys and contingent valuation (Goulder andKennedy 1997) (Table 16.4). Although themethodology is still developing, economic valuations of the provision of goods or ser-vices have successfully contributed to decisionmaking about ecosystem management at avariety of scales (Daily et al. 2000) (Box 16.3).

Valuation of goods and services requiresboth their identification and an assessment of their temporal and spatial variability, theirvulnerability to stresses, and the extent to which they can be predicted using simulationmodels or indicators. Ecologists often have asense for what allows an ecosystem or land-scape to provide a given service, but we have alimited ability to quantify that service and topredict it in the future. When, for example,and under what conditions, will a watershedprovide clean water? Do all mangroves protectshorelines?

Valuation of ecosystem services is oftenapplied to a single variable or resource (e.g.,

Table 16.3. Examples of services provided by rivers,lakes, aquifers, and wetlands.

Water supplyDrinking, cooking, washing, and other household usesManufacturing, thermoelectric power, and other

industrial usesIrrigation of crops, lawns, etc.Aquaculture

Supply of goods other than waterFishWaterfowlClams and mussels

Nonextractive or in-stream benefitsFlood controlTransportationRecreational swimming, boating, etc.Pollution dilution and water-quality protectionHydroelectric generationBird and wildlife habitatEnhanced property valuesNonuser values

Data from Postel and Carpenter (1997).

Service Valuation method

Inputs that support productionPest control avoided costFlood control avoided costSoil fertility avoided costWater filtration avoided cost

Sustenance of plants and animalsConsumptive uses direct valuations based on market pricesNonconsumptive uses indirect valuations (travel cost or

contingent aluation methods)Provision of existence values indirect valuations (contingent valuation

method)Provision of option values empirical assessments of individual risk

aversion

Data from Goulder and Kennedy (1997).

Table 16.4. Examples of eco-system services and valuationmethods.


fresh-water supply) and does not integrateacross the numerous goods and services thatoccur in an ecosystem. A more comprehensiveapproach is essential to estimate the trade-offsamong management strategies that maximizedifferent ecosystem goods and services. Thisproblem is clearly illustrated in efforts to limitthe environmental costs of fertilization in agriculture. In the midwestern United States,wetlands are often used to reduce fertilizer lossfrom agriculture, due to their capacity to filternutrients and sediments from laterally movinggroundwater and surface water. A singularfocus on wetlands and the provision of cleanwater in rivers, however, prevents a completeanalysis of the environmental effects of fertil-ization and the steps needed to reduce them.Fertilizer nitrogen is also lost to the atmos-phere, causing global and regional air pollutionand is lost to aquifers, where it can affect thequality of drinking water. Researchers are onlybeginning to ask which good or service is beingvalued and protected, and at what cost to others(Naylor and Drew 1998).

Despite these and other limitations, the val-uation of ecosystem services is becoming animportant tool that contributes to sustaining

ecosystems.The protection of highly valued andwell-understood services (such as clean water)through the protection of ecosystems is increas-ingly viewed as a wise alternative to expensiveconstruction and engineering projects (Box16.3). With increasing knowledge, the benefitsof protecting the less-known ecosystem serviceswill become more widely recognized (Daily andEllison 2002).

SummaryHuman activities influence all ecosystems onEarth. Ecosystems are directly impacted byactivities such as resource harvests, land con-version, and management and are indirectlyinfluenced by human-caused changes in atmospheric chemistry, hydrology, and climate.Because human activities strongly influencemost of Earth’s ecosystems, it follows that weshould also take responsibility for their careand protection. Part of that responsibility mustbe to slow the rate and extent of global changesin climate, biogeochemical cycles, and land use.In addition, active management of all ecosys-tems is required to maintain populations,species, and ecosystem functions in the face of

Box 16.3. Water Purification for New York City

New York City has a long tradition of cleanwater. This water, which originates in theCatskill Mountains, was once bottled andsold because of its high purity. In recentyears, the Catskills natural ecological purifi-cation system has been overwhelmed bysewage and agricultural runoff, causing thewater to drop below accepted health stan-dards. The cost of a filtration plant to purifythis water was estimated at $6 to $8 billionin capital costs, plus annual operating costsof $300 million, a high price to pay for whatonce could be obtained for free.

This high cost prompted investigation ofthe cost of restoring the integrity of the

watershed’s natural purification services.Thecost of this environmental solution wasapproximately $1 billion to purchase andhalt development on critical lands within thewatershed, to compensate land owners forrestrictions on private development, and tosubsidize the improvement of septic systems.The great cost savings provided by ecosys-tem services was selected by the city as thepreferred alternative. This choice providedadditional valuable services including floodcontrol and sequestration of carbon in plantsand soils.

Excerpted from the President’s Committee of Advisors on Science and Technology (1998).


anthropogenic change and to sustain the provi-sion of goods and services that humans receivefrom them.

State factors and interactive controls exertsuch strong control over ecosystem processesthat changes in these controlling factorsinevitably alter ecosystems and reduce theextent to which their current properties can besustained. Management practices can, however,strongly influence the degree of sustainability.If the goal of management is to enhance sus-tainability of managed and unmanaged ecosys-tems, then state factors and interactive controlsmust be conserved as much as possible andnegative feedbacks, which contribute to main-taining these controls, must be strengthenedwithin and among ecosystems. Directionalchanges in many of these ecosystem controlsheighten the challenge of sustainably managingnatural resources and threaten the sustainabil-ity of natural ecosystems everywhere.

The ecosystem approach to managementapplies ecological understanding to resourcemanagement to promote long-term sustainabil-ity of ecosystems and the delivery of essentialecosystem goods and services to society. Thisrequires a landscape or regional perspective toaccount for interactions among ecosystems andexplicitly includes humans as components ofthis regional system. Ecosystem managementacknowledges the importance of stochasticevents and our inability to predict future con-ditions with certainty. Adaptive management is a commonly used approach to ecosystemmanagement that takes actions based onhypotheses of how management will affect theecosystem. According to the results of theseexperiments, management policies are modifiedto improve sustainability.

ICDPs apply adaptive management to con-servation in the developing world. ICDPs focusequally on biological conservation and humandevelopment.The main goal of ICDPs is to linkthese often previously opposing goals, based onthe assumption that local human populationswill place immediate socioeconomic securitybefore conservation concerns. Fundamentalprinciples underlying ecosystem managementin general, and ICDPs in particular, are thatpeople are an integral component of regional

systems and that planning for a sustainablefuture requires solutions that are ecologically,economically, and culturally sustainable.

Review Questions

1. What are the major direct and indirecteffects of human activities on ecosystems?Give examples of the magnitude of humanimpacts on ecosystems.

2. How does the resilience of an ecosysteminfluence its sustainability in the face ofhuman-induced environmental change?What ecological properties of ecosystemsinfluence their sustainability?

3. Describe a management approach thatwould maximize ecosystem sustainability.What factors or events are most likely tocause this management approach to fail?

4. What are ecosystem goods and services?How can an understanding of ecosystemservices be used in management decisions?

5. What is ecosystem management? What isthe role of humans in ecosystems in thecontext of ecosystem management?

6. What are the advantages and disadvantagesof adaptive management as an approach tomanaging ecosystems?

7. What have ICDPs taught us about the advis-ability of including humans as componentsof ecosystems?

Additional Reading

Alpert, P. 1996. Integrated conservation and devel-opment projects: Examples from Africa. Bio-Science 46:845–855.

Carpenter, S. R., and J.F. Kitchell, editors. 1993. TheTrophic Cascade in Lakes. Cambridge UniversityPress, Cambridge, UK.

Chapin, F.S. III, M.S. Torn, and M. Tateno. 1996.Principles of ecosystem sustainability. AmericanNaturalist 148:1016–1037.

Christensen, N.L., A.M. Bartuska, J.H. Brown, S.Carpenter, C. D’Antonio, R. Francis, J.F. Franklin,J.A. MacMahon, R.F. Noss, D.J. Parsons, C.H.Peterson, M.G. Turner, and R.G. Woodmansee.1996. The report of the Ecological Society ofAmerica committee on the scientific basis for


ecosystem management. Ecological Applications6:665–691.

Daily, G.C. 1997. Nature’s Services: Societal De-pendence on Natural Ecosystems. Island Press,Washington, DC.

Holling, C.S. 1986. Resilience of ecosystems: Localsurprise and global change. Pages 292–317 in W.C.Clark, and R.E. Munn, editors. Sustainable Devel-opment and the Biosphere. Cambridge UniversityPress, Cambridge, UK.

Lubchenco, J., A.M. Olson, L.B. Brubaker, S.R.Carpenter, M.M. Holland, S.P. Hubbell, S.A.

Levin, J.A. MacMahon, P.A. Matson, J.M. Melillo,H.A. Mooney, C.H. Peterson, H.R. Pulliam, L.A.Real, P.J. Regal, and P.G. Risser. 1991. The sustain-able biosphere initiative: An ecological researchagenda. Ecology 72:371–412.

Matson, P.A.,W.J. Parton,A.G. Power, and M.J. Swift.1997. Agricultural intensification and ecosystemproperties. Science 227:504–509.

Sutherland, W.J. 2000. The Conservation Handbook:Research, Management and Policy. Blackwell Scientific, Oxford, UK.

an nutrient productivityAn quantity of energy or material

assimilated by a trophic levelABA abscisic acidADP adenosine diphosphateAM arbuscular mycorrhizaeATP adenosine triphosphateB bacteriab buffer capacity of soilBanimal biomass of animalsBmicrob biomass of soil microbesBplant biomass of plantsBS bundle sheath cellC degrees Celsuis; carnivorec bulk soil phosphorus concentra-

tionC1 primary carnivoreC2 secondary carnivoreC3 related to the photosynthetic path-

way whose initial carboxylationproducts are three-carbon sugars

C4 related to the photosyntheticpathway whose initial carboxyla-tion products are four-carbonacids

Cp specific heat at constant pressure%CS1 percentage of soil carbon derived

from the initial vegetation typeCS2

13C content of second soil type13Cstd

13C content of a standardCV1

13C content of vegetation from theinitial vegetation type

CV213C content of vegetation from thesecond vegetation type

CAM crassulacean acid metabolism

CEC cation exchange capacityCFC chlorofluorocarboncm centimeter (10-2 m)C:N carbon :nitrogen ratioCPOM coarse particulate organic matterd dayD deuteriumD diffusion coefficient of phosphorusDDT an insecticideDIC dissolved inorganic carbonDMS dimethylsulfoxideDNA deoxyribonucleic acidDOC dissolved organic carbonDON dissolved organic nitrogenE rate of evapotranspiration of an

ecosysteme root elongation rateEassim assimilation efficiencyEconsump consumption efficiencyEprod production efficiencyEtroph trophic efficiencyENSO El Niño southern oscillationF fungiFanim-soil flux of carbon from animals to the

soil in feces and dead animalsFCH4

methane emission from the eco-system to the atmosphere

Fdisturb flux of carbon from an ecosys-tem to the atmosphere due to disturbance

Femiss emission to the atmosphere of vol-atile organic compounds by plants

Fharv flux of carbon from an ecosystemto the atmosphere due to humanharvest

Abbreviations

371

372 Abbreviations

Fherbiv consumption of plants by animalsFlateral lateral flux of carbon into or out of

an ecosystemFleach flux of carbon from an ecosystem

to groundwater by leachingFmicro-anim consumption of microbial biomass

by animalsFn component of the gravitational

force that is normal to the slopeand therefore contributes to fric-tion that resists erosion

Fp component of the gravitationalforce that is parallel to the slopeand therefore drives erosion

Fpl-fire flux from plants to the atmospheredue to combustion during fires

Fpl-soil flux from plants to soilFsoil-fire flux of carbon from dead organic

matter to the atmosphere due tocombustion during fires

Ft total gravitational forceFPAR fraction of photosynthetically

active radiationFPOM fine particulate organic matterg gramg gravitational accelerationG ground heat fluxGBRMPA Great Barrier Reef Marine Park

AuthorityGIS geographic information systemGPP gross primary productionh hourh height (m)H sensible heat fluxH herbivoreHNLC high-nitrogen, low-chlorophyllI irradiance at any point in the

canopyI0 irradiance at the top of the canopyImax maximum uptake rateIn quantity of energy or material

ingested by trophic level nICDP integrated conservation and

development projectITCZ intertropical convergence zoneJ jouleJp rate of water flow through plantsJs rate of water flow through the soilk extinction coefficient; decomposi-

tion constant

K degrees KelvinK equilibrium constant; shortwave

radiationKin incoming shortwave radiationKm affinity of roots for phosphorusKout outgoing shortwave radiationkg kilogram (103 g)kJ kilojoule (103 J)km kilometer (103 m)kPa kilopascal (103 Pa)L longwave radiation; latent heat of

vaporizationL literl lengthLin incoming longwave radiationLout outgoing longwave radiationLp hydraulic conductivity of plant

xylemLs hydraulic conductivity of soilLt litter mass at time tL0 litter mass at time zeroLAI leaf area indexLE latent heat fluxLUE light use efficiencym meterm massM microbivoreMa angular momentumMes mesophyllMg megagram (106 g)mg milligram (10-3 g)MJ megajoule (106 J)mL milliliter (10-3 L)mm millimeter (10-3 m)MPa megapascal (106 Pa)MRT mean residence timeN northn sample sizeNavail available nitrogen; available

nutrientsNADP nicotinamide adenine dinu-

cleotide phosphate (in its oxidizedform)

NADPH nicotinamide adenine dinu-cleotide phosphate (in its reducedform)

NAO North Atlantic oscillationNBP net biome productionNDVI normalized difference vegetation

index

Abbreviations 373

NEE net ecosystem exchange of CO2

between the ecosystem and theatmosphere

NEEdark net ecosystem exchange measuredin the dark

NEElight net ecosystem exchange measuredin the light

NEP net ecosystem productionNIR near infrared radiationnm nanometer (10-9 m)nmol nanomole (10-9 mole)NOx nitric oxides in general (includes

NO and NO2)N : P nitrogen to phosphorus ratioNPP net primary productionNUE nitrogen use efficiencyO organicOa highly decomposed organic

horizon of soilOe moderately decomposed organic

horizon of soilOi slightly decomposed organic

horizon of soilP pressure; precipitationP productionParea photosynthetic rate (per unit leaf

area)Pmass photosynthetic rate (per unit leaf

mass)Pa pascalPAR photosynthetically active radia-

tionPBL planetary boundary layerPCB polychlorinated biphenyl (an

industrial class of compounds con-taining chlorine)

PEP phosphoenolpyruvatePg pedagram (1015 g)pH negative log of H+ activityPNA Pacific North America patternPOC particulate organic carbonPON particulate organic nitrogenppmv parts per million by volumeppt parts per thousandProdn production by trophic level nProdn-1 Production at the preceeding

tropic levelQ10 proportional increase in the rate

of a process with a 10°C increasein temperature

r radiusR runoff; radiation; respirationR regolith; universal gas constantR* phosporus uptake thresholdRanimal animal respirationRecosyst ecosystem respirationRgrowth growth respirationRheterotr heterotrophic respirationRion respiration associated with ion

uptakeRmi mitochondrial respirationRmicrob microbial respirationRmaint maintenance respirationRnet net radiationRplant plant respirationRsam isotope ratio of a sampleRstd isotope ratio of a standardRe Reynolds numberRH relative humidityRubisco ribulose-bis-phosphate carboxy-

laseRuBP ribulose-bis-phosphates secondS heat storage by a surfaceS water storage by an ecosystem;

southSE standard errorSLA specific leaf areaSOM soil organic matterSRL specific root lengtht timetr residence timeT temperature; transpiration rate;

absolute temperatureTEM Terrestrial Ecosystem modelTg teragram (1012 g)UV ultravioletv velocityVk kinematic viscosityVAM vesicular arbuscular mycorrhizaeVIS visible radiationVPD vapor pressure deficitW wattWCA water conservation areaWUE water use efficiencyyr yeara albedod del; difference in isotope concen-

tration relative to a standardD change in a quantity

374 Abbreviations

e emissivitymg microgram (10-6 g)mL microliter (10-6 L)mm micrometer (10-6 m)mmol micromole (10-6 moles)r densitys Stefan-Boltzman constant

Ym matric potentialYo osmotic potentialYp pressure potentialYplant total plant water potentialYsoil total soil water potentialYt total water potential

A horizon. Uppermost mineral horizon ofsoils.

Abiotic. Not directly caused or induced byorganisms.

Abiotic condensation. Nonenzymatic reactionof quinones with other organic materials insoil.

Abscisic acid. Plant hormone that is trans-ported from roots to leaves and causes areduction in stomatal conductance.

Absolute humidity. Vapor density.Absorbence. Fraction of the global solar irra-

diance incident on a surface that is absorbed.Absorbed photosynthetically active radiation.

Visible light (400 to 700nm) absorbed byplants.

Acclimation. Morphological or physiologicaladjustment by an individual plant to com-pensate for the change in performancecaused by a change in one environmentalfactor (e.g., temperature).

Accumulation. Buildup of storage productsresulting from an excess supply over demand.

Acid rain. Rain that has low pH, due to highconcentrations of sulfuric and nitric acidreleased from combustion of fossil fuels.

Active transport. Energy-requiring transportof ions or molecules across a membraneagainst an electrochemical gradient.

Activity budget. Proportion of time that ananimal spends in various activities.

Actual evapotranspiration. Annual evapotran-spiration at a site; a climate index that inte-grates temperature and moisture availability.

Actual vegetation. Vegetation that actuallyoccurs on a site.

Adaptation. Genetic adjustment by a popula-tion to maximize performance in a particularenvironment.

Adaptive management. Management involv-ing experimentation in the design and implementation of policies so subsequentmanagement can be modified based on learn-ing from these experiments.

Adiabatic lapse rate. Change in temperatureexperienced by a parcel of air as it moves ver-tically in the atmosphere due to a change inatmospheric pressure.The dry adiabatic lapserate is the change in temperature that occursif the air does not exchange energy with thesurrounding air (about 9.8°Ckm-1). Themoist adiabatic lapse rate also includes temperature changes due to release of latentheat as water vapor condenses. The observedlapse rate varies regionally, depending onsurface heating and atmospheric moisturebut averages about 6.5°Ckm-1.

Advection. Net horizontal transfer of gases orwater.

Aerobic. Occurring in the presence of oxygen.Aerodynamic conductance. Boundary layer

conductance of a canopy.Aerosol. Small (0.1 to 10mm) particles sus-

pended in air.Aggregate. Clumps of soil particles bound

together by polysaccharides, fungal hyphae,or minerals.

Albedo. Fraction of the incident shortwaveradiation reflected from a surface.

Glossary

375

376 Glossary

Alfisol. Soil order that develops beneath tem-perate and subtropical forests, characterizedby less leaching than spodosol.

Allocation. Proportional distribution of photo-synthetic products or newly acquired nutri-ents among different organs or functions in aplant.

Allochthonous input. Input of energy andnutrients from outside the ecosystem; syn-onymous with subsidy.

Allometric relationship. Regression relation-ship that describes the biomass of some partof an organism as a function of some easilymeasured parameter (e.g., plant biomass as afunction of stem diameter and height).

Ammonification. Conversion of organic nitro-gen to ammonium due to the breakdown oflitter and soil organic matter; synonymouswith nitrogen mineralization.

Amorphous minerals. Minerals with noregular arrangements of atoms.

Anadromous. Life cycle in which reproductionoccurs in lakes, streams, or rivers while theadult phase occurs primarily in the ocean.

Anaerobic. Occurring in the absence ofoxygen.

Andisol. Soil order characterized by youngsoils on volcanic substrates.

Angular momentum. Force possessed by arotating body.

Anion. Negatively charged ion.Anion exchange capacity. Capacity of a soil to

hold exchangeable anions on positivelycharged sites at the surface of soil mineralsand organic matter.

Anthropocene. Geologic epoch characterizedby human impacts, initiated with the Industrial Revolution.

Anthropogenic. Resulting from or caused bypeople.

Arbuscular mycorrhizae. Mycorrhizae thatexchange carbohydrates between plant rootsand fungal hyphae via arbuscules; alsotermed vesicular arbuscular mycorrhizae orendomycorrhizae.

Arbuscules. Exchange organs between plantand mycorrhizal fungus that occur withinplant cells.

Aridisol. Soil order that develops in arid climates.

Aspect. Compass direction in which a slopefaces.

Assimilation. Incorporation of an inorganicresource (e.g., CO2 or NH4

+) into organiccompounds; transfer of digested food from the intestine to the bloodstream of an animal.

Assimilation efficiency. Proportion of ingestedenergy that is assimilated into the blood-stream of an animal.

Assimilatory nitrate reduction. Conversion ofnitrate to amino acids by soil microbes.

Autochthonous production. Production occur-ring within the ecosystem.

Autotroph. Organism that produces organicmatter from CO2 and environmental energyrather than by consuming organic matterproduced by other organisms. Most produceorganic matter by photosynthesis; synony-mous with primary producer.

Available energy. Absorbed energy that is notstored or conducted into the ground; it is theenergy available for turbulent exchange withthe atmosphere.

B horizon. Soil horizon with maximum accu-mulation of iron and aluminum oxides andclays.

Backscatter. Reflection from small particles.Base cations. Nonhydrogen, nonaluminum

cations.Base flow. Background stream flow from

groundwater input in the absence of recentstorm events.

Base saturation. Percentage of the totalexchangeable cation pool that is accountedfor by base cations.

Benthic. Associated with aquatic sediments.Biofilm. Microbial community embedded in

a matrix of polysaccharides secreted by bacteria.

Biogenic. Biologically produced.Biogeochemical cycling. Biologically mediated

cycling of materials in ecosystems.Biogeochemistry. Biologically influenced

chemical processes in ecosystems.Biological pump. Flux of carbon and nutrients

in feces and dead organisms from theeuphotic zone to deeper waters and the sediments of the ocean.

Glossary 377

Biomass. Quantity of living material (e.g.,plant biomass).

Biomass burning. Combustion of plants andsoil organic matter following forest clearing.

Biomass pyramid. Quantity of biomass in different trophic levels of an ecosystem.

Biome. General class of ecosystems (e.g.,tropical rain forest, arctic tundra).

Biosphere. Biotic component of Earth, includ-ing all ecosystems and living organisms.

Biotic. Caused or induced by organisms.Bloom. Rapid increase in phytoplankton

biomass.Bottom–up controls. Regulation of consumer

populations by quantity and quality of food.

Bottom water. Deep ocean water below about1000m depth.

Boundary layer. Thin layer around a leaf orroot in which the conditions differ from those in the bulk atmosphere or soil,respectively.

Boundary layer conductance. Conductanceof water vapor across the boundary of a leaf or canopy; canopy boundary layer conductance is also termed aerodynamicconductance.

Bowen ratio. Ratio of sensible to latent heatflux.

Brine rejection. Exclusion of salt during formation of ice crystals in sea ice.

Bulk density. Mass of dry soil per unit volume.Buffering capacity. Capacity of the soil to

release cations to replace ions lost by uptakeor leaching.

Bulk density. Mass of soil per unit volume.Bulk soil. Soil outside the rhizosphere.Bundle sheath cells. Cells surrounding the

vascular bundle of a leaf; site of C3

photosynthesis in C4 plants.

C horizon. Soil horizon that is relatively unaf-fected by the soil forming processes.

C3 photosynthesis. Photosynthetic pathway inwhich CO2 is initially fixed by Rubisco,producing three-carbon sugars.

C4 photosynthesis. Photosynthetic pathway inwhich CO2 is initially fixed by PEP carboxy-lase during the day, producing four-carbonorganic acids.

Calcic horizon. Hard calcium (or magnesium)carbonate-rich horizon formed in deserts;formerly termed caliche.

Caliche. Calcic horizon.Canopy closure. Time during succession at

which crowns of adjacent trees overlap toproduce a relatively uniform canopy.

Canopy conductance. Measure of the physiological controls over water vaportransfer from the ecosystem to the at-mosphere. It equals the average stomatalconductance of individual leaves times LAI.

Canopy interception. Fraction of precipitationthat does not reach the ground.

Carbon-based defense. Organic compoundsthat contain no nitrogen and defend plantsagainst pathogens and herbivores.

Carbon-fixation reactions. Those reactions inphotosynthesis that use the products of thelight-harvesting reactions to reduce CO2 tosugars.

Carboxylase. Enzyme that catalyzes the reac-tion of a substrate with CO2.

Carboxylation. Attachment of CO2 to anacceptor molecule.

Carnivore. Organism that eats live animals.Carrier. Protein involved in ion transport

across a membrane.Catalyst. Molecule that speeds the conversion

of substrates to products.Catena. Sequence of soils or ecosystems

between crests of hills and floors of adjacentdrainages, whose characteristics change frompoint to point, depending on drainage andother land-surface processes.

Cation. Positively charged ion.Cation exchange capacity. Capacity of a soil to

hold exchangeable cations on negativelycharged sites at the surface of soil mineralsand organic matter.

Cavitation. Breakage of water columns undertension in the xylem.

Cellobiase. Enzyme that breaks down cel-lobiose to form glucose.

Cellobiose. Organic compound composed of two glucose units formed by cellulosebreakdown.

Charge density. Charge per unit hydratedvolume of the ion.

378 Glossary

Chelation. Reversible combination, usuallywith high affinity, with a metal ion (e.g., iron,copper).

Chemical alteration. Chemical changes in deadorganic matter during decomposition.

Chemical weathering. Changes due to chemi-cal reactions between the materials and theatmosphere or water.

Chemodenitrification. Abiotic conversion ofnitrite to nitric oxide (NO).

Chemosynthesis. Synthesis of organic matterfueled by oxidation–reduction reactionsunrelated to photosynthesis.

Chlorofluorocarbon. Organic chemicals con-taining chlorine and/or fluorine; gases thatdestroy stratospheric ozone.

Chlorophyll. Green pigment involved in lightcapture by photosynthesis.

Chloroplast. Organelles that carry out photosynthesis.

Chronosequence. Sites that are similar to oneanother with respect to all state factorsexcept time since disturbance.

Circadian rhythms. Innate physiological cyclesin organisms that have a period of about 24h.

Clay. Soil particles less than 0.002mm diameter.

Climate modes. Relatively stable configura-tions of global atmospheric circulation.

Climate system. Interactive system made up ofthe atmosphere, hydrosphere, biosphere,cryosphere, and land surface.

Climatic climax. End point of succession thatis determined only by climate.

Climax. End point of succession where thestructure and rates of ecosystem processesreach steady state and where resource con-sumption by vegetation is balanced by therate of resource supply.

Closed system. System in which the internaltransfers of substances are much greater thaninputs and outputs.

Cloud condensation nuclei. Aerosols aroundwhich water vapor condenses to form clouds.

C:N ratio. Ratio of carbon mass to nitrogenmass.

CO2 compensation point. CO2 concentrationat which net photosynthesis equals zero.

Coarse particulate organic matter. Organicmatter in aquatic ecosystems, includingleaves and wood, that is larger than 1mmdiameter.

Collector. Benthic macroinvertebrate thatfeeds on fine organic particles; includes filtering collectors that consume suspendedparticles and gathering collectors thatconsume deposited particles.

Co-metabolism. Breakdown of a substrate by a series of enzymes that are produced bydifferent microbes.

Community. Group of co-existing organisms inan ecosystem.

Compensation. Increased growth of somespecies in a community, due to release ofresources, in response to reduced growth byother species.

Compensation point. Temperature, CO2 con-centration or light level at which net carbonexchange by a leaf is zero (i.e., photosynthe-sis equals respiration).

Competition. Interactions among organismsthat use the same limiting resources(resource competition) or that harm oneanother in the process of seeking a resource(interference competition).

Complementary resource use. Use of re-sources that differ in type, depth, or timing byco-occurring species.

Conductance. Flux per unit driving force (e.g., concentration gradient); inverse ofresistance.

Configuration. Spatial arrangement of patchesin a landscape.

Connectivity. Degree of connectedness amongpatches in a landscape.

Consortium. Group of genetically unrelatedbacteria, each of which produce only some ofthe enzymes required to break downcomplex molecules.

Consumer. Organism that meets its energeticand nutritional needs by eating other livingorganisms.

Consumption efficiency. Proportion of the pro-duction at one trophic level that is ingestedby the next tropic level.

Convection. Heat transfer by turbulent move-ment of a fluid (e.g., air or water).

Glossary 379

Coriolis effect. Tendency, due to Earth’s rota-tion, of objects to be deflected to the right inthe Northern Hemisphere and to the left inthe Southern Hemisphere.

Cortex. Layers of root cells outside the endo-dermis involved in nutrient uptake.

Coupling. Effectiveness of atmospheric mixingbetween the canopy and the atmosphere.

Crassulacean acid metabolism. Photosyntheticpathway in which stomates open and carbonis fixed at night into four-carbon acids.During the day stomates close, C4 acids aredecarboxylated, and CO2 is fixed by C3

photosynthesis.Crystalline minerals. Minerals with highly

regular arrangements of atoms.Cytoplasm. Contents of a cell that are

contained within its plasma membrane butoutside the vacuole and the nucleus.

Deciduous. Shedding leaves in response tospecific environmental cues.

Decomposer. Organism that breaks downdead organic matter and consumes theresulting energy and nutrients for its ownproduction.

Decomposition. Breakdown of dead organicmatter through fragmentation, chemicalalteration, and leaching.

Decomposition constant. Constant thatdescribes the exponential breakdown of atissue.

Decoupling coefficient. Measure of the extentto which the canopy is decoupled from theatmosphere.

Deep water. Ocean water greater than 1000mdepth.

Deforestation. Conversion of forest to a nonforest ecosystem type.

Demand. Requirement; used in the context of the control of the rate of a process (e.g., nutrient uptake) by the amount needed.

Denitrification. Conversion of nitrate togaseous forms (N2, NO, and N2O).

Deposition. Atmospheric input of materials toan ecosystem.

Detritivore. Organism that derives energyfrom breakdown of dead organic matter.

Detritus. Dead plant and animal material,including leaves, stems, roots, dead animals,and animal feces.

Detritus-based trophic system. Organisms thatconsume detritus or energy derived fromdetritus.

Diffuse radiation. Radiation that is scatteredby particles and gases in the atmosphere.

Diffusion. Net movement of molecules or ionsalong a concentration gradient due to theirrandom kinetic activity.

Diffusion shell. Zone of nutrient depletionaround individual roots caused by activenutrient uptake at the root surface and dif-fusion to the root from the surrounding soil.

Direct radiation. Radiation that comes directlyfrom the sun without scattering or reradia-tion by the atmosphere or objects in the environment.

Discrimination. Preferential reaction with thelighter isotope of an element or compoundcontaining that element.

Dissolved organic carbon. Water-solubleorganic carbon compounds.

Dissolved organic nitrogen. Water-solubleorganic nitrogen compounds.

Disturbance. Relatively discrete event in timeand space that alters the structure of popula-tions, communities, and ecosystems andcauses changes resource availability or thephysical environment.

Disturbance regime. The range of severity,frequency, type, size, timing, and intensity ofdisturbances characteristic of an ecosystemor region.

Disturbance severity. Magnitude of change inresource supply or environment caused by adisturbance.

Doldrums. Region near the equator with lightwinds and high humidity.

Down regulation. Decrease in capacity to carryout a reaction; for example down regulationof CO2 uptake in response to elevated CO2.

Downwelling. Downward movement ofsurface ocean water, due to high densityassociated with high salinity and low temperature.

Drift. Invertebrates that move downstream inflowing water.

380 Glossary

E horizon. Heavily leached horizon be-neath the A horizon; formed in humid climates.

Eccentricity. Degree of ellipticity of Earth’sorbit around the sun.

Ecosystem. Ecological system consisting of all the organisms in an area and the physical environment with which they interact.

Ecosystem approach. Management of eco-system goods and services provided byecosystems in light of their interactions andtrade-offs.

Ecosystem ecology. Study of the interactionsbetween organisms and their environment asan integrated system.

Ecosystem engineer. Organisms that alterresource availability by modifying the physi-cal properties of soils and litter.

Ecosystem good. Substance produced by anecosystem and used by people (e.g., oxygen,food, or fiber).

Ecosystem management. Application of eco-logical science to resource management topromote long-term sustainability of ecosys-tems and the delivery of essential ecosystemgoods and services to society.

Ecosystem processes. Inputs or losses of mate-rials and energy to and from the ecosystemand the transfers of these substances amongcomponents of the system.

Ecosystem respiration. Sum of plant respira-tion and heterotrophic respiration.

Ecosystem service. Societally important conse-quences of ecosystem processes (e.g., waterpurification, mitigation of floods, pollinationof crops).

Ectomycorrhizae. Mycorrhizal association insome woody plants in which a large part of the fungal tissue is found outside the root.

Efficiency. Rate of a process per unit plantresource.

El Niño. Warming of surface water throughoutthe central and eastern tropical PacificOcean.

Electron-transport chain. Series of membrane-bound enzymes that produce ATP andNADPH as a result of passing electronsdown an electropotential gradient.

Emergent properties. Properties of organisms,communities, or ecosystems that are notimmediately obvious from study of processesat finer levels of organization.

Emissivity. Coefficient that describes themaximum rate at which a body emits radia-tion, relative to a perfect (black body) radia-tor, which has a value of 1.0.

Endocellulase. Enzyme that breaks down theinternal bonds to disrupt the crystallinestructure of cellulose.

Endodermis. Layer of suberin-coated cellsbetween the cortex and xylem of roots; waterpenetrates this layer only by moving throughthe cytoplasm of these cells.

Endomycorrhizae. Mycorrhizal association inmany herbaceous species and some trees inwhich a large part of the fungal tissue isfound inside the root; also termed arbuscularmycorrhizae.

Energy pyramid. Quantity of energy trans-ferred between successive trophic levels.

Entisols. Soil order characterized by minimalsoil development.

Enzyme. Organic molecule produced by an organism that catalyzes a chemical reaction.

Epidermis. Layer of cells on the surface of aleaf or root.

Equilibrium. State of balance between oppos-ing forces.

Estuary. Coastal ecosystem where a rivermixes with seawater.

Euphotic zone. Uppermost layer of water inaquatic ecosystems where there is enoughlight to support photosynthesis.

Eutrophic. Nutrient rich.Eutrophication. Nutrient-induced increase in

productivity.Evapotranspiration. Water loss from an

ecosystem by transpiration and surface evaporation.

Evergreen. Retention of green leaves through-out the year.

Exocellulase. Enzyme that cleaves off disac-charide units from the ends of cellulosechains, forming cellobiose.

Exodermis. Layer of suberin-coated cells justbeneath the epidermis of roots of somespecies.

Glossary 381

Exoenzyme. Enzyme that is secreted by anorganism into the environment.

Extensification. Expansion of the aerial extent of land cover change due to humanactivities.

Extinction coefficient. Constant that describesthe exponential decrease in irradiancethrough a canopy.

Exudation. Secretion of soluble organic com-pounds by roots into the soil.

Facilitation. Processes by which some speciesmake the environment more favorable forthe growth of other species.

Fast variable. Variable that changes rapidly.Feedback. Response in which the product of

one of the final steps in a chain of eventsaffects one of the first steps in this chain;fluctuations in rate or concentration are min-imized with negative feedbacks or amplifiedwith positive feedbacks.

Fermentation. Anaerobic process that breaksdown labile organic matter to produceorganic acids and CO2.

Ferrell cell. Atmospheric circulation cellbetween 30° and 60° N or S latitude.

Field capacity. Water held by a soil after gravitational water has drained.

Filter feeder. Aquatic animal that feeds on suspended particles.

Fine particulate organic matter. Particulateorganic matter in aquatic ecosystems that issmaller than 1mm diameter.

Fire intensity. Rate of heat production.Fixation. Covalent binding of an ion to a

mineral surface.Flux. Flow of energy or materials from one

pool to another.Food chain. Group of organisms that are

linked together by the linear transfer ofenergy and nutrients from one organism toanother.

Food web. Group of organisms that are linked together by the transfer of energy andnutrients that originates from the samesource.

Forward modeling. Modeling that estimatesthe outputs of a simulation model based on the temporal and spatial patterns ofinputs.

Fractionation. Preferential incorporation of alight isotope (e.g., 12C vs. 13C).

Fragmentation. Breaking up of intact litterinto small pieces.

Fulvic acids. Humic compounds that are relatively water soluble due to their ex-tensive side chains and many charged groups.

Functional type. A group of species that issimilar with respect to their impacts on community or ecosystem processes (effectsfunctional type); functional types have alsobeen defined with respect to their similarityof response to a given environmental change,such as elevated CO2 (response functionaltypes).

Gap-phase succession. Succession that occursin small patches within a stand due to deathof individual plants or plant parts.

Gelisol. Soil order characterized by presenceof permafrost.

Generalist herbivore. Herbivore that is rela-tively nonselective in its choice of plantspecies.

Geotropism. Growth response of plant organswith respect to gravity.

Gley soil. Blue-gray soil due to loss of ferric iron; formed under anaerobic conditions.

Graminoid. Grasslike plant (grasses, sedges,and rushes).

Grazer. Herbivore that consumes herbaceousplants (terrestrial ecosystems) or periphyton(aquatic ecosystems).

Grazing lawn. Productive grassland or wetlandecosystem in which plants are heavily grazedbut supported by large nutrient inputs fromgrazers.

Greenhouse effect. Warming of the atmos-phere due to atmospheric absorption ofinfrared radiation.

Greenhouse gas. Atmospheric gas that absorbsinfrared radiation.

Gross primary production. Net carbon input to ecosystems—that is, net photosyn-thesis expressed at the ecosystem scale (gCm-2 yr-1).

Ground heat flux. Heat transferred from thesurface into the soil.

382 Glossary

Groundwater. Water in soil and rocks beneaththe rooting zone.

Growth. Production of new biomass.Growth respiration. Respiration to support

biosynthesis (the production of newbiomass).

Guano. Large accumulations of seabird feces.Gyre. Large circulation systems in surface

ocean waters.

Hadley cell. Atmospheric circulation cellbetween the equator and 30° N or S latitude,driven by rising air where the sun’s rays areperpendicular to Earth’s surface.

Halocline. Relatively sharp vertical gradient insalinity in a lake or ocean.

Halophyte. Plant species that typically growson saline soils.

Hard pan. Soil horizon with low hydraulic conductivity.

Hartig net. Hyphae that penetrate cell walls ofroot cortical cells in ectomycorrhizae.

Heat capacity. Amount of energy required toraise the temperature of unit volume of abody by 1°C.

Heat of fusion. Energy required to change asubstance from a solid to a liquid without a change in temperature.

Heat of vaporization. Energy required tochange a gram of a substance from a liquidto a vapor without change in temperature.

Heat storage. Energy stored by an object dueto an increase in temperature.

Herbivore. Organism that eats live plants.Herbivory. Consumption of plants by animals.Heterocyst. Specialized nonphotosynthetic

cells of phototrophs that protect nitrogenasefrom denaturation by oxygen.

Heterotrophic respiration. Respiration bynonautotrophic organisms.

Heterotroph. Organism that consumes organicmatter produced by other organisms ratherthan producing organic matter from CO2 andenvironmental energy; includes decom-posers, consumers, and parasites.

Histosol. Soil order characterized by highlyorganic soils due to poor drainage and lowoxygen.

Homeothermy. Maintenance of a constantbody temperature.

Horizon. Layer in a soil profile. The horizons,from top to bottom, are the O horizon, whichconsists of organic matter above mineral soil;the A horizon, a dark layer with substantialorganic matter; the E horizon, which isheavily leached; a B horizon, where iron andaluminum oxides and clays accumulate; anda C horizon, which is relatively unaffected bysoil-forming processes.

Horse latitudes. Latitudes 30° N and S,characterized by weak winds and high temperatures.

Hot spot. Zone of high rates of biogeochemi-cal processes in a soil or landscape.

Humic acid. Relatively insoluble humic com-pounds with extensive networks of aromaticrings and few side chains.

Humification. Nonenzymatic process by which recalcitrant breakdown products ofdecomposition are complexed to formhumus.

Humin. Relatively insoluble humic com-pounds with extensive networks of long-chain, nonpolar groups.

Humus. Amorphous soil organic matter that isthe final product of decomposition.

Hydraulic conductivity. Capacity of a givenvolume of a substance (such as soil) toconduct water; this defines the relationshipbetween discharge and the hydraulic gradient causing it.

Hydraulic lift. Vertical movement of waterthrough roots from moist to dry soils along agradient in water potential.

Hydrothermal vent. Vent that emits reducedgases such as H2S in zones of sea-floorspreading.

Hyphae. Filamentous structures that make upthe vegetative body of fungi.

Hyporrheic zone. Zone of flowing groundwa-ter within the streambed or riverbed.

Ice-albedo feedback. Atmospheric warmingcaused by warming-induced decrease inalbedo due to earlier melting of sea ice.

Igneous rocks. Rocks formed when magmafrom Earth’s core cools near the surface.

Immobilization. Removal of inorganic nutri-ents from the available pool by microbialuptake and chemical fixation.

Glossary 383

Inceptisol. Soil order characterized by weaksoil development.

Infiltration. Movement of water into the soil.Integrated conservation and development

project. Project in developing nation thatfocuses simultaneously on biological conser-vation and human development.

Intensification. Intensive application of water,energy, and fertilizers to agricultural ecosys-tems to enhance their productivity.

Intensity. Energy released by a disturbance perunit area and time.

Interactive controls. Factors that control andrespond to ecosystem characteristics, includ-ing resource supply, modulators, major functional types of organisms, disturbanceregime, and human activities.

Interception. Contact of nutrients with rootsdue to the growth of roots to the nutrients;fraction of precipitation that does not reachthe ground (canopy interception).

Intermediate water. Middle layer of oceanwater between about 200 and 1000m depth.

Intertropical convergence zone. Region of lowpressure and rising air where surface air fromthe Northern and Southern Hemispheresconverge.

Inverse modeling. Modeling that estimates thetemporal and spatial patterns of inputsrequired to produce the observed temporaland spatial patterns of model outputs.

Inversion. Increase in atmospheric tempera-ture with height.

Inverted biomass pyramid. Biomass pyramidin which there is a smaller biomass ofprimary producers than of upper trophiclevels; typical of pelagic ecosystems of lakes,streams, and oceans.

Ionic binding. Electrostatic attraction betweenoppositely charged ions or surfaces.

Irradiance. Radiant energy flux densityreceived at a surface—that is, the quantity ofradiant energy received at a surface per unittime.

Jet stream. Strong winds over a broad heightrange in the upper troposphere.

Katabatic winds. Downslope winds that occurat night when air cools, becomes more dense,and flows downhill.

Kelvin waves. Large-scale ocean waves thattravel back and forth across the ocean.

Keystone species. Species that has a muchgreater impact on ecosystem processes thanwould be expected from its biomass;functional type represented by a singlespecies.

La Niña. Sea surface temperatures in the equa-torial Pacific Ocean associated with strongupwelling of cold water off South Americaand warm currents in the western Pacific.

Labile. Easily decomposed.Land breeze. Night breeze from the land to the

ocean caused by the higher surface tempera-ture over the ocean at night.

Landscape. Mosaic of patches that differ inecologically important properties.

Land use conversion. Human-induced changeof an ecosystem to one that is dominated bya different physical environment or differentplant functional types.

Land use modification. Human alteration of an ecosystem in ways that significantly affect ecosystem processes, community structure and population dynamics withoutchanging the physical environment or the dominant plant functional type of theecosystem.

Latent heat flux. Energy transferred between asurface and the atmosphere by the evapora-tion of water or the condensation of watervapor.

Latent heat of vaporization. Heat absorbed byevaporation or released by condensation ofwater (or of other substances) when thephase changes.

Laterite. Iron-rich layer in tropical soils thathave hardened irreversibly on exposure torepeated saturation and drying cycles; alsotermed plinthite layers.

Law of the minimum. Plant growth is limitedby a single resource at any one time; anotherresource becomes limiting only when thesupply of the first resource is increased abovethe point of limitation.

Leaching. Downward movement of materialsin solution. This can occur from the canopyto the soil, from soil organic matter to the soilsolution, from one soil horizon to another,

384 Glossary

or from the ecosystem to groundwater oraquatic ecosystems.

Leaf area index. Leaf area per unit groundarea. Projected LAI is the leaf area projectedonto a horizontal plane.Total LAI is the totalsurface area of leaves, including the upperand lower surface of flat leaves and the cylindrical surface of conifer needles; it isapproximately twice the value of projectedLAI, except in the case of conifer needles,where the projected leaf area is multiplied byp (3.1416) to get total leaf area.

Leaf mass ratio. Ratio of leaf mass to totalplant mass.

Legacy. Effect of past events on the currentfunctioning of an ecosystem.

Life history traits. Traits (e.g., seed size andnumber, potential growth rate, maximumsize, and longevity) of an organism thatdetermine how quickly a species can get to asite, how quickly it grows, how tall it gets, andhow long it survives.

Light compensation point. Light intensity atwhich net photosynthesis equals zero.

Light-harvesting reactions. Reactions of photosynthesis that transform light energyinto chemical energy.

Light saturation. Range of light intensitiesabove which the rate of photosynthesis isinsensitive to light intensity.

Light use efficiency. Ratio of gross primaryproduction to absorbed photosyntheticallyactive radiation at the leaf or ecosystemscale.

Limitation. Reduced rate of a process (e.g., netprimary production, growth or photosynthe-sis) due to inadequate supply of a resource(e.g., light) or low temperature.

Lithosphere. Hard outermost shell of Earth.Litter. Dead plant material that is sufficiently

intact to be recognizable.Litterbag. Mesh bag used to measure decom-

position rate of detritus.Litterfall. Shedding of aboveground plant

parts and death of plants.Littoral zone. Shore of a lake or ocean.Loam. Soil with substantial proportions of at

least two size classes of soil particles.Loess. Soil derived from wind-blown silt

particles.

Longwave radiation. Radiation with wave-lengths 3000 to 30,000nm.

Macrofauna. Soil animals larger than 10mm inlength.

Macronutrients. Nutrients that are required inlarge quantities by organisms.

Macropores. Large pores between soil aggre-gates that allow rapid movement of water,roots, and soil animals.

Maintenance respiration. Respiration used tosupport maintenance of live biomass.

Mantle. Fungal hyphae that surround the root in ectomycorrhizae; also termed sheath.

Mass flow. Bulk transport of solutes due to themovement of soil solution.

Mass wasting. Downslope movement of soil orrock material under the influence of gravitywithout the direct aid of other media such aswater, air, or ice.

Matric potential. Component of water poten-tial caused by adsorption of water to sur-faces; it is considered a component ofpressure potential in some treatments.

Matrix. Predominant patch type in a land-scape.

Mean residence time. Mass divided by the flux into or out of the pool over a given time period; synonymous with turnover time.

Mechanical weathering. Physical fragmenta-tion of the rock without chemical change.

Mesofauna. Soil animals 0.2 to 10mm inlength.

Mesopause. Boundary between the mesos-phere and thermosphere.

Mesophyll cells. Photosynthetic cells in a leaf.Mesosphere. Atmospheric layer between the

stratosphere and the thermosphere, which ischaracterized by a decrease in temperaturewith height.

Metamorphic rocks. Sedimentary or igneousrocks that are modified by exposure to heator pressure.

Metapopulations. Populations of a species that consist of partially isolated subpopulations.

Methanogen. Methane-producing bacteria.Methanotroph. Methane-consuming bacteria.

Glossary 385

Microbial loop. Microbial food web (includingboth plant- and detritus-based organic material) that recycles carbon and nutrientswithin the euphotic zone.

Microbial transformation. Transformation ofplant-derived substrates into microbial-derived substrates as a result of microbialturnover.

Microbivore. Organism that eats microbes.Microfauna. Soil animals less than 0.2mm in

length.Micronutrients. Nutrients that are required in

small quantities by organisms.Milankovitch cycles. Cycles of solar input

to Earth caused by regular variations inEarth’s orbit (eccentricity, tilt, and pre-cession).

Mineralization. Conversion of carbon andnutrients from organic to inorganic formsdue to the breakdown of litter and soilorganic matter. Gross mineralization is thetotal amount of nutrients released via mineralization (regardless of whether it issubsequently immobilized or not). Net mineralization is the net accumulation ofinorganic nutrients in the soil solution over agiven time interval.

Modulator. Factor that influences growth ratebut is not consumed in the growth process(e.g., temperature, ozone).

Mollisol. Soil order characterized by anorganic-rich, fertile A horizon that gradesinto a B horizon.

Monsoon. Tropical or subtropical system of airflow characterized by a seasonal shift betweenprevailing onshore and offshore winds.

Mutualism. Symbiotic relationship betweentwo species that benefits both partners.

Mycorrhizae. Symbiotic relationship betweenplant roots and fungal hyphae, in which theplant acquires nutrients from the fungus inreturn for carbohydrates that constitute themajor carbon source for the fungus.

Mycorrhizosphere. Zone of soil that is directlyinfluenced by mycorrhizal hyphae.

Negative feedback. Interaction in which twocomponents of a system have oppositeeffects on one another; this reduces the rateof change in the system.

Net biome production. Net ecosystem produc-tion at the regional scale; includes patchesthat have accumulated carbon and those thathave lost carbon through disturbance andother processes during the time period ofmeasurement.

Net ecosystem exchange. Net carbon exchangebetween the land or ocean and the atmos-phere; equals net ecosystem productionminus transport of carbon to groundwater orto deep ocean water.

Net ecosystem production. Net annual carbonaccumulation by the ecosystem.

Net primary production. Quantity of new plantmaterial produced annually (gross primaryproduction minus plant respiration); includesnew biomass, hydrocarbon emissions, rootexudates, and transfers to mycorrhizae.

Net radiation. Balance between the inputs and outputs of shortwave and longwave radiation.

Niche. Ecological role of an organism in anecosystem.

Nitrification. Conversion of ammonium tonitrate in the soil. Autotrophic nitrifiers usethe energy yield from NH4

+ oxidation to fixcarbon used in growth and maintenance,analogous to the way plants use solar energy to fix carbon via photosynthesis.Heterotrophic nitrifiers gain their energyfrom breakdown of organic matter.

Nitrogenase. Enzyme that converts dinitrogento ammonium.

Nitrogen-based defense. Plant defensive compound containing nitrogen.

Nitrogen fixation. Conversion of dinitrogengas to ammonium.

Nonoccluded phosphorus. Exchangeable pho-sphate that is loosely adsorbed to surfaces of iron and aluminum oxides or calcium carbonate.

Non–steady state mosaic. Landscape that isnot in equilibrium with the current environ-ment because large-scale disturbances causelarge proportions of the landscape to be inone or a few successional stages.

Normalized difference vegetation index. Indexof vegetation greenness.

Nutrient cycling. Mineralization and uptake ofnutrients within an ecosystem patch.

386 Glossary

Nutrient productivity. Instantaneous rate ofcarbon gain per unit nutrient.

Nutrient spiraling. Mineralization and uptakeof nutrients that occurs as dead organicmatter, dissolved nutrients, and organismsmove along a section of a stream or river.

Nutrient uptake. Nutrient absorption by plantroots.

Nutrient use efficiency. Growth per unit ofplant nutrient; ratio of nutrients to biomasslost in litterfall; also calculated as nutrientproductivity times residence time.

O horizon. Organic horizon above mineralsoil.

Occluded phosphorus. Unavailable phosphatethat is most tightly bound to oxides of ironand aluminum.

Oligotrophic. Nutrient poor.Omnivore. Organism that eats food from

several trophic levels.Orographic effects. Effects due to presence of

mountains.Osmotic potential. Component of water

potential due to the presence of substancesdissolved in water.

Overland flow. Movement of water over thesoil surface.

Oxidation. Loss of electrons by an electrondonor in oxidation–reduction reactions.

Oxisol. Soil order found in the wet tropicscharacterized by highly weathered, leachedsoils.

Oxygenase. Enzyme that catalyzes a reactionwith oxygen.

Ozone hole. Zone of destruction of stratos-pheric ozone at high southern and highnorthern latitudes. This hole allows increasedpenetration of UV radiation to Earth’ssurface.

Parent material. Rocks or other substrates thatgenerate soils through weathering.

Patch. Relatively homogeneous stand of anecosystem in a landscape.

PEP carboxylase. Initial carboxylating enzymein C4 photosynthesis.

Pelagic. Open water.Percolation. Saturated flow of water through a

soil.

Periphyton. Algae that attach to rocks, vascu-lar plants, and any other stable surfaces.

Permafrost. Permanently frozen ground—thatis, soil that remains frozen for at least 2 years.

Permanent wilting point. Water held by a soil that cannot be extracted by plant uptake.

Perturbation. An external force that displacesa system from equilibrium.

pH. Negative log of the hydrogen ion concen-tration; denotes the activity of H+ ions andthus the acidity of the system.

Phagocytosis. Consumption of material by acell by enclosing it in a membrane-boundstructure that enters the cell.

Phenology. Time course of periodic events inorganisms that are correlated with climate(e.g., budbreak).

Phloem. Long-distance transport system inplants for flow of carbohydrates and othersolutes.

Phosphatase. Enzyme that hydrolyzes phos-phate from a phosphate-containing organiccompound.

Photo-oxidation. Oxidation of compounds bylight energy; photosynthetic enzymes can bephoto-oxidized under conditions of highlight.

Photoperiod. Daylength.Photoprotection. Protection of photosynthetic

pigments from destruction by high light.Photorespiration. Production of CO2 due to

the oxygenation reaction catalyzed byRubisco.

Photosynthesis. Biochemical process that useslight energy to convert CO2 to sugars.Net photosynthesis is the net carbon input to ecosystems; synonymous at the ecosystem level with gross primary production.

Photosynthetic capacity. Photosynthetic rateper unit leaf mass measured under favorableconditions of light, moisture, and temperature.

Photosynthetic light use efficiency. Rate ofphotosynthesis per unit light.

Photosynthetic nitrogen use efficiency. Rate ofphotosynthesis per unit nitrogen.

Photosynthetically active radiation. Visiblelight; radiation with wavelengths between400 and 700nm.

Glossary 387

Phototroph. Nitrogen-fixing microorganismthat produces its own organic carbon throughphotosynthesis.

Phreatophyte. Deep-rooted plant that tapsgroundwater.

Phyllosphere decomposition. Decompositionthat occurs on leaves before leaf fall.

Phytoplankton. Microscopic algae suspendedin the surface water of aquatic ecosystems.

Pixel. Individual cell of a satellite image thatprovides a generalized spectral response forthat area.

Planetary boundary layer. The layer of theatmosphere that is directly affected by thefluxes and friction of Earth’s surface.

Planetary wave. Large (greater than 1500kmlength) wave in the atmosphere.

Plankton. Microscopic organisms suspendedin the surface water of aquatic ecosystems.

Plant-based trophic system. Plants, herbivores,and organisms that consume herbivores andtheir predators.

Plant defense. Chemical or physical propertyof plants that deters herbivores.

Plasmodesmata. Cytoplasmic connections be-tween adjacent cortical cells.

Plinthite layers. Laterite layers in tropical soils.Podzol. Spodosol.Poikilothermic. Organism whose body tem-

perature depends on the environment.Polar cell. Atmospheric circulation cell

between 60° and the pole driven by subsidence at the poles.

Polar front. Boundary between the polar andsubtropical air masses characterized by risingair and frequent storms.

Polyphenol. Soluble organic compound withmultiple phenolic groups.

Pool. Quantity of energy or material in anecosystem compartment such as plants orsoil.

Positive feedback. Interaction in which twocomponents of a system have a positive effecton the other or in which both have a nega-tive effect on one another; this amplifies therate of change in the system.

Potential biota. Organisms that are present in aregion and could potentially occupy the site.

Potential vegetation. Vegetation that wouldoccur in the absence of human disturbance.

Precession. A “wobbling” in Earth’s axis ofrotation with respect to the stars, determin-ing the date during the year when solsticesand equinoxes occur.

Precipitation. Water input to an ecosystem asrain and snow.

Pressure potential. Component of waterpotential generated by gravitational forcesand by physiological processes of organisms.

Prevailing wind. Most frequent wind direction.Primary forest. Forest that has never been

cleared.Primary minerals. Minerals present in the rock

or unconsolidated parent material beforechemical changes have taken place.

Primary producers. Organisms that convertCO2, water, and solar energy into biomass(i.e., plants); synonymous with autotroph.

Primary production. Conversion of CO2,water, and solar energy into biomass. Grossprimary production is the net carbon input to ecosystems, or the net photosynthesis ex-pressed at the ecosystem scale (gCm-2 yr-1).Net primary production is the net carbonaccumulation by vegetation (GPP minusplant respiration).

Primary succession. Succession followingsevere disturbances that remove or bury mostproducts of ecosystem processes, leaving littleor no organic matter or organisms.

Production efficiency. Proportion of assimi-lated energy that is converted to animal production, including both growth and reproduction.

Profile. Vertical cross-section of soil.Protease. Protein-hydrolyzing enzyme.Proteoid roots. Dense clusters of fine roots

produced by certain families such as the Proteaceae.

Protozoan. Single-celled animal.

Quality. Chemical nature of live or deadorganic matter that determines the ease withwhich it is broken down by herbivores ordecomposers, respectively.

Quantum yield. Moles of CO2 fixed per moleof light quanta absorbed; the initial slope ofthe light-response curve.

Quinone. Highly reactive class of compoundsproduced from polyphenols.

388 Glossary

R horizon. Unweathered bedrock at the baseof a soil profile.

Rain shadow. Zone of low precipitation down-wind of a mountain range.

Recalcitrant. Not readily decomposed.Recovery. Extent to which a system returns to

its original state following perturbation.Redfield ratio. Ratio of nitrogen to phospho-

rus (approximately 14) giving optimal growthof algae.

Radiatively active gases. Gases that absorbinfrared radiation.

Redox potential. Electrical potential of asystem due to the tendency of substances init to lose or accept electrons.

Reduction. The gain of electrons by an electron acceptor in oxidation–reductionreactions.

Regolith. Unweathered bedrock layer.Relative accumulation rate. Nutrient uptake

per unit plant nutrient.Relative growth rate. Growth per unit plant

biomass.Relative humidity. Ratio of the actual amount

of water held in the atmosphere compared to maximum that could be held at that temperature.

Release. Sudden increase in growth, whenresource availability increases in response to death or reduced growth of neighboringindividuals.

Residence time. Average time that an elementor tissue remains in a system, calculated asthe pool size divided by the input; synony-mous with turnover time.

Resilience. Rate at with which a system returnsto its reference state after a perturbation.

Resistance. Tendency of a system to remain in its reference state in the face of a perturbation.

Resorption. Withdrawal of nutrients fromtissues during their senescence.

Resorption efficiency. Proportion of themaximum leaf nutrient pool that is resorbedbefore leaf fall.

Resource. Substance that is taken up from theenvironment and consumed in growth (e.g.,light, CO2, water, nutrients).

Respiration. Biochemical process that con-verts carbohydrates into CO2 and water,

releasing energy that can be used for growthand maintenance. Respiration can be associ-ated with trophic groups (plant respiration,animal respiration, microbial respiration) or combinations of groups (heterotrophicrespiration: animal plus microbial respira-tion; ecosystem respiration: heterotrophicplus plant respiration). Alternatively, can bedefined by the way in which the resultantenergy is used (maintenance respiration,growth respiration, respiration to support ionuptake).

Response. Direction and magnitude of changein the system following a perturbation.

Rhizosphere. Zone of soil that is directly influenced by roots.

River continuum concept. Idealized transitionin ecosystem structure and function fromnarrow headwater streams to broad rivers.

Rock cycle. Formation, transformation, andweathering of rocks.

Root cap. Cells at the tips of roots that pro-duce mucilaginous carbohydrates that lubri-cate the movement of roots through soil.

Root exudation. Diffusion and secretion oforganic compounds from roots into the soil.

Root hair. Elongate epidermal cell of the rootthat extends out into the soil.

Root:shoot ratio. Ratio of root biomass toshoot biomass.

Roughness element. Obstacle to air flow (e.g., a tree) that creates mechanical turbulence.

Rubisco. Ribulose bisphosphate carboxylase;photosynthetic enzyme that catalyzes theinitial carboxylation in C3 photosynthesis.

Runoff. Water loss from an ecosystem instreams and rivers.

Saline. Salty.Salinization. Salt accumulation due to evapo-

ration of surface water.Salt flat. Depression in an arid area that

receives runoff but has no outlet.Salt lick. Mineral-rich springs or outcrops

that are used by animals as a source of minerals.

Sampling effect. Increased probability ofencountering a species with particular traits in a species-rich community due

Glossary 389

simply to the greater number of speciespresent.

Sand. Soil particles 0.05 to 2mm diameter.Saprovore. Organism that eats other live

organisms in a detritus-based food chain.Sapwood. Total quantity of functional con-

ducting tissue of the xylem.Saturated flow. Drainage of water under the

influence of gravity.Savanna. Grassland with scattered trees or

shrubs.Sea breeze. Daytime onshore breeze that

occurs on coastlines due to differentialheating of the land and water.

Secondary forest. Forest that has regrown afterearlier clearing.

Secondary metabolites. Compounds producedby plants that are not essential for normalgrowth and development.

Secondary minerals. Crystalline and amor-phous products that are formed through the reaction of materials released duringweathering.

Secondary producers. Herbivores and carnivores.

Secondary succession. Succession that occurson previously vegetated sites after a dis-turbance in which there are residual effects of organisms and organic matter from organisms present before the dis-turbance.

Sedimentary rocks. Rocks formed from sediments.

Seed bank. Seeds produced after previous disturbances that remain dormant in the soiluntil postdisturbance conditions (light, widetemperature fluctuations, and/or high soilnitrate) trigger germination.

Seedling bank. Seedlings beneath a canopythat show negligible growth beneath thedense shade of a forest canopy but growrapidly in treefall gaps.

Selective preservation. Increase in concentra-tion of recalcitrant material as a result ofdecomposition of labile substrates.

Senescence. Programmed breakdown of planttissues.

Sensible heat. Heat energy that can be sensed(e.g., by a thermometer) and involves nochange in state.

Sensible heat flux. Energy transferred betweena surface and the near-surface atmosphere by conduction and movement to the bulkatmosphere by convection.

Seston. Particles suspended in the watercolumn, including algae, bacteria, detritus,and mineral particles.

Severity. Proportion of the organic matter lostfrom the vegetation and surface soils due todisturbance.

Shade leaf. Leaf that is acclimated to shade or is produced by a plant adapted to shade.

Shifting agriculture. Clearing of forest forcrops followed by a fallow period duringwhich forests regrow, after which the cyclerepeats; synonymous with slash-and-burn orswidden agriculture.

Shifting steady-state mosaic. Landscape inwhich patches differ in successional stage, butthe landscape as a whole is at steady state(i.e., there is no directional change in the rel-ative proportions of different successionalstages).

Shortwave radiation. Radiation with wave-lengths 300 to 3000nm.

Shredder. Invertebrate that breaks leaves andother detritus into pieces and digests themicrobial jam on the surface of these particles.

Siderophore. Organic chelate produced byplant roots.

Silt. Soil particles 0.002 to 0.05mm diameter.Sink. Part of the plant that shows a net import

of a compound.Sink strength. Demand of a plant organ or

process for carbohydrates.Slash-and-burn agriculture. Shifting agricul-

ture.Slow variable. Variable that changes slowly.Snow–albedo feedback. Atmospheric warming

caused by warming-induced decrease inalbedo due to earlier snowmelt.

Soil creep. Downhill movement of soil;dubious character covered with soil.

Soil order. Major soil groupings in the U.S.soil taxonomic classification.

Soil organic matter. Dead organic matter in thesoil that has decomposed to the point that itsoriginal identity is uncertain.

390 Glossary

Soil phase. Soils belonging to the same soiltype that differ in landscape position, stoni-ness, or other soil properties.

Soil resources. Water and nutrients available inthe soil.

Soil series. Soils belonging to the same orderthat differ in profile characteristics, such asnumber and types of horizons, thickness, andhorizon properties.

Soil structure. Binding together of soil parti-cles to form aggregates.

Soil types. Soils belonging to the same soilseries but having different textures of the Ahorizon.

Solifluction. Downslope flow of saturated soilsabove a frozen layer.

Solubility pump. Downward flux of carbonfrom surface to deep waters due to the downwelling of CO2-rich North Atlantic orAntarctic waters.

Sorption. Binding of an ion to a mineralsurface, ranging from electrostatic attractionto covalent binding.

Source. Part of a plant that shows a net exportof a compound.

Southern oscillation. Atmospheric pressurechanges over the southeastern Pacific andIndian Ocean.

Specialist herbivore. Herbivore that special-izes on consumption of one or a few plantspecies or tissues.

Species composition. Identity of species in anecosystem.

Species diversity. Number, evenness, and com-position of species in an ecosystem; the totalrange of biological attributes of all speciespresent in an ecosystem.

Species evenness. Relative abundances ofspecies in an ecosystem.

Species richness. Number of species in anecosystem.

Specific heat. Energy required to warm a gramof a substance by 1°C.

Specific leaf area. Ratio of leaf area to leafmass.

Specific root length. Root length per unit rootmass.

Spiraling length. Average horizontal distancethat a nutrient moves between successiveuptake events.

Spodosol. Soil order characterized by highlyleached soils that develop in cold climates;Formerly termed podzols.

Stand-replacing disturbance. Large distur-bances that affect entire stands of vegetation.

State factors. Independent variables thatcontrol the characteristics of soils and ecosys-tems (climate, parent material, topography,potential biota, and time).

Steady state. State of a system in which incre-ments are approximately equal to losses,when averaged over a long time (e.g., theturnover time of the system); there are nodirectional changes in the major pools in asystem at steady state.

Stem flow. Water that flows down stems to theground.

Stomata. Pores in the leaf surface throughwhich water and CO2 are exchanged betweenthe leaf and the atmosphere.

Stomatal conductance. Flux of water vapor orCO2 per unit driving force between the leafand the atmosphere.

Stratopause. Boundary between the stratos-phere and the mesosphere.

Stratosphere. Atmospheric layer above thetroposphere, which is characterized by anincrease in temperature with height.

Strength of soil. Amount of force required toinitiate slope failure.

Stress. Environmental factor that reducesplant performance; physical force that pro-motes mass wasting of soils.

Stroma. Gel matrix within the chloroplast inwhich the carbon-fixation reactions occur.

Subduction. Downward movement of a platemargin beneath another plate.

Suberin. Hydrophobic waxy substance thatoccurs in the cell walls of the endodermis andexodermis of plant roots.

Sublimation. Vaporization of a solid.Subsidy. Energy or nutrient transfers from one

ecosystem to another; synonymous withallochthonous input.

Succession. Directional change in ecosystemstructure and functioning resulting from biotically driven changes in resource supply.

Sunfleck. Short period of high irradiance thatinterrupts a general background of lowdiffuse radiation.

Glossary 391

Sun leaf. Leaf that is acclimated to high light oris produced by a plant adapted to high light.

Supply rate. Rate of input of a resource (e.g.,nitrate supply rate).

Surface conductance. Potential of the leaf andsoil surfaces in the ecosystem to lose water.

Surface water. Surface layer of the oceanheated by the sun and mixed by winds, typi-cally 75 to 200m deep.

Swidden agriculture. Shifting agriculture.Systems ecology. Study of the ecosystem as a

group of components linked by fluxes ofmaterials or energy.

Taiga. Boreal forest.Teleconnections. Dynamic interactions that

interconnect distant regions of the atmos-phere.

Temporal scaling. Extrapolation of measure-ments made at one time interval to longer (oroccasionally shorter) time intervals.

Texture. Particle size distribution of soils.Thermocline. Relatively sharp vertical temper-

ature gradient in a lake or ocean.Thermohaline circulation. Global circulation

of deep and intermediate ocean watersdriven by downwelling of cold saline surfacewater off of Greenland and Antarctica.

Thermosphere. Outermost layer of the atmos-phere, which is characterized by an increasein temperature with height.

Throughfall. Water that drops from the canopyto the ground.

Thylakoids. Membrane-bound vesicles inchloroplasts in which the light-harvestingreactions of photosynthesis occur.

Tilt. Angle of Earth’s axis of rotation and theplane of its orbit around the sun.

Time step. Shortest unit of time simulated by amodel.

Top–down controls. Regulation of populationdynamics by predation.

Toposequence. Series of ecosystems that aresimilar except with respect to their topo-graphic position.

Trade winds. Easterly winds between 30° Nand S latitudes.

Transformation. Conversion of the organiccompounds contained in litter to recalcitrantorganic compounds in soil humus.

Transpiration. Water movement through stomates from plants to the atmosphere.

Transporter. Membrane-bound protein thattransports ions across cell membranes.

Trophic cascade. Top–down effect of predatorson the biomass of organisms at lower trophiclevels; results in alternation of high and lowbiomass of organisms in successive trophiclevels.

Trophic efficiency. Proportion of production of prey that is converted to production ofconsumers at the next trophic level.

Trophic interactions. Feeding relationshipsamong organisms.

Trophic level. Organisms that obtain theirenergy with the same number of stepsremoved from plants or detritus.

Trophic transfer. Flux of energy or materialsdue to consumption of one organism byanother.

Tropopause. Boundary between the tropos-phere and the stratosphere.

Troposphere. Lowest layer of the atmosphere,which is continually mixed by weathersystems and is characterized by a decrease intemperature with height.

Tundra. Ecosystem type that is too cold tosupport growth of trees.

Turbulence. State of air or water movement inwhich velocities exhibit irregular fluctuationscapable of transporting heat and materialsmuch more rapidly than by diffusion.Mechanical turbulence is caused by theuneven slowing of air by a rough surface.Convective turbulence is caused by theincreased buoyancy of surface air caused byheat transfer from the surface.

Turnover. Replacement of a pool; ratio of theflux to the pool size; lake mixing that occurswhen surface waters become more densethan deep waters.

Turnover length. Downstream distance moved in a stream while an element is inorganisms.

Turnover time. Average time that an elementspends in a system (pool/input); synonymouswith residence time.

Ultisol. Soil order characterized by substantialleaching in a warm, humid environment.

392 Glossary

Unsaturated flow. Water movement throughsoils with a water content less than fieldcapacity.

Uplift. Upward movement of Earth’s surface.Uptake. Absorption of water or mineral by an

organism or tissue.Uptake length. Average distance that an

atom moves down stream from the time it is released by mineralization until it isabsorbed again.

Upwelling. Upward movement of deep andintermediate ocean water, usually driven byoffshore winds near coasts.

Validation. Comparison of model predictionswith data.

Vapor density. Mass of water per volume ofair; absolute humidity.

Vapor pressure. Partial pressure exerted bywater molecules in the air.

Vapor pressure deficit. Difference in actualvapor pressure and the vapor pressure in airof the same temperature and pressure that issaturated with water vapor; loosely used todescribe the difference in vapor pressure inair immediately adjacent to an evaporatingsurface and the bulk atmosphere, althoughstrictly speaking the air masses are at differ-ent temperatures.

Vertisol. Soil order characterized by swellingand shrinking clays.

Vesicular arbuscular mycorrhizae. Synony-mous with arbuscular mycorrhizae.

Water-holding capacity. Difference in soilwater content between field capacity andpermanent wilting point.

Water potential. Potential energy of water relative to pure water at the soil surface.

Water saturated. All soil pores filled withwater.

Watershed. Drainage area of a stream, river, orlake leading to a single outlet for its runoff;synonymous with catchment. In England, theterm refers to a ridge that separates twodrainages.

Water use efficiency. Ratio of gross primaryproduction to water loss; also sometimes calculated as the ratio of net primary production to cumulative transpiration(growth water use efficiency).

Water vapor feedback. Additional greenhouseeffect provided by water vapor, when theatmosphere warms and increases its watervapor content.

Weathering. Processes by which parent rocksand minerals are altered to more stableforms. Physical weathering breaks rocks intosmaller fragments with greater surface area.Chemical weathering results from chemicalreactions between rock minerals and theatmosphere or water.

Westerlies. Winds that blow from the west.

Xanthophyll cycle. Transfer of absorbedenergy to xanthophyll and eventually to heatat times when electron acceptors are notavailable to transfer electrons to carbon-fixation reactions.

Xeric. Characterized by plants that are tolerant of dry conditions.

Xylem. Water-conducting tissue of plants.

Zooplankton. Microscopic animals suspendedin the surface water of aquatic ecosystems.

Color Plate I

Pla

te

1.T

he

glob

al

dist

ribu

tion

of

E

arth

’s

maj

or

biom

es

and

the

seas

onal

pat

tern

s of

mon

thly

ave

rage

tem

pera

ture

and

pre

cipi

tati

on a

ton

e re

pres

enta

tive

sit

e fo

r ea

ch b

iom

e (B

aile

y 19

98).

Clim

ate

data

are

mon

thly

av

erag

es

of

the

enti

re

peri

od

of

reco

rd

for

sele

cted

si

tes

thro

ugh

the

year

20

00

(htt

p://w

ww

.ncd

c.no

aa.g

ov/o

l/clim

ite/

rese

arch

/ghc

n/gh

cn.h

lml)

.

Color Plate II

Plate 2. The global pattern of mean annual temperature and total annual precipitation (New et al. 1999).Temperature is highest at the equator and lowest at the poles and at high elevations. (Reproduced with per-mission from the Atlas of the Biosphere <http://atlas.sage.wisc.edu>.)

Color Plate III

Plate 3. The global pattern of net primary produc-tiviy (Foley et al. 1996, Kucharik et al. 2000).The pat-terns of productivity correlates more closely withprecipitation than with temperature, including a

strong role of moisture in regulating the productiv-ity of the biosphere. (Reproduced with permissionfrom the Atlas of the Biosphere <http://atlas.sage.wisc.edu>.)

Color Plate IV

Plate 4. Global distribution of species richnessbased on observations and on model simulations,which use climate as a filter to reduce the number

of allocation strategies (Kleiden and Mooney 2000). Reprinted with permission from GlobalChange Biology.

-180 -120 -60 0 60 120 180

-90

-60

-30

0

30

60

90

ice<2% 2- 4% 4-10%10-20%20-30%30-40%40-60%60-80% 80%

Date post:	12-Sep-2021
Category:	Documents
Upload:	others
View:	3 times
Download:	0 times

Principles of terrestrial ecosystem

Documents