8.11 The Global Oxygen Cycle - University of California...

8.11The Global Oxygen CycleS. T. Petsch

University of Massachusetts, Amherst, MA, USA

8.11.1 INTRODUCTION 515

8.11.2 DISTRIBUTION OF O2 AMONG EARTH SURFACE RESERVOIRS 5168.11.2.1 The Atmosphere 5168.11.2.2 The Oceans 5168.11.2.3 Freshwater Environments 5198.11.2.4 Soils and Groundwaters 520

8.11.3 MECHANISMS OF O2 PRODUCTION 5208.11.3.1 Photosynthesis 5208.11.3.2 Photolysis of Water 522

8.11.4 MECHANISMS OF O2 CONSUMPTION 5228.11.4.1 Aerobic Cellular Respiration 5228.11.4.2 Photorespiration 5238.11.4.3 C1 Metabolism 5238.11.4.4 Inorganic Metabolism 5238.11.4.5 Macroscale Patterns of Aerobic Respiration 5238.11.4.6 Volcanic Gases 5248.11.4.7 Mineral Oxidation 5248.11.4.8 Hydrothermal Vents 5258.11.4.9 Iron and Sulfur Oxidation at the Oxic–Anoxic Transition 5258.11.4.10 Abiotic Organic Matter Oxidation 526

8.11.5 GLOBAL OXYGEN BUDGETS AND THE GLOBAL OXYGEN CYCLE 526

8.11.6 ATMOSPHERIC O2 THROUGHOUT EARTH’S HISTORY 5268.11.6.1 Early Models 5268.11.6.2 The Archean 528

8.11.6.2.1 Constraints on the O2 content of the Archean atmosphere 5288.11.6.2.2 The evolution of oxygenic photosynthesis 5308.11.6.2.3 Carbon isotope effects associated with photosynthesis 5318.11.6.2.4 Evidence for oxygenic photosynthesis in the Archean 533

8.11.6.3 The Proterozoic Atmosphere 5338.11.6.3.1 Oxygenation of the Proterozoic atmosphere 5338.11.6.3.2 Atmospheric O2 during the Mesoproterozoic 5368.11.6.3.3 Neoproterozoic atmospheric O2 536

8.11.6.4 Phanerozoic Atmospheric O2 5398.11.6.4.1 Constraints on Phanerozoic O2 variation 5398.11.6.4.2 Evidence for variations in Phanerozoic O2 5398.11.6.4.3 Numerical models of Phanerozoic oxygen concentration 542

8.11.7 CONCLUSIONS 550

REFERENCES 552

8.11.1 INTRODUCTION

One of the key defining features of Earth as aplanet that houses an active and diverse biology isthe presence of free molecular oxygen (O2) in theatmosphere. Biological, chemical, and physical

processes interacting on and beneath the Earth’ssurface determine the concentration of O2 andvariations in O2 distribution, both temporal andspatial. In the present-day Earth system,the process that releases O2 to the atmosphere(photosynthesis) and the processes that consume

515

O2 (aerobic respiration, sulfide mineral oxidation,oxidation of reduced volcanic gases) result inlarge fluxes of O2 to and from the atmosphere.Even relatively small changes in O2 productionand consumption have the potential to generatelarge shifts in atmospheric O2 concentrationwithin geologically short periods of time. Yet allavailable evidence supports the conclusion thatstasis in O2 variation is a significant feature ofthe Earth’s atmosphere over wide spans of thegeologic past. Study of the oxygen cycle istherefore important because, while an equableO2 atmosphere is central to life as we know it, ourunderstanding of exactly why O2 concentrationsremain nearly constant over large spans ofgeologic time is very limited.

This chapter begins with a review of distri-bution of O2 among various reservoirs on Earth’ssurface: air, sea, and other natural waters. The keyfactors that affect the concentration of O2 in theatmosphere and surface waters are next con-sidered, focusing on photosynthesis as the majorprocess generating free O2 and various biologicaland abiotic processes that consume O2. Thechapter ends with a synopsis of current modelson the evolution of an oxygenated atmospherethrough 4.5 billion years of Earth’s history,including geochemical evidence constrainingancient O2 concentrations and numerical modelsof atmospheric evolution.

8.11.2 DISTRIBUTION OF O2 AMONGEARTH SURFACE RESERVOIRS

8.11.2.1 The Atmosphere

The partial pressure of oxygen in the present-day Earth’s atmosphere is ,0.21 bar, correspond-ing to a total mass of ,34 £ 1018 mol O2 (0.20946bar (force/area) multiplied by the surface area ofthe Earth (5.1 £ 1014 m2), divided by averagegravitational acceleration g (9.8 m s22) and theformula weight for O2 (32 g mol21) yields,34 £ 1018 mol O2). There is a nearly uniformmixture of the main atmospheric gases (N2, O2,Ar) from the Earth’s surface up to ,80 km altitude(including the troposphere, stratosphere, andmesosphere), because turbulent mixing dominatesover molecular diffusion at these altitudes.Because atmospheric pressure (and thus gasmolecule density) decreases exponentially withaltitude, the bulk of molecular oxygen in theatmosphere is concentrated within several kilo-meters of Earth’s surface. Above this, in thethermosphere, gases become separated based ontheir densities. Molecular oxygen is photodisso-ciated by UV radiation to form atomic oxygen(O), which is the major form of oxygen above,120 km altitude.

Approximately 21% O2 in the atmosphererepresents an average composition. In spite ofwell-developed turbulent mixing in the loweratmosphere, seasonal latitudinal variations in O2

concentration of ^15 ppm have been recorded.These seasonal variations are most pronounced athigh latitudes, where seasonal cycles of primaryproduction and respiration are strongest (Keelingand Shertz, 1992). In the northern hemisphere, theseasonal variations are anticorrelated with atmos-pheric pCO2

; summers are dominated by high O2

(and high inferred net photosynthesis), whilewinters are dominated by lower O2. In addition,there has been a measurable long-term decline inatmospheric O2 concentration of ,1014 mol yr21,attributed to oxidation of fossil fuels. Thisdecrease has been detected in both long-termatmospheric monitoring stations (Keeling andShertz, 1992, Figure 1(a)) and in atmosphericgases trapped in Antarctic firn ice bubbles(Figure 1(b)). The polar ice core records extendthe range of direct monitoring of atmosphericcomposition to show that a decline in atmosphericO2 linked to oxidation of fossil fuels has beenoccurring since the Industrial Revolution (Benderet al., 1994b; Battle et al., 1996).

8.11.2.2 The Oceans

Air-saturated water has a dissolved O2

concentration dependent on temperature, theHenry’s law constant kH, and ionic strength.In pure water at 0 8C, O2 saturation is 450 mM;at 25 8C, saturation falls to 270 mM. Othersolutes reduce O2 solubility, such that at normalseawater salinities, O2 saturation is reduced by,25%. Seawater is, of course, rarely at perfectO2 saturation. Active photosynthesis may locallyincrease O2 production rates, resulting in super-saturation of O2 and degassing to the atmos-phere. Alternately, aerobic respiration below thesea surface can consume dissolved O2 and leadto severe O2-depletion or even anoxia.

Lateral and vertical gradients in dissolved O2

concentration in seawater reflect balancesbetween O2 inputs from air–sea gas exchange,biological processes of O2 production and con-sumption, and advection of water masses. Ingeneral terms, the concentration of O2 with depthin the open ocean follows the general structuresdescribed in Figure 2. Seawater is saturated tosupersaturated with O2 in the surface mixed layer(,0–60 m water depth). Air–sea gas exchangeand trapping of bubbles ensures constant dissol-ution of atmospheric O2. Because gas solubility istemperature dependent, O2 concentrations aregreater in colder high-latitude surface watersthan in waters near the equator. Oxygen concen-trations in surface waters also vary strongly with

The Global Oxygen Cycle516

season, especially in high productivity waters.Supersaturation is strongest in spring and summer(time of greatest productivity and strongest watercolumn stratification) when warming of surfacelayers creates a shallow density gradient thatinhibits vertical mixing. Photosynthetic O2 pro-duction exceeds consumption and exchange, andsupersaturation can develop. O2 concentrationsdrop below the surface mixed layer to form O2

minimum zones (OMZs) in many ocean basins.O2 minima form where biological consumption ofO2 exceeds resupply through advection anddiffusion. The depth and thickness of O2 minimavary among ocean basins. In the North Atlantic,the OMZ extends several hundred meters. O2

concentrations fall from an average of ,300 mMin the surface mixed layer to ,160 mM at 800 mdepth. In the North Pacific, however, the O2

minimum extends deeper, and O2 concentrationsfall to ,100 mM. Along the edges of oceanbasins, where OMZs impinge on the seafloor,aerobic respiration is restricted, sediments areanoxic at or near the sediment–water interface,

and burial of organic matter in sediments may beenhanced. Below the oxygen minima zones in theopen ocean, O2 concentrations gradually increaseagain from 2000 m to the seafloor. This increase inO2 results from the slow progress of globalthermohaline circulation. Cold, air-saturated sea-water sinks to the ocean depths at high-latitudes inthe Atlantic, advecting in O2-rich waters belowthe O2 minimum there. Advection of O2-rich deepwater from the Atlantic through the Indian Oceaninto the Pacific is the source of O2 in deep Pacificwaters. However, biological utilization of thisdeep-water O2 occurs along the entire path fromthe North Atlantic to the Pacific. For this reason,O2 concentrations in deep Atlantic water areslightly greater (,200 mM) than in the deepPacific (,150 mM).

In some regions, dissolved O2 concentrationfalls to zero. In these regions, restricted watercirculation and ample organic matter supply resultin biological utilization of oxygen at a rate thatexceeds O2 resupply through advection anddiffusion. Many of these are temporary zones of

0

–20

–60

–80

–100

–120

–140

–160in situ dataFlask data

La Jolla (32.9˚ N. 117.3˚W)

Scripps dataFitted curves

–180

–200

–220

1989(a)

(b)

1990 1991 1992

3,000

dO2/

N2

(meg

–1)

dO2/

N2

(meg

–1)

2,000

1,000

0

1401201008060

Depth (mbs)

40200

Jan Apr Jul Oct Jan Jul Oct JanApr Jul Oct JanApr Jul Oct JanApr

–40

Figure 1 (a) Interannual variability of atmospheric O2/N2 ratio, measured at La Jolla, California. 1 ppm O2 isequivalent to ,4.8 meg21 (source Keeling and Shertz, 1992). (b) Variability in atmospheric O2/N2 ratio measuredin firn ice at the South Pole, as a function of depth in meters below surface (mbs). The gentle rise in O2/N2 between40 m and 100 m reflects a loss of atmospheric O2 during the last several centuries due to fossil fuel burning.Deeper than 100 m, selective effusion of O2 out of closing bubbles into firn air artificially boosts O2/N2 ratios

(source Battle et al., 1996).

Distribution of O2 among Earth Surface Reservoirs 517

anoxia that form in coastal regions during summer,when warming facilitates greatest water columnstratification, and primary production and organicmatter supply are high. Such O2-depletion is nowcommon in the Chesapeake Bay and Scheldeestuaries, off the mouth of the Mississippi Riverand other coastal settings. However, there areseveral regions of the world oceans wherestratification and anoxia are more permanentfeatures (Figure 3). These include narrow, deep,

and silled coastal fjords, larger restricted basins(e.g., the Black Sea, Cariaco Basin, and thechain of basins along the southern CaliforniaBorderlands). Lastly, several regions of the openocean are also associated with strong O2-depletion.These regions (the equatorial Pacific along Centraland South America and the Arabian Sea) areassociated with deep-water upwelling, high rates ofsurface water primary productivity, and highdissolved oxygen demand in intermediate waters.

Pacific Atlantic0

1,000

8˚ N

12 N

50 N 53˚ N

45˚ S

45˚ S

Dep

th (

m)

Dep

th (

m) 2,000

3,000

4,000

5,000

0

1,000

2,000

3,000

4,000

5,0000 2 4 6 8 100 2 4 6 8 10

Oxygen concentration (mL L–1) Oxygen concentration (mL L–1)

(a) (b)

Figure 2 Depth profiles of oxygen concentration dissolved in seawater for several latitudes in the Pacific(a) and Atlantic (b). Broad trends of saturation or supersaturation at the surface, high dissolved oxygen demand atmid-depths, and replenishment of O2 through lateral advection of recharged deep water are revealed, althoughregional influences of productivity and intermediate and deep-water heterotropy are also seen (source Ingmanson

and Wallace, 1989).

Figure 3 Map detailing locations of extensive and permanent oxygen-deficient intermediate and deep waters(Deuser, 1975) (reproduced by permission of Elsevier from Chemical Oceanography 1975, p.3).


Oxygen concentrations in the pore fluids ofsediments are controlled by a balance of entrain-ment of overlying fluids during sediment depo-sition, diffusive exchange between the sedimentand the water column, and biological utilization. Inmarine sediments, there is a good correlationbetween the rate of organic matter supply and thedepth of O2 penetration in the sediment (Hartnettet al., 1998). In coastal sediments and on thecontinental shelf, burial of organic matter issufficiently rapid to deplete the sediment of oxygenwithin millimeters to centimeters of the sediment–water interface. In deeper abyssal sediments,where organic matter delivery is greatly reduced,O2 may penetrate several meters into the sedimentbefore being entirely consumed by respiration.

There is close coupling between surface waterand atmospheric O2 concentrations and air–seagas exchange fluxes (Figure 4). High rates ofmarine primary productivity result in net out-gassing of O2 from the oceans to the atmosphere inspring and summer, and net ingassing of O2 duringfall and winter. These patterns of air–sea O2

transfer relate to latitude and season: outgassing ofO2 during northern hemisphere high productivitymonths (April through August) are accompaniedby simultaneous ingassing in southern latitudes

when and where the productivity is lowest(Najjar and Keeling, 2000). Low-latitude oceansurface waters show very little net air–sea O2

exchange and minimal change in outgassing oringassing over an annual cycle.

8.11.2.3 Freshwater Environments

Oxygen concentrations in flowing freshwaterenvironments closely match air-saturated values,due to turbulent mixing and entrainment of air bub-bles. In static water bodies, however, O2-depletioncan develop much like in the oceans. This isparticularly apparent in some ice-covered lakes,where inhibited gas exchange and wintertimerespiration can result in O2-depletion and fishkills. High productivity during spring and summerin shallow turbid aquatic environments can resultin extremely sharp gradients from strong O2

supersaturation at the surface to near O2-depletionwithin a few meters of the surface. The highconcentration of labile dissolved and particulateorganic matter in many freshwater environmentsleads to rapid O2-depletion where advectiveresupply is limited. High rates of O2 consumptionhave been measured in many temperature andtropical rivers.

90

60

0

–20

–20

20

2020

10

1010

–10

–10

0

0

0

0

0

0

0

30

0

Lat

itude

–30

–60

–90J F M A M

Sea-to-air oxygen flux (mol m–2 yr–1)zonal average of world ocean

Month

J J A S O N D

–10–20

Figure 4 Zonal average monthly sea-to-air oxygen flux for world oceans. Outgassing and ingassing areconcentrated at mid- to high latitudes. Outgassing of oxygen is strongest when primary production rates are greatest;ingassing is at a maximum during net respiration. These patterns oscillate during an annual cycle from northern to

southern hemisphere (source Najjar and Keeling, 2000).

Distribution of O2 among Earth Surface Reservoirs 519

8.11.2.4 Soils and Groundwaters

In soil waters, oxygen concentrations dependon gas diffusion through soil pore spaces,infiltration and advection of rainwater andgroundwater, air–gas exchange, and respirationof soil organic matter (see review by Hinkle,1994). In organic-matter-rich temperate soils,dissolved O2 concentrations are reduced, but notentirely depleted. Thus, many temperate shallowgroundwaters contain some dissolved oxygen.Deeper groundwaters, and water-saturated soilsand wetlands, generally contain little dissolvedO2. High-latitude mineral soils and groundwaterscontain more dissolved O2 (due to lowertemperature and lesser amounts of soil organicmatter and biological O2 demand). Dry tropicalsoils are oxidized to great depths, with dissolvedO2 concentration less than air saturated, but notanoxic. Wet tropical forests, however, mayexperience significant O2-depletion as rapidoxidation of leaf litter and humus occurs nearthe soil surface. Soil permeability also influencesO2 content, with more clay-rich soils exhibitinglower O2 concentrations.

In certain environments, localized anomalouslylow concentrations of soil O2 have been used byexploration geologists to indicate the presence of alarge body of chemically reduced metal sulfides inthe subsurface. Oxidation of sulfide mineralsduring weathering and soil formation drawsdown soil gas pO2

below regional average.Oxidation of sulfide minerals generates solid andaqueous-phase oxidation products (i.e., sulfate

anion and ferric oxyhydroxides in the case ofpyrite oxidation). In some instances, the volumeof gaseous O2 consumed during mineral oxidationgenerates a mild negative pressure gradient,drawing air into soils above sites of sulfidemineral oxidation (Lovell, 2000).

8.11.3 MECHANISMS OF O2 PRODUCTION

8.11.3.1 Photosynthesis

The major mechanism by which molecularoxygen is produced on Earth is through thebiological process of photosynthesis. Photosyn-thesis occurs in higher plants, the eukaryoticprotists collectively called algae, and in twogroups of prokaryotes: the cyanobacteria and theprochlorophytes. In simplest terms, photosyn-thesis is the harnessing of light energy tochemically reduce carbon dioxide to simpleorganic compounds (e.g., glucose). The overallreaction (Equation (1)) for photosynthesis showscarbon dioxide and water reacting to produceoxygen and carbohydrate:

6CO2 þ 6H2O ! 6O2 þ C6H12O6 ð1Þ

Photosynthesis is actually a two-stage process,with each stage broken into a cascade of chemicalreactions (Figure 5). In the light reactions ofphotosynthesis, light energy is converted tochemical energy that is used to dissociate waterto yield oxygen and hydrogen and to form thereductant NADPH from NADPþ.

ATP

ADP

NADPHLight

reactions

Light

NADP

Calvincycle

CO2

H2OO2 + e–

S0 H2SH2

S2O32–

SO42–

H+

S0

Sugars

Figure 5 The two stages of photosynthesis: light reaction and the Calvin cycle. During oxygenic photosynthesis,H2O is used as an electron source. Organisms capable of anoxygenic photosynthesis can use a variety of otherelectron sources (H2S, H2, S0, S2O3

22) during the light reactions, and do not liberate free O2. Energy in the form ofATP and reducing power in the form of NADPH are produced by the light reactions, and subsequently used in the

Calvin cycle to deliver electrons to CO2 to produce sugars.


The next stage of photosynthesis, theCalvin cycle, uses NADPH to reduce CO2 tophosphoglyceraldehyde, the precursor for a varietyof metabolic pathways, including glucose syn-thesis. In higher plants and algae, the Calvin cycleoperates in special organelles called chloroplasts.However, in bacteria the Calvin cycle occursthroughout the cytosol. The enzyme ribulose-1,5-biphosphate carboxylase (rubisco) catalyzesreduction of CO2 to phosphoglycerate, which iscarried through a chain of reactions that consumeATP and NADPH and eventually yield phospho-glyceraldehyde. Most higher plants are termed C3

plants, because the first stable intermediate formedduring the carbon cycle is a three-carbon com-pound. Several thousand species of plant, spreadamong at least 17 families including the grasses,precede the carbon cycle with a CO2-concentratingmechanism which delivers a four-carbon com-pound to the site of the Calvin cycle and rubisco.These are the C4 plants. This four-carbon com-pound breaks down inside the chloroplasts, supply-ing CO2 for rubisco and the Calvin cycle. TheC4-concentrating mechanisms is an advantage inhot and dry environments where leaf stomata arepartially closed, and internal leaf CO2 concen-trations are too low for rubisco to efficientlycapture CO2. Other plants, called CAM plants,which have adapted to dry climates utilize anotherCO2-concentrating mechanisms by closing sto-mata during the day and concentrating CO2 atnight. All higher plants, however, produce O2

and NADPH from splitting water, and use theCalvin cycle to produce carbohydrates. Someprokaryotes use mechanisms other than the Calvincycle to fix CO2 (i.e., the acetyl-CoA pathway orthe reductive tricarboxylic acid pathway), but noneof these organisms is involved in oxygenicphotosynthesis.

Global net primary production estimates havebeen derived from variations in the abundanceand isotopic composition of atmospheric O2.These estimates range from 23 £ 1015 mol yr21

(Keeling and Shertz, 1992) to 26 £ 1015 mol yr21,distributed between 14 £ 1015 mol yr21 O2

production from terrestrial primary productionand 12 £ 1015 mol yr21 from marine primaryproduction (Bender et al., 1994a). It is estimatedthat ,50% of all photosynthetic fixation of CO2

occurs in marine surface waters. Collectively,free-floating photosynthetic microorganisms arecalled phytoplankton. These include the algaleukaryotes (dinoflagellates, diatoms, and thered, green, brown, and golden algae), variousspecies of cyanobacteria (Synechococcus andTrichodesmium), and the common prochlorophyteProchlorococcus. Using the stoichiometry of thephotosynthesis reaction, this equates to half ofall global photosynthetic oxygen productionresulting from marine primary production.

Satellite-based measurements of seasonal andyearly average chlorophyll abundance (for marinesystems) and vegetation greenness (for terrestrialecosystems) can be applied to models thatestimate net primary productivity, CO2 fixation,and O2 production. In the oceans, there aresignificant regional and seasonal variations inphotosynthesis that result from limitations bylight, nutrients, and temperature. Yearly averagesof marine chlorophyll abundance show concen-trated primary production at high-latitudes inthe North Atlantic, North Pacific, and coastalAntarctica, in regions of seawater upwelling offof the west coasts of Africa, South America, andthe Arabian Sea, and along the Southern Sub-tropical Convergence. At mid- and high latitudes,marine productivity is strongly seasonal, withprimary production concentrated in spring andsummer. At low latitudes, marine primaryproduction is lower and varies little with seasonor region. On land, primary production alsoexhibits strong regional and seasonal patterns.Primary production rates (in g C m22 yr21) aregreatest year-round in the tropics. Tropicalforests in South America, Africa, and SoutheastAsia are the most productive ecosystems onEarth. Mid-latitude temperate forests and high-latitude boreal forests are also highly productive,with a strong seasonal cycle of greatest pro-duction in spring and summer. Deserts (concen-trated at ,308N and S) and polar regions are lessproductive. These features of seasonal variabilityin primary production on land and in the oceansare clearly seen in the seasonal variations inatmospheric O2 (Figure 4).

Transfer of O2 between the atmosphere andsurface seawater is controlled by air–sea gasexchange. Dissolved gas concentrations trendtowards thermodynamic equilibrium, but otherfactors may complicate dissolved O2 concen-trations. Degassing of supersaturated waters canonly occur at the very surface of the water. Thus, inregions of high primary production, concen-trations of O2 can accumulate in excess of therate of O2 degassing. In calm seas, where theair–sea interface is a smooth surface, gas exchangeis very limited. As seas become more rough, andespecially during storms, gas exchange is greatlyenhanced. This is in part because of entrainment ofbubbles dispersed in seawater and water dropletsentrained in air, which provide much more surfacefor dissolution or degassing. Also, gas exchangedepends on diffusion across a boundary layer.According to Fick’s law, the diffusive flux dependson both the concentration gradient (degree ofsuper- or undersaturation) and the thickness ofthe boundary layer. Empirically it is observed thatthe boundary layer thickness decreases withincreasing wind speed, thus enhancing diffusionand gas exchange during high winds.

Mechanisms of O2 Production 521

8.11.3.2 Photolysis of Water

In the upper atmosphere today, a small amountof O2 is produced through photolysis of watervapor. This process is the sole source of O2 to theatmospheres on the icy moons of Jupiter(Ganymede and Europa), where trace concen-trations of O2 have been detected (Vidal et al.,1997). Water vapor photolysis may also have beenthe source of O2 to the early Earth before theevolution of oxygenic photosynthesis. However,the oxygen formed by photolysis would have beenthrough reactions with methane and carbonmonoxide, preventing any accumulation in theatmosphere (Kasting et al., 2001).

8.11.4 MECHANISMS OF O2 CONSUMPTION

8.11.4.1 Aerobic Cellular Respiration

In simple terms, aerobic respiration is theoxidation of organic substrates with oxygen toyield chemical energy in the form of ATP andNADH. In eukaryotic cells, the respiration path-way follows three steps (Figure 6). Glycolysisoccurs throughout the cytosol, splitting glucoseinto pyruvic acid and yielding some ATP.Glycolysis does not directly require free O2, andthus occurs among aerobic and anaerobic organ-isms. The Krebs citric acid (tricarboxylic acid)cycle and oxidative phosphorylation are localizedwithin the mitochondria of eukaryotes, and alongthe cell membranes of prokaryotes. The Krebscycle completes the oxidation of pyruvate to CO2

from glycolysis, and together glycolysis and theKrebs cycle provide chemical energy and reduc-tants (in the form of ATP, NADH, and FADH2)for the third step—oxidative phosphorylation.

Oxidative phosphorylation involves the transferof electrons from NADH and FADH2 through acascade of electron carrying compounds tomolecular oxygen. Compounds used in theelectron transport chain of oxidative phosphoryl-ation include a variety of flavoproteins, quinones,Fe–S proteins, and cytochromes. Transfer ofelectrons from NADH to O2 releases considerableenergy, which is used to generate a protongradient across the mitochondrial membraneand fuel significant ATP synthesis. In part, thisgradient is created by reduction of O2 to H2Oas the last step of oxidative phosphorylation.While eukaryotes and many prokaryotes use theKrebs cycle to oxidize pyruvate to CO2, thereare other pathways as well. For example, prokar-yotes use the glyoxalate cycle to metabolizefatty acids. Several aerobic prokaryotes also canuse the Entner–Doudoroff pathway in place ofnormal glycolysis. This reaction still producespyruvate, but yields less energy in the form ofATP and NADH.

Glycolysis, the Krebs cycle, and oxidativephosphorylation are found in all eukaryotes(animals, plants, and fungi) and many of theaerobic prokaryotes. The purpose of these reactionpathways is to oxidize carbohydrates with O2,yielding CO2, H2O, and chemical energy in theform of ATP. While macrofauna generally requirea minimum of ,0.05–0.1 bar (,10 mM) O2 tosurvive, many prokaryotic microaerophilic organ-isms can survive and thrive at much lower O2

concentrations. Because most biologicallymediated oxidation processes occur through theactivity of aerotolerant microorganisms, it isunlikely that a strict coupling between limitedatmospheric O2 concentration and limited globalrespiration rates could exist.

Figure 6 The three components of aerobic respiration: glycolysis, the Krebs cycle, and oxidative phosphorylation.Sugars are used to generate energy in the form of ATP during glycolysis. The product of glycolyis, pyruvate, isconverted to acetyl-CoA, and enters the Krebs cycle. CO2, stored energy as ATP, and stored reducing power asNADH and FADH2 are generated in the Krebs cycle. O2 is only directly consumed during oxidative phosphorylation

to generate ATP as the final component of aerobic respiration.


8.11.4.2 Photorespiration

The active site of rubisco, the key enzymeinvolved in photosynthesis, can accept either CO2

or O2. Thus, O2 is a competitive inhibitor ofphotosynthesis. This process is known as photo-respiration, and involves addition of O2 toribulose-biphosphate. Products of this reactionenter a metabolic pathway that eventuallyproduces CO2. Unlike cellular respiration, photo-respiration generates no ATP, but it does consumeO2. In some plants, as much as 50% of the carbonfixed by the Calvin cycle is respired throughphotorespiration. Photorespiration is enhanced inhot, dry environments when plant cells closestomata to slow water loss, CO2 is depleted and O2

accumulates. Photorespiration does not occur inprokaryotes, because of the much lower relativeconcentration of O2 versus CO2 in water com-pared with air.

8.11.4.3 C1 Metabolism

Beyond metabolism of carbohydrates, there areseveral other biological processes common inprokaryotes that consume oxygen. For example,methylotrophic organisms can metabolize C1

compounds such as methane, methanol, formal-dehyde, and formate, as in

CH4þNADHþHþþO2!CH3OHþNADþþH2O

CH3OHþPQQ!CH2OþPQQH2

CH2OþNADþþH2O!HCOOHþNADHþHþ

HCOOHþNADþ!CO2þNADHþHþ

These compounds are common in soils andsediments as the products of anaerobic fermenta-tion reactions. Metabolism of these compoundscan directly consume O2 (through monooxygen-ase enzymes) or indirectly, through formation ofNADH which is shuttled into oxidative phos-phorylation and the electron transport chain.Oxidative metabolism of C1 compounds is animportant microbial process in soils and sedi-ments, consuming the methane produced bymethanogenesis.

8.11.4.4 Inorganic Metabolism

Chemolithotrophic microorganisms are thosethat oxidize inorganic compounds rather thanorganic substrates as a source of energy andelectrons. Many of the chemolithotrophs are also

autotrophs, meaning they reduce CO2 to generatecellular carbon in addition to oxidizing inorganiccompounds. In these organisms, CO2 fixation isnot tied to O2 production by the stoichiometry ofphotosynthesis. Hydrogen-oxidizing bacteriaoccur wherever both O2 and H2 are available.While some H2 produced by fermentation inanoxic environments (deep soils and sediments)may escape upward into aerobic environments,most H2 utilized by hydrogen-oxidizing bacteriaderives from nitrogen fixation associated withnitrogen-fixing plants and cyanobacteria. Nitrify-ing bacteria are obligate autotrophs that oxidizeammonia. Ammonia is produced in many environ-ments during fermentation of nitrogen compoundsand by dissimilatory nitrate reduction. Nitrifyingbacteria are common at oxic–anoxic interfaces insoils, sediments, and the water column. Otherchemolithoautotrophic bacteria can oxidizenitrite. Non-photosynthetic bacteria that canoxidize reduced sulfur compounds form a diversegroup. Some are acidophiles, associated withsulfide mineral oxidation and tolerant of extre-mely low pH. Others are neutrophilic and occur inmany marine sediments. These organisms canutilize a wide range of sulfur compounds producedin anaerobic environments, including H2S, thio-sulfate, polythionates, polysulfide, elemental sul-fur, and sulfite. Many of the sulfur-oxidizingbacteria are also autotrophs. Some bacteria canlive as chemoautotrophs through the oxidation offerrous-iron. Some of these are acidophilesgrowing during mining and weathering of sulfideminerals. However, neutrophilic iron-oxidizingbacteria have also been detected associated withthe metal sulfide plumes and precipitates at mid-ocean ridges. Other redox-sensitive metals thatcan provide a substrate for oxidation includemanganese, copper, uranium, arsenic, and chro-mium. As a group, chemolithoautotrophic micro-organisms represent a substantial flux of O2

consumption and CO2 fixation in many commonmarine and terrestrial environments. Because thenet reaction of chemolithoautotrophy involvesboth O2 and CO2 reduction, primary productionresulting from chemoautotrophs has a verydifferent O2/CO2 stoichiometry than does photo-synthesis. Chemolithoautotrophs use the electronflow from reduced substrates (metals, H2, andreduced sulfur) to O2 to generate ATP andNAD(P)H, which in turn are use for CO2-fixation.It is believed that most aerobic chemoautotrophsutilize the Calvin cycle for CO2-fixation.

8.11.4.5 Macroscale Patterns of AerobicRespiration

On a global scale, much biological O2 con-sumption is concentrated where O2 is abundant.

Mechanisms of O2 Consumption 523

This includes surficial terrestrial ecosystems andmarine surface waters. A large fraction ofterrestrial primary production is consumed byaerobic degradation mechanisms. Although mostof this is through aerobic respiration, somefraction of aerobic degradation of organicmatter depends on anaerobic breakdown oflarger biomolecules into smaller C1 compounds,which, if transported into aerobic zones of soilor sediment, can be degraded by aerobic C1

metabolizing microorganisms. Partially degradedterrestrial primary production can be incorporatedinto soils, which slowly are degraded and eroded.Research has shown that soil organic matter (OM)can be preserved for up to several millennia, andriverine export of aged terrestrial OM may be asignificant source of dissolved and particulateorganic carbon to the oceans.

Aside from select restricted basins and special-ized environments, most of the marine watercolumn is oxygenated. Thus, aerobic respirationdominates in open water settings. Sediment trapand particle flux studies have shown that sub-stantial fractions of marine primary production arecompletely degraded (remineralized) prior todeposition at the sea floor, and thus by impli-cation, aerobic respiration generates an O2 con-sumption demand nearly equal to the release ofoxygen associated with photosynthesis. The bulkof marine aerobic respiration occurs within thewater column. This is because diffusion andmixing (and thus O2 resupply) are much greaterin open water than through sediment pore fluids,and because substrates for aerobic respiration (andthus O2 demand) are much less concentrated in thewater column than in the sediments. O2 consump-tion during aerobic respiration in the watercolumn, coupled with movement of deep-water masses from O2-charged sites of deep-water formation, generates the vertical and lateralprofiles of dissolved O2 concentration observed inseawater.

In most marine settings, O2 does not penetratevery far into the sediment. Under regions of highprimary productivity and limited water columnmixing, even if the water column is oxygenated,O2 may penetrate 1 mm or less, limiting theamount of aerobic respiration that occurs in thesediment. In bioturbated coastal sediments, O2

penetration is facilitated by the recharging of porefluids through organisms that pump overlyingwater into burrows, reaching several centimetersinto the sediment in places. However, patterns ofoxic and anoxic sediment exhibit a great degreeof spatial and temporal complexity as a result ofspotty burrow distributions, radial diffusion of O2

from burrows, and continual excavation andinfilling through time. Conversely, in deep-sea(pelagic) sediments, where organic matter deliv-ery is minor and waters are cold and charged with

O2 from sites of deep-water formation, O2 maypenetrate uniformly 1 m or more into thesediment.

8.11.4.6 Volcanic Gases

Gases emitted from active volcanoes andfumaroles are charged with reduced gases,including CO, H2, SO2, H2S, and CH4. Duringexplosive volcanic eruptions, these gases areejected high into the atmosphere along withH2O, CO2, and volcanic ash. Even the relativelygentle eruption of low-silicon, low-viscosity-shield volcanoes is associated with the release ofreduced volcanic gases. Similarly, reduced gasesare released dissolved in waters associated withhot springs and geysers. Oxidation of reducedvolcanic gases occurs in the atmosphere, in naturalwaters, and on the surfaces of minerals. This ispredominantly an abiotic process, although manychemolithoautotrophs have colonized the wallsand channels of hot springs and fumaroles,catalyzing the oxidation of reduced gases withO2. Much of biological diversity in hyperthermo-philic environments consists of prokaryotesemploying these unusual metabolic types.

Recent estimates of global average volcanic gasemissions suggest that volcanic sulfur emissionsrange, 0.1–1 £ 1012 mol S yr21, is nearly equallydistributed between SO2 and H2S (Halmer et al.,2002; Arthur, 2000). This range agrees well withthe estimates used by both Holland (2002) andLasaga and Ohmoto (2002) for average volcanic Semissions through geologic time. Other reducedgas emissions (CO, CH4, and H2) are estimated tobe similar in magnitude (Morner and Etiope,2002; Arthur, 2000; Delmelle and Stix, 2000). Allof these gases have very short residence times inthe atmosphere, revealing that emission andoxidative consumption of these gases are closelycoupled, and that O2 consumption through volca-nic gas oxidation is very efficient.

8.11.4.7 Mineral Oxidation

During uplift and erosion of the Earth’scontinents, rocks containing chemically reducedminerals become exposed to the oxidizingconditions of the atmosphere. Common rock-forming minerals susceptible to oxidation includeolivine, Fe2þ-bearing pyroxenes and amphiboles,metal sulfides, and graphite. Ferrous-iron oxi-dation is a common feature of soil formation.Iron oxides derived from oxidation of Fe2þ in theparent rock accumulate in the B horizon oftemperate soils, and extensive laterites consistingof iron and aluminum oxides develop in tropicalsoils to many meters of depth. Where erosion


rates are high, iron-bearing silicate minerals maybe transported short distances in rivers; however,iron oxidation is so efficient that very fewsediments show deposition of clastic ferrous-iron minerals. Sulfide minerals are extremelysusceptible to oxidation, often being completelyweathered from near-surface rocks. Oxidation ofsulfide minerals generates appreciable acidity,and in areas where mining has brought sulfideminerals in contact with the atmosphere or O2-charged rainwater, low-pH discharge has becomea serious environmental problem. Althoughferrous silicates and sulfide minerals such aspyrite will oxidize under sterile conditions, agrowing body of evidence suggests that in manynatural environments, iron and sulfur oxidation ismediated by chemolithoautotrophic microorgan-isms. Prokaryotes with chemolithoautotrophicmetabolic pathways have been isolated frommany environments where iron and sulfuroxidation occurs. The amount of O2 consumedannually through oxidation of Fe2þ and sulfur-bearing minerals is not known, but based onsulfur isotope mass balance constraints is on theorder of (0.1–1) £ 1012 mol yr21 each for ironand sulfur, similar in magnitude to the flux ofreduced gases from volcanism.

8.11.4.8 Hydrothermal Vents

The spreading of lithospheric plates along mid-ocean ridges is associated with much underseavolcanic eruptions and release of chemicallyreduced metal sulfides. Volcanic gases releasedby subaerial volcanoes are also generated bysubmarine eruptions, contributing dissolvedreduced gases to seawater. Extrusive lava flowsgenerate pillow basalts, which are composed inpart of ferrous-iron silicates. Within concentratedzones of hydrothermal fluid flow, fracture-fillingand massive sulfide minerals precipitate withinthe pillow lavas, large chimneys grow from theseafloor by rapid precipitation of Fe–Cu–Znsulfides, and metal sulfide-rich plumes of high-temperature “black smoke” are released intoseawater. In cooler zones, more gradual emana-tions of metal and sulfide rich fluid diffuse upwardthrough the pillow basalts and slowly mix withseawater. Convective cells develop, in which coldseawater is drawn down into pillow lavas offthe ridge axis to replace the water released at thehydrothermal vents. Reaction of seawater withbasalt serves to alter the basalt. Some sulfide isliberated from the basalt, entrained seawatersulfate precipitates as anhydrite or is reduced tosulfide, and Fe2þ and other metals in basalt arereplaced with seawater-derived magnesium. Oxy-gen dissolved in seawater is consumed duringalteration of seafloor basalts. Altered basalts

containing oxidized iron-mineral can extend asmuch as 500 m below the seafloor. Much more O2

is consumed during oxidation of black smoker andchimney sulfides. Chimney sulfide may be onlypartially oxidized prior to transport off-axisthrough spreading and burial by sediments.However, black smoker metal-rich fluids are fairlyrapidly oxidized in seawater, forming insolubleiron and manganese oxides, which slowly settleout on the seafloor, generating metalliferoussediments.

Because of the diffuse nature of reducedspecies in hydrothermal fluids, it is not knownwhat role marine chemolithoautotrophic micro-organisms may play in O2 consumption associ-ated with fluid plumes. Certainly such organismslive within the walls and rubble of coolerchimneys and basalts undergoing seafloor weath-ering. Because of the great length of mid-oceanridges throughout the oceans, and the abundanceof pyrite and other metal sulfides associated withthese ridges, colonization and oxidation of metalsulfides by chemolithoautotrophic organisms mayform an unrecognized source of primaryproduction associated with consumption, not netrelease, of O2.

8.11.4.9 Iron and Sulfur Oxidation atthe Oxic–Anoxic Transition

In restricted marine basins underlying highlyproductive surface waters, where consumption ofO2 through aerobic respiration near the surfaceallows anoxia to at least episodically extendbeyond the sediment into the water column,oxidation of reduced sulfur and iron may occurwhen O2 is present. Within the Black Sea andmany anoxic coastal fjords, the transition fromoxic to anoxic environments occurs within thewater column. At other locations, such as themodern Peru Shelf and coastal California basins,the transition occurs right at the sediment–waterinterface. In these environments, small concen-trations of sulfide and O2 may coexist within avery narrow band where sulfide oxidation occurs.In such environments, appreciable sulfur recy-cling may occur, with processes of sulfatereduction, sulfide oxidation, and sulfur dispropor-tionation acting within millimeters of eachother. These reactions are highly mediated bymicroorganisms, as evidenced by extensivemats of chemolithoautotrophic sulfur-oxidizingBeggiatoa and Thioploca found where the oxic–anoxic interface and seafloor coincide.

Similary, in freshwater and non-sulfidic brack-ish environments with strong O2 demand,dissolved ferrous-iron may accumulate ingroundwaters and anoxic bottom waters. Signifi-cant iron oxidation will occur where the water

Mechanisms of O2 Consumption 525

table outcrops with the land surface (i.e., ground-water outflow into a stream), or in lakes andestuaries, at the oxic–anoxic transition within thewater column. Insoluble iron oxides precipitateand settle down to the sediment. Recycling of ironmay occur if sufficient organic matter exists forferric-iron reduction to ferrous-iron.

The net effect of iron and sulfur recycling onatmospheric O2 is difficult to constrain. In mostcases, oxidation of sulfur or iron consumes O2

(there are some anaerobic chemolithoautotrophicmicroorganisms that can oxidize reduced sub-strates using nitrate or sulfate as the electronacceptor). Reduction of sulfate or ferric-iron arealmost entirely biological processes, coupled withthe oxidation of organic matter; sulfate andferric-iron reduction individually have no effecton O2. However, the major source of organicsubstrates for sulfate and ferric-iron reduction isultimately biomass derived from photosyntheticorganisms, which is associated with O2 gener-ation. The net change derived from summingthe three processes (S22- or Fe2þ-oxidation, SO4

22-or Fe3þ- reduction, and photosynthesis), is net gainof organic matter with no net production orconsumption of O2.

8.11.4.10 Abiotic Organic Matter Oxidation

Aerobic respiration is the main means bywhich O2 is consumed on the Earth. Thispathway occurs throughout most Earth-surfaceenvironments: soils and aquatic systems, marinesurface waters supersaturated with O2, within thewater column and the upper zones of sediments.However, reduced carbon materials are alsoreacted with O2 in a variety of environmentswhere biological activity has not been demon-strated. Among these are photo-oxidation ofdissolved and particulate organic matter andfossil fuels, fires from burning vegetation andfossil fuels, and atmospheric methane oxidation.Olefins (organic compounds containing doublebonds) are susceptible to oxidation in thepresence of transition metapls, ozone, UV light,or gamma radiation. Low-molecular-weight oxi-dized organic degradation products form fromoxidation reactions, which in turn may provideorganic substrates for aerobic respiration or C1

metabolism. Fires are of course high-temperaturecombustion of organic materials with O2.Research on fires has shown that O2 concen-tration has a strong influence on initiation andmaintenance of fires (Watson, 1978; Lenton andWatson, 2000). Although the exact relationshipbetween pO2

and initiation of fire in real terrestrialforest communities is debated (see Robinson,1989, 1991), it is generally agreed that, at lowpO2

, fires cannot be started even on dry wood,

although smoldering fires with inefficient oxi-dation can be maintained. At high pO2

, even wetwood can support flame, and fires are easilyinitiated with a spark discharge such as lightning.

Although most methane on Earth is oxidizedduring slow gradual transport upwards throughsoils and sediments, at select environments thereis direct injection of methane into the atmosphere.Methane reacts with O2 in the presence of light ormetal surface catalysts. This reaction is fastenough, and the amount of atmospheric O2 largeenough, that significant concentrations of atmos-pheric CH4 are unlikely to accumulate. Cata-strophic calving of submarine, CH4-rich hydratesduring the geologic past may have liberatedlarge quantities of methane to the atmosphere.However, isotopic evidence suggests that thismethane was oxidized and consumed in ageologically short span of time.

8.11.5 GLOBAL OXYGEN BUDGETS ANDTHE GLOBAL OXYGEN CYCLE

One window into the global budget of oxygen isthe variation in O2 concentration normalized to N2.O2/N2 ratios reflect changes in atmospheric O2

abundance, because N2 concentration is assumedto be invariant through time. Over an annual cycle,d(O2/N2) can vary by 100 per meg or more,especially at high-latitude sites (Keeling andShertz, 1992). These variations reflect latitudinalvariations in net O2 production and consumptionrelated to seasonal high productivity duringsummer. Observations of O2/N2 variation havebeen expanded beyond direct observation (limitedto the past several decades) to records of atmos-pheric composition as trapped in Antarctic firn iceand recent ice cores (Battle et al., 1996; Sowerset al., 1989; Bender et al., 1985, 1994b). Theserecords reflect a slow gradual decrease in atmos-pheric O2 abundance over historical times, attrib-uted to the release and oxidation of fossil fuels.

As yet, details of the fluxes involved in theprocesses that generate and consume molecularoxygen are too poorly constrained to establish abalanced O2 budget. A summary of the processesbelieved to dominate controls on atmospheric O2,and reasonable best guesses for the magnitude ofthese fluxes, if available, are shown in Figure 7(from Keeling et al., 1993).

8.11.6 ATMOSPHERIC O2 THROUGHOUTEARTH’S HISTORY

8.11.6.1 Early Models

Starting with Cloud (1976), two key geologicformations have been invoked to constrain the


history of oxygenation of the atmosphere: bandediron formations (BIFs) and red beds (Figure 8).BIFs are chemical sediments containing verylittle detritus, and consist of silica laminaeinterbedded with layers of alternately high andlow ratios of ferric- to ferrous-iron. As chemicalsediments, BIFs imply the direct precipitation offerrous-iron from the water column. A ferrous-iron-rich ocean requires anoxia, which by impli-cation requires an O2-free atmosphere and ananaerobic world. However, the ferric-iron layersin BIFs do reflect consumption of molecular O2

(oxidation of ferrous-iron) at a rate much greaterthan supply of O2 through prebiotic H2Ophotolysis. Thus, BIFs may also record theevolution of oxygenic photosynthesis and atleast localized elevated dissolved O2 concen-trations (Walker, 1979). Red beds are sandysedimentary rocks rich in coatings, cements, andparticles of ferric-iron. Red beds form during andafter sediment deposition, and thus require thatboth the atmosphere and groundwater are oxidiz-ing. The occurrence of the oldest red beds(,2.0 Ga) coincides nearly with the disappear-ance of BIFs (Walker, 1979), suggesting thatsome threshold of atmospheric O2 concentrationwas reached at this time. Although the generalconcept of a low-O2 atmosphere before ,2.0 Gaand accumulated O2 since that time has been

agreed on for several decades, the details andtexture of oxygenation of the Earth’s atmosphereare still being debated.

Geochemists and cosmochemists initiallylooked to models of planetary formation andcomparison with other terrestrial planets tounderstand the earliest composition of Earth’satmosphere. During planetary accretion and coreformation, volatile components were liberated

Atmosphere34,000

Hydrogen escape~10–5

~140 ~140

Respiration11

Primary production12

Primary production

9.2Autotrophrespiration

4.6Heterotroph

and soilrespiration

4.6

Weathering oforganic matter and sulfide minerals,

volcanic gas oxidation0.01

Respiration0.4

0.6 0.2

Surface waters6

Intermediate anddeep water

219

Figure 7 Global budget for molecular oxygen, including gas and dissolved O2 reservoirs (sources Keeling et al.,1993; Bender et al., 1994a,b).

Time (Ga)

ProterozoicPaleo- Meso- Neo-

Phanero-zoic

Archean

Banded iron formations

Red beds

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.0

Figure 8 Archean distribution of banded iron for-mations, with short reoccurrence associated with wide-spread glaciation in the Neoproterozoic, and theProterozoic and Phanerozoic distribution of sedimen-tary rocks containing ferric-iron cements (red beds).The end of banded iron formation and beginning of redbed deposition at ,2.2 Ga has been taken as evidencefor a major oxygenation event in Earth’s atmosphere.

Atmospheric O2 throughout Earth’s History 527

from a molten and slowly convecting mixtureof silicates, metals, and trapped gases. Thegravitational field of Earth was sufficient toretain most of the gases released from theinterior. These include CH4, H2O, N2, NH3,and H2S. Much H2 and He released from theinterior escaped Earth’s gravitational field intospace; only massive planets such as Jupiter,Saturn, Neptune, and Uranus have retained anH2–He rich atmosphere. Photolysis of H2O,NH3, and H2S produced free O2, N2, and S,respectively. O2 was rapidly consumed byoxidation of CH4 and H2S to form CO2, CO,and SO2. High partial pressures of CO2 and CH4

maintained a strong greenhouse effect and warmaverage Earth surface temperature (,90 8C), inspite of much lower solar luminosity. Recogniz-ing that the early Earth contained an atmos-pheric substantially richer in strong greenhousegases compared with the modern world provideda resolution to subfreezing average Earth surfacetemperatures predicted for the early Earth due toreduced solar luminosity (Kasting et al., 2001,1983; Kiehl and Dickinson, 1987).

Liquid water on the early planet Earth alloweda hydrologic cycle and silicate mineral weath-ering to develop. Fairly quickly, much of theatmosphere’s CO2 was reacted with silicates toproduce a bicarbonate-buffer ocean, while CH4

was rapidly consumed by oxygen producedthrough photolysis of H2O. Early microorgan-isms (and many of the most primitive organismsin existence today) used inorganic substrates toderive energy, and thrive at the high tempera-tures expected to be widespread during Earth’searly history. These organisms include Archeathat oxidize H2 using elemental sulfur. Oncephotolysis and CH4 oxidation generated sufficientpCO2

, methanogenic Archea may have evolved.These organisms reduce CO2 to CH4 using H2.However, sustainable life on the planet isunlikely to have developed during the firstseveral hundred million years of Earth’s history,due to large and frequent bolide impacts thatwould have sterilized the entire Earth’s surfaceprior to ,3.8 Ga (Sleep et al., 1989; Sleep andZahnle, 1998; Sleep et al., 2001; Wilde et al.,2001), although recent work by Valley et al. (2002)suggests a cool early Earth that continuallysupported liquid water as early as 4.4 Ga.

Much of present understanding of the earliestevolution of Earth’s atmosphere can tracedescent from Walker (1979) and referencestherein. The prebiological atmosphere (beforethe origin of life) was controlled principally bythe composition of gases emitted from volca-noes. Emission of H2 in volcanic gases hascontributed to net oxidation of the planetthrough time. This is achieved through severalmechanisms. Simplest is hydrodynamic escape

of H2 from Earth’s gravity. Because H2 is astrongly reducing gas, loss of H2 from the Earthequates to loss of reducing power or net increasein whole Earth oxidation state (Walker, 1979).Today, gravitational escape of H2 and thusincrease in oxidation state is minor, becauselittle if any H2 manages to reach the upper levelsof the atmosphere without oxidizing. Early inEarth’s history, sources of H2 included volcanicgases and water vapor photolysis. The small fluxof O2 produced by photolysis was rapidlyconsumed by reaction with ferrous-iron andsulfide, contributing to oxidation of the crust.

Today, volcanic gases are fairly oxidized,consisting mainly of H2O and CO2, with smalleramounts of H2, CO, and CH4. The oxidation stateof volcanic gases derives in part from theoxidation state of the Earth’s mantle. Mantleoxygen fugacity today is at or near the quartz-fayalite-magnesite buffer (QFM), as is the fO2

oferuptive volcanic gases. Using whole-rock andspinel chromium abundance from volcanogenicrocks through time, Delano (2001) has argued thatthe average oxidation state of the Earth’s mantlewas set very early in Earth’s history (,3.6–3.9 Ga) to fO2

at or near the QFM buffer. Magmaswith this oxidation state release volcanic gasesrich in H2O, CO2, and SO2, rather than morereducing gases. Thus, throughout much of Earth’shistory, volcanic gases contributing to the atmos-phere have been fairly oxidized. More reducedmagma compositions have been detected indiamond-bearing assemblages likely Hadeanin age (.4.0 Ga) (Haggerty and Toft, 1985).The increase in mantle oxidation withinseveral hundred million years of early Earth’shistory reveals very rapid “mantle þ crust” over-turn and mixing at this time, coupled withsubduction and reaction of the mantle withhydrated and oxidized crustal minerals (gene-rated from reaction with O2 produced throughH2O photolysis).

8.11.6.2 The Archean

8.11.6.2.1 Constraints on the O2 contentof the Archean atmosphere

Several lines of geochemical evidence supportlow to negligible concentrations of atmosphericO2 during the Archean and earliest Proterozoic,when oxygenic photosynthesis may have evolved.The presence of pyrite and uraninite in detritalArchean sediments reveals that the atmospherein the earliest Archean contained no free O2

(Cloud, 1972). Although Archean-age detritalpyrites from South Africa may be hydrothermalin origin, Australian sediments of the Pilbaracraton (3.25–2.75 Ga) contain rounded grains of


pyrite, uraninite, and siderite (Rasmussen andBuick, 1999), cited as evidence for an anoxicatmosphere at this time. Although disputed(Ohmoto, 1999), it is difficult to explain detritalminerals that are extremely susceptible to dissol-ution and oxidation under oxidizing conditionunless the atmosphere of the Archean wasessentially devoid of O2.

Archean paleosols provide other geochemicalevidence suggesting formation under reducingconditions. For example, the 2.75 Ga Mount Roe#2 paleosol of Western Australia contains up to0.10% organic carbon with isotope ratiosbetween 233‰ and 255‰ (Rye and Holland,2000). These isotope ratios suggest that metha-nogenesis and methanotrophy were importantpathways of carbon cycling in these soils. Formodern soils in which the bulk organic matter isstrongly 13C-depleted (,240‰), the methanefueling methanotrophy must be derived fromsomewhere outside the soil, because reasonablerates of fermentation and methanogenesis cannotsupply enough CH4. By extension, Rye andHolland (2000) argue that these soils formedunder an atmosphere rich in CH4, with any O2

consumed during aerobic methanotrophy havingbeen supplied by localized limited populations ofoxygenic photoautotrophs. Other paleosol studieshave used lack of cerium oxidation during soilformation as an indicator of atmospheric anoxiain the Archean (Murakami et al., 2001). In abroader survey of Archean and Proterozoicpaleosols, Rye and Holland (1998) observe thatall examined paleosols older than 2.4 Ga indicateloss of iron during weathering and soil formation.This chemical feature is consistent with soildevelopment under an atmosphere containing,1024 atm O2 (1 atm ¼ 1.01325 £ 105 Pa),although some research has suggested that anoxicsoil development in the Archean does notnecessarily require an anoxic atmosphere(Ohmoto, 1996).

Other evidence for low Archean atmosphericoxygen concentrations come from studies ofmass-independent sulfur isotope fractionation.Photochemical oxidation of volcanic sulfurspecies, in contrast with aqueous-phase oxidationand dissolution that characterizes the modernsulfur cycle, may have been the major source ofsulfate to seawater in the Archean (Farquharet al., 2002; Farquhar et al., 2000). Distinct shiftsin d 33S and d34S in sulfide and sulfate fromArchean rocks occurred between 2.4–2.0 Ga,consistent with a shift from an O2-free earlyatmosphere in which SO2 photochemistry coulddominate among seawater sulfate sources to anO2-rich later atmosphere in which oxidativeweathering of sulfide minerals predominatedover photochemistry as the major source ofseawater sulfate. Sulfur isotope heterogeneities

detected in sulfide inclusions in diamonds alsoare believed to derive from photochemical SO2

oxidation in an O2-free atmosphere at 2.9 Ga(Farquhar et al., 2002). Not only do these isotoperatios require an O2-free atmosphere, but theyalso imply significant contact between themantle, crust, and atmosphere as recently as2.9 Ga.

Nitrogen and sulfur isotope ratios in Archeansedimentary rocks also indicate limited ornegligible atmospheric O2 concentrations(Figure 9). Under an O2-free environment,nitrogen could only exist as N2 and reducedforms (NH3, etc.). Any nitrate or nitrite producedby photolysis would be quickly reduced, likelywith Fe2þ. If nitrate is not available, thendenitrification (reduction of nitrate to free N2)cannot occur. Denitrification is associated with asubstantial nitrogen-isotope discrimination, gen-erating N2 that is substantially depleted in 15Nrelative to the NO3

2 source. In the modernsystem, this results in seawater nitrate (andorganic matter) that is 15N-enriched relative toair. Nitrogen in kerogen from Archean sedimen-tary rocks is not enriched in 15N (as is found inall kerogen nitrogen from Proterozoic age to thepresent), but instead is depleted relative tomodern atmospheric N2 by several ‰ (Beaumontand Robert, 1999). This is consistent with anArchean nitrogen cycle in which no nitrate and nofree O2 was available, and nitrogen cycling waslimited to N2-fixation, mineralization and ammo-nia volatilization. Bacterial sulfate reduction isassociated with a significant isotope discrimi-nation, producing sulfide that is depleted in 34Srelative to substrate sulfate. The magnitude ofsulfur isotope fractionation during sulfatereduction depends in part on available sulfateconcentrations. Very limited differences insulfur isotopic ratios among Archean sedimentarysulfide and sulfate minerals (,2‰ d 34S) indicateonly minor isotope fractionation during sulfatereduction in the Archean oceans (Canfield et al.,2000; Habicht et al., 2002), best explained byextremely low SO4

22 concentrations (,200mM incontrast with modern concentrations of ,28 mM)(Habicht et al., 2002). The limited supply of sulfatein the Archean ocean suggests that the major sourceof sulfate to Archean seawater was volcanic gas,because oxidative weathering of sulfide mineralscould not occur under an O2-free atmosphere.The limited sulfate concentration would havesuppressed the activity of sulfate-reducingbacteria and facilitated methanogenesis. How-ever, by ,2.7 Ga, sedimentary sulfides that are34S-depleted relative to sulfate are detected,suggesting at least sulfate reduction, and byimplication sources of sulfate to seawater throughsulfide oxidation, may have developed (Canfieldet al., 2000).


8.11.6.2.2 The evolution of oxygenicphotosynthesis

In the early Archean, methanogenesis (reac-tion of H2 þ CO2 to yield CH4) was likely asignificant component of total primary pro-duction. O2 concentrations in the atmospherewere suppressed due to limited O2 productionand rapid consumption with iron, sulfur, andreduced gases, while CH4 concentrations werelikely very high (Kasting et al., 2001). The highmethane abundance is calculated to have

generated a hydrocarbon-rich smog that couldscreen UV light and protect early life in theabsence of O2 and ozone (Kasting et al., 2001).Other means of protection from UV damage inthe O2-free Archean include biomineralizationof cyanobacteria within UV-shielded iron–silicasinters (Phoenix et al., 2001).

Biological evolution may have contributedto early Archean oxidation of the Earth. Catlinget al., (2001) have recognized that CO2 fixationassociated with early photosynthesis may havebeen coupled with active fermentation and

60

40

20

0

–20

–60

3.5

Tim

e be

fore

pre

sent

(Ga)

Mod

ern

atm

osph

eric

val

ue

PAL

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

–10 0 10 10–12 10–5 10–4 10–3 10–2 10–1

O2 concentration / (PAL)d15 Nair ‰

d34 S

(‰)

Reactionsinvolved in

nitrogencycle

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

NH+4 + 2O2

NH+4 → NO–

3

R-NH2 → NH+4

NO–3 → N2

NO–3 + H2O + H+

R-NH2 → NH+4

1

3.0 2.5 2.0

Time (Ga)

Evolution of nitrogen isotopic composition compared tooxygen atmospheric level evolution

(a)

(b)

1.5 1.0 0.5 0.0

–40

Figure 9 (a) The isotopic composition of sedimentary sulfate and sulfides through geologic time. The two upperlines show the isotopic composition of seawater sulfate (5‰ offset indicates uncertainty). The lower line indicatedd 34S of sulfate displaced by 255‰, to mimic average Phanerozoic maximum fractionation during bacterial sulfatereduction. Sulfide isotopic data (circles) indicated a much reduced fractionation between sulfate and sulfide in theArchean and Proterozoic (source Canfield, 1998). (b) Geologic evolution of the nitrogen isotopic composition ofkerogen (left), estimated atmospheric O2 content, and representative reactions in the biogeochemical nitrogen cycle at

those estimated O2 concentrations (sources Beaumont and Robert, 1999; Kasting, 1992).


methanogenesis. Prior to any accumulation ofatmospheric O2, CH4 may have been a largecomponent of Earth’s atmosphere. A high flux ofbiogenic methane is supported by coupled eco-system–climate models of the early Earth (Pavlovet al., 2000; Kasting et al., 2001) as a supplementto Earth’s greenhouse warming under reducedsolar luminosity in the Archean. CH4 in the upperatmosphere is consumed by UV light to yieldhydrogen (favored form of hydrogen in the upperatmosphere), which escapes to space. Because H2

escape leads to net oxidation, biological pro-ductivity and methanogenesis result in slowoxidation of the planet. Net oxidation may be inthe form of direct O2 accumulation (if CO2

fixation is associated with oxygenic photosyn-thesis) or indirectly (if CO2 fixation occurs viaanoxygenic photosynthesis or anaerobic che-moautotrophy) through production of oxidizediron or sulfur minerals in the crust, which uponsubduction are mixed with other crustal rocks andincrease the crustal oxidation state. One systemthat may represent a model of early Archeanbiological productivity consists of microbial matsfound in hypersaline coastal ponds. In these mats,cyanobacteria produce H2 and CO that can be usedas substrates by associated chemoautotrophs, andsignificant CH4 fluxes out of the mats have beenmeasured (Hoehler et al., 2001). Thus, mats mayrepresent communities of oxygenic photosyn-thesis, chemoautotrophy, and methanogenesisoccurring in close physical proximity. Suchcommunities would have contributed to elevatedatmospheric CH4 and the irreversible escape ofhydrogen from the Archean atmosphere, contri-buting to oxidation of the early Earth.

The earliest photosynthetic communities mayhave contributed to oxidation of the early Earththrough production of oxidized crustal mineralswithout requiring production of O2. Crustal rockstoday are more oxidized than the mantle, withcompositions ranging between the QFM andhematite-magnesite fO2

buffers. This is bestexplained as an irreversible oxidation of thecrust associated with methanogenesis and hydro-gen escape (Catling et al., 2001). It is unlikely thatO2 produced by early photosynthesis ever directlyentered the atmosphere.

Maintaining low O2 concentrations in theatmosphere while generating oxidized crustalrocks requires oxidation mechanisms that do notinvolve O2. Among these may be serpentinizationof seafloor basalts (Kasting and Siefert, 2002).During seafloor weathering, ferrous-iron isreleased to form ferric oxyhydroxides. UsingH2O or CO2 as the ferrous-iron oxidant, H2 andCH4 are generated. This H2 gas (either produceddirectly, or indirectly through UV decompositionof CH4) can escape the atmosphere, resulting innet oxidation of the crust and accumulation of

oxidized crustal rocks. BIF formation in theArchean may be related to dissolved Fe2þ thatwas oxidized in shallow-water settings associatedwith local oxygenic photosynthesis or chemoau-totrophy (Kasting and Siefert, 2002). After thebiological innovation of oxygenic photosynthesisevolved, there was still a several hundred millionyear gap until O2 began to accumulate in theatmosphere, because O2 could only accumulateonce the supply of reduced gases (CO and CH4)and ferrous-iron fell below rates of photosyntheticO2 supply.

8.11.6.2.3 Carbon isotope effects associatedwith photosynthesis

The main compound responsible for harvestinglight energy to produce NADPH and splittingwater to form O2 is chlorophyll. There are severaldifferent structural variants of chlorophyll, includ-ing chlorophyll a, chlorophyll b, and bacterio-chlorophylls a–e and g. Each of these showsoptimum excitation at a different wavelength oflight. All oxygenic photosynthetic organismsutilize chlorophyll a and/or chlorophyll b. Othernon-oxygenic photoautotrophic microorganismsemploy a diverse range of chlorophylls.

The earliest evidence for evolution of oxygenicphotosynthesis comes from carbon isotopic signa-tures preserved in Archean rocks (Figure 10). CO2

fixation through the Calvin cycle is associated witha significant carbon isotope discrimination, suchthat organic matter produced through CO2 fixationis depleted in 13C relative to 12C by several per mil.In a closed or semi-closed system (up to andincluding the whole ocean–atmosphere system),isotope discrimination during fixation of CO2 thenresults in a slight enrichment in the 13C/12C ratioof CO2 not taken up during photosynthesis. Thus, abiosignature of CO2 fixation is an enrichmentin 13C/12C ratio in atmospheric CO2 and seawaterbicarbonate over a whole-Earth averaged isotoperatio. When carbonate minerals precipitate fromseawater, they record the seawater isotope value.Enrichment of the 13C/12C ratio in early Archeancarbonate minerals, and by extension seawaterbicarbonate, is taken as early evidence for CO2

fixation. Although anoxygenic photoautotrophsand chemoautotrophs fix CO2 without generatingO2, these groups either do not employ the Calvincycle (many anoxygenic photoautotrophs use theacetyl-CoA or reverse tricarboxylic acid pathways)or require O2 (most chemolithoautotrophs). Thus,the most likely group of organisms responsible forthis isotope effect are oxygenic photoautotrophssuch as cyanobacteria. Kerogen and graphite that isisotopically depleted in 13C is a common andcontinuous feature of the sedimentary record,extending as far back as ,3.8 Ga to the Isua


Supracrustal Suite of Greenland (Schidlowski,1988, 2001; Nutman et al., 1997). Although someisotopically depleted graphite in the metasedi-ments from the Isua Suite may derive from abiotichydrothermal processes (van Zuilen et al., 2002),rocks interpreted as metamorphosed turbiditedeposits retain 13C-depleted graphite believed tobe biological in origin. Moreover, the isotopicdistance between coeval carbonate and organicmatter (in the form of kerogen) can be used toestimate biological productivity through time.With a few exceptions, these isotope mass balanceestimates reveal that, since Archean times, globalscale partitioning between inorganic and organiccarbon, and thus global productivity and carbonburial, have not varied greatly over nearly 4 Gyr ofEarth’s history (Schidlowski, 1988, 2001).Approximately 25% of crustal carbon burial is inthe form of organic carbon, and the remainder isinorganic carbonate minerals. This estimatederives from the mass- and isotope-balanceequation:

d13Cavg ¼ fd13Corganic matter þ ð1 2 f Þ

� d13Ccarbonate ð2Þ

where d 13Cavg is the average isotopic compositionof crustal carbon entering the oceans fromcontinental weathering and primordial carbonemitted from volcanoes, d 13Corganic matter is theaverage isotopic composition of sedimentaryorganic matter, d 13Ccarbonate is the average isotopiccomposition of carbonate sediments, and f is thefraction of carbon buried as organic matter insediments.

The observation that the proportion of carbonburied in sediments as organic matter versus

carbonate has not varied throughout geologic timeraises several intriguing issues. First, biogeochem-ical cycling of carbon exhibits remarkable con-stancy across 4 Gyr of biological evolution, inspite of large-scale innovations in primary pro-duction and respiration (including anoxygenic andoxygenic photosynthesis, chemoautotrophy, sul-fate reduction, methanotrophy, and aerobic res-piration). Thus, with a few notable exceptionsexpressed in the carbonate isotope record, burialflux ratios between organic matter and carbonatehave remained constant, in spite of varyingdominance of different modes of carbon fixationand respiration through time, not to mention otherpossible controls on organic matter burial andpreservation commonly invoked for Phanerozoicsystems (anoxia of bottom waters, sedimentationrate, selective preservation, or cumulative oxygenexposure time). Second, the constancy of organicmatter versus carbonate burial through timereveals that throughout geologic time, the relativecontributions of various sources and sinks ofcarbon to the “ocean þ atmosphere” system haveremained constant. In other words, to maintain aconstant carbonate isotopic composition throughtime, not only must the relative proportion oforganic matter versus carbonate burial haveremained nearly constant, but the relative intensityof organic matter versus carbonate weatheringalso must have remained nearly constant. In theearliest stages of Earth’s history, when continentswere small and sedimentary rocks were sparse,inputs of carbon from continental weatheringmay have been small relative to volcanic inputs.However, the several billion year sedimentaryrecord of rocks rich in carbonate mineralsand organic matter suggests that continental

+10

Lomagundi–Jatulianevent (2.0–2.3 Ga)

Terminal Neoproterozoicevent (~0.6 Ga)

Late Permianevent (~0.26 Ga)

00.511.5Age (Ga = 109 yr)

d13 C

(‰

, PD

B)

Corg

Ccarb

A1

A2

Isua metasedi·mentary suite

(max. 3.85 Ga)

Francevillian excursion(~2.0 Ga)

Fortescue excursion(~2.7 Ga)

Archean Proterozoic Phanerozoic

22.534 3.5

0

–10

–20

–30

–40

–50

Approximate mean

Figure 10 Isotopic composition of carbonates and organic carbon in sedimentary and metasedimentary rocksthrough geologic time. The negative excursions in organic matter d13C at 2.7 Ga and 2.0 Ga may relate to extensivemethanogenesis as a mechanism of carbon fixation (Schidlowski, 2001) (reproduced by permission of Elsevier from

Precanb. Res. 2001, 106, 117–134).


weathering must have formed a significantcontribution to total oceanic carbon inputs fairlyearly in the Archean. Intriguingly, this indicatesthat oxidative weathering of ancient sedimentaryrocks may have been active even in the Archean,prior to accumulation of O2 in the atmosphere.This runs counter to traditional interpretations ofgeochemical carbon–oxygen cycling, in whichorganic matter burial is equated to O2 productionand organic matter weathering is equated to O2

consumption.

8.11.6.2.4 Evidence for oxygenicphotosynthesis in the Archean

Fossil evidence for photosynthetic organismsfrom the same time period can be traced to theexistence of stromatolites (Schopf, 1992, 1993;Schopf et al., 2002; Hofmann et al., 1999),although evidence for these early oxygenicphotosynthetic communities is debated (Buick,1990; Brasier et al., 2002). Stromatolites arelaminated sediments consistently of alternatingorganic matter-rich and organic matter-leanlayers; the organic matter-rich layers are largelycomposed of filamentous cyanobacteria. Stroma-tolites and similar mat-forming cyanobacterialcrusts occur today in select restricted shallowmarine environments. Fossil evidence from manylocales in Archean rocks (Greenland, Australia,South Africa) and Proterozoic rocks (Canada,Australia, South America) coupled with carbonisotope geochemistry provides indirect evidencefor the evolution of oxygenic photosynthesis earlyin Earth’s history (,3.5 Ga).

Other evidence comes from molecular fossils.All cyanobacteria today are characterized by thepresence of 2a-methylhopanes in their cellmembranes. Brocks et al., (1999) have demon-strated the existence of this taxon-specific bio-marker in 2.7 Ga Archean rocks of the PilbaraCraton in NW Australia. Also in these rocks isfound a homologous series of C27 to C30 steranes.Steranes today derive from sterols mainly pro-duced by organisms in the domain Eukarya.Eukaryotes are obligate aerobes that requiremolecular O2, and thus Brocks and colleaguesargue that the presence of these compoundsprovides strong evidence for both oxygenicphotosynthesis and at least localized utilizationof accumulated O2. Of course, coexistence of twotraits within a biological lineage today revealsnothing about which evolved first. It is uncertainwhether 2a-methylhopane lipid biosynthesis pre-ceded oxygenic photosynthesis among cyanobac-teria. Sterol synthesis certainly occurs in modernprokaryotes (including some anaerobes), but noexisting lineage produces the distribution ofsteranes found in the Brocks et al., (1999) studyexcept eukaryotes.

8.11.6.3 The Proterozoic Atmosphere

8.11.6.3.1 Oxygenation of the Proterozoicatmosphere

Although there is evidence for the evolution ofoxygenic photosynthesis several hundred millionyears before the Huronian glaciation (e.g., Brockset al., 1999) or earlier (Schidlowski, 1988, 2001),high fluxes of UV light reacting with O2 derivedfrom the earliest photosynthetic organisms wouldhave created dangerous reactive oxygen speciesthat severely suppressed widespread developmentof large populations of these oxygenic photoauto-trophs. Cyanobacteria, as photoautotrophs, needbe exposed to visible light and have evolvedseveral defense mechanisms to protect cell con-tents and repair damage. However, two keymetabolic pathways (oxygenic photosynthesisand nitrogen fixation) are very sensitive to UVdamage. Indisputably cyanobacterial fossil occurin 1,000 Ma rocks, with putative fossils occurring2,500 Ma and possibly older (Schopf, 1992).Sediment mat-forming cyanobacteria and stroma-tolites are least ambiguous and oldest. Terrestrialencrusting cyanobacteria are only known in thePhanerozoic. Planktonic forms are not known forthe Archean and early Proterozoic. They may nothave existed, or they may not be preserved.Molecular evolution, specifically coding forproteins that build gas vesicles necessary forplanktonic life, are homologous and conserved inall cyanobacteria. Thus, perhaps planktonic cya-nobacteria existed throughout Earth’s historysince the late Archean.

Today, ozone forms in the stratosphere byreaction of O2 with UV light. This effectivelyscreens much incoming UV radiation. Prior to theaccumulation of atmospheric O2, no ozone couldform, and thus the UV flux to the Earth’s surfacewould be much greater (with harmful effects onDNA and proteins, which adsorb and are alteredby UV). A significant ozone shield could developat ,1022 PAL O2 (Kasting, 1987). However, afainter young Sun would have emitted somewhatlower UV, mediating the lack of ozone. Althoughseawater today adsorbs most UV light by 6–25 m(1% transmittance cutoff), seawater with abun-dant dissolved Fe2þ may have provided aneffective UV screen in the Archean (Olson andPierson, 1986). Also, waters rich in humicmaterials, such as modern coastal oceans, arenearly UV opaque. If Archean seawater containedDOM, this could adsorb some UV. Iron oxidationand BIF at ,2.5–1.9 Ga would have removedthe UV screen in seawater. Thus, a significantUV stress may have developed at this time.This would be mediated coincidentally byaccumulation of atmospheric O2.

In the early Archean before oxygenic photo-synthesis evolved, cyanobacteria were limited.


Planktonic forms were inhibited, and limited bydissolved iron content of water, and existence ofstratified, UV-screen refuges. Sedimentary mat-forming and stromatolite-forming commu-nities were much more abundant. Iron-oxideprecipitation and deposition of screen enzymesmay have created UV-free colonies under a shieldeven in shallow waters.

Advent of oxygenic photosynthesis in theArchean generated small oxygen oases (wheredissolved O2 could accumulate in the water)within an overall O2-free atmosphere waterscontaining oxygenic photoautotrophs might havereached 10% air saturation (Kasting, 1992). Atthis time, both unscreened UV radiation and O2

may have coincided within the water columnand sediments. This would lead to increased UVstress for cyanobacteria. At the same time,precipitation of iron from the water wouldmake the environment even more UV-transpar-ent. To survive, cyanobacteria would need toevolve and optimize defense and repair mech-anisms for UV damage (Garcia-Pichel, 1998).Perhaps this explains the ,500 Myr gap betweenorigin of cyanobacteria and accumulation of O2

in the atmosphere. For example, the synthesis ofscytonemin (a compound found exclusively incyanobacteria) requires molecular oxygen(implying evolution in an oxic environment); itoptimally screens UV-a, the form of UVradiation only abundant in an oxygenatedatmosphere.

Oxygenation of the atmosphere at ,2.3–2.0 Ga may derive from at least three separatecauses. First, discussed earlier, is the titration ofO2 with iron, sulfur, and reduced gases. The otheris global rates of photosynthesis and organiccarbon burial in sediments. If the rate ofatmospheric O2 supply (oxygenic photosynthesis)exceeds all mechanisms of O2 consumption(respiration, chemoautotrophy, reduced mineraloxidation, etc.), then O2 can accumulate in theatmosphere. One means by which this can beevaluated is through seawater carbonated 13C. Because biological CO2 fixation is associ-ated with significant carbon isotope discrimi-nation, the magnitude of carbon fixation isindicated by the isotopic composition of seawatercarbonate. At times of more carbon fixation andburial in sediments, relatively more 12C isremoved from the atmosphere þ ocean inorganiccarbon pool than is supplied through respiration,organic matter oxidation, and carbonate mineraldissolution. Because carbon fixation is dominatedby oxygenic photosynthesis (at least since the lateArchean), periods of greater carbon fixation andburial of organic matter in sediments (observedas elevated seawater carbonate d 13C) are equatedto periods of elevated O2 production throughoxygenic photosynthesis. The early Proterozoic

Lomagundi event (,2.3–2.0 Ga) is recorded inthe sediment record as a prolonged period ofelevated seawater carbonate d 13C, with carbon-ate d 13C values reaching nearly 10‰ in severalsections around the world (Schidlowski, 2001).This represents an extended period of time(perhaps several hundred million years) duringwhich removal of carbon from the “ocean þatmosphere” system as organic matter greatlyexceeded supply. By implication, release of O2

through photosynthesis was greatly acceleratedduring this time.

A third mechanism for oxygenation of theatmosphere at ,2.3 Ga relies on the slow, gradualoxidation of the Earth’s crust. Irreversible H2

escape and basalt-seawater reactions led to agradual increase in the amount of oxidized andhydrated minerals contained in the Earth’s crustand subducted in subduction zones throughout theArchean. Gradually this influenced the oxidationstate of volcanic gases derived in part fromsubducted crustal rocks. Thus, although mantleoxygen fugacity may not have changed since theearly Archean, crustal and volcanic gas oxygenfugacity slowly increased as the abundance ofoxidized and hydrated crust increased (Holland,2002; Kasting et al., 1993; Kump et al., 2001).Although slow to develop, Holland (2002)estimates that an increase in fO2

of less than 1 logunit is all that would have been required fortransition from an anoxic to an oxic atmosphere,assuming rates of oxygenic photosynthesis con-sistent with modern systems and the sedimentisotope record. Once a threshold volcanic gas fO2

had been reached, O2 began to accumulate.Oxidative weathering of sulfides released largeamounts of sulfate into seawater, facilitatingbacterial sulfate reduction.

There are several lines of geochemical evidencethat suggest a rise in oxygenation of theatmosphere ,2.3–2.0 Ga, beyond the coincidentlast occurrence of BIFs (with one late Proterozoicexception) and first occurrence of red bedsrecognized decades ago, and carbon isotopicevidence suggesting ample burial of sedimentaryorganic matter (Karhu and Holland, 1996; Bekkeret al., 2001; Buick et al., 1998). The Huronianglaciation (,2.3 Ga) is the oldest known glacialepisode recorded in the sedimentary record. Oneinterpretation of this glaciation is that the coolerclimate was a direct result of the rise ofphotosynthetically derived O2. The rise of O2

scavenged and reacted with the previously highatmospheric concentration of CH4. Methaneconcentrations dropped, and the less-effectivegreenhouse gas product CO2 could not maintainequable surface temperatures. Kasting et al.,(1983) estimate that a rise in pO2

above,1024 atm resulted in loss of atmospheric CH4

and onset of glaciation.


Paleosols have also provided evidence for achange in atmosphere oxygenation at some timebetween 2.3–2.0 Ga. Evidence from rare earthelement enrichment patterns and U/Th fraction-ation suggests a rise in O2 to ,0.005 bar by thetime of formation of the Flin Flon paleosol ofManitoba, Canada, 1.85 Ga (Pan and Stauffer,2000; Holland et al., 1989; Rye and Holland,1998). Rye and Holland (1998) examined severalearly Proterozoic paleosols and observed thatnegligible iron loss is a consistent feature fromsoils of Proterozoic age through the present. Theseauthors estimate that a minimum pO2

of.0.03 atm is required to retain iron during soilformation, and thus atmospheric O2 concentrationhas been 0.03 or greater since the early Paleozoic.However, re-evaluation of a paleosol crucial to theargument of iron depletion during soil formationunder anoxia, the Hekpoort paleosol dated at2.2 Ga, has revealed that the iron depletiondetected by previous researchers may in fact bethe lower zone of a normal oxidized lateritic soil.Upper sections of the paleosol that are notdepleted in iron have been eroded away in theexposure examined by Rye and Holland (1998),but have been found in drill core sections. Thedepletion of iron and occurrence of ferrous-ironminerals in the lower sections of this paleosolhave been reinterpreted by Beukes et al., (2002) toindicate an abundant soil surface biomass at thetime of deposition that decomposed to generatereducing conditions and iron mobilization duringthe wet season, and precipitation of iron oxidesduring the dry season.

Sulfur isotope studies have also providedinsights into the transition from Archean lowpO2

to higher values in the Proterozoic. In thesame studies that revealed extremely lowArchean ocean sulfate concentrations, it wasfound that by ,2.2 Ga, isotopic compositions ofsedimentary sulfates and sulfides indicate bac-terial sulfate reduction under more elevatedseawater sulfate concentrations compared withthe sulfate-poor Archean (Habicht et al., 2002;Canfield et al., 2000). As described above,nitrogen isotope ratios in sedimentary kerogensshow a large and permanent shift at ,2.0 Ga,consistent with denitrification, significant sea-water nitrate concentrations, and thus availableatmospheric O2.

Prior to ,2.2 Ga, low seawater sulfate concen-trations would have limited precipitation andsubduction of sulfate-bearing minerals. Thiswould maintain a lower oxidation state in volcanicgases derived in part from recycled crust (Holland,2002). Thus, even while the oxidation state of themantle has remained constant since ,4.0 Ga(Delano, 2001), the crust and volcanic gasesderived from subduction of the crust could only

achieve an increase in oxidation state onceseawater sulfate concentrations increased.

A strong model for oxygenation of theatmosphere has developed based largely on thesulfur isotope record and innovations in microbialmetabolism. The classical interpretation of thedisappearance of BIFs relates to the rise ofatmospheric O2, oxygenation of the oceans, andremoval of dissolved ferrous-iron by oxidation.However, another interpretation has developed,based largely on the evolving Proterozoic sulfurisotope record. During the oxygenation of theatmosphere at ,2.3–2.0 Ga, the oceans may nothave become oxidized, but instead remainedanoxic and became strongly sulfidic as well(Anbar and Knoll, 2002; Canfield, 1998 andreferences therein). Prior to ,2.3 Ga, the oceanswere anoxic but not sulfidic. Ferrous-iron wasabundant, as was manganese, because both arevery soluble in anoxic, sulfide-free waters. Thehigh concentration of dissolved iron and manga-nese facilitated nitrogen fixation by early cyano-bacteria, such that available nitrogen wasabundant, and phosphorus became the nutrientlimiting biological productivity (Anbar andKnoll, 2002). Oxygenation of the atmosphere at,2.3 Ga led to increased oxidative weathering ofsulfide minerals on the continents and increasedsulfate concentration in seawater. Bacterialsulfate reduction generated ample sulfide, andin spite of limited mixing, the deep oceans wouldhave remained anoxic and now also sulfidic aslong as pO2

remained below ,0.07 atm (Canfield,1998), assuming reasonable rates of primaryproduction. Both iron and manganese forminsoluble sulfides, and thus were effectivelyscavenged from seawater once the oceans becamesulfidic. Thus, the Proterozoic oxygenation led tosignificant changes in global oxygen balances,with the atmosphere and ocean mixed layerbecoming mildly oxygenated (probably,0.01 atm O2), and the deep oceans becomingstrongly sulfidic in direct response to the rise ofatmospheric O2.

In addition to increased oxygen fugacity ofvolcanic gases, and innovations in biologicalproductivity to include oxygenic photosynthesis,the oxygenation of the Proterozoic atmospheremay be related to large-scale tectonic cycles.There are several periods of maximum depositionof sedimentary rocks rich in organic matterthrough geologic time. These are ,2.7 Ga,2.2 Ga, 1.9 Ga, and 0.6 Ga (Condie et al., 2001).The increased deposition of black shales at 2.7 Gaand 1.9 Ga are associated with superplume events:highly elevated rates of seafloor volcanism,oceanic crust formation. Superplumes lead toincreased burial of both organic matter andcarbonates (through transgression, increasedatmospheric CO2, accelerated weathering, and


nutrient fluxes to the oceans), with no net effect oncarbonate isotopic composition. Thus, periods ofincreased absolute rates of organic matter burialand O2 production may be masked by a lackof carbon isotopic signature. Breakup ofsupercontinents may be related to the black shaledepositional events at 2.2 Ga and 0.6 Ga. Breakupof supercontinents may lead to more sedimentaccommodation space on continental shelves, aswell as accelerated continental weathering anddelivery of nutrients to seawater, fertilizingprimary production and increasing organic matterburial. These supercontinent breakup events at2.2 Ga and 0.6 Ga are clearly observed on thecarbonate isotopic record as increases in relativeburial of organic matter versus carbonate. Model-ing efforts examining the evolution of the carbon,sulfur, and strontium isotope records have shownthat gradual growth of the continents duringthe Archean and early Proterozoic may in factplay a very large role in controlling the onsetof oxygenation of the atmosphere, and thatbiological innovation may not be directlycoupled to atmospheric evolution (Godderıs andVeizer, 2000).

8.11.6.3.2 Atmospheric O2 duringthe Mesoproterozoic

The sulfur-isotope-based model of Canfieldand colleagues and the implications for limitingnutrient distribution proposed by Anbar andKnoll (2002) suggest that for nearly thousandmillion years (,2.2–1.2 Ga), oxygenation of theatmosphere above pO2

,0.01 atm was held instasis. Although oxygenic photosynthesis wasactive, and an atmospheric ozone shield haddeveloped to protect surface-dwelling organismsfrom UV radiation, much of the deep ocean wasstill anoxic and sulfidic. Removal of iron andmanganese from seawater as sulfides generatedsevere nitrogen stress for marine communities,suggesting that productivity may have beenlimited throughout the entire Mesoproterozoic.The sluggish but consistent primary productivityand organic carbon burial through this time isseen in the carbonate isotope record. For severalhundred million years, carbonate isotopic com-position varied by no more than ^2‰, revealingvery little change in the relative carbonate/organic matter burial in marine sediments.Anbar and Knoll (2002) suggest that thisindicates a decoupling of the link betweentectonic events and primary production, becausealthough variations in tectonic activity (andassociated changes in sedimentation, generationof restricted basins, and continental weathering)occurred during the Mesoproterozoic, these arenot observed in the carbonate isotope record. This

decoupling is a natural result of a shift in thesource of the biological limiting nutrient fromphosphorus (derived from continental weather-ing) to nitrogen (limited by N2 fixation).Furthermore, the isotopic composition of carbon-ates throughout the Mesoproterozoic is 1–2‰depleted relative to average carbonates from theearly and late Proterozoic and Phanerozoic. Thisis consistent with a decrease in the relativeproportion of carbon buried as organic matterversus carbonate during this time. It appears thatafter initial oxygenation of the atmosphere fromcompletely anoxic to low pO2

in the earlyProterozoic, further oxygenation was halted forseveral hundred million years.

Global-scale reinvigoration of primary pro-duction, organic matter burial and oxygenationof the atmosphere may be observed in the latestMesoproterozoic. Shifts in carbonate d 13C of upto 4‰ are observed in sections around the globe at,1.3 Ga (Bartley et al., 2001; Kah et al., 2001).These positive carbon isotope excursions areassociated with the formation of the Rodiniansupercontinent, which led to increased continentalmargin length, orogenesis, and greater sedimen-tation (Bartley et al., 2001). The increased organicmatter burial and atmospheric oxygenation associ-ated with this isotope excursion may be related tothe first occurrence of laterally extensive CaSO4

evaporites (Kah et al., 2001). Although the rapid10‰ increase in evaporate d 34S across thesesections is taken to indicate a much reducedseawater sulfate reservoir with much more rapidturnover times than found in the modern ocean,the isotope fractionation between evaporate sul-fates and sedimentary sulfides indicates that theoceans were not sulfate-limiting. Atmospheric O2

was of sufficient concentration to supply amplesulfate for bacterial sulfate reduction throughoutthe Mesoproterozoic.

8.11.6.3.3 Neoproterozoic atmospheric O2

Elevated carbonate isotopic compositions(,4‰) and strong isotope fractionation betweensulfate and biogenic sulfide minerals through theearly Neoproterozoic indicates a period ofseveral hundred million years of elevated bio-logical productivity and organic matter burial.This may relate in part to the oxygenation of theatmosphere, ammonia oxidation, and increasedseawater nitrate availability. Furthermore, oxi-dative weathering of molybdenum-bearing sul-fide minerals, and greater oxygenation of theoceans increased availability of molybdenumnecessary for cyanobacterial pathways ofN2-fixation (Anbar and Knoll, 2002). Much ofthe beginning of the Neoproterozoic, like thelate Mesoproterozoic, saw gradual increases in


oxygenation of the atmosphere, and possibly ofthe surface oceans. Limits on the oxygenation ofthe atmosphere are provided by sulfur isotopicand molecular evidence for the evolution ofsulfide oxidation and sulfur disproportionation.The increase in sulfate–sulfide isotope fraction-ation in the Neoproterozoic (,1.0 – 0.6 Ga)reflects a shift in sulfur cycling from simpleone-step reduction of sulfate to sulfide, to asystem in which sulfide was oxidized to sulfurintermediates such as thiosulfate or elementalsulfur, which in turn were disproportionated intosulfate and sulfide (Canfield and Teske, 1996).Sulfur disproportionation is associated withsignificant isotope effects, generating sulfidethat is substantially more 34S-depleted than canbe achieved through one-step sulfate reduction.The sulfur isotope record reveals that an increasein sulfate–sulfide isotope fractionation occurredbetween 1.0 Ga and 0.6 Ga, consistent withevolution of sulfide-oxidation at this time asderived from molecular clock based divergenceof non-photosynthetic sulfide-oxidizing bacteria(Canfield and Teske, 1996). These authorsestimated that innovation of bacterial sulfideoxidation occurred when much of the coastalshelf sediment (,200 m water depth) wasexposed to water with 13–46 mM O2, whichcorresponds to 0.01–0.03 atm pO2

. Thus, after theinitial oxygenation of the atmosphere in the earlyProterozoic (,2.3–2.0 Ga) to ,0.01 atm, pO2

wasconstrained to this level until a second oxygen-ation in the late Proterozoic (,1.0–0.6 Ga) to0.03 atm or greater.

During the last few hundred million years of theProterozoic (,0.7–0.5 Ga), fragmentation of theRodinian supercontinent was associated with atleast two widespread glacial episodes (Hoffmanet al., 1998; Hoffman and Schrag, 2002). Neo-proterozoic glacial deposits are found in Canada,Namibia, Australia, and other locations worldwide(Evans, 2000). In some of these, paleomagneticevidence suggests that glaciation extended com-pletely from pole to equator (Sumner, 1997; seealso Evans (2000) and references therein). The“Snowball Earth”, as these events have come to beknown, is associated with extreme fluctuations incarbonate isotopic stratigraphy, reoccurrence ofBIFs after an ,1.0 Ga hiatus, and precipitation ofenigmatic, massive cap carbonate sedimentsimmediately overlying the glacial deposits(Figure 11). It has been proposed that the particularconfiguration of continents in the latest Neopro-terozoic, with land masses localized within themiddle–low latitudes, lead to cooling of climatethrough several mechanisms. These includehigher albedo in the subtropics (Kirschvink,1992), increased silicate weathering as the bulkof continents were located in the warm tropics,resulting in a drawdown of atmospheric CO2

(Hoffman and Schrag, 2002), possibly acceleratedthrough high rates of OM burial, and reducedmeridional Hadley cell heat transport, becausetropical air masses were drier due to increasedcontinentality (Hoffman and Schrag, 2002).Growth of initially polar ice caps would havecreated a positive ice–albedo feedback such thatgreater than half of Earth’s surface became icecovered, global-scale growth of sea ice becameinevitable (Pollard and Kasting, 2001; Baum andCrowley, 2001). Kirschvink (1992) recognizedthat escape from the Snowball Earth becomespossible because during extreme glaciation, thecontinental hydrologic cycle would be shut down.He proposed that sinks for atmospheric CO2,namely, photosynthesis and silicated weathering,would have been eliminated during the glaciation.Because volcanic degassing continued during theglaciation, CO2 in the atmosphere could rise tovery high concentrations. It was estimated that anincrease in pCO2

to 0.12 bar would be sufficient toinduce a strong enough greenhouse effect to beginwarming the planet and melting the ice (Caldeiraand Kasting, 1992). Once meltback began, it

–5 0 5 10

–5

–50

–30

–20

–10

0

0

10

20

30

40

m

–40

0 5 10

d13C (‰)

Ghaub Glaciation

Om

baat

JIE

FM

Mai

eber

g C

AP

7

6

Figure 11 Composite section of carbonate isotopestratigraphy before and after the Ghaub glaciation fromNamibia. Enriched carbonates prior to glaciationindicate active primary production and organic matterburial. Successive isotopic depletion indicates a shutdown in primary production as the world became

covered in ice (source Hoffman et al., 1998).


would subsequently accelerate through thepositive ice–albedo feedback. Intensely warmtemperatures would follow quickly after theglaciation, until accelerated silicate weatheringunder a reinvigorated hydrologic cycle couldconsume excess CO2 and restore pCO2

to moreequable values. Carbonate mineral precipitationwas inhibited during the global glaciation, soseawater became enriched in hydrothermallyderived cations. At the end of glaciation, mixingof cation-rich seawater with high alkalinitysurface runoff under warming climates led to theprecipitation of massive cap carbonates.

Geochemical signatures before and after theSnowball episodes reveal rapid and short-livedchanges in global biogeochemical cycles thatimpacted O2 concentrations in the atmosphere andoceans. Prior to the Snowball events, the Neopro-terozoic ocean experienced strong primary pro-duction and organic matter burial, as seen in thecarbon isotopic composition of Neoproterozoicseawater in which positive isotope excursionsreached 10‰ in some sections (Kaufman andKnoll, 1995). These excursions are roughlycoincident with increased oxygenation of theatmosphere in the Neoproterozoic (Canfield andTeske, 1996; Des Marais et al., 1992). Althoughsulfur isotopic evidence suggests that much of theocean may have become oxygenated during theNeoproterozoic, it is very likely that this was atemporary phenomenon. The global-scale glacia-tions of the late Neoproterozoic would havedriven the oceans to complete anoxia throughseveral mechanisms (Kirschvink, 1992). Ice coverinhibited air–sea gas exchange, and thus surfacewaters and sites of deep-water formation were cutoff from atmospheric O2 supplies. Intense oxi-dation of organic matter in the water column andsediment would have quickly consumed anyavailable dissolved O2. Extreme positive sulfur-isotope excursions associated with snowball-succession deposits reflect nearly quantitative,closed-system sulfate reduction in the oceansduring glaciation (Hurtgen et al., 2002). Asseawater sulfate concentrations were reduced,hydrothermal inputs of ferrous-iron exceededsulfide supply, allowing BIFs to form (Hoffmanand Schrag, 2002). Closed-system sulfatereduction and iron formations require oceananoxia throughout the water column, althoughrestricted oxygenic photosynthesis beneath thintropical sea ice may be responsible for ferrous-iron oxidation and precipitation (Hoffman andSchrag, 2002).

Once global ice cover was achieved, it isestimated to have lasted several million years,based on the amount of time required toaccumulate 0.12 bar CO2 at modern rates ofvolcanic CO2 emission (Caldeira and Kasting,1992). This must have placed extreme stress on

eukaryotes and other organisms dependent onaerobic respiration. Obviously, oxygenated refu-gia must have existed, because the rise of manyeukaryotes predates the Neoproterozoic Snowballevents (Butterfield and Rainbird, 1998; Porter andKnoll, 2000). Nonetheless, carbon isotopic evi-dence leading up to and through the intense glacialintervals suggests that overall biological pro-ductivity was severely repressed during Snowballevents. Carbonate isotope compositions fall fromextremely enriched values to depletions of,(26‰) at glacial climax (Hoffman et al.,1998; Kaufman et al., 1991; Kaufman andKnoll, 1995). These carbonate isotope valuesreflect primordial (volcanogenic) carbon inputs,and indicate that effectively zero organic matterwas buried in sediments at this time. Sustainedlack of organic matter burial and limited if anyoxygenic photosynthesis under the glacial ice forseveral million years would have maintained ananoxic ocean, depleted in sulfate, with low sulfideconcentrations due to iron scavenging, andextremely high alkalinity and hydrothermally-derived ion concentrations. Above the ice, theinitially mildly oxygenated atmosphere wouldslowly lose its O2. Although the hydrologic cycleswere suppressed, and thus oxidative weathering ofsulfide minerals and organic matter exposed onthe continents was inhibited, the emission ofreduced volcanic gases would have been enoughto fully consume 0.01 bar O2 within severalhundred thousand years. Thus, it is hypothesizedthat during Snowball events, the Earth’s atmos-phere returned to pre-Proterozoic anoxic con-ditions for at least several million years. Gradualincreases in carbonate d 13C after glaciation anddeposition of cap carbonates suggests slowrestored increases in primary productivity andorganic carbon burial.

Carbon isotopic evidence suggests that reox-ygenation, at least of seawater, was extremelygradual throughout the remainder of the Neopro-terozoic. Organic matter through the terminalProterozoic derives largely from bacterialheterotrophs, particularly sulfate reducing bac-teria, as opposed to primary producers (Loganet al., 1995). These authors suggested thatthroughout the terminal Neoproterozoic, anaero-bic heterotrophy dominated by sulfate reductionwas active throughout the water column, and O2

penetration from surface waters into the deepocean was inhibited. Shallow-water oxygen-deficient environments became widespread atthe Precambrian–Cambrian boundary (Kimuraand Watanabe, 2001), corresponding to negativecarbonate d 13C excursions and significant bio-logical evolution from Ediacaran-type metazoansto emergence of modern metazoan phyla in theCambrian.


8.11.6.4 Phanerozoic Atmospheric O2

8.11.6.4.1 Constraints on PhanerozoicO2 variation

Oxygenation of the atmosphere during thelatest Neoproterozoic led to a fairly stable andwell-oxygenated atmosphere that has persistedthrough the present day. Although directmeasurement or a quantifiable proxy for Phaner-ozoic paleo-pO2

concentrations have not beenreported, multiple lines of evidence point toupper and lower limits on the concentration of O2

in the atmosphere during the past several hundredmillion years. Often cited is the nearly continuousrecord of charcoal in sedimentary rocks since theevolution of terrestrial plants some 350 Ma. Thepresence of charcoal indicates forest firesthroughout much of the Phanerozoic, which areunlikely to have occurred below pO2 ¼ 0.17 atm(Cope and Chaloner, 1985). Combustion andsustained fire are difficult to achieve at lower pO2

.The existence of terrestrial plants themselves alsoprovides a crude upper bound on pO2

for tworeasons. Above the compensation point ofpO2

/pCO2ratios, photorespiration outcompetes

photosynthesis, and plants experience zero ornegative net growth (Tolbert et al., 1995).Although plants have developed various adaptivestrategies to accommodate low atmospheric pCO2

or aridity (i.e. C4 and CAM plants), net terrestrialphotosynthesis and growth of terrestrial ecosys-tems are effectively inhibited if pO2

rises too high.This upper bound is difficult to exactly constrain,but is likely to be ,0.3–0.35 atm. Woody tissueis also extremely susceptible to combustion athigh pO2

, even if the tissue is wet. Thus, terrestrialecosystems would be unlikely above some upperlimit of pO2

because very frequent reoccurrenceof wildfires would effectively wipe out terrestrialplant communities, and there is no evidence ofthis occurring during the Phanerozoic on a globalscale. What this upper pO2

limit is, however,remains disputed. Early experiments used combus-tion of paper under varying humidity and pO2

(Watson, 1978) but paper may not be the mostappropriate analog for inception of fires in woodytissue with greater moisture content and thermalthickness (Robinson, 1989; Berner et al., 2002).Nonetheless, the persistence of terrestrial plantcommunities from the middle Paleozoic throughthe present does imply that pO2

concentrationshave not risen too high during the past 350 Myr.

Prior to evolution of land plants, even circum-stantial constraints on early Paleozoic pO2

aredifficult to obtain. Invertebrate metazoans, whichhave a continous fossil record from the Cambrianon, require some minimum amount of dissolvedoxygen to support their aerobic metabolism. Theabsolute minimum dissolved O2 concentrationable to support aerobic metazoans varies from

species to species, and is probably impossible toreconstruct for extinct lineages, but moderninfaunal and epifaunal metazoans can accommo-date dissolved O2 dropping to the tens ofmicromolar in concentration. If these concen-trations are extrapolated to equilibrium of theatmosphere with well-mixed cold surface waters,they correspond to ,0.05 atm pO2

. Although localand regional anoxia occurred in the oceans atparticular episodes through Phanerozoic time, thecontinued presence of aerobic metazoans suggeststhat widespread or total ocean anoxia did notoccur during the past ,600 Myr, and that pO2

hasbeen maintained at or above ,5% O2 since theearly Paleozoic.

8.11.6.4.2 Evidence for variationsin Phanerozoic O2

Several researchers have explored possiblelinks between the final oxygenation of the atmos-phere and explosion of metazoan diversity at thePrecambrian–Cambrian boundary (McMenaminand McMenamin, 1990; Gilbert, 1996).While theCambrian explosion may not record the origin ofthese phyla in time, this boundary does record thedevelopment of large size and hard skeletonsrequired for fossilization (Thomas, 1997). Indeed,molecular clocks for diverse metazoan lineagestrace the origin of these phyla to ,400 Myr beforethe Cambrian explosion (Doolittle et al., 1996;Wray et al., 1996). The ability of lineages todevelop fossilizable hardparts may be linked withincreasing oxygenation during the latest Precam-brian. Large body size requires elevated pO2

anddissolved O2 concentrations, so the diffusion cansupply O2 to internal tissues. Large body sizeprovides several advantages, and may haveevolved rather quickly during the earliest Cam-brian (Gould, 1995), but large size also requiresgreater structural support. The synthesis ofcollagen (a ubiquitous structural protein amongall metazoans and possible precursor to inorganicstructural components such as carbonate orphosphate biominerals) requires elevated O2.The threshold for collagen biosynthesis, andassociated skeletonization and development oflarge body size, could not occur until pO2

reachedsome critical threshold some time in the latestPrecambrian (Thomas, 1997).

There is evidence to suggest that the latePaleozoic was a time of very elevated pO2

, toconcentrations substantially greater than observedin the modern atmosphere. Coals from theCarboniferous and Permian contain a greaterabundance of fusain, a product of woody tissuecombustion and charring, than observed for anyperiod of the subsequent geologic past (Robinson,1989, 1991; Glasspool, 2000), suggesting moreabundant forest fires and by implication possibly


higher pO2at this time. Less ambiguous is the

biological innovation of gigantism at this timeamong diverse arthropod lineages (Graham et al.,1995). All arthropods rely on tracheal networksfor diffusion of O2 to support their metabolism;active pumping of O2 through a vascular systemas found in vertebrates does not occur.This sets upper limits on body size for a givenpO2

concentration. In comparison with arthropodcommunities of the Carboniferous and Permian,modern terrestrial arthropods are rather small.Dragonflies at the time reached 70 cm wingspan,mayflies reached 45 cm wingpans, millipedesreached 1 m in length. Even amphibians, whichdepend in part on diffusion of O2 through theirskin for aerobic respiration, reached giganticsize at this time. These large body sizes couldnot be supported by today’s 21% O2 atmosphere,and instead require elevated pO2

between 350 Maand 250 Ma. Insect taxa that were giants in theCarboniferous do not survive past the Permian,suggesting that declining pO2

concentrations inthe Permian and Mesozoic led to extinction(Graham et al., 1995; Dudley, 1998).

Increases in atmospheric O2 affect organisms inseveral ways. Increased O2 concentration facili-tates aerobic respiration, while elevated pO2

against a constant pN2increases total atmospheric

pressure, with associated changes in atmosphericgas density and viscosity (Dudley, 1998, 2000). Intandem these effects may have played a strong rolein the innovation of insect flight in the Carbon-iferous (Dudley, 1998, 2000), with secondarypeaks in the evolution of flight among birds, batsand other insect lineages corresponding to timesof high O2 in the late Mesozoic (Dudley, 1998).

In spite of elevated pO2in the late Paleozoic,

leading up to this time was an extended period ofwater column anoxia and enhanced burial oforganic matter in the Devonian. Widespreaddeposition of black shales, fine-grained laminatedsedimentary rocks rich in organic matter, duringthe Devonian indicate at least partial stratificationof several ocean basins around the globe, withoxygen deficiency throughout the water column.One example of this is from the Holy CrossMountains of Poland, from which particularmolecular markers for green sulfur bacteria havebeen isolated (Joachimski et al., 2001). Theseorganisms are obligately anaerobic chemopho-toautotrophs, and indicate that in the Devonianbasin of central Europe, anoxia extended upwardsthrough the water column well into the photiczone. Other black shales that indicate at leastepisodic anoxia and enhanced organic matterburial during the Devonian are found at severalsites around the world, including the ExshawShale of Alberta (Canada), the Bakken Shale ofthe Williston Basin (Canada/USA), the WoodfordShale in Oklahoma (USA), and the many

Devonian black shales of the Illinois, Michigan,and Appalachian Basins (USA). Widespreadburial of organic matter in the late Devonian hasbeen linked to increased fertilization of the surfacewaters through accelerated continental weatheringdue to the rise of terrestrial plant communities(Algeo and Scheckler, 1998). High rates ofphotosynthesis with relatively small rates ofglobal respiration led to accumulation of organicmatter in marine sediments, and the beginnings ofa pulse of atmospheric hyperoxia that extendedthrough the Carboniferous into the Permian.Coalescence of continental fragments to formthe Pangean supercontinent at this time led towidespread circulation-restricted basins thatfacilitated organic matter burial and net oxygenrelease, and later, to extensive infilling to generatenear-shore swamps containing terrestrial veg-etation which was often buried to form coaldeposits during the rapidly fluctuating sea levelsof the Carboniferous and Permian.

The largest Phanerozoic extinction occurred atthe end of the Permian (,250 Ma). A noticeabledecrease in the burial of organic matter in marinesediments across the Permian–Triassic boundarymay be associated with a global decline in primaryproductivity, and thus, with atmospheric pO2

. Thegigantic terrestrial insect lineages, thought torequire elevated pO2

, do not survive across thisboundary, further suggesting a global drop in pO2

,and the sedimentary and sulfur isotope recordsindicate an overall increase in sulfate reduction andburial of pyritic shales (Berner, 2002; Beerling andBerner, 2000). Although a long-duration deep-seaanoxic event has been proposed as a cause for thePermian mass extinction, there are competingmodels to explain exactly how this might haveoccurred. Hotinski et al., (2001) has shown thatwhile stagnation of the water column to generatedeep-water anoxia might at first seem attractive,global thermohaline stagnation would starve theoceans of nutrients, extremely limiting primaryproductivity, and thus shutting down dissolved O2

demand in the deep oceans. Large negativeexcursions in carbon, sulfur, and strontium iso-topes during the late Permian may indicatestagnation and reduced ventilation of seawater forextended periods, coupled with large-scale over-turn of anoxic waters. Furthermore, sluggishthermohaline circulation at this time could derivefrom a warmer global climate and warmer water atthe sites of high-latitude deep-water formation(Hotinski et al., 2001). The late Permian paleogeo-graphy of one supercontinent (Pangea) and onesuperocean (Panthalassa) was very different fromthe arrangement of continents and oceans on themodern Earth. Coupled with elevated pCO2

at thetime (Berner, 1994; Berner and Kothavala, 2001),GCM models predict warmer climate, weaker windstress, and low equator to pole temperature


gradients. Although polar deep-water formationstill occurred, bringing O2 from the atmosphere tothe deep oceans, anoxia was likely to develop atmid-ocean depths (Zhang et al., 2001), andthermohaline circulation oscillations betweenthermally versus salinity driven modes of circula-tion were likely to develop. During salinity drivenmodes, enhanced bottom-water formation inwarm, salty low-latitude regions would limitoxygenation of the deep ocean. Thus, althoughsustained periods of anoxia are unlikely to havedeveloped during the late Permian, reducedoxygenation of deep water through sluggishthermohaline circulation, coupled with episodicanoxia driven by low-latitude warm salty bottom-water formation, may have led to reoccurringepisodes of extensive ocean anoxia over period ofseveral million years.

Other researchers have invoked extraterrestrialcauses for the End-Permian extinction and anoxia.Fullerenes (cage-like hydrocarbons that effec-tively trap gases during formation and heating)have been detected in late Permian sedimentsfrom southern China. The noble gas complementin these fullerenes indicates an extraterrestrialorigin, which has been interpreted by Kaiho et al.,(2001) to indicate an unrecognized bolide impactat the Permian–Triassic boundary. The abruptdecrease in d 34S across this boundary (from 20‰to 5‰) implies an enormous and rapid release of34S-depleted sulfur into the ocean–atmospheresystem. These authors propose that volatilizedbolide- and mantle-derived sulfur (,0‰) oxi-dized in air, consumed atmospheric and dissolvedO2, and generated severe oxygen and acid stress inthe oceans. Isotope mass balance estimates require,1019 mol sulfur to be released, consuming asimilar mass of oxygen. 1019 mol O2 representssome 10–40% of the total available inventory ofatmospheric and dissolved O2 at this time,removal of which led to immediate anoxia, asthese authors propose.

Other episodes of deep ocean anoxia andextensive burial of organic matter are knownfrom the Jurassic, Cretaceous, Miocene, andPleistocene, although these have not been linkedto changes in atmospheric O2 and instead serve asexamples of the decoupling between atmosphericand deep-ocean O2 concentrations through muchof the geologic past. Widespread Jurassic blackshale facies in northern Europe (PosidonienSchiefer, Jet Rock, and Kimmeridge Clay) weredeposited in a restricted basin on a shallowcontinental margin. Strong monsoonal circulationled to extensive freshwater discharge and a low-salinity cap on basin waters during summer, andintense evaporation and antiestuarine circulationduring winter (Rohl et al., 2001), both of whichcontributed to water column anoxia and blackshale deposition. Several oceanic anoxic events

(OAEs) are recognized from the Cretaceous in allmajor ocean basins, suggesting possible globaldeep-ocean anoxia. Molecular markers of greensulfur bacteria, indicating photic zone anoxia,have been detected from Cenomanian–Turonianboundary section OAE sediments from the NorthAtlantic (Sinninghe Damste and Koster, 1998).The presence of these markers (namely,isorenieratene, a diaromatic carotenoid accessorypigment used during anoxygenic photosynthesis)indicates that the North Atlantic was anoxic andeuxinic from the base of the photic zone (,50–150 m) down to the sediment. High concentrationsof trace metals scavenged by sulfide and anabsence of bioturbation further confirm anoxiathroughout the water column. Because mid-Cretaceous oceans were not highly productive,accelerated dissolved O2 demand from high ratesof respiration and primary production cannot bethe prime cause of these OAEs. Most likely, thewarm climate of the Cretaceous led to low O2

bottom waters generated at warm, high salinityregions of low-latitude oceans. External forcing,perhaps through Milankovic-related precession-driven changes in monsoon intensity and strength,influenced the rate of salinity-driven deep-waterformation, ocean basin oxygenation, and OAEformation (Wortmann et al., 1999). Sapropels areorganic matter-rich layers common to late Cen-ozoic sediments of the eastern Mediterranean.They are formed through a combination ofincreased primary production in surface waters,and increased organic matter preservation in thesediment likely to be associated with changes inventilation and oxygenation of the deep easternMediterranean basins (Stratford et al., 2000). Thewell-developed OMZ located off the coast ofSouthern California today may have been moreextensive in the past. Variations in climateaffecting intensity of upwelling and primaryproduction, coupled with tectonic activity alteringthe depth of basins and height of sills along theCalifornia coast, have generated a series of anoxia-facies organic-matter-rich sediments along thewest coast of North America, beginning withthe Monterey Shale and continuing through to themodern sediments deposited in the Santa Barbaraand Santa Monica basins.

As shown by the modeling efforts of Hotinskiet al., (2001) and Zhang et al., (2001), extensivedeep ocean anoxia is difficult to achieve forextended periods of geologic time during thePhanerozoic when pO2

were at or near modernlevels. Thus, while localized anoxic basins arecommon, special conditions are required togenerate widespread, whole ocean anoxia suchas observed in the Cretaceous. Deep-waterformation in highly saline low-latitude waterswas likely in the geologic past when climates werewarmer and equator to pole heat gradients were


reduced. Low-latitude deep-water formation has asignificant effect on deep-water oxygenation, notentirely due to the lower O2 solubility in warmerwaters, but also the increased efficiency of nutrientuse and recycling in low-latitude surface waters(Herbert and Sarmiento, 1991). If phytoplanktonwere 100% efficient at using and recyclingnutrients, even with modern high-latitude modesof cold deep-water formation, the deep oceanswould likely become anoxic.

8.11.6.4.3 Numerical models of Phanerozoicoxygen concentration

Although photosynthesis is the ultimate sourceof O2 to the atmosphere, in reality photosynthesisand aerobic respiration rates are very closelycoupled. If they were not, major imbalances inatmospheric CO2, O2, and carbon isotopes wouldresult. Only a small fraction of primary production(from photosynthesis) escapes respiration in thewater column or sediment to become buried indeep sediments and ultimately sedimentary rocks.This flux of buried organic matter is in effect“net photosynthesis”, or total photosynthesisminus respiration. Thus, while over timescales ofdays to months, dissolved and atmospheric O2 mayrespond to relative rates of photosynthesis or res-piration, on longer timescales it is burial of organicmatter in sediments (the “net photosynthesis”) thatmatters. Averaged over hundreds of years orlonger, burial of organic matter equates to releaseof O2 into the “atmosphere þ ocean” system:

Photosynthesis:

6CO2 þ 6H2O ! C6H12O6 þ 6O2 ð3Þ

Respiration:

C6H12O6 þ 6O2 þ 6CO2 þ 6H2O ð4Þ

If for every 1,000 rounds of photosynthesis thereare 999 rounds of respiration, the net result is oneround of organic matter produced by photosyn-thesis that is not consumed by aerobic respiration.The burial flux of organic matter in sedimentsrepresents this lack of respiration, and as such is anet flux of O2 to the atmosphere. In geochemists’shorthand, we represent this by the reaction

Burial of organic matter in sediments:

CO2 þ H2O ! O2 þ “CH2O” ð5Þ

where “CH2O” is not formaldehyde, or even anyspecific carbohydrate, but instead representssedimentary organic matter. Given the elementalcomposition of most organic matter in sedimentsand sedimentary rocks, a more reduced organicmatter composition might be more appropriate,i.e., C10H12O, which would imply release of12.5 mol O2 for every mole of CO eventually

buried as organic matter. However, the simplifiedstoichiometry of (5) is applied for most geochem-ical models of C–S–O cycling.

If burial of organic matter equates to O2 releaseto the atmosphere over long timescales, then theoxidative weathering of ancient organic matter insedimentary rocks equates to O2 consumption.This process has been called “georespiration” bysome authors (Keller and Bacon, 1998). It can berepresented by Equation (6), the reverse of (5):

Weathering of organic matter from rocks:

O2 þ “CH2O” ! CO2 þ H2O ð6Þ

Both Equations (3) and (4) contain terms foraddition and removal of O2 from the atmosphere.Thus, if we can reconstruct the rates of burial andweathering of OM into/out of sedimentary rocksthrough time, we can begin to quantify sourcesand sinks for atmospheric O2. The physical mani-festation of this equation is the reaction of organicmatter with O2 during the weathering and erosionof sedimentary rocks. This is most clearly seen inthe investigation of the changes in OM abundanceand composition in weathering profiles developedon black shales (Petsch et al., 2000, 2001).

In addition to the C–O system, the coupledC–S–O system has a strong impact on atmos-pheric oxygen (Garrels and Lerman, 1984; Kumpand Garrels, 1986; Holland, 1978, 1984). This isthrough the bacterial reduction of sulfate to sulfideusing organic carbon substrates as electrondonors. During bacterial sulfate reduction (BSR),OM is oxidized and sulfide is produced fromsulfate. Thus, BSR provides a means of resupply-ing oxidized carbon to the “ocean þ atmosphere”system without consuming O2. The net reactionfor BSR shows that for every 15 mol of OMconsumed, 8 mol sulfate and 4 mol ferric-iron arealso reduced to form 4 mol of pyrite (FeS2):

4 FeðOHÞ3 þ 8 SO224 þ 15 CH2O ! 4 FeS2

þ 15 HCO23 þ 13 H2O þ ðOHÞ2 ð7Þ

Oxidation of sulfate using organic substrates aselectron donors provides a means of restoringinorganic carbon to the ocean þ atmospheresystem without consuming free O2. Every 4 molof pyrite derived from BSR buried in sedimentsrepresents 15 mol of O2 produced by photosyn-thesis to generate organic matter that will not beconsumed through aerobic respiration. In effect,pyrite burial equates to net release of O2 to theatmosphere, as shown by (8), obtained by theaddition of Equation (7) to (5):

4FeðOHÞ3 þ 8SO224 þ 15CH2O

! 4FeS2 þ 15HCO23 þ 13H2O þ ðOHÞ2

þ15CO2 þ 15H2O ! 15O2 þ 15“CH2O”


4FeðOHÞ3 þ 8SO224 þ 15CO2 þ 2H2O

! 4FeS2 þ 15HCO23 þ ðOHÞ2 þ 15O2 ð8Þ

Oxidative weathering of sedimentary sulfideminerals during exposure and erosion on the con-tinents results in consumption of O2 (Equation (9):

4FeS2 þ 15O2 þ 14H2O ! 4FeðOHÞ3

þ 8SO224 þ 16Hþ ð9Þ

Just as for the C–O geochemical system, if wecan reconstruct the rates of burial and oxidativeweathering of sedimentary sulfide mineralsthrough geologic time, we can use these toestimate additional sources and sinks for atmos-pheric O2 beyond organic matter burial andweathering.

In total, then, the general approach taken inmodeling efforts of understanding Phanerozoic O2

variability is to catalog the total sources and sinksfor atmospheric O2, render these in the form of arate of change equation in a box model, andintegrate the changing O2 mass through timeimplied by changes in sources and sinks:

dMO2=dt ¼

XFO2

into the atmosphere

2X

FO2out of the atmosphere

¼ Fburial of organic matter

þ ð15=8ÞFburial of pyrite

2 Fweathering of organic matter

2 ð15=8ÞFweathering of pyrite ð10Þ

One approach to estimate burial and weath-ering fluxes of organic matter and sedimentarysulfides through time uses changes in the relativeabundance of various sedimentary rock typesestimated over Phanerozoic time. Some sedi-mentary rocks are typically rich in both organicmatter and pyrite. These are typically marineshales. In contrast, coal basin sediments containmuch organic carbon, but very low amounts ofsedimentary sulfides. Non-marine coarse-grainedclastic sediments contain very little of eitherorganic matter or sedimentary sulfides. Bernerand Canfield (1989) simplified global sedimen-tation through time into one of three categories:marine shales þ sandstone, coal basin sediments,and non-marine clastic sediments. Using rockabundance estimates derived from the data ofRonov and others (Budyko et al., 1987; Ronov,1976), these authors estimated burial rates fororganic matter and pyrite as a function of timefor the past ,600 Myr (Figure 12). Weatheringrates for sedimentary organic matter and pyritewere calculated as first order dependents on thetotal mass of sedimentary organic matter orpyrite, respectively. Although highly simplified,this model provided several new insights into

global-scale coupling C–S–O geochemistry.First, the broad-scale features of PhanerozoicO2 evolution were established. O2 concentrationsin the atmosphere were low in the earlyPaleozoic, rising to some elevated pO2

levelsduring the Carboniferous and Permian (probablyto a concentration substantially greater thantoday’s 0.21 bar), and then falling through theMesozoic and Cenozoic to more modern values.This model confirmed the suspicion that incontrast with the Precambrian, Phanerozoic O2

evolution was a story of relative stabilitythrough time, with no great excursions in pO2

.Second, by linking C–S–O cycles with sedi-ments and specifically sedimentation rates, thismodel helped fortify the idea that a strongcontrol on organic matter burial rates globally,and thus ultimately on release of O2 to theatmosphere, may be rates of sedimentation innear-shore environments. These authorsextended this idea to propose that the closelinkage between sedimentation and erosion (i.e.,the fact that global rates of sedimentation arematched nearly exactly to global rates ofsediment production—in other words, erosion)may in fact be a stabilizing influence onatmospheric O2 fluctuations. If higher sedimen-tation rates result in greater burial of organicmatter and pyrite, and greater release of O2 tothe atmosphere, at the same time there will begreater rates of erosion on the continents, someof which will involve oxidative weathering ofancient organic matter and/or pyrite.

The other principal approach towards modelingthe Phanerozoic evolution of atmospheric O2 restson the isotope systematics of the carbon andsulfur geochemical cycles. The significant isotopediscriminations associated with biological fix-ation of CO2 to generate biomass and withbacterial reduction of sulfate to sulfide havebeen mentioned several times previously in thischapter. Given a set of simplifications of theexogenic cycles of carbon, sulfur and oxygen,these isotopic discriminations and the isotopiccomposition of seawater through time (d13C,d 34S) can be used to estimate global rates ofburial and weathering of organic matter, sedi-mentary carbonates, pyrite, and evaporativesulfates (Figure 13).

(i) The first required simplification is that thetotal mass of exogenic carbon is constant throughtime (carbon in the oceans, atmosphere, andsedimentary rocks). This of course neglects inputsof carbon and sulfur from volcanic activity andmetamorphic degassing, and outputs into themantle at subduction zones. However, if thesefluxes into and out of the exogenic cycle are smallenough (or have no effect on bulk crustal carbonor sulfur isotopic composition), then this simpli-fication may be acceptable.


–600

10

Org C Burial

Time (Myr)

8

6

4

2

Org

C B

uria

l (10

–18 m

ol M

yr–1

)

0–500 –300 –200 –100 0–400

(a)

Pyr S Burial

–600

2

Time (Myr)

1.5

0.5

1

Pyr

S B

uria

l (10

–18 m

ol M

yr–1

)

0–500 –300 –200 –100 0–400

(b)

–600

100

Time (Myr)

80Oxygen

60

20

40

Oxy

gen

(10–1

8 mol

)

0–500

O S D C P J KTR T

–300 –200 –100 00

5

10

15

20

25

30

35

40

O2 (%)

–400(c)

Figure 12 Estimated organic carbon burial (a), pyrite sulfur burial (b), and atmospheric oxygen concentrations(c) through Phanerozoic time, derived from estimates of rock abundance and their relative organic carbon and

sulfide content (source Berner and Canfield, 1989).


(ii) The second simplification is that the totalmass of carbon and sulfur dissolved in seawaterplus the small reservoir of atmospheric carbonand sulfur gases remains constant through time.For carbon, this may be a realistic simplification.Dissolved inorganic carbon in seawater is

strongly buffered by carbonate mineral precipi-tation and dissolution, and thus it is unlikely thatextensive regions of the world ocean could havebecome significantly enriched or depleted ininorganic carbon during the geologic past. Forsulfur, this assumption may not be completelyaccurate. Much of the interpretation of sulfurisotope records (with implications for atmos-pheric O2 evolution) in the Precambrian dependon varying, but generally low dissolved sulfateconcentrations. Unlike carbonate, there is nogreat buffering reaction maintaining stable sulfateconcentrations in seawater. And while thesources of sulfate to seawater have likely variedonly minimally, with changes in the sulfate fluxto seawater increasing or decreasing smoothlythrough time as the result of broad-scale tectonicactivity and changes in bulk continental weath-ering rates, removal of sulfate through excessiveBSR or rapid evaporate formation may be muchmore episodic through time, possibly resulting infairly extensive shifts in seawater sulfate con-centration, even during the Phanerozoic. None-theless, using this simplification allows us toestablish that for C–S–O geochemical models,total weathering fluxes for carbon and sulfur mustequal total burial fluxes for carbon and sulfur,respectively.

Referring to Figure 14, we can see that thesesimplifications allow us to say that

dMoc=dt ¼ Fwg þ Fwc 2 Fbg 2 Fbc ¼ 0 ð11Þ

dMos=dt ¼ Fws þ Fwp 2 Fbs 2 Fbp ¼ 0 ð12Þ

5

4

3

2

1

0

–1

–2

30

25

30

15

10

5

0

500 400 300 200

Time (Ma)

d34S(‰)

d13C(‰)

100 0

Figure 13 Globally averaged isotopic composition ofcarbonates (d13C) and sulfates (d34S) through Phaner-

ozoic time (source Lindh, 1983).

Weathering

Burial

Weathering

Burial

OC

Ocean + atmosphere

∑CO2

OS

Dissolved seawater

sulfate

E

Evaporite

sulfates

w wbb

P

Sedimentary

sulfides

G

Organic matter

in rocks

C

Carbonate

rocks

Figure 14 The simplified geochemical cycles of carbon and sulfur, including burial and weathering of sedimentarycarbonates, organic matter, evaporites, and sulfides. The relative fluxes of burial and weathering of organic matter and

sulfide minerals plays a strong role in controlling the concentration of atmospheric O2.


The full rate equations for each reservoir massin Figure 14 are as follows:

dMc=dt ¼ Fbc 2 Fwc ð13Þ

dMg=dt ¼ Fbg 2 Fwg ð14Þ

dMs=dt ¼ Fbs 2 Fws ð15Þ

dMp=dt ¼ Fbp 2 Fwp ð16Þ

dMO2=dt ¼ Fbg 2 Fwg þ 15=8ðFbp 2 FwpÞ ð17Þ

This system of equations has four unknowns:two burial fluxes and two weathering fluxes.

(iii) At this point, a third simplification of thecarbon and sulfur systems is often applied to theweathering fluxes of sedimentary rocks. As a firstapproach, it is not unreasonable to guess that therate of weathering of a given type of rock relates insome sense to the total mass of that rock typeavailable on Earth’s surface. If that relation isassumed to be first order with respect to rockmass, an artificial weathering rate constant foreach rock reservoir can be derived. Such constantshave been derived by assuming that the weath-ering rate equation has the form Fwi ¼ ki Mi: If wecan establish the mass of a sedimentary rockreservoir i, and also the average global river fluxto the oceans due to weathering of reservoir i,then ki can easily be calculated. For example, ifthe total global mass of carbonate in sedimentaryrock is 5000 £ 1018 mol C, and annually thereare 20 £ 1012 mol C discharged from rivers to theoceans from carbonate rock weathering, thenkcarbonate becomes (20 £ 1012 mol C yr21)/(5,000 £ 1018 mol C), i.e., 4 £ 1023 Myr21.These simple first-order weathering rate constantshave been calculated for each sedimentary rock re-servoir in the C–S–O cycle, derived entirely fromestimated preanthropogenic carbon and sulfurfluxes from continental weathering. Lack of atrue phenomenological relationship relating mic-roscale and outcrop-scale rock weathering reac-tions to regional- and global-scale carbon andsulfur fluxes remains one of the primary weak-nesses limiting the accuracy of numerical modelsof the coupled C–S–O geochemical cycles.

If weathering fluxes from eroding sedimentaryrocks are independently established, such asthrough use of mass-dependent weatheringfluxes, then all of the weathering fluxes in oursystem of equations above become effectivelyknown. This leaves only the burial fluxes as un-knowns in solving the evolution of the C–S–Osystem.

The exact isotope discrimination that occursduring photosynthesis and during BSR is depen-dent on many factors. For carbon, these include

cell growth rate, geometry and nutrient avail-ability (Rau et al., 1989, 1992), CO2 availabilityspecies-specific effects, modes of CO2 sequestra-tion (i.e., C3, C4, and CAM plants). For sulfur,these can include species–specific effects, thedegree of closed-system behavior and sulfateconcentration, sulfur oxidation and disproportio-nation. However, as a simplification in geochem-ical modeling of the C–S–O cycles, variability incarbon and sulfur isotopic fractionation is limited.In the simplest case, fractionations are constantthrough all time for all environments. As a result,for example, the isotopic composition of organicmatter buried at a given time is set at a constant25‰ depletion relative to seawater dissolvedcarbonate at that time, and pyrite isotopiccomposition is set at a constant 35‰ depletionrelative to seawater sulfate. Of course, in reality,ac and as (the isotopic discriminations assignedbetween inorganic–organic carbon and sulfate–sulfide, respectively) vary greatly in both time andspace. Regardless of how ac and as are set,however, once fractionations have been defined,the mass balance equations given in (13)–(17)above can be supplemented with isotope massbalances as well.

Based on the first simplification listed above,the exogenic cycles of carbon and sulfur areregarded as closed systems. As such, the bulkisotopic composition of average exogenic carbonand sulfur do not vary through time. We can writea rate equation for the rate of change in (mass £isotopic composition) of each reservoir inFigure 14 to reflect this isotope mass balance.For example:

dðdocMocÞ

dt;

docdMoc

dtþ

Mocddoc

dt

¼ dcFwc þ dcFwg 2 docFbg

2 ðdoc 2 acÞFbgð18Þ

Using the simplification that dMoc=dt ¼ 0;Equation (18) reduces to an equation relating theorganic matter burial flux Fbg in terms of otherknown entities:

Fbg ¼� 1

ac

�hMoc

� ddoc

dt

�þ Fwcðdoc 2 dcÞ

þ Fwgðdoc 2 dgÞi

ð19Þ

Using (11) above,

Fbc ¼ Fwg þ Fwc 2 Fbg ð20Þ

A similar pair of equations can be written forthe sulfur system:


Fbp ¼

�1

as

��Mos

�ddos

dt

�þ Fwsðdos 2 dsÞ

þ Fwpðdos 2 dpÞ

ð21Þ

Fbs ¼ Fwp þ Fws 2 Fbp ð22Þ

Full rate equations for the rate of change insedimentary rock reservoir isotopic compositioncan be written as

ddc

dt¼

Fbcðdoc 2 dcÞ

Mc

ð23Þ

ddg

dt¼

Fbgðdoc 2 ac 2 dgÞ

Mg

ð24Þ

dds

dt¼

Fbwðdos 2 dsÞ

Ms

ð25Þ

ddp

dt¼

Fbpðdos 2 as 2 dpÞ

Mp

ð26Þ

as well as the rate of change in the mass of O2. Theisotopic composition of seawater carbonate andsulfate through time comes directly from thesedimentary rock record. However, it is uncertainhow to average globally as well as over substantialperiod of geologic time. The rates of change inseawater carbonate and sulfate isotopic compo-sitions are fairly small terms, and have been leftout of many modeling efforts directed at thegeochemical C–S–O system; however, for com-pleteness sake these terms are included here.

One implication of the isotope-driven modelingapproach to understanding the C–S–O coupledcycle is that burial fluxes of organic matter andpyrite (which are the sources of atmospheric O2 inthese models) are nearly proportional to seawaterisotopic composition of inorganic carbon andsulfate. Recalling the equation for organic matterburial above, it is noted that because sedimentarycarbonate and organic matter mass are nearlyconstant through time, weathering fluxes do notvary greatly through time. Likewise, the averageisotopic composition of carbonates and organicmatter are also nearly constant, as is the mass ofdissolved carbonate and the rate of change indissolved inorganic carbon isotopic composition.Assuming that Fwc, Fwg, Moc, ddoc/dt, dc, and dg areconstant or nearly constant, Fbg becomes linearlyproportional to the isotopic composition of sea-water dissolved inorganic carbon. The samerationale applies for the sulfur system, with pyriteburial becoming linearly proportional to theisotopic composition of seawater sulfate. Althoughthese relationships are not strictly true, even withinthe constraints of the model simplifications, these

relationships provide a useful guide for evaluatingchanges in organic matter and pyrite burial fluxesand the impact these have on atmospheric O2,simply by examining the isotopic records ofmarine carbonate and sulfate.

Early efforts to model the coupled C–S–Ocycles yielded important information. The work ofGarrels and Lerman (1984) showed that theexogenic C and S cycles can be treated as closedsystems over at least Phanerozoic time, withoutexchange between sedimentary rocks and the deep“crust þ mantle.” Furthermore, over timescales ofmillions of years, the carbon and sulfur cycles wereseen to be closely coupled, with increase insedimentary organic carbon mass matched byloss of sedimentary pyrite (and vice versa). Othermodels explored the dynamics of the C–S–Osystem. One important advance was promoted byKump and Garrels (1986). In these authors’ model,a steady-state C–S–O system was generated andperturbed by artificially increasing rates of organicmatter burial. These authors tracked the shifts inseawater carbon and sulfur isotopic compositionthat resulted, and compared these results with thetrue sedimentary record. Importantly, these authorsrecognized that although there is a general inverserelationship between seawater d 13C and d 34S, theexact path along an isotope–isotope plot throughtime is not a straight line. Instead, because of thevastly different residence times of sulfate versuscarbonate in seawater, any changes in the C–S–Osystem are first expressed through shifts in carbonisotopes, then sulfur isotopes (Figure 15).The authors also pointed out large-scale divisionsin C–S isotope coupling through Phanerozoictime. In the Paleozoic, organic matter and pyriteburial were closely coupled (largely because thesame types of depositional environment favorburial of marine organic matter and pyrite). Duringthis time, seawater carbonate and sulfur isotopesco-varied positively, indicating concomitantincreases (or decreases) in burial of organic matterand pyrite. During the late Paleozoic, Mesozoic,and Cenozoic, terrestrial depositional environ-ments became important settings for burial oforganic matter. Because pyrite formation andburial in terrestrial environments is extremelylimited, organic matter and pyrite burial becamedecoupled at this time, and seawater carbonate andsulfur isotopes co-varied negatively, indicatingclose matching of increased sedimentary organicmatter with decreased sedimentary pyrite (and viceversa), perhaps suggesting a net balance in O2

production and consumption, and maintenance ofnearly constant, equable pO2

throughout muchof the latter half of the Phanerozoic. The modelof Berner (1987) introduced the concept of rapidrecycling: the effort to numerically representthe observation that younger sedimentary rocksare more likely to be eroded and weathered than are


older sedimentary rocks. Because young sedimen-tary rocks are likely to be isotopically distinct fromolder rocks (because they are recording any recentshifts in seawater carbon or sulfur isotopiccomposition), restoring that isotopically distinctcarbon or sulfur more quickly back into seawaterprovides a type of negative feedback, dampeningexcessively large or small burial fluxes required forisotope mass balance. This negative feedbackserves to reduce calculated fluctuations in organicmatter and pyrite burial rates, which in turn reducefluctuations in release of O2 to the atmosphere.Results from this study also predict large increasesin OM burial fluxes during the Permocarboniferous(,300 Ma) above values present earlier in thePaleozoic. This increase, likely associated withproduction and burial of refractory terrigenousorganic matter (less easily degraded than OMproduced by marine organisms), led to elevatedconcentrations of O2 ,300 Ma.

One flaw with efforts to model the evolution ofPhanerozoic O2 using the carbon and sulfurisotope records is that unreasonably large fluctu-ations in organic matter and pyrite burial fluxes(with coincident fluctuations in O2 productionrates) would result. Attempts to model the wholePhanerozoic generated unreasonably low and highO2 concentrations for several times in thePhanerozoic (Lasaga, 1989), and applications ofwhat seemed to be a realistic feedback based onreality (weathering rates of sedimentary organicmatter and pyrite dependent on the concentrationof O2) were shown to actually become positivefeedbacks in the isotope-driven C–S–O models(Berner, 1987; Lasaga, 1989).

Phosphorus is a key nutrient limiting primaryproductivity in many marine environments.If phosphorus supply is increased, primaryproduction and perhaps organic matter burialwill also increase. Degradation and remineraliza-tion of OM during transit from surface waters intosediments liberates phosphorus, but most of this isquickly scavenged by adsorption onto the surfacesof iron oxyhydroxides. However, work by Ingalland Jahnke (1994) and Van Cappellen and Ingall(1996) has shown that phosphorus recycling andrelease into seawater is enhanced under low O2 oranoxic conditions. This relationship provides astrong negative feedback between primary pro-duction, bottom water anoxia, and atmosphericO2. As atmospheric O2 rises, phosphorus scaven-ging on ferric-iron is enhanced, phosphorusrecycling back into surface waters is reduced,primary production rates are reduced, and O2

declines. If O2 concentrations were to fall,phosphorus scavenging onto ferric-iron would beinhibited, phosphorus recycling back intosurface waters would be accelerated, fuelingincreased primary production and O2 release tothe atmosphere. Van Cappellen and Ingall (1996)applied these ideas to a mathematical model of theC–P–Fe–O cycle to show how O2 concentrationscould be stabilized by phosphorus recycling rates.

Petsch and Berner (1998) expanded the modelof Van Cappellen and Ingall (1996) to include thesulfur system, as well as carbon and sulfur isotopeeffects. This study examined the response of theC–S–O–Fe–P system, and in particular carbonand sulfur isotope ratios, to perturbation in globalocean overturn rates, changes in continental

Figure 15 Twenty-million-year average values of seawater d 34S plotted against concomitant carbonate d 13C forthe last 700 Myr (source Kump and Garrels, 1986).

20 Myr averages

1 0 – 20 Myr2 20 – 40 Myr

34 660 – 680 Myr33 640 – 660 Myr

. .... .

14

1315

16

65

9

84 3

2117

10

12

7

1119

20

34

18

33

22

23

32

31

302621

24

2725 28

3020

d34S(‰)

d13C(‰)

10–1

0

1

2

3

29


weathering, and shifts in the locus of organicmatter burial from marine to terrestrial depocen-ters. Confirming the idea promoted by Kump andGarrels (1986), these authors showed that pertur-bations of the exogenic C–S–O cycle result inshifts in seawater carbon and sulfur isotopiccomposition similar in amplitude and duration toobserved isotope excursions in the sedimentaryrecord.

Other proposed feedbacks stabilizing the con-centration of atmospheric O2 over Phanerozoictime include a fire-regulated PO4 feedback(Kump, 1988, 1989). Terrestrial primary pro-duction requires much less phosphate per moleCO2 fixed during photosynthesis than marineprimary production. Thus, for a given supply ofPO4, much more CO2 can be fixed as biomass andO2 released from photosynthesis on land versus inthe oceans. If terrestrial production proceeds toorapidly, however, pO2

levels may rise slightly andlead to increased forest fires. Highly weatherable,PO4-rich ash would then be delivered throughweathering and erosion to the oceans. Primaryproduction in the oceans would lead to less CO2

fixed and O2 released per mole of PO4.Hydrothermal reactions between seawater and

young oceanic crust have been proposed asan influence on atmospheric O2 (Walker, 1986;Carpenter and Lohmann, 1999; Hansen andWallmann, 2002). While specific periods ofoceanic anoxia may be associated with acceler-ated hydrothermal release of mantle sulfide (i.e.,the Mid-Cretaceous, see Sinninghe-Damste andKoster, 1998), long-term sulfur and carbonisotope mass balance precludes substantial inputsof mantle sulfur to the Earth’s surface of adifferent net oxidation state and mass flux thanwhat is subducted at convergent margins (Petsch,1999; Holland, 2002).

One recent advance in the study of isotope-driven models of the coupled C–S–O cycles isre-evaluation of isotope fractionations. Hayes et al.,(1999) published a compilation of the isotopiccomposition of inorganic and organic carbon forthe past 800 Myr. One feature of this dual record isa distinct shift in the net isotopic distance betweencarbonate and organic carbon, occurring during thepast ,100 million years. When carbon isotopedistance is compared to estimates of Cenozoic andMesozoic pO2

, it becomes apparent that there maybe some relationship between isotopic fraction-ation associated with organic matter productionand burial and the concentration of O2 in theatmosphere. The physiological underpinningbehind this proposed relationship rests on compe-tition between photosynthesis and photorespirationin the cells of photosynthetic organisms. BecauseO2 is a competitive inhibitor of CO2 for attachmentto the active site of Rubisco, as ambient O2

concentrations rise relative to CO2, so will rates

of photorespiration. Photorespiration is a netconsumptive process for plants; previously fixedcarbon is consumed with O2 to produce CO2 andenergy. CO2 produced through photorespirationmay diffuse out of the cell, but it is also likely to betaken up (again) for photosynthesis. Thus, in cellsundergoing fairly high rates of photorespiration inaddition to photosynthesis, a significant fraction oftotal CO2 available for photosynthesis derives fromoxidized, previously fixed organic carbon. Theeffect of this on cellular carbon isotopic compo-sition is that because each round of photosynthesisresults in 13C-depletion in cellular carbon relativeto CO2, cells with high rates of photorespirationwill contain more 13C-depleted CO2 and thus willproduce more 13C-depleted organic matter.

In controlled-growth experiments using bothhigher plants and single-celled marine photosyn-thetic algae, a relationship between ambient O2

concentration and net isotope discrimination hasbeen observed (Figure 16) (Berner et al., 2000;Beerling et al., 2002). The functional form of thisrelationship has been expressed in several ways.The simplest is to allow isotope discrimination tovary linearly with changing atmospheric O2 mass:ac ¼ 25 £ ðMO2

=38Þ: More complicated relation-ships have also been derived, based on curve-fitting the available experimental data on isotopicfractionation as a function of [O2]. O2-dependentisotopic fractionation during photosynthesis hasprovided the first mathematically robust isotope-driven model of the C–S–O cycle consistent withgeologic observations (Berner et al., 2000).Results of this model show that allowing isotopefractionation to respond to changes in ambient O2

+

6

4

2

0

–2

–60 200 400

Growth atmospheric O2/CO2 ratio

∆ (∆

13C

) (p

er m

il)

600 800 1,000 1,200 1,400

–4

Figure 16 Relationship between change in D(D13C) ofvascular land plants determined experimentally inresponse to growth under different O2/CO2 atmosphericmixing ratios. D(D13C) is the change in carbon isotopefractionation relative to fractionation for the controls atpresent day conditions (21% O2, 0.036% CO2). Thesolid line shows the nonlinear curve fitted to the data,given by D(D13C) ¼ 219.94 þ 3.195 £ ln(O2/CO2);(†) Phaseolus vulgaris; (V) Sinapis alba; (A) fromBerner et al. (2000); (þ) from Berry et al. (1972)(Beerling et al., 2002) (reproduced by permission ofElsevier from Geochem. Cosmochim. Acta 2002,

66, 3757–3767).


provides a strong negative feedback dampeningexcessive increases or decreases in organic matterburial rates. Rates of organic matter burial in thismodel are no longer simply dependent on seawatercarbonate d 13C, but now also vary with 1/ac. Asfractionation becomes greater (through elevatedO2), less of an increase in organic matter burialrates is required to achieve the observed increase inseawater d 13C than if ac were constant.

The same mathematical argument can beapplied to sulfate–sulfide isotope fractionationduring BSR. As O2 concentrations increase, sodoes sulfur isotope fractionation, resulting in astrong negative feedback on pyrite burial rates.This is consistent with the broad-scale changes insulfur isotope dynamics across the Proterozoic,reflecting a large increase in D34S (between sulfateand sulfide) when atmospheric O2 concentrationswere great enough to facilitate bacterial sulfideoxidation and sulfur disproportionation. Perhapsduring the Phanerozoic, when O2 concentrationswere greater, sulfur recycling (sulfate to sulfidethrough BSR, sulfur oxidation, and sulfur dis-proportionation) was increased, resulting ingreater net isotopic distance between sulfate andsulfide. Another means of changing net sulfurisotope discrimination in response to O2 may bethe distribution of reduced sulfur between sulfideminerals and organic matter-associated sulfides.Work by Werne et al., 2000, 2003) has shown thatorganic sulfur is consistently ,10 ‰ enriched in34S relative to associated pyrite. This is believed toresult from different times and locations oforganic sulfur versus pyrite formation. Whilepyrite may form in shallow sediments or evenanoxic portions of the water column, reflecting

extreme sulfur isotope depletion due to severalcycles of BSR, sulfide oxidation andsulfur disproportionation, organic matter is sulfur-ized within the sediments. Closed, or nearlyclosed, system behavior of BSR in the sedimentsresults in late-stage sulfide (the source of sulfurin sedimentary organic matter) to be moreenriched compared with pyrite in the samesediments. It is known that burial of sulfide asorganic sulfur is facilitated in low O2 or anoxicwaters. If lower atmospheric O2 in the pastencouraged development of more extensiveanoxic basins and increased burial of sulfide asorganic sulfur instead of pyrite, the 10‰ offsetbetween pyrite and organic sulfur would becomeeffectively a change in net sulfur fractionation inresponse to O2.

Applying these newly recognized modificationsof carbon and sulfur isotope discrimination inresponse to changing O2 availability has alloweddevelopment of new numerical models of theevolution of the coupled C – S – O systemsand variability of Phanerozoic atmospheric O2

concentration (Figure 17).

8.11.7 CONCLUSIONS

Molecular oxygen is generated and consumedby a wide range of processes. The net cycling ofO2 is influenced by physical, chemical, andmost importantly, biological processes acting onand beneath the Earth’s surface. The exactdistribution of O2 concentrations depends on thespecific interplay of these processes in time andspace. Large inroads have been made towards

70

60

50

40

30

20

10

0500

1018

mol

O2

400

Time (Ma)

300 200 100 0

Figure 17 Evolution of the mass of atmospheric O2 through Phanerozoic time, estimated using an isotope massbalance described in Equations (11)–(26). The model employs the isotope date of Figure 13, and includes newadvances in understanding regarding dependence of carbon isotope discrimination during photosynthesis and sulfurisotope discrimination during sulfur disproportionation and organic sulfur formation. The system of coupleddifferential equations were integrated using an implicit fourth-order Kaps–Rentrop numerical integration algorithm

appropriate for this stiff set of equations (sources Petsch, 2000; Berner et al., 2000).


understanding the processes that control the con-centration of atmospheric O2, especially regardingO2 as a component of coupled biogeochemicalcycles of many elements, including carbon, sulfur,nitrogen, phosphorus, iron, and others.

Earth’s modern oxygenated atmosphere is theproduct of over four billion years of its history(Figure 18). The early anoxic atmosphere wasslowly oxidized (although not oxygenated) as theresult of slow H2 escape. Evolution of oxygenicphotosynthesis accelerated the oxidation of Earth’scrust and atmosphere, such that by,2.2 Ga a smallbut significant concentration of O2 was likelypresent in Earth’s atmosphere. Limited primaryproduction and oxygen production compared withthe flux of reduced volcanic gases maintained thislow pO2

atmosphere for over one billion years untilthe Neoproterozoic. Rapid oscillations in Earth’scarbon and sulfur cycles associated with globalSnowball glaciation may also have expression in areturn to atmospheric anoxia at this time, butsubsequent to the late Proterozoic isotope excur-sions, oxygenation of the atmosphere to near-modern concentrations developed such that by thePrecambrian–Cambrian boundary, O2 concen-trations were high enough to support widespreadskeletonized metazoans. Phanerozoic seawaterand atmospheric O2 concentrations have fluctuatedin response to tectonic forcings, generatingregional-scale anoxia in ocean basins at certain

times when biological productivity and oceancirculation facilitate anoxic conditions, but in theatmosphere, O2 concentrations have remainedwithin ,0.05–0.35 bar pO2

for the past ,600 Myr.Several outstanding unresolved gaps in our

understanding remain, in spite of a well-developed understanding of the general featuresof the evolution of atmospheric O2 through time.These gaps represent potentially meaningfuldirections for future research, including:

(i) assessing the global importance of mineraloxidation as a mechanism of O2 consumption;

(ii) the flux of reduced gases from volcanoes,metamorphism, and diffuse mantle/lithospheredegassing;

(iii) the true dependence of organic matteroxidation on availability of O2, in light of the greatabundance of microaerophilic and anaerobicmicroorganisms utilizing carbon respiration as ametabolic pathway, carbon isotopic evidencesuggesting continual and essentially constantorganic matter oxidation as part of the sedimen-tary rock cycle during the entire past four billionyears, and the inefficiency of organic matteroxidation during continental weathering;

(iv) stasis in the oxygenation of the atmosphereduring the Proterozoic;

(v) contrasting biochemical, fossil, and mol-ecular evidence for the antiquity of the innovationof oxygenic photosynthesis; and

10–10

10–9

10–8

10–7

10–6

10–5

10–4

10–3

10–2

10–1

100

H escape fromCH4 photolysis

Evolution ofphotosynthesis

Huronian glaciation,oxidation of crust

oxygenationof atmosphere

Rodiniansupercontinent

Snowballanoxia?

[O2](bar)

Phanerozoic

4,0004,500 3,500 3,000 2,500 2,000 1,500 1,000

Time (Ma)

500 0

Figure 18 Composite estimate of the evolution of atmospheric oxygen through 4.5 Gyr of Earth’s history.Irreversible oxidation of the Earth resulted from CH4 photolysis and hydrogen escape during early Earth’s history.Evolution of oxygenic photosynthesis preceded substantial oxygenation of the atmosphere by several hundred millionyears. Relative stasis in atmospheric pO2

typified much of the Proterozoic, with a possible pulse of oxygenationassociated with formation of the Rodinian supercontinent in the Late Mesoproterozoic, and possible return to anoxiaassociated with snowball glaciation in the Neoproterozoic (sources Catling et al., 2001; Kasting, 1992; Rye and

Holland, 1998; Petsch, 2000; Berner et al., 2000).

Conclusions 551

(vi) evaluating the relative strength of biologi-cal productivity versus chemical evolution of theEarth’s crust and mantle in controlling the earlystages of oxygenation of the atmosphere.

Thus, study of the global biogeochemical cycleof oxygen, the component of our atmosphereintegral and crucial for life as we know it, remainsa fruitful direction for Earth science research.

REFERENCES

Algeo T. J. and Scheckler S. E. (1998) Terrestrial-marineteleconnections in the Devonian: links between the evolutionof land plants, weathering processes, and marine anoxicevents. Phil. Trans. Roy. Soc. London B 353, 113–128.

Anbar A. D. and Knoll A. H. (2002) Proterozoic oceanchemistry and evolution: a bioinorganic bridge? Science297, 1137–1142.

Arthur M. A. (2000) Volcanic contributions to the carbon andsulfur geochemical cycles and global change. In Encyclo-pedia of Volcanoes (eds. H. Sigurdsson, B. F. Houghton,S. R. McNutt, H. Rymer, and J. Stix). Academic Press, SanDiego, pp. 1045–1056.

Bartley J. K., Semikhatov M. A., Kaufman A. J., Knoll A. H.,Pope M. C., and Jacobsen S. B. (2001) Global events acrossthe Mesoproterozoic–Neoproterozoic boundary: C and Srisotopic evidence from Siberia. Precamb. Res. 111,165–202.

Battle M., Bender M., Sowers R., Tans P. P., Butler J. H.,Elkins J. W., Ellis J. T., Conway T., Zhang N., Pang P., andClarke A. D. (1996) Atmospheric gas concentrations overthe past century measured in air from firn at the South Pole.Nature 383, 231–235.

Baum S. K. and Crowley T. J. (2001) GCM response to latePrecambrian (,590 Ma) ice-covered continents. Geophys.Res. Lett. 28, 583–586.

Beaumont V. and Robert F. (1999) Nitrogen isotope ratios ofkerogens in Precambrian cherts: a record of the evolution ofatmospheric chemistry? Precamb. Res. 96, 63–82.

Beerling D. J. and Berner R. A. (2000) Impact of a Permo–Carboniferous high O2 event on the terrestrial carbon cycle.Proc. Natl. Acad. Sci. 97, 12428–12432.

Beerling D. J., Lake J. A., Berner R. A., Hickey L. J., TaylorD. W. and Royer D. L. (2002) Carbon isotope evidenceimplying high O2/CO2 ratios in the Permo–Carboniferousatmosphere. Geochim. Cosmochim. Acta 66, 3757–3767.

Bekker A., Kaufman A. J., Karhu J. A., Beukes N. J., QuintenS. D., Coetzee L. L., and Kenneth A. E. (2001) Chemos-tratigraphy of the Paleoproterozoic Duitschland Formation,South Africa. Implications for coupled climate change andcarbon cycling. Am. J. Sci. 301, 261–285.

Bender M., Labeyrie L. D., Raynaud D., and Lorius C. (1985)Isotopic composition of atmospheric O2 in ice linked withdeglaciation and global primary productivity. Nature 318,349–352.

Bender M., Sowers T., and Labeyrie L. (1994a) The Dole effectand its variations during the last 130,000 years as measured inthe Vostok ice core. Global Biogeochem. Cycle 8, 363–376.

Bender M. L., Sowers T., Barnola J.-M., and Chappellaz J.(1994b) Changes in the O2/N2 ratio of the atmosphereduring recent decades reflected in the composition of airin the firn at Vostok Station, Antarctica. Geophys. Res.Lett. 21, 189–192.

Berner R. A. (1987) Models for carbon and sulfur cycles andatmospheric oxygen: application to Paleozoic geologichistory. Am. J. Sci. 287, 177–196.

Berner R. A. (1994) GEOCARB II: a revised model ofatmospheric CO2 over Phanerozoic time. Am. J. Sci. 294,56–91.

Berner R. A. (2002) Examination of hypotheses for thePermo–Triassic boundary extinction by carbon cyclemodeling. Proc. Natl. Acad. Sci. 99, 4172–4177.

Berner R. A. and Canfield D. E. (1989) A new model foratmospheric oxygen over Phanerozoic time. Am. J. Sci. 289,333–361.

Berner R. A. and Kothavala Z. (2001) GEOCARB III: a revisedmodel of atmospheric CO2 over Phanerozoic time. Am. J. Sci.301, 182–204.

Berner R. A., Petsch S. T., Lake J. A., Beerling D. J., PoppB. N., Lane R. S., Laws E. A., Westley M. B., Cassar N.,Woodward F. I., and Quick W. P. (2000) Isotopefractionation and atmospheric oxygen: Implications forPhanerozoic O2 evolution. Science 287, 1630–1633.

Berry J. A., Troughton J. H., and Bjorkman O. (1972) Effect ofoxygen concentration during growth on carbon isotopediscrimination in C3 and C4 species of Atriplex. CarnegieInst. Yearbook 71, 158–161.

Beukes N. J., Dorland H., Gutzmer J., Nedachi M., and OhmotoH. (2002) Tropical laterites, life on land, and the history ofatmospheric oxygen in the Paleoproterozoic. Geology 30,491–494.

Brasier M. D., Green O. R., Jephcoat A. P., Kleppe A. K., vanKranendonk M. J., Lindsay J. F., Steele A., and GrassineauN. V. (2002) Questioning the evidence for Earth’s earliestfossils. Nature 416, 76–81.

Brocks J. J., Logan G. A., Buick R., and Summons R. E. (1999)Archean molecular fossils and the early rise of eukaryotes.Science 285, 1033–1036.

Budyko M. I., Ronov A. B., and Yanshin A. L. (1987) Historyof the Earth’s Atmosphere. Springer, Berlin.

Buick R. (1990) Microfossil recognition in Archean rocks: anappraisal of spheroids and filaments from a 3500 M. Y. oldchert-barite unit at North Pole, Western Australia. Palaios 5,441–459.

Buick I. S., Uken R., Gibson R. L., and Wallmach T. (1998)High-d 13C Paleoproterozoic carbonates from the TransvaalSupergroup, South Africa. Geology 26, 875–878.

Butterfield N. J. and Rainbird R. H. (1998) Diverse organic-walled fossils, including possible dinoflagellates, from earlyNeoproterozoic of arctic Canada. Geology 26, 963–966.

Caldeira K., and Kasting J. F. (1992) Susceptibility of the earlyEarth to irreversible glaciation caused by carbon dioxideclouds. Nature 359, 226–228.

Canfield D. E. (1998) A new model for Proterozoic oceanchemistry. Nature 396, 450–453.

Canfield D. E. and Teske A. (1996) Late Proterozoic rise inatmospheric oxygen concentration inferred from phylo-genetic and sulphur-isotope studies. Nature 382,127–132.

Canfield D. E., Habicht K. S., and Thamdrup B. (2000) TheArchean sulfur cycle and the early history of atmosphericoxygen. Science 288, 658–661.

Carpenter S. J. and Lohmann K. C. (1999) Carbon isotoperatios of Phanerozoic marine cements: re-evaluating globalcarbon and sulfur systems. Geochim. Cosmochim. Acta 61,4831–4846.

Catling D. C., Zahnle K. J., and McKay C. P. (2001) Biogenicmethane, hydrogen escape, and the irreversible oxidation ofthe early Earth. Science 293, 839–843.

Cloud P. (1972) A working model of the primitive Earth. Am.J. Sci. 272, 537–548.

Cloud P. E. (1976) Beginnings of biospheric evolution andtheir biogeochemical consequences. Paleobiol. 2,351–387.

Condie K. C., Des Marais D. J., and Abbott D. (2001)Precambrian superplumes and supercontinents: a record inblack shales, carbon isotopes, and paleoclimates? Precamb.Res. 106, 239–260.

Cope M. J. and Chaloner W. G. (1985) Wildfire: an interactionof biological and physical processes. In Geological Factorsand the Evolution of Plants (ed. B. H. Tiffney). YaleUniversity Press, New Haven, CT, pp. 257–277.


Delano J. W. (2001) Redox history of the Earth’s interior since,3900 Ma: implications for prebiotic molecules. OriginsLife Evol. Biosphere 31, 311–341.

Delmelle P. and Stix J. (2000) Volcanic gases. In Encyclopediaof Volcanoes (eds. H. Sigurdsson, B. F. Houghton, S. R.McNutt, H. Rymer, and J. Stix). Academic Press, SanDiego, pp. 803–816.

Des Marais D. J., Strauss H., Summons R. E., and Hayes J. M.(1992) Carbon isotope evidence for stepwise oxidation of theProterozoic environment. Nature 359, 605–609.

Deuser W. G. (1975) Chemical Oceanography. AcademicPress, Orlando, FL, p. 3.

Doolittle R. F., Feng D. F., Tsang S., Cho G., and Little E.(1996) Determining divergence times of the major kingdomsof living organisms with a protein clock. Science 271,470–477.

Dudley R. (1998) Atmospheric oxygen, giant Paleozoic insectsand the evolution of aerial locomotor performance. J. Exp.Biol. 201, 1043–1050.

Dudley R. (2000) The evolutionary physiology of animal flight:Paleobiological and present perspectives. Ann. Rev. Phys.62, 135–155.

Evans D. A. D. (2000) Stratigraphic, geochronological andpaleomagnetic constraints upon the Neoproterozoic climaticparadox. Am. J. Sci. 300, 347–433.

Farquhar J., Bao H., and Thiemens M. (2000) Atmosphericinfluence of Earth’s earliest sulfur cycle. Science 289,756–758.

Farquhar J., Wing B. A., McKeegan K. D., Harris J. W.,Cartigny P., and Thiemens M. H. (2002) Mass-independentsulfur of inclusions in diamond and sulfur recycling on earlyEarth. Science 297, 2369–2372.

Garcia-Pichel F. (1998) Solar ultraviolet and the evolutionaryhistory of cyanobacteria. Origins Life Evol. Biosphere 28,321–347.

Garrels R. M. and Lerman A. (1984) Coupling of thesedimentary sulfur and carbon cycles—an improvedmodel. Am. J. Sci. 284, 989–1007.

Gilbert D. L. (1996) Evolutionary aspects of atmosphericoxygen and organisms. In Environmental Physiology: 2 (eds.M. J. Fregly, and C. M. Blatteis). Oxford University Press,Oxford, UK, pp. 1059–1094.

Glasspool I. (2000) A major fire event recorded in themesofossils and petrology of the late Permian, LowerWhybrow coal seam, Sydney Basin, Australia. Palaeogeogr.Palaeoclimat. Palaeoecol. 164, 373–396.

Godderıs Y. and Veizer J. (2000) Tectonic control of chemicaland isotopic composition of ancient oceans: the impact ofcontinental growth. Am. J. Sci. 300, 434–461.

Gould S. J. (1995) Of it and not above it. Nature 377, 681–682.Graham J. B., Dudley R., Anguilar N., and Gans C. (1995)

Implications of the late Paleozoic oxygen pulse forphysiology and evolution. Nature 375, 117–120.

Habicht K. S., Gade M., Thamdrup B., Berg P., and CanfieldD. E. (2002) Calibration of sulfate levels in the Archeanocean. Science 298, 2372–2374.

Haggerty S. E. and Toft P. B. (1985) Native iron in thecontinental lower crust: petrological and geophysicalimplications. Science 229, 647–649.

Halmer M. M., Schmincke H.-U., and Graf H.-F. (2002) Theannual volcanic gas input into the atmosphere, in particularinto the stratosphere: a global data set for the past 100years. J. Volcanol. Geotherm. Res. 115, 511–528.

Hansen K. W. and Wallmann K. (2003) Cretaceous andCenozoic evolution of seawater composition, atmosphericO2 and CO2: a model perspective. Am. J. Sci. 303, 94–148.

Hartnett H. E., Keil R. G., Hedges J. I., and Devol A. H. (1998)Influence of oxygen exposure time on organic carbonpreservation in continental margine sediments. Nature 391,572–574.

Hayes J. M., Strauss H., and Kaufman A. J. (1999) Theabundance of 13C in marine organic matter andisotopic fractionation in the global biogeochemical cycle

of carbon during the past 800 Myr. Chem. Geol. 161,103–125.

Herbert T. D. and Sarmiento J. L. (1991) Ocean nutrientdistribution and oxygenation: limits on the formation ofwarm saline bottom water over the past 91 m.y. Geology 19,702–705.

Hinkle M. E. (1994) Environmental conditions affecting theconcentrations of He, CO2, O2 and N2 in soil gases. Appl.Geochem. 9, 53–63.

Hoehler T. M., Bebout B. M., and Des Marais D. J. (2001) Therole of microbial mats in the production of reduced gases onthe early Earth. Nature 412, 324–327.

Hoffman P. F. and Schrag D. P. (2002) The snowball Earthhypothesis: testing the limits of global change. Terra Nova14, 129–155.

Hoffman P. F., Kaufman A. J., Halverson G. P., and SchragD. P. (1998) A Neoproterozoic snowball Earth. Science 281,1342–1346.

Hofmann H. J., Gery K., Hickman A. H., and Thorpe R. I.(1999) Origin of 2.45 Ga coniform stromatolites in Warra-woona Group, Western Australia. Geol. Soc. Am. Bull. 111,1256–1262.

Holland H. D. (1978) The Chemistry of the Atmosphere andOceans. Wiley, New York.

Holland H. D. (1984) The Chemical Evolution of theAtmosphere and Oceans. Princeton University Press,Princeton, NJ.

Holland H. D. (2002) Volcanic gases, black smokers and thegreat oxidation event. Geochim. Cosmochim. Acta 66,3811–3826.

Holland H. D., Feakes C. R., and Zbinden E. A. (1989) The FlinFlon paleosol and the composition of the atmosphere 1.8BYBP. Am. J. Sci. 289, 362–389.

Hotinski R. M., Bice K. L., Kump L. R., Najjar R. G., andArthur M. A. (2001) Ocean stagnation and End-Permiananoxia. Geology 29, 7–10.

Hurtgen M. T., Arthur M. A., Suits N. S., and Kaufman A. J.(2002) The sulfur isotopic composition of Neoproterozoicseawater sulfate: implications for a snowball Earth? EarthPlanet. Sci. Lett. 203, 413–429.

Ingall E. and Jahnke R. (1994) Evidence for enhancedphosphorus regeneration from marine sediments overlainby oxygen depleted waters. Geochim. Cosmochim. Acta 58,2571–2575.

Ingmanson D. E. and Wallace W. J. (1989) Oceanography: AnIntroduction. Wadworth, Belmont, CA, pp. 99.

Joachimski M. M., Ostertag-Henning C., Pancost R. D.,Strauss H., Freeman K. H., Littke R., Sinninghe-DamsteJ. D., and Racki G. (2001) Water column anoxia,enhanced productivity and concomitant changes in d 13Cand d 34S across the Frasnian–Famennian boundary(Kowala—Holy Cross Mountains/Poland). Chem. Geol.175, 109–131.

Kah L. C., Lyons T. W., and Chesley J. T. (2001) Geochemistryof a 1.2 Ga carbonate-evaporite succession, northern Baffinand Bylot Islands: implications for Mesoproterozoic marineevolution. Precamb. Res. 111, 203–234.

Kaiho K., Kajiwara Y., Nakano T., Miura Y., Kawahata H.,Tazaki K., Ueshima M., Chen Z., and Shi G. (2001)End-Permian catastrophe by a bolide impact: evidence ofa gigantic release of sulfur from the mantle. Geology 29,815–818.

Karhu J. and Holland H. D. (1996) Carbon isotopes and therise of atmospheric oxygen. Geology 24, 867–870.

Kasting J. F. (1987) Theoretical constraints on oxygen andcarbon dioxide concentrations in the Precambrian atmos-phere. Precamb. Res. 34, 205–229.

Kasting J. F. (1992) Models relating to Proterozoic Atmos-pheric and Oceanic Chemistry. In The Proterozoic Bio-sphere (eds. J. W. Schopf, and C. Klein). CambridgeUniversity Press, Cambridge, pp. 1185–1187.

Kasting J. F. and Siefert J. L. (2002) Life and the evolution ofEarth’s atmosphere. Science 296, 1066–1068.

References 553

Kasting J. F., Zahnle K. J., and Walker J. C. G. (1983)Photochemistry of methane in the Earth’s early atmosphere.Precamb. Res. 20, 121–148.

Kasting J. F., Eggler D. H., and Raeburn S. P. (1993) Mantleredox evolution and the oxidation state of the Archeanatmosphere. J. Geol. 101, 245–257.

Kasting J. F., Pavlov A. A., and Siefert J. L. (2001) A coupledecosystem-climate model for predicting the methane con-centration in the Archean atmosphere. Orig. Life Evol.Biosph. 31, 271–285.

Kaufman A. J. and Knoll A. H. (1995) Neoproterozoicvariations in the C-isotopic composition of seawater:stratigraphic and biogeochemical implications. Precamb.Res. 73, 27–49.

Kaufman A. J., Hayes J. M., Knoll A. H., and Germs G. J. B.(1991) Isotopic composition of carbonates and organiccarbon from upper Proterozoic successions in Namibia:stratigraphic variation and the effects of diagenesis andmetamorphism. Precamb. Res. 49, 301–327.

Keeling R. F. and Shertz S. R. (1992) Seasonal and interannualvariations in atmospheric oxygen and implications for thecarbon cycle. Nature 358, 723–727.

Keeling R. F., Najjar R. P., Bender M. L., and Tans P. P. (1993)What atmospheric oxygen measurements can tell us aboutthe global carbon cycle. Global Biogeochem. Cycles 7,37–67.

Keller C. K. and Bacon D. H. (1998) Soil respiration andgeorespiration distinguished by transport analyses of vadoseCO2, 13CO2, and 14CO2. Global Biogeochem. Cycles 12,361–372.

Kiehl J. T. and Dickinson R. E. (1987) A study of theradiative effects of enhanced atmospheric CO2 and CH4 onearly Earth surface temperatures. J. Geophys. Res. 92,2991–2998.

Kimura H. and Watanabe Y. (2001) Oceanic anoxia at thePrecambrian–Cambrian boundary. Geology 29, 995–998.

Kirschvink J. L. (1992) Late Proterozoic low-latitude globalglaciation: the snowball earth. In The Proterozoic Biosphere(eds. J. W. Schopf, and C. Klein). Cambridge UniversityPress, Cambridge, pp. 51–52.

Kump L. R. (1988) Terrestrial feedback in atmospheric oxygenregulation by fire and phosphorus. Nature 335, 152–154.

Kump L. R. (1989) Chemical stability of the atmosphereand ocean. Palaeogeogr. Palaeoclimat. Paleoecol. 75,123–136.

Kump L. R. and Garrels R. M. (1986) Modeling atmosphericO2 in the global sedimentary redox cycle. Am. J. Sci. 286,337–360.

Kump L. R., Kasting J. F., and Barley M. E. (2001) Rise ofatmospheric oxygen and the upside down Archean mantle.Geochem. Geophys. Geosys. 2 No. 2000 GC 000114.

Lasaga A. C. (1989) A new approach to isotopic modeling ofthe variation in atmospheric oxygen through the Phaner-ozoic. Am. J. Sci. 289, 411–435.

Lasaga A. C. and Ohmoto H. (2002) The oxygen geochemicalcycles: dynamics and stability. Geochim. Cosmochim. Acta66, 361–381.

Lenton T. M. and Watson A. J. (2000) Redfield revisited 2.What regulates the oxygen content of the atmosphere?Global Biogeochem. Cycles 14, 249–268.

Lindh T. B. (1983) Temporal variations in 13C, 34S and globalsedimentation during the Phanerozoic. MS Thesis, Univer-sity of Miami.

Logan G. A., Hayes J. M., Hleshima G. B., and Summons R. E.(1995) Terminal Proterozoic reorganization of biogeochem-ical cycles. Nature 376, 53–56.

Lovell J. S. (2000) Oxygen and carbon dioxide in soil air.In Handbook of Exploration Geochemistry, GeochemicalRemote Sensing of the Subsurface (ed. M. Hale). Elsevier,Amsterdam, vol. 7, pp. 451–469.

McMenamin M. A. S. and McMenamin D. L. S. (1990) TheEmergence of Animals—The Cambrian Breakthrough.Columbia University Press, New York.

Morner N.-A. and Etiope G. (2002) Carbon degassing from thelithosphere. Global. Planet. Change 33, 185–203.

Murakami T., Utsunomiya S., Imazu Y., and Prasad N. (2001)Direct evidence of late Archean to early Proterozoic anoxicatmosphere from a product of 2.5 Ga old weathering. EarthPlanet. Sci. Lett. 184, 523–528.

Najjar R. G. and Keeling R. F. (2000) Mean annual cycle of theair–sea oxygen flux: a global view. Global Biogeochem.Cycles 14, 573–584.

Nutman A. P., Mojzsis S. J., and Friend C. R. L. (1997)Recognition of ^3850 Ma water-lain sediments inWest Greenland and their significance for the early ArcheanEarth. Geochim. Cosmochim. Acta 61, 2475–2484.

Ohmoto H. (1996) Evidence in pre-2.2 Ga paleosols for theearly evolution of atmospheric oxygen and terrestrial biota.Geology 24, 1135–1138.

Ohmoto H. (1999) Redox state of the Archean atmosphere:evidence from detrital heavy minerals in ca. 3250–2750 Ma sandstones from the Pilbara Craton, Australia:comment and reply. Geology 27, 1151–1152.

Olson J. M. and Pierson B. K. (1986) Photosynthesis 3.5thousand million years ago. Photosynth. Res. 9, 251–259.

Pan Y. and Stauffer M. R. (2000) Cerium anomaly and Th/Ufractionation in the 1.85 Ga Flin Flon Paleosol: clues fromREE- and U-rich accessory minerals and implications forpaleoatmospheric reconstruction. Am. Mineral. 85, 898–911.

Pavlov A. A., Kasting J. F., Brown L. L., Rages K. A., andFreedman R. (2000) Greenhouse warming by CH4 in theatmosphere of early Earth. J. Geophys. Res. 105,11981–11990.

Petsch S. T. (1999) Comment on Carpenter and Lohmann(1999). Geochim Cosmochim Acta 63, 307–310.

Petsch S. T. (2000) A study on the weathering of organic matterin black shales and implications for the geochemical cyclesof carbon and oxygen. PhD Dissertation. Yale University.

Petsch S. T. and Berner R. A. (1998) Coupling the geochemicalcycles of C, P, Fe, and S: the effect on atmospheric O2 andthe isotopic records of carbon and sulfur. Am. J. Sci. 298,246–262.

Petsch S. T., Berner R. A., and Eglinton T. I. (2000) A fieldstudy of the chemical weathering of ancient sedimentaryorganic matter. Org. Geochem. 31, 475–487.

Petsch S. T., Smernik R. J., Eglinton T. I., and Oades J. M.(2001) A solid state 13C-NMR study of kerogen degradationduring black shale weathering. Geochim. Cosmochim. Acta65, 1867–1882.

Phoenix V. R., Konhauser K. O., Adams D. G., and Bottrell S. H.(2001) Role of biomineralization as an ultraviolet shield:implications for Archean life. Geology 29, 823–826.

Pollard D. and Kasting J. K. (2001) Coupled GCM-ice sheetsimulations of Sturtian (750–720 Ma) glaciation: when inthe snowball-earth cycle can tropical glaciation occur? EOS82, S8.

Porter S. M. and Knoll A. H. (2000) Testate amoebae in theNeoproterozoic Era: evidence from vase-shaped microfos-sils in the Chuar Group, Grand Canyon. Paleobiology 26,360–385.

Rasmussen B. and Buick R. (1999) Redox state of the Archeanatmosphere: evidence from detrital heavy minerals in ca.3250–2750 Ma sandstones from the Pilbara Craton, Aus-tralia. Geology 27, 115–118.

Rau G. H., Takahashi T., and Des Marais D. J. (1989) Latitudinalvariations in plankton d 13C: implications for CO2 andproductivity in past oceans. Nature 341, 516–518.

Rau G. H., Takahashi T., Des Marais D. J., Repeta D. J.,and Martin J. H. (1992) The relationship between d13Cof organic matter and [CO2(aq)] in ocean surface water:Data from a JGOFS site in the northeast AtlanticOcean and a model. Geochim. Cosmochim. Acta 56,1413–1419.

Robinson J. M. (1989) Phanerozoic O2 variation, fire, andterrestrial ecology. Palaeogeogr. Palaeoclimat. Palaeoecol.(Global Planet Change) 75, 223–240.


Robinson J. M. (1991) Phanerozoic atmospheric reconstruc-tions: a terrestrial perspective. Palaeogeogr. Palaeoclimat.Palaeoecol. 97, 51–62.

Rohl H.-J., Schmid-Rohl A., Oschmann W., Frimmel A., andSchwark L. (2001) The Posidonia Shale (Lower Toarcian) ofSW-Germany: an oxygen-depleted ecosystem controlled bysea level and paleoclimate. Palaeogeogr. Palaeoclimat.Palaeocol. 165, 27–52.

Ronov A. B. (1976) Global carbon geochemistry, volcanism,carbonate accumulation, and life. Geochem. Int. 13,172–195.

Rye R. and Holland H. D. (1998) Paleosols and the evolution ofatmospheric oxygen: a critical review. Am. J. Sci. 298,621–672.

Rye R. and Holland H. D. (2000) Life associated with a2.76 Ga ephemeral pond? Evidence from Mount Roe #2paleosol. Geology 28, 483–486.

Schidlowski M. (1988) A 3,800-million year isotopic record oflife from carbon in sedimentary rocks. Nature 333, 313–318.

Schidlowski M. (2001) Carbon isotopes as bio-geochemical recorders of life over 3.8 Ga of Earth history:evolution of a concept. Precamb. Res. 106, 117–134.

Schopf J. W. (1992) Paleobiology of the Archean. In TheProterozoic Biosphere (eds. J. W. Schopf, and C. Klein).Cambridge University Press, pp. 25–39.

Schopf J. W. (1993) Microfossils of the early Archean ApexChert: new evidence of the antiquity of life. Science 260,640–646.

Schopf J. W., Kudryavtvev A. B., Agresti D. G., Wdowiak T. J.,and Czaja A. D. (2002) Laser-Raman imagery of Earth’searliest fossils. Nature 416, 73–76.

Sinninghe-Damste J. S. and Koster J. (1998) A euxinic southernNorth Atlantic Ocean during the Cenomanian/Turonianoceanic anoxic event. Earth Planet. Sci. Lett. 158, 165–173.

Sleep N. H. and Zahnle K. (1998) Refugia from asteroidimpacts on early Mars and the early Earth. J. Geophys. Res.E—Planets. 103, 28529–28544.

Sleep N. H., Zahnle K. J., Kasting J. F., and Morowitz H. J.(1989) Annihilation of ecosystems by large asteroid impactson the early earth. Nature 342, 139–142.

Sleep N. H., Zahnle K. J., and Neuhoff P. S. (2001) Initiation ofclement surface conditions on the earliest Earth. Proc. Natl.Acad. Sci. 98, 3666–3672.

Sowers T., Bender M., and Raynaud D. (1989) Elemental andisotopic composition of occluded O2 and N2 in polar ice.J. Geophys. Res. 94, 5137–5150.

Stratford K., Williams R. G., and Myers P. G. (2000) Impactof the circulation on sapropel formation in theeastern Mediterranean. Global Biogeochem. Cycles 14,683–695.

Sumner D. Y. (1997) Carbonate precipitation and oxygenstratification in late Archean seawater as deduced from

facies and stratigraphy of the Gamohaan and FriscoFormations, Transvaal Supergroup, South Africa. Am. J.Sci. 297, 455–487.

Thomas A. L. R. (1997) The breath of life—did increasedoxygen levels trigger the Cambrian Explosion? Trends Ecol.Evol. 12, 44–45.

Tolbert N. E., Benker C., and Beck E. (1995) The oxygen andcarbon dioxide compensation points of C3 plants: possiblerole in regulating atmospheric oxygen. Proc. Natl. Acad. Sci.92, 11230–11233.

Valley J. W., Peck W. H., King E. M., and Wilde S. A. (2002)A cool early Earth. Geology 30, 351–354.

Van Cappellen P. and Ingall E. D. (1996) Redox stabilizationof the atmosphere and oceans by phosphorus-limited marineproductivity. Science 271, 493–496.

van Zuilen M., Lepland A., and Arrhenius G. (2002)Reassessing the evidence for the earliest traces of life.Nature 418, 627–630.

Vidal R. A., Bahr D., Baragiola R. A., and Peters M. (1997)Oxygen on Ganymede: laboratory studies. Science 276,1839–1842.

Walker J. C. G. (1979) The early history of oxygen and ozonein the atmosphere. Pure Appl. Geophys. 117, 498–512.

Walker J. C. G. (1986) Global geochemical cycles of carbon,sulfur, and oxygen. Mar. Geol. 70, 159–174.

Watson A. J. (1978) Consequences for the Biosphere ofGrassland and Forest Fires. PhD Dissertation. ReadingUniversity, UK.

Werne J. P., Hollander D. J., Behrens A., Schaeffer P., AlbrechtP., and Sinninghe-Damste J. S. (2000) Timing of earlydiagenetic sulfurization of organic matter: a precursor-product relationship in Holocene sediments of the anoxicCariaco Basin, Venezuela. Geochim. Cosmochim. Acta 64,1741–1751.

Werne J. P., Lyons T. W., Hollander D. J., Formolo M., andSinninghe-Damste J. S. (2003) Reduced sulfur in euxinicsediments of the Cariaco Basin: sulfur isotope constraints onorganic sulfur formation. Chem. Geol. 195, 159–179.

Wilde S. A., Valley J. W., Peck W. H., and Graham C. M.(2001) Evidence from detrital zircons for the existence ofcontinental crust and oceans on the Earth 4.4 Gyr ago.Nature 409, 175–178.

Wortmann U. G., Hesse R., and Zacher W. (1999) Major-element analysis of cyclic black shales: paleoceanographicimplications for the early Cretaceous deep western Tethys.Paleoceanography 14, 525–541.

Wray G. A., Levinton J. S., and Shapiro L. H. (1996) Molecularevidence for deep Precambrian divergences amongMetazoan phyla. Science 274, 568–573.

Zhang R., Follows M. J., Grotzinger J. P., and Marshall J.(2001) Could the late Permian deep ocean have been anoxic?Paleoceanography 16, 317–329.

q 2003, Elsevier Ltd. All rights reservedNo part of this publication may be reproduced, stored in a retrieval system ortransmitted in any form or by any means, electronic, mechanical, photocopying,recording or otherwise, without prior written permission of the Publisher.

Treatise on GeochemistryISBN (set): 0-08-043751-6

Volume 8; (ISBN: 0-08-044343-5); pp. 515–555

References 555

Date post:	19-Jul-2018
Category:	Documents
Upload:	lyngoc
View:	216 times
Download:	0 times

8.11 The Global Oxygen Cycle - University of California...

Documents