Julia G. Lazar, Arthur J. Gold, Kelly Addy, Paul M. Mayer, Kenneth J. Forshay, and Peter M. Groffman2
ABSTRACT: We examined the effect of instream large wood on denitrification capacity in two contrasting, lowerorder streams — one that drains an agricultural watershed with no riparian forest and minimal stores of in-stream large wood and another that drains a forested watershed with an extensive riparian forest and abundantinstream large wood. We incubated two types of wood substrates (fresh wood blocks and extant streambed wood)and an artificial stone substrate for nine weeks in each stream. After in situ incubation, we collected the sub-strates and their attached biofilms and established laboratory-based mesocosm assays with stream wateramended with 15N-labeled nitrate-N. Wood substrates at the forested site had significantly higher denitrificationthan wood substrates from the agricultural site and artificial stone substrates from either site. Nitrate-Nremoval rates were markedly higher on woody substrates compared to artificial stones at both sites. Nitrate-Nremoval rates were significantly correlated with biofilm biomass. Denitrification capacity accounted for only aportion of nitrate-N removal observed within the mesocosms in both the wood controls and instream substrates.N2 accounted for 99.7% of total denitrification. Restoration practices that generate large wood in streams shouldbe encouraged for N removal and do not appear to generate high risks of instream N2O generation.
Lazar, Julia G., Arthur J. Gold, Kelly Addy, Paul M. Mayer, Kenneth J. Forshay, and Peter M. Groffman, 2014.Instream Large Wood: Denitrification Hotspots with Low N2O Production. Journal of the American WaterResources Association (JAWRA) 50(3): 615-625. DOI: 10.1111/jawr.12202
INTRODUCTION
Restoring riparian forests to reduce waterbornenitrogen (N) pollution has been an objective of manywatershed management efforts (Schultz et al., 2004;Hassett et al., 2005; Mitsch, 2007). Riparian forestscan reduce groundwater N loading to streams
through denitrification, plant uptake, and microbialimmobilization (Gold et al., 2001; Mayer et al., 2007;Vidon et al., 2010). Riparian forests also supportfunctions that enhance the ecosystems of lower orderstreams by modulating stream temperature throughshading, increasing stream width and habitat com-plexity through geomorphic effects on streambanks(Sweeney et al., 2004), increasing species richness,
2Doctoral Researcher (Lazar), Professor (Gold), and Research Associate (Addy), Department of Natural Resources Science, University ofRhode Island, Coastal Institute, 1 Greenhouse Road, Kingston, Rhode Island 02881; Ecological Effects Branch Chief (Mayer), NationalHealth and Environmental Effects Research Laboratory, U.S. Environmental Protection Agency, Corvallis, Oregon 97330; Research Ecologist(Forshay), National Risk Management Research Lab, U.S. Environmental Protection Agency, Ada, Oklahoma 74820; and Senior Scientist(Groffman), Cary Institute of Ecosystem Studies, Millbrook, New York 12545 (E-Mail/Gold: [email protected]).
JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION JAWRA615
JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION
Vol. 50, No. 3 AMERICAN WATER RESOURCES ASSOCIATION June 2014
and affecting biogeochemical functions through addi-tions of large wood and organic carbon (C) (Welsh,1991; Naiman and Decamps, 1997; Bilby, 2003).
Increases in anthropogenic N inputs have led toincreased N in riverine systems (Howarth et al.,1996; Galloway et al., 2004), accelerating rates ofeutrophication in coastal areas (Turner and Rabalais,1994). Much effort has been made to understand andmanage N loads within aquatic systems to improvewater quality and other ecosystem services (Gallowayet al., 2003). Evidence has pointed to relationshipsbetween riparian forests and increased soil denitrifi-cation, an anaerobic microbial process that perma-nently removes nitrate from fluvial systems byreturning N to the atmosphere (Alexander et al.,2000; Gold et al., 2001; Mulholland et al., 2008).
Higher fluvial denitrification rates have been foundto be associated with increases in the organic contentof benthic sediments, respiration rates, and opportu-nities for contact with the stream bed (e.g., shallow,wide streams, and streams with extensive hyporheicflow) (Hinkle et al., 2001; Mulholland et al., 2008,2009; Hall et al., 2009). These conditions can be fos-tered by large wood (e.g., sticks, branches or treetrunks) and debris dams in streams that are derivedfrom a riparian forest. Large wood in streams from ariparian forest can be a direct source of labile C tostreams to fuel microbial processes. Large wood alsoadds structure to the stream channel and createsobstacles that slow the flow of water and extend theresidence time of surface water in the stream andfacilitate accumulation of finer organic sedimentsthat support biofilms. Instream large wood has thepotential to function as microsites or “hotspots” of ele-vated biogeochemical cycling including denitrification(McClain et al., 2003; Groffman et al., 2005, 2009).Downstream declines in nutrient concentrations havealso been attributed to biofilms (Sabater et al., 1991;Ryhiner et al., 1994; Mulholland et al., 1995). Biofilmstructure, composition, and capacity for biogeochemi-cal cycling is influenced by substrate composition,light penetration, nutrient concentration, flow rates,seasonality, sediment composition, and the commu-nity of grazers in the vicinity (Sabater et al., 1988,2002; Rott et al., 1998). Biofilms formed on wood sub-strates have been found to have higher respirationrates and greater N demand than biofilms developedon rock substrates (Sabater et al., 1998).
Here we examine the effect of instream large woodon denitrification capacity in two contrasting, lowerorder streams — one that drains an agriculturalwatershed with no riparian forest and minimal storesof instream large wood and another that drains a for-ested watershed with an extensive riparian forestand abundant instream large wood. The agriculturalstream was scheduled for extensive riparian restora-
tion that is expected to increase the extent of largewood in the channel. We incubated two types of woodsubstrates (fresh wood blocks and extant streambedwood) and an artificial stone substrate for nine weeksin each stream. After in situ incubation, we collectedthe substrates and their attached biofilms andestablished laboratory-based mesocosm assays withstream water amended with 15N-labeled nitrate-N.We hypothesized that mesocosms containing woodsubstrates would have higher denitrification capacityrates than other mesocosms as we expected the labileC to promote conditions that would enhance denitrifi-cation. While the agricultural stream had less in-stream large wood, we hypothesized that the lack ofshade and elevated nutrients associated with theagricultural stream would yield higher rates ofnitrate removal on substrates in response toincreased autotrophic communities that can formunder those conditions. This research aims to furtherour understanding of the effects of riparian forests onfluvial denitrification.
METHODS
We used a mesocosm approach to examine denitri-fication and nitrate-N removal rates of substratesand associated biofilm that were placed within a spe-cific reach of each stream (i.e., the study sites) fornine weeks as well as from bare substrates withoutbiofilm development which were never subjected tofield conditions as controls.
Study Sites
During the summer of 2009, we used two streamsthat differed markedly in nitrate concentration,watershed land use, and riparian cover: Big SpringRun, located in Lancaster County, Pennsylvania, andMawney Brook, located in Kent County, Rhode Island(Table 1). The Big Spring Run and Mawney Brookwatersheds that drain to the study sites are 4.3 and4.8 km2, respectively. On the basis of NLCD geospa-tial data (Fry et al., 2011), land use in the Big SpringRun watershed is 41% agricultural, 4% forest, and55% developed (Table 1). Agricultural land cover bor-ders the Big Spring Run riparian zone adjacent tothe study reach. Land use in the Mawney Brookwatershed is 62% forested, 11% wetlands, and 27%developed (Table 1). A mature riparian forest bordersthe entire length of Mawney Brook along the studyreach and upstream from the study site. Hereafter,Big Spring Run is referred to as the “agricultural”
JAWRA JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION616
site and Mawney Brook is referred to as the“forested” site. U.S. Geological Survey (USGS) gagestations provided all flow data: gage 015765195 and01116905 for the agricultural and forested sites,respectively.
We computed sinuosity of the 300 m reachupstream of the study sites from 1:24,000 USGStopography maps (Cushing and Allan, 2001; USGS,2011) as 1.39 and 1.27 for the agricultural and for-ested sites, respectively (Table 1). Acidic stratifieddrift deposits and limestone bedrock dominated soilsat the forested site and the agricultural site, respec-tively; these surficial geology differences are reflectedin the hardness levels of the two streams (Table 1).Ambient nitrate concentrations in the agriculturalstream were more than 10-fold higher than in theforested stream (Table 1).
In Situ Incubation and Harvesting of Substrates
Treatment substrates of similar size and shape con-sisted of (1) wood blocks (26 cm 9 4.5 cm 9 2.2 cm)made from red maple (Acer rubrum); (2) artificialstone made from unglazed clay-fired blocks (25 cm 9
5 cm 9 1.25 cm); and (3) bundles of sticks (~5 cm 9
25 cm bundle composed of ~1.5 cm diameter sticks)collected within 25-75 m of each stream site. Sinceextant wood at the forested site was widely available,we bundled sticks from within 25 m of the sitewhereas at the agricultural site, we compiled sticksfrom approximately 75 m from the site due to lesswood being available. In both streams, we placed sub-strates within a 25 m reach of the stream in earlysummer. We monitored nitrate concentrations for ayear at the forested stream and used flow and nitrate
data collected by the U.S. Environmental ProtectionAgency for the agricultural stream. At each site, weanchored 16 wood blocks, 10 artificial stone substrates,and 10 extant wood bundles to individual bricks viaplastic zip ties. The anchors also kept the substratessubmerged in water. After nine weeks, we collectedthe substrates, their associated biofilms and bricks in35 cm 9 25 cm 9 12.8 cm clear plastic bins underwa-ter to minimize exposure to air, taking care to avoiddisturbance of biofilms associated with the substratestructure. Nevertheless, slight turbidity in both thestream reach and mesocosm bin was inevitable. Theseplastic bins containing the extracted substrates arehereafter referred to as “mesocosms.” We added ambi-ent stream water to the mesocosms until they were fulland sealed them with dark lids to limit photosynthesisand to minimize exposure of the blocks to air duringthe 30 min transportation to the laboratory. We sam-pled the ambient stream water, transported sampleson ice and stored samples at 4°C until analysis.
Upon arrival at the laboratory, we removed lids,and sampled a 9 cm2 area (<3.0% of the total surfacearea of the artificial substrates) of one corner of eachwood block and artificial stone. We placed the har-vested biofilm into 450 ml of deionized water andstored it at 4°C for further analyses. We did notcollect biofilm from the extant wood bundles due todifficulties in quickly establishing a firm estimate ofsubstrate area before beginning the sealed mesocosmexperiment.
Mesocosm Experiments
Each mesocosm contained one substrate, attachedbiofilm, brick, and lid fitted with a #37 Suba-SealTM
TABLE 1. Land Use, Sinuosity, Soil Parent Material, and Ambient Stream Characteristics for the Study Sites.
Agricultural Forested
Stream name Big Spring Run Mawney BrookLocation Lancaster, Pennsylvania East Greenwich, Rhode IslandLatitude, longitude 39°59035.75″N, 76°15041.73″W 41°38037.93″N, 71°31016.73″W% Wetland 0 10.6% Agriculture 41 0.2% Forest 4.1 62.0% Developed 54.7 27.1Sinuosity 1.39 1.27Dominant soil parent material Carbonate limestone Acidic stratified driftAverage summer NO3
1Single data point.2Flow rates of Mawney Brook were estimated from the USGS gage (01116905) located at Fry Brook that was down gradient of the study site.Flow rates were adjusted based on the ratio of the watershed area of the study reach to the watershed area of USGS gage.
JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION JAWRA617
INSTREAM LARGE WOOD: DENITRIFICATION HOTSPOTS WITH LOW N2O PRODUCTION
rubber septa (Sigma Aldrich, Saint Louis, Missouri)placed in a drilled sampling port in the center of thelid. We used a 15N tracer technique to estimate deni-trification capacity (Nishio et al., 1983; Jenkins andKemp, 1984). We amended the forested mesocosmswith KNO3
� and 15N-KNO3� to 20 atom% for a final
concentration of 8 mg N/l of NO3�. Average summer
nitrate concentrations at the agricultural site(Table 1) were much higher, 9.69 mg/l, and therefore,only 99% 15N-KNO3
� was added, for a final concen-tration of 11 mg N/l of NO3
� at 20 atom%. These con-centrations ensured NO3
� was available in excessthroughout the incubation. The agricultural site hadhigher final N concentration than the forested sitedue to elevated background concentrations in ambi-ent stream water. The computed rates represent deni-trification capacity where nitrate is abundant andother factors, such as electron donors or redox condi-tions control the observed rates (Addy et al., 2005).
Prior to sealing, we collected well-mixed watersamples for analysis of initial conditions and storedthem at 4°C until analysis. A headspace of 1.5 cmremained between the top of the water and the lid(total headspace volume per mesocosm ~1,800 ml) tofacilitate gas sampling. We steadily bubbled helium(He) gas into half of the wood block mesocosms withsparge stones until dissolved oxygen (DO) concentra-tions were below 2 mg/l to optimize conditions fordenitrification. DO levels in mesocosms without Headded averaged 8.5 mg/l. We recorded DO and tem-perature (°C) before securing the dark mesocosm lids.We secured lids onto the bins and sealed with siliconeto prevent air from leaving or entering the meso-cosms. In mesocosms where He was bubbled into thewater before sealing, we added additional He via nee-dle through the sampling port for five minutes toreplace headspace gases with He. During this addi-tion, a second needle placed in sampling port ventedexcess gas.
After 1 min of shaking to equilibrate headspace,we extracted 20 ml of initial headspace samples viasyringe and placed these samples into 12 ml pre-evacuated ExetainterTM vials (Labco 839W; Labco,Lampeter, United Kingdom). We repeated 20 mlheadspace samples at 1.5, 3, and 18 h. To preventnegative pressure, we added 20 ml of He back intoeach mesocosm via the septa after samples weretaken. After the last gas sample was taken, weremoved lids, measured DO, and collected a finalwater sample which we stored at 4°C until analysis.
“Blank” mesocosms consisted of ambient streamwater from each site, with 20 atom% 15NO3
� asKNO3
� added to reach a desired concentration of8 mg and 11 mg N/l of NO3
�, for the forested siteand the agricultural site, respectively. Control meso-cosms consisted of wood blocks or artificial stone sub-
strates, attached to individual bricks, which werekept dry in the laboratory while the other blockswere submerged in the stream. Biofilms did notdevelop on these controls. Table 2 provides definitionsof each of the mesocosm incubation types for easy ref-erence. On the day when the mesocosms were estab-lished, we placed control blocks and attached bricksin mesocosms with fresh stream water. We treatedand sampled blanks and controls as described abovefor the instream substrates.
Analyses
We filtered the biofilm samples removed from eachwood and artificial stone block onto Whatman 42 ash-less 90 mm pre-weighed filters, dried the filters andthen reweighed the filters to quantify biomass. TheUniversity of California Davis Stable Isotope Facilityanalyzed these filters for natural abundance of d13Cand d15N isotopes and bulk C and N compositionusing a PDZ Europa ANCA-GSL elemental analyzerinterfaced to a PDZ Europa 20-20 isotope ratio massspectrometer (Sercon Ltd., Cheshire, United King-dom).
The University of California Davis Stable IsotopeFacility analyzed the mesocosm headspace samplesfor concentrations and isotope ratios of N2 and N2Ousing a ThermoFinnigan GasBench + PreCon tracegas concentration system interfaced to a Thermo-Scientific Delta V Plus isotope-ratio mass spectrome-ter (Bremen, Germany).
We analyzed water samples from the beginningand end of mesocosm incubations using the opentubular cadmium reduction method (American PublicHealth Association et al., 1995) on an Astoria PacificModel 303A Segmented Continuous Flow Autoanalyz-er (Astoria-Pacific Inc., Clackamas, Oregon). Samplesfrom Mawney Brook were analyzed for alkalinityusing a Hanna Instruments 902 Color Automatic
TABLE 2. Mesocosm Terminology Defined.
MesocosmTerminology Definition
Blank Streamwater only, no substratesControl wood Streamwater + wood block that has not been
incubated in the streamControl stone Streamwater + artificial stone that has not been
incubated in the streamExtant wood Streamwater + bundle of sticks found in the
stream site attached to bricks and incubated inthe stream
Wood blocks Streamwater + fresh red maple wood blocksattached to bricks and incubated in the stream
Artificialstones
Streamwater + clay-fired blocks attached to bricksand incubated in the stream
JAWRA JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION618
Potentiometric Titrator (Woonsocket, Rhode Island).Samples from Big Spring Run were analyzed for alka-linity by manual titration (American Public HealthAssociation et al., 1999). A LaMotte Total Calcium &Magnesium Hardness test kit (Code 4824 DR-LT)(LaMotte, Chestertown, Maryland) was used to deter-mine hardness at Mawney Brook, while hardnesssamples from Big Spring Run were analyzed usingthe hardness by calculation method after mineralanalysis was performed (American Public HealthAssociation et al., 1999). DO and temperature weremeasured using a YSI DO-temperature meter, model55 (YSI, Yellow Springs, Ohio). We measured thelength and width of extant large wood after the meso-cosm experiment was completed to calculate surfacearea.
Data Analyses
We used a 15N tracer technique to estimate denitrifi-cation capacity (Nishio et al., 1983; Jenkins and Kemp,1984). Denitrification masses of N2O and N2 gases(lmol) in headspace samples were extrapolated to thewhole mesocosm scale using Bunsen coefficients fromTiedje (1982) and equation constants from Mosier andKlemedtsson (1994) following the formulas used inKellogg et al. (2005). The total masses of N2O-N andN2 generated during the incubation period were calcu-lated by dividing the masses of 15N2O-N and 15N2 bythe dosed NO3
�-N atom%. The mass of 15N2O-N and15N2 generated was divided by the number of hoursthat have passed since the last sample time. Sampleswere taken at time 0, 1.5, 3, and 18. The average deni-trification rate from those three time periods isrecorded. Gas production rates (N2O-N and N2) wereexpressed as lg N/m2 of substrate per hour.
We use the term nitrate-N removal to reflectreduction in total nitrate per unit time within eachmesocosm, calculated by subtracting the post-incuba-tion nitrate concentration from the pre-incubationnitrate concentration. Because the mesocosms weresealed throughout the incubation period, we did notobtain estimates of uptake kinetics. The data from“blank” mesocosms estimated denitrification andnitrate-N removal in the stream water itself. We sub-tracted the rates of the blanks in all substrate ratesgiven in the results to highlight the rates associatedwith the addition of substrates.
Statistical Analyses
Unless otherwise noted, instream mesocosmresults are based on the following n values for theforested site: thirteen wood blocks, eight extant wood
bundles, and five artificial stone blocks. For the agri-cultural site there were nine wood blocks, four extantwood bundles, and five artificial stone blocks. Thenumbers differ by site because some substrates werelost, presumably during high flows. We also employedthe same mesocosm setup to evaluate denitrificationcapacity and nitrate-N removal on five forested con-trol wood blocks, four agricultural control woodblocks, four control artificial stone blocks, and fiveblanks (mesocosms without substrates).
We tested for differences in biofilm dry mass, deni-trification, and nitrate-N removal rates, and biomassN and C between substrates and sites. Aside frombiomass C, data were normally distributed. We pooleddata within a site if they were not significantly differ-ent. For biofilm dry mass, denitrification rates, andnitrate-N removal rates from the agricultural site, weused Student’s t-tests to test for differences betweensubstrates within each site. Nitrate-N removal ratesfrom the two woody substrates at the forested sitecould not be pooled and analysis of variance(ANOVA) was used to test for differences between thethree substrates. For pairwise comparisons of denitri-fication and nitrate-N removal rates between sub-strates and across sites we used ANOVA with aTukey’s post hoc test. Biomass C data for the forestedsite and all biomass d13C and d15N data were not nor-mal so we used the Kruskal-Wallis test to determinesignificant differences between sites and substrates.
To test for differences between substrates and sitesof the percent of nitrate-N removal that can beascribed to denitrification we used ANOVA with aTukey’s post hoc test. We evaluated correlationbetween denitrification rates, nitrate-N removalrates, and biofilm mass using Pearson product-moment correlation coefficients, after log transform-ing the data. Stream sites were treated separatelyand only pooled for correlation statistics. Statisticalsignificance was set at a < 0.05 for all analyses. Allstatistical analyses were performed with Analyse-itversion 2.26 (Leeds, United Kingdom).
RESULTS
Ambient Water Quality
The agricultural site had an average summernitrate concentration of 9.69 mg/l, several orders ofmagnitude greater than concentrations in the for-ested site, with mean summer nitrate of 0.05 mg/l(Table 1). Average stream DO concentration duringthat same time period was 9.0 and 8.0 mg/l for agri-cultural and forested sites, respectively. Median flow
JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION JAWRA619
INSTREAM LARGE WOOD: DENITRIFICATION HOTSPOTS WITH LOW N2O PRODUCTION
from June to October was 0.013 (n = 4) and 0.015(n = 131) m3/s/km2 for the agricultural and the for-ested sites, respectively. The agricultural site had anaverage summer temperature of 16.7°C. Stream tem-perature data were not obtained at the forested site,but summer stream temperatures at the gaging sta-tion at the Beaver River, a neighboring (within10 km) forested watershed with similar physiography,had an average summer (June-August) temperatureof 17.5°C.
Quality and Quantity of Biofilm
Artificial stone substrates at the forested site hadsignificantly less (p ≤ 0.05) biofilm mass than woodblocks (Table 3a), while the biofilm masses were notsignificantly different between the artificial stone andwood block substrates at the agricultural site(Table 3b). Biomass was not measured on extantwood at either site.
Within each site, the appearance of the biofilmswas similar across the artificial stone and wood blocksubstrates; however, between sites, the biofilms werevisibly different. Green biofilms developed on sub-strates at the agricultural site, which was withoutshade, while the substrates at the forested site devel-oped darker biofilms (Table 3). Biofilm C and N didnot differ between wood and artificial stone sub-strates within each site, so the data were pooled forstatistical comparison. The mean C:N ratio of thebiofilms at the forested site was 16.2 (SD: 5.9) andsignificantly greater (p ≤ 0.05) than at the agricul-tural site (mean: 8.0; SD: 1.6). The mean biomass N/cm2 was not significantly different between the twosites, but the mean biomass C/cm2 at the forested sitewas significantly higher (p ≤ 0.05) than the agricul-tural site, 531 lg C/cm2 (SD: 410) and 213 lg C/cm2
(SD: 84), respectively.Biofilms at the agricultural site had significantly
more enriched d13C values (p ≤ 0.05) than at the for-ested site. There were no significant differences inbiofilm d15N between sites or substrates.
Mesocosm Denitrification Capacity
Control wood block substrates generated signifi-cantly (p ≤ 0.05) higher denitrification rates than thecontrol artificial stone substrates (Figure 1). No sig-nificant differences in denitrification rates were foundbetween sites for each type of control and blanks, andresults were pooled for statistical tests.
Instream artificial stones at the agricultural sitehad significantly higher denitrification rates than thecontrol artificial stones that were not incubated inthe stream. In contrast, at the forested site, the deni-trification capacity rates of instream artificial stoneswere much lower and did not significantly differ fromthe control artificial stones.
Within each site, denitrification rates from in-stream extant wood bundles were not significantlydifferent from instream wood block substrates. Thedenitrification rates from both wood sources weretherefore combined within each site for further statis-tical comparisons; hereafter referred to as “instreamwood substrates.” The instream wood substrates atthe forested site had significantly higher denitrifica-tion rates (p ≤ 0.05) than those at the agriculturalsite (Figure 2). At the forested site, mesocosms of in-stream wood substrates had significantly higher deni-trification rates than instream artificial stones(Figure 2). However, at the agricultural site, denitri-fication rates from instream wood and artificialstones were not significantly different (p > 0.05) (Fig-ure 2).
Wood blocks subject to hypoxic and oxic mesocosmswere not significantly different. Hypoxic mesocosmsgenerally remained below 2.2 mg/l of DO for the mes-ocosm assays. Oxic mesocosms which started withDO over 7.0 mg/l, ended below 3 mg/l. Although it isexpected that the oxygen levels decreased over time,the N2 and N2O production rates did not significantlydiffer between sampling times.
Low levels of nitrous oxide were generated throughdenitrification. Rates of N2O-N were consistently<0.02 lg N/m2/h. N2:N2O ratios were >99.7 in allmeasurements.
TABLE 3. Biofilm Biomass and Characteristics on Substrates at the Forested (a) and Agricultural (b) Study Reaches.At the forested site wood blocks had significantly higher biofilm masses than artificial stones blocks, whereas the agriculturalsite had similar biofilm masses on both substrates. Biofilm was only measured on wood blocks and artificial stones, not on
extant large wood. Significant differences within a site are noted by superscripts, p ≤ 0.05 using a Student’s t-test.
Site Substrate Mean Biofilm Mass (g) Standard Deviation n Value Color Description
(a) Forested Wood block 0.530a 0.32 8 Dark brown MattedArtificial stone 0.068b 0.05 5 Dark brown Matted
(b) Agricultural Wood block 0.304 0.20 8 Bright green FilamentousArtificial stone 0.132 0.17 5 Bright green Filamentous
Note: The n-values for the wood block biofilm mass do not match the n-values for denitrification rates. This is because six blocks from thesite were inadvertently not sampled for biofilm mass.
JAWRA JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION620
LAZAR, GOLD, ADDY, MAYER, FORSHAY, AND GROFFMAN
Nitrate-N Removal
Blanks displayed no evidence of nitrate-N removal(limit of detection on instrument is 0.02 mg N/l).Nitrate-N removal trends at the agricultural site fol-lowed the denitrification results; extant wood andwood blocks had similar rates and were significantlyhigher (p ≤ 0.05) in nitrate-N removal than instreamartificial stones (Figure 3). Ending nitrate-N concen-tration for the wood blocks and extant wood at theagricultural site was 10.80 mg/l, a reduction of1.40 mg/l. At the forested site, wood blocks had sig-nificantly higher (p ≤ 0.05) nitrate-N removal than
instream extant wood. The wood blocks at the for-ested site had an average ending nitrate concentra-tion of 8.28 mg/l, a reduction of 1.75 mg/l. Combinedextant wood and wood blocks at the agricultural sitehad significantly (p ≤ 0.05) higher nitrate-N removalrates than extant wood from the forested site. Woodblocks from the forested site had significantly(p ≤ 0.05) higher nitrate-N removal than instreamartificial stones (p ≤ 0.05), which had no nitrate-Nremoval. Denitrification rates of wood blocks with bio-film were significantly correlated with nitrate-Nremoval rates (r = 0.57, p ≤ 0.01). Nitrate-N removalrates were also significantly correlated with biofilmmass (r = 0.69, p ≤ 0.01); however, no significant cor-relation was found between biofilm mass and denitri-fication rates.
DISCUSSION
Wood substrates were found to promote denitrifica-tion and nitrate removal in starkly contrasting siteswith different levels of riparian forest cover, ambientnutrient enrichment, alkalinity, and hardness (Fig-ure 2). This evidence follows other studies, whichhave shown that organic substrates such as riparianforests, organic debris dams, and C bioreactors can behotspots of denitrification (Groffman et al., 2005; Hallet al., 2009; Schipper et al., 2010; Reisinger et al.,2013). Most of the mesocoms with woody substratesdisplayed high N transformation rates. This studysupports the importance of instream large wood forpromoting conditions that stimulate N cycling withinstreams.
The significantly higher denitrification generatedby the control wood block mesocosms (not subjectedto instream incubation) compared to the artificialstone substrate controls was expected as wood sub-strates have been found to generate labile C, promotedenitrification, and are used in denitrifying C biore-actors — where a C substrate is added to the flowpath of nitrate enriched water to stimulate denitrifi-cation in groundwater and agricultural runoff (Bern-hardt and Likens, 2002; Robertson, 2010; Schipperet al., 2010).
Instream wood blocks and extant large wood sub-strates generated comparable denitrification at bothsites implying that the wood blocks created for thismesocosm experiment are comparable to the woodthat is already found at these two stream sites. Thesignificantly higher denitrification rates of the in-stream wood substrates than the wood block controls,suggest the importance of biofilm development for in-stream cycling of N. Although no significant correla-
a
b
FIGURE 1. Denitrification Capacity of Wood Block and ArtificialStone Substrates without Biofilms (controls) Pooled across Sites.
Different letters above bars indicate significant differences,p ≤ 0.05 using a Student’s t-test.
a
bb
b
FIGURE 2. Denitrification Rates of Instream Wood and ArtificialStone Substrates in the Two Study Sites. Treatments with
different letters above bars are significantly different at p ≤ 0.05using an ANOVA, Tukey’s post hoc comparison test.
JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION JAWRA621
INSTREAM LARGE WOOD: DENITRIFICATION HOTSPOTS WITH LOW N2O PRODUCTION
tion was found between biofilm mass and denitrifica-tion rates of instream substrates, there was a signifi-cant correlation between biomass and nitrate-Nremoval capacity. The lowest biomass was found onthe forest artificial stone, which corresponded withlower nitrate-N removal rates. The forest wood blockshad significantly higher biofilm mass than the forestartificial stones, corresponding with the highestnitrate-N removal rates. In contrast, the biofilmmasses at the agricultural site were quite similarbetween wood and stones, potentially obscuring sub-strate differences in denitrification rates. The biofilmat the agricultural site without a riparian forestreceived more sunlight and may thus have been moreproductive and had higher N turnover rates due tointense sunlight and nutrient availability.
Although we did not identify the type and extentof algal vs. bacterial biomass, we note that theappearance and color of the biofilms contrasted sharplybetween sites and that the composition of biofilmshas been found to alter N cycling (Romani and Sabat-er, 2000). Measures of d13C and C:N ratios also showthat the biofilm compositions differed by site. Pub-lished C:N ratios for epilithon match the C:N ratiosfound at the forested site and published ratios forfilamentous green algae coincide with biofilm resultsat the agricultural site (Kemp and Dodds, 2002).
The high denitrification capacity of the forestedwood blocks compared to the agricultural wood blocksis noteworthy given the high nitrate concentrations inthe agricultural stream. Peterson et al. (2011) com-pared biofilm growth in two streams that differed innitrate concentrations by an order of magnitude andsuggest that in un-enriched nitrate conditions algaeinfluence the denitrifying community due to theirdependence on dissolved organics, while in enrichedconditions this relationship is disconnected. A clear
separation between the two biofilm communities wasnoted, and the low nitrate stream had increased spe-cies diversity, which they suggest leads to increaseddenitrification rates (Peterson et al., 2011). Similar tothis study, Peterson et al. (2011) found no differencein biofilm mass between the enriched and un-enrichedbiofilm communities. The biofilm at the forested sitemay have had a more robust denitrifying communityleading to higher denitrification rates. Another possi-bility is that oxygen generated by photosynthesizingalgae at the agricultural site could create conditionsthat limited the extent of denitrifiers in the biofilm.
The high nitrate levels in the agricultural streamare reflected in higher biofilm N content compared tothe forested biofilm. The low levels of biomass C inthe agricultural biofilm may be due to macroinverte-brate grazing (Hillebrand and Kahlert, 2001) or toenhanced rates of microbial degradation. In contrastto our results, Romani et al. (2004) found that C:Nmolar ratios of biofilms at enriched and nonenrichedstream sites were not different. Biofilm d13C valuesseen at both study sites fall in the normal range forC3 plants, which ranges from �32 to �22 (Rounickand Winterbourn, 1986). In sites with greater periph-yton productivity and less canopy cover d13C tends tobe enriched relative to those with more canopy (Ishik-awa et al., 2012). Our agricultural site is exposed tomore sunlight and likely supports greater algalstanding stock than the shaded forest site and theenriched biofilm d13C observed in this study. Thisagrees with a phenomenon observed in Canada whereperiphyton grown in high light conditions had moreenriched d13C values than in low light (MacLeod andBarton, 1998) and in New Zealand where algae inunshaded pasture streams (especially filamentousgreen algae), were more enriched than algae (dia-toms) in shaded forest streams (Hicks, 1997).
0
500
1000
1500
2000
2500
forestwoodblock
agricwoodblock
forestextantwood
agricextantwood
controlwoodblock
foreststone
agricstone
controlstone
ug N
m-2
hr-1
Denitrification
Other Nitrate-NRemovalProcesses
FIGURE 3. Net Rate of Nitrate-N Removal (represented by the value equivalent to the total heightof each vertical bar) and the Denitrification Rate for Each Mesocosm Type.
JAWRA JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION622
LAZAR, GOLD, ADDY, MAYER, FORSHAY, AND GROFFMAN
Although significantly correlated, denitrificationcapacity accounted for only a portion of nitrate-Nremoval observed within the mesocosms in both thewood controls and instream substrates. Assimilation,both autotrophic and heterotrophic, generallyaccounts for a higher proportion of N removal thandenitrification (Peterson et al., 2001; Mulhollandet al., 2008).
The oxic and hypoxic mesocosms did not have sig-nificantly different denitrification rates. However,both oxic and hypoxic mesocosms were hypoxic at theend of the 18 h incubation. Mesocosms were coveredwith dark lids in an effort to limit photosynthesis,thereby decreasing oxygen production. One drawbackof this mesocosm technique is that by creating a darkenvironment we may have increased net respirationrates, decreasing oxygen, and increasing denitrifica-tion rates. Microbial respiration was likely responsi-ble for decreasing O2 concentrations.
N2 gas accounted for 99.7% of the total denitrifica-tion indicating complete denitrification. Therefore,wood in these two stream ecosystems are not a sub-stantial source for N2O generation. Similarly, in alarge study comparing denitrification rates of 49streams across varying landscapes, median N2 pro-duction rate was 99.4% of the sum of N2 and N2O(Mulholland et al., 2009), and Mosier et al. (1998)suggested comparable results.
Controlling N loads from watersheds is a hugeproblem that will likely require multiple activities,including management of both sources and sinks(Kellogg et al., 2010; Groffman et al., 2011). Plant-ing woody species in riparian buffers next to agri-cultural lands can be an important component of Nmanagement. Riparian forests have been shown toincrease hydrological connectivity, increasing deni-trification in groundwater before it enters thestream (Gold et al., 2001). This study furtheremphasizes the value of restoring mature riparianforests for N management since wood substrates,regardless of the extent of biofilm development, tendto generate higher denitrification than stone sub-strates.
ACKNOWLEDGMENTS
We thank Bob Walter and Dorothy Merritts, Franklin and Mar-shall College for assistance and support; URI Coastal Fellows MatLautenberger, Lauren Creamer, Aimee Welch, and Molly Welsh fortheir help and support; and anonymous reviewers for insightfulcomments on drafts of this manuscript. This project was supportedby grants from USDA-NRCS, RI Agricultural Experiment Station(contribution no. 5400), and NSF EPSCoR Grant No. 0554548. Theresearch has not been subjected to U.S. Environmental ProtectionAgency review and, therefore, does not necessarily reflect the viewsof any of the funding agencies, and no official endorsement shouldbe inferred.
LITERATURE CITED
Addy, K., A. Gold, B. Nowicki, J. McKenna, M. Stolt, and P. Groff-man, 2005. Denitrification Capacity in a Subterranean Estuarybelow a Rhode Island (USA) Fringing Salt Marsh. Estuaries28:896-908.
Alexander, R.B., R.A. Smith, and G.E. Schwarz, 2000. Effect ofStream Channel Size on the Delivery of Nitrogen to the Gulf ofMexico. Nature 403:758-761.
American Public Health Association (APHA), American WaterWorks Association (AWWA), and Water Environment Federa-tion (WEF), 1995. Standard Methods for the Examination ofWater and Wastewater (Nineteenth Edition). Method refer-enced: Automated Cadmium Reduction (4500-NO3–F), Auto-mated Ascorbic Acid Reduction Method (4500-P-F)875532356.Washington, D.C. ISBN: 978-0875532233.
American Public Health Association (APHA), American WaterWorks Association (AWWA), and Water Environment Federa-tion (WEF), 1999. Standard Methods for Examination of Waterand Wastewater (SMEWW) (Twentieth Edition). Section 2320B,Alkalinity Titrimetric Method. Washington, D.C. ISBN-13: 978-0875532356.
Bernhardt, E.S. and G.E. Likens, 2002. Dissolved Organic CarbonEnrichment Alters Nitrogen Dynamics in a Forest Stream. Ecol-ogy 83:1689-1700.
Bilby, R.E., 2003. Decomposition and Nutrient Dynamics of Woodin Streams and Rivers. American Fisheries Society Symposium37:135-147.
Cushing, C.E. and J.D. Allan, 2001. Streams: Their Ecology andLife. Academic Press, San Diego, California, 11 pp.
Fry, J., G. Xian, S. Jin, J. Dewitz, C. Homer, L. Yang, C. Barnes,N. Herold, and J. Wickham, 2011. Completion of the 2006National Land Cover Database for the Conterminous UnitedStates. PE&RS 77(9):858-864.
Galloway, J.N., J.D. Aber, J.W. Erisman, S.P. Seitzinger, R.W. Ho-warth, E.B. Cowling, and B.J. Cosby, 2003. The NitrogenCascade. BioScience 53:341-356.
Gold, A.J., P.M. Groffman, K. Addy, D.Q. Kellogg, M. Stolt, andA. Rosenblatt, 2001. Landscape Attributes as Controls onGroundwater Nitrate Removal Capacity of Riparian Zones.Journal of the American Water Resources Association 37:1457-1464.
Groffman, P.M., K. Butterbach-Bahl, R.W. Fulweiler, A.J. Gold,J.L. Morse, E.K. Stander, C.L. Tague, C. Tonitto, and P. Vidon,2009. Challenges to Incorporating Spatially and TemporallyExplicit Phenomena (hotspots and hot moments) in Denitrifica-tion Models. Biogeochemistry 93:49-77.
Groffman, P.M., A.M. Dorsey, and P.M. Mayer, 2005. Nitrogen Pro-cessing within Geomorphic Features in Urban Streams. Journalof North American Benthological Society 24:613-625.
Groffman, P.M., A.J. Gold, L. Duriancik, and R.R. Lowrance, 2011.Connecting the Dots” to “Threading the Needle”: The ChallengesAhead in Managing Agricultural Landscapes for EnvironmentalQuality. In: Managing Agricultural Landscapes to Achieve MoreEffective Conservation, P. Nowak and M. Schnepf (Editors). Soiland Water Conservation Society, Ankeny, Iowa, pp. 1-12.
JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION JAWRA623
INSTREAM LARGE WOOD: DENITRIFICATION HOTSPOTS WITH LOW N2O PRODUCTION
A.J. Burgin, M.J. Bernot, J.J. Beaulieu, and C.P. Arango,2009. Nitrate Removal in Stream Ecosystems Measured by15N Addition Experiments: Total Uptake. Limnology andOceanography 54(3):653-665.
Hassett, B., M. Palmer, E. Bernhardt, S. Smith, J. Carr, and D.Hart, 2005. Restoring Watersheds Project by Project: Trends inChesapeake Bay Tributary Restoration. Frontiers in Ecologyand the Environment 3:259-267.
Hicks, B.J., 1997. Food Webs in Forest and Pasture Streams in theWaikato Region, New Zealand: A Study Based on the Analysesof Stable Isotopes of Carbon and Nitrogen, and Fish GutContents. New Zealand Journal of Marine and FreshwaterResearch 31:651-664.
Hillebrand, H. and M. Kahlert, 2001. Effect of Grazing and Nutri-ent Supply on Periphyton Biomass and Nutrient Stoichiometryin Habitats of Different Productivity. Limnology and Oceanogra-phy 46(8):1881-1898.
Hinkle, S.R., J.H. Duff, F.J. Triska, A. Laenen, E.B. Gates, K.E.Bencala, D.A. Wentz, and S.R. Silva, 2001. Linking HyporheicFlow and Nitrogen Cycling near the Willamette River — ALarge River in Oregon, USA. Journal of Hydrology 244:157-180.
Howarth, R.W., G. Billen, D. Swaney, A. Townsend, N. Jaworski, K.Lajtha, J.A. Downing, R. Elmgren, N. Caraco, T. Jordan, F. Ber-endse, J. Freney, V. Kudeyarov, P. Murdoch, and Z. Zhao-Liang,1996. Regional Nitrogen Budgets and Riverine N & P Fluxes forthe Drainages to the North Atlantic Ocean: Natural and HumanInfluences. In: Nitrogen Cycling in the North Atlantic Ocean andIts Watersheds. Springer, Netherlands, pp. 75-139.
Ishikawa, N.F., H. Doi, and J.C. Finlay, 2012. Global Meta-Analy-sis for Controlling Factors on Carbon Stable Isotope Ratios ofLotic Periphyton. Oecologia 170:541-549.
Jenkins, M.C. and W.M. Kemp, 1984. The Coupling of Nitrificationand Denitrification in Two Estaurine Sediments. Limnology andOceanography 29:609-619.
Kellogg, D.Q., A.J. Gold, S. Cox, K. Addy, and P.V. August,2010. A Geospatial Approach for Assessing DenitrificationSinks within Lower-Order Catchments. Ecological Engineering36:1596-1606.
Kellogg, D.Q., A.J. Gold, P.M. Groffman, K. Addy, M. Stolt, and G.Blazejewski, 2005. In Situ Groundwater Denitrification in Strat-ified, Permeable Soils Underlying Riparian Wetlands. Journal ofEnvironmental Quality 34:524-533.
Kemp, M.J. and W.K. Dodds, 2002. Comparison of Nitrification andDenitrification in Prairie and Agriculturally Influenced Streams.Ecological Applications 12(4):998-1009.
MacLeod, N.A. and D.R. Barton, 1998. Effects of Light Intensity,Water Velocity and Species Composition on Carbon and Nitro-gen Stable Isotopes in Periphyton. Canadian Journal of Fisher-ies and Aquatic Science 55:1919-1925.
Mayer, P., S. Reynolds, M. McCutchen, and T. Canfield, 2007.Meta-Analysis of Nitrogen Removal in Riparian Buffers. Journalof Environmental Quality 36:1172-1180.
McClain, M.E., E.W. Boyer, C.L. Dent, S.E. Gergel, N.B. Grimm,P.M. Groffman, S.C. Hart, J.W. Harvey, C.A. Johnston, E. May-orga, W.H. McDowell, and G. Pinay, 2003. Biogeochemical HotSpots and Hot Moments at the Interface of Terrestrial andAquatic Ecosystems. Ecosystems 6(4):301-312.
Mitsch, W.J., 2007. Ecological and Social Complexity in RestoringWater Resources: Coastal Louisiana, the Mississippi RiverBasin, and Mesopotamia. Stockholm Water Prize LaureatesSeminar: Challenges and Opportunities within the WaterSector, Stockholm International Water Institute, Stockholm,Sweden, pp. 19-23.
Mosier, A., C. Kroeze, C. Nevison, O. Oenema, S. Seitzinger, andO. Van Cleemput, 1998. Closing the Global N2O Budget:Nitrous Oxide Emissions through the Agricultural NitrogenCycle. Nutrient Cycling in Agroecosystems 52(2-3):225-248.
Mosier, A.R. and L. Klemedtsson, 1994. Measuring Denitrificationin the Field. In: Methods of Soil Analysis Part 2, P.S. Bottomley,J.S. Angle, and R.W. Weaver (Editors). American Society ofAgronomy, Madison, Wisconsin, pp. 1048-1062. ISBN: 978-0-89118-865-0.
Mulholland, P.J., R.O. Hall, D.J. Sobota, W.K. Dodds, S.G. Findlay,N.B. Grimm, S.K. Hamilton, W.H. McDowell, J.M. O’Brien, J.L.Tank, L.R. Ashkenas, L.W. Cooper, C.N. Dahm, S.V. Gregory,S.L. Johnson, J.L. Meyer, B.J. Peterson, G.C. Poole, H.M. Va-lett, J.R. Webster, C. Arango, J.J. Beaulieu, M.J. Bernot, A.J.Burgin, C. Crenshaw, A.M. Helton, L. Johnson, B.R. Niederleh-ner, J.D. Potter, R.W. Sheibley, and S.M. Thomas, 2009. NitrateRemoval in Stream Ecosystems Measured by 15N AdditionExperiments: Denitrification. Limnology and Oceanography54:666-680.
Mulholland, P.J., E.R. Marzolf, S.P. Hendricks, R.V. Wilkerson,and A.K. Baybayan, 1995. Longitudinal Patterns of NutrientCycling and Periphyton Characteristics in Streams: A Test ofUpstream-Downstream Linkage. Journal of North AmericanBenthological Society 14:357-370.
Naiman, R.J. and H. Decamps, 1997. The Ecology of Interfaces:Riparian Zones. Annual Review of Ecology and Systematics28:621-658.
Nishio, T., I. Koike, and A. Hattorri, 1983. Denitrification, NitrateReduction and Oxygen Consumption in Coastal and EstuarineSediments. Applied and Environmental Microbiology 43:648-653.
Peterson, B.J., et al., 2001. Control of Nitrogen Export from Water-sheds by Headwater Streams. Science 292:86-90.
Peterson, C.G., A.D. Daley, S.M. Pechauer, K.N. Kalscheur, M.J.Sullivan, S.L. Kufta, M. Rojas, K.A. Gray, and J.J. Kelly, 2011.Development of Associations between Microalgae and Denitrify-ing Bacteria in Streams of Contrasting Anthropogenic Influence.FEMS Microbiology Ecology 77:477-492.
Reisinger, A.J., J.M. Blair, C.W. Rice, and W.K. Dodds, 2013.Woody Vegetation Removal Stimulates Riparian and BenthicDenitrification in Tallgrass Prairie. Ecosystems 16:547-560.
Romani, A. and S. Sabater, 2000. Influence of Algal Biomass onExtracellular Enzyme Activity in River Biofilms. Microbial Ecol-ogy 40:16-24.
Romani, A.M., A. Giorgi, V. Acuna, and S. Sabater, 2004. The Influ-ence of Substratum Type and Nutrient Supply on Biofilm OrganicMatter Utilization in Streams. Limnology and Oceanography49:1713-1721.
Rott, E., H.C. Duthie, and E. Pipp, 1998. Monitoring Organic Pollu-tion and Eutrophication in the Grand River, Ontario, by Meansof Diatoms. Canadian Journal of Fisheries and Aquatic Science55:1443-1453.
Rounick, J.S. and M.J. Winterbourn, 1986. Stable Carbon Isotopesand Carbon Flow in Ecosystems. BioScience 36:171-177.
Ryhiner, G., K. Sorensen, B. Birou, and H. Gros, 1994. BiofilmReactors Configuration for Advanced Nutrient Removal. WaterScience and Technology 29:111-117.
Sabater, F., J. Armengol, and S. Sabater, 1991. Physico-ChemicalDisturbances Associated with Spatial and Temporal Variationin a Mediterranean River. Journal of North American Bentho-logical Society 10:2-13.
JAWRA JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION624
LAZAR, GOLD, ADDY, MAYER, FORSHAY, AND GROFFMAN
Sabater, S., S.V. Gregory, and J.R. Sedell, 1998. CommunityDynamics and Metabolism of Benthic Algae Colonizing Woodand Rock Substrata in Forest Stream. Journal of Phycology34:561-567.
Sabater, S., H. Guasch, A. Roman�ı, and I. Mu~noz, 2002. The Effectof Biological Factors on the Efficiency of River Biofilms inImproving Water Quality. Hydrobiologia 469:149-156.
Sabater, S., F. Sabater, and J. Armengol, 1988. Relationshipsbetween Diatom Assemblages and Physico-Chemical Variablesin the River Ter (NE Spain). International Revue der GesamtenHydrobiologie und Hydrographie 73:171-179.
Schipper, L.A., W.D. Robertson, A.J. Gold, D.B. Jaynes, and S.C.Cameron, 2010. Denitrifying Bioreactors — An Approach forReducing Nitrate Loads to Receiving Waters. Ecological Engi-neering 36:1532-1543.
Schultz, R.C., T.M. Isenhart, W.W. Simpkins, and J.P. Colletti,2004. Riparian Forest Buffers in Agroecosystems — LessonsLearned from the Bear Creek Watershed, Central Iowa, USA.Agroforestry Systems 61:35-50.
Sweeney, B.W., T.L. Bott, B.K. Jackson, L.A. Kaplan, J.D. New-bold, L.J. Standley, W.C. Hession, and R.J. Horwitz, 2004.Riparian Deforestation, Stream Narrowing, and Loss of StreamEcosystem Services. Proceedings of the National Academy ofSciences USA 101:14132-14137.
Turner, R.E. and N.N. Rabalais, 1994. Coastal Eutrophication nearthe Mississippi River Delta. Nature 368:619-621.
USGS (U.S. Geological Survey), 2011. Streamstats. http://streamstats09.cr.usgs.gov/, accessed November 2011.
Vidon, P., C. Allan, D. Burns, T. Duval, N. Gurwick, S. Inamdar, R.Lowrance, J. Okay, D. Scott, and S. Sebestyen, 2010. Hot Spotsand Hot Moments in Riparian Zones: Potential for ImprovedWater Quality Management. Journal of the American WaterResources Association 46:278-298, doi: 10.1111 ⁄ j.1752-1688.2010.00420.x.
Welsh, D.J., 1991. Riparian Forest Buffers: Function and Designfor Protection and Enhancement of Water Resources. NA-PR-07-91. U.S. Department of Agriculture, Forest Service, NorthernArea State & Private Forestry.
JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION JAWRA625
INSTREAM LARGE WOOD: DENITRIFICATION HOTSPOTS WITH LOW N2O PRODUCTION
