Nitrous oxide emission from denitrification in stream and river networks

Jake J. Beaulieu; Jennifer L. Tank; Stephen K. Hamilton; Wilfred M. Wollheim; Robert O. Hall; Patrick J. Mulholland; Bruce J. Peterson; Linda R. Ashkenas; Lee W. Cooper; Clifford N. Dahm; Walter K. Dodds; Nancy B. Grimm; Sherri L. Johnson; William H. McDowell; Geoffrey C. Poole; H. Maurice Valett; Clay P. Arango; Melody J. Bernot; Amy J. Burgin; Chelsea L. Crenshaw; Ashley M. Helton; Laura T. Johnson; Jonathan M. O'Brien; Jody D. Potter; Richard W. Sheibley; Daniel J. Sobota; Suzanne M. Thomas

doi:10.1073/pnas.1011464108

New Research In

Edited* by William H. Schlesinger, Cary Institute of Ecosystem Studies, Millbrook, NY, and approved November 11, 2010 (received for review August 4, 2010)

Abstract

Nitrous oxide (N₂O) is a potent greenhouse gas that contributes to climate change and stratospheric ozone destruction. Anthropogenic nitrogen (N) loading to river networks is a potentially important source of N₂O via microbial denitrification that converts N to N₂O and dinitrogen (N₂). The fraction of denitrified N that escapes as N₂O rather than N₂ (i.e., the N₂O yield) is an important determinant of how much N₂O is produced by river networks, but little is known about the N₂O yield in flowing waters. Here, we present the results of whole-stream ¹⁵N-tracer additions conducted in 72 headwater streams draining multiple land-use types across the United States. We found that stream denitrification produces N₂O at rates that increase with stream water nitrate (NO₃⁻) concentrations, but that <1% of denitrified N is converted to N₂O. Unlike some previous studies, we found no relationship between the N₂O yield and stream water NO₃⁻. We suggest that increased stream NO₃⁻ loading stimulates denitrification and concomitant N₂O production, but does not increase the N₂O yield. In our study, most streams were sources of N₂O to the atmosphere and the highest emission rates were observed in streams draining urban basins. Using a global river network model, we estimate that microbial N transformations (e.g., denitrification and nitrification) convert at least 0.68 Tg·y⁻¹ of anthropogenic N inputs to N₂O in river networks, equivalent to 10% of the global anthropogenic N₂O emission rate. This estimate of stream and river N₂O emissions is three times greater than estimated by the Intergovernmental Panel on Climate Change.

Humans have more than doubled the availability of fixed nitrogen (N) in the biosphere, particularly through the production of N fertilizers and the cultivation of N-fixing crops (1). Increasing N availability is producing unintended environmental consequences including enhanced emissions of nitrous oxide (N₂O), a potent greenhouse gas (2) and an important cause of stratospheric ozone destruction (3). The Intergovernmental Panel on Climate Change (IPCC) estimates that the microbial conversion of agriculturally derived N to N₂O in soils and aquatic ecosystems is the largest source of anthropogenic N₂O to the atmosphere (2). The production of N₂O in agricultural soils has been the focus of intense investigation (i.e., >1,000 published studies) and is a relatively well constrained component of the N₂O budget (4). However, emissions of anthropogenic N₂O from streams, rivers, and estuaries have received much less attention and remain a major source of uncertainty in the global anthropogenic N₂O budget.

Microbial denitrification is a large source of N₂O emissions in terrestrial and aquatic ecosystems. Most microbial denitrification is a form of anaerobic respiration in which nitrate (NO₃⁻, the dominant form of inorganic N) is converted to dinitrogen (N₂) and N₂O gases (5). The proportion of denitrified NO₃⁻ that is converted to N₂O rather than N₂ (hereafter referred to as the N₂O yield and expressed as the mole ratio) partially controls how much N₂O is produced via denitrification (6), but few studies provide information on the N₂O yield in streams and rivers because of the difficulty of measuring N₂ and N₂O production in these systems. Here we report rates of N₂ and N₂O production via denitrification measured using whole-stream ¹⁵NO₃⁻-tracer experiments in 72 headwater streams draining different land-use types across the United States. This project, known as the second Lotic Intersite Nitrogen eXperiment (LINX II), provides unique whole-system measurements of the N₂O yield in streams.

Although N₂O emission rates have been reported for streams and rivers (7, 8), the N₂O yield has been studied mostly in lentic freshwater and marine ecosystems, where it generally ranges between 0.1 and 1.0%, although yields as high as 6% have been observed (9). These N₂O yields are low compared with observations in soils (0–100%) (10), which may be a result of the relatively lower oxygen (O₂) availability in the sediments of lakes and estuaries. However, dissolved O₂ in headwater streams is commonly near atmospheric equilibrium and benthic algal biofilms can produce O₂ at the sediment–water interface, resulting in strong redox gradients more akin to those in partially wetted soils. Thus, streams may have variable and often high N₂O yields, similar to those in soils (11). The N₂O yield in headwater streams is of particular interest because much of the NO₃⁻ input to rivers is derived from groundwater upwelling into headwater streams. Furthermore, headwater streams compose the majority of stream length within a drainage network and have high ratios of bioreactive benthic surface area to water volume (12).

Results and Discussion

The ¹⁵N-NO₃⁻ tracer was detected in the dissolved N₂O pool in 53 of 72 streams and we assume that direct denitrification of stream water NO₃⁻ to N₂O (N₂O_DN) was the source of this ¹⁵N-N₂O. It is unlikely that nitrification was an important source of labeled ¹⁵N-N₂O because the 24-h duration of the experiments was too short for the tracer to be assimilated by stream biota, mineralized, and subsequently nitrified. Rates of N₂O_DN varied by land use with the highest rates observed in high NO₃⁻ urban streams and the lowest in reference streams (i.e., those with little land conversion in their watersheds) (Fig. 1A). A positive relationship between N₂O_DN and stream water NO₃⁻ concentration (Fig. 1B) suggests anthropogenic N loading to streams stimulates denitrification and concomitant N₂O production. The N₂O_DN rates reported here are lower than most published reports (Fig. 1A), possibly because our in situ measurements are not affected by the experimental artifacts and scaling problems associated with sediment slurries, cores, and chambers used in most previously published estimates (13).

Fig. 1.

(A) Box plots of stream N₂O production rates via denitrification of water column NO₃⁻ by catchment land use (reference, agricultural and urban). Benthic N₂O production rates reported in other studies are also shown. Significant differences between land-use types were determined with a one-way ANOVA followed by Tukey's post hoc test (P = 0.004) and are displayed as different lowercase letters above the box plots. See SI Materials and Methods for references. (B) Relationship between stream water NO₃⁻ and rates of N₂O production via denitrification (r² = 0.68, P < 0.001). (C) Nitrous oxide emission rates from streams. Significant differences between land use types were determined with a one-way ANOVA (P = 0.002) followed by Tukey's post hoc test and are displayed as different lowercase letters above the box plots. (D) Relationship between stream water NO₃⁻ concentrations and N₂O emission rates. The vertical dashed line represents a NO₃⁻ threshold (95 μg N·L⁻¹) below which N₂O emission rates are unrelated to NO₃⁻ (two-dimensional Kolmogorov–Smirnov test). Above the threshold N₂O emission rates are positively related to NO₃⁻ concentrations as represented by the least-squares best-fit line (solid black). (E) Percentages of stream N₂O emissions attributed to direct denitrification. Values >100% indicate N₂O was accumulating in the water column. There was no effect of land use (P = 0.13). (F) Variation in the percentage of stream N₂O emissions attributed to direct denitrification is partially explained by stream water NO₃⁻ concentration (r² = 0.32, P < 0.001).

The ¹⁵N-NO₃⁻ tracer was detected in both the dissolved N₂ and N₂O pools in 40 of the 72 study sites and we assume all ¹⁵N-N₂ was produced via direct denitrification. The only other potential source of ¹⁵N-N₂ production is anammox, a process by which chemolithoautotrophic bacteria convert ammonium (NH₄⁺) and nitrite (NO₂⁻) to N₂, but available evidence suggests that anammox is unimportant relative to denitrification in streams and rivers (14). Furthermore, any N₂ produced via anammox during the ¹⁵N tracer additions would have contained little ¹⁵N tracer because stream water NH₄⁺ was minimally labeled with the ¹⁵N tracer.

The N₂O yield ranged from 0.04% to 5.6% across the 53 streams; however, the interquartile range (0.3–1.0%) was well constrained despite substantial variation in NO₃⁻ availability, dissolved O₂, primary productivity, sediment organic matter, and stream geomorphology across our study sites (Fig. 2, Table S1). Denitrification proceeds by sequentially reducing NO₃⁻ to NO₂⁻, nitric oxide (NO), N₂O, and finally N₂. Each reduction is performed by a different enzyme and the N₂O yield is determined by the relative activities of the N₂O-producing and reducing enzymes. There is a positive relationship between the N₂O yield and NO₃⁻ concentration in soils (15, 16) and estuarine sediments (9), possibly because higher NO₃⁻ availability suppresses nitrous oxide reductase (nos), the enzyme that reduces N₂O to N₂ (16). However, we did not find a significant relationship between stream water NO₃⁻ concentration and the N₂O yield (P = 0.09), despite NO₃⁻ concentrations spanning five orders of magnitude. Our findings suggest increased NO₃⁻ loading to streams stimulates overall denitrification rates and concomitant N₂O production, but does not increase the N₂O yield.

Fig. 2.

Denitrification N₂O yields (percentage of denitrified N released as N₂O) measured in this study in comparison with other ecosystems. Data are displayed in box plots unless there were fewer than nine observations, in which case each observation is represented by a solid circle. See SI Materials and Methods for references.

The N₂O yield in soils is related to the relative availability of oxidants (NO₃⁻) and reductants (organic carbon). When the availability of NO₃⁻ greatly exceeds that of organic carbon, NO₃⁻ is preferred over N₂O as a terminal electron acceptor and N₂O accumulates (5, 17–19). The N₂O yield was not related to the ratio of stream water NO₃⁻ concentration to dissolved or particulate organic carbon concentration (P ≥ 0.17), but was negatively related to stream ecosystem respiration (P = 0.04, r² = 0.11), suggesting factors promoting aerobic respiration (e.g., labile carbon availability) may decrease the N₂O yield.

Our data suggest that denitrification in aquatic ecosystems produces lower and less variable N₂O yields than in terrestrial ecosystems (Fig. 2). This finding may be explained by differences in oxygen (O₂) availability and molecular diffusion rates between aquatic sediments and the partially water-filled pore spaces of soils. Because nos is the most O₂ sensitive denitrifying enzyme (20), minor amounts of O₂ can suppress the reduction of N₂O without inhibiting its production, resulting in an elevated N₂O yield. Nitrous oxide is produced as a free intermediate that can escape reduction to N₂ by diffusing away from the denitrification zone (16). Partially wet soils may present air-filled routes through which the N₂O could more readily evade from soil solutions and ultimately escape to the atmosphere, whereas in aquatic sediments there may be a much greater likelihood of interception of the dissolved N₂O by nos before it can diffuse to the overlying water column. Overall, lower O₂ availability and gas diffusion rates in aquatic sediments compared with soils may account for the low aquatic N₂O yield.

Resource managers have used stream restoration to reduce watershed N export to estuaries and coastal oceans where it can contribute to eutrophication (21). This approach has been criticized on the grounds that stream denitrification alone cannot alleviate watershed N pollution (22) and that enhanced stream denitrification may lead to increased N₂O emission (23). Our data demonstrate that the N₂O yield in headwater streams is no larger than in other aquatic ecosystems and much lower than in soils (Fig. 2), indicating that measures to promote stream denitrification may have a relatively lower impact on climate change than the promotion of an equivalent amount of denitrification in terrestrial environments.

Other sources of N₂O to streams include in-stream nitrification and upwelling groundwater. The sum of all these N₂O sources determines the total amount of N₂O emitted by a stream. We investigated the potential for these additional sources to contribute to total N₂O emissions by estimating N₂O emission rates from the dissolved N₂O concentration and the air–water gas exchange rates in each of the 72 streams. The majority of streams were net sources of N₂O to the atmosphere (55 of 72) and only 2 streams showed a diel pattern in emission rates so we did not further consider diel variations in constructing N₂O budgets (cf. ref. 24). Stream N₂O emission rates were related to watershed land use, with highest emission rates in urban streams, intermediate rates in agricultural streams, and lowest rates in reference streams (Fig. 1C). Stream NO₃⁻ concentrations predicted N₂O emission rates when NO₃⁻-N exceeded 95 μg·L⁻¹ (P = 0.01, r² = 0.16), but below this concentration N₂O emission rates were uniformly low and unrelated to NO₃⁻ concentration (Fig. 1D). This finding suggests that stream N₂O emission rates are not solely controlled by direct denitrification within the stream, but are likely enhanced by other sources including inputs of dissolved N₂O from groundwater.

We compared N₂O_DN (Fig. 1A) to N₂O emission rates (Fig. 1C) and found that the direct denitrification of stream water NO₃⁻ accounted for an average of 26% of N₂O emissions (Fig. 1E). This is a conservative estimate of in-stream N₂O production via denitrification because our method does not detect N₂O produced from the denitrification of NO₃⁻ regenerated within sediments and biofilms (e.g., indirect denitrification following organic N mineralization and nitrification), which can be the dominant source of NO₃⁻ supporting denitrification when stream water NO₃⁻ concentration is low (<140 μg N·L⁻¹) (25). The relative importance of N₂O_DN as a source of N₂O was positively related to stream water NO₃⁻ concentrations (Fig. 1F), reflecting the greater importance of direct denitrification with increasing stream water NO₃⁻ concentrations.

Nitrification is a potentially large source of N₂O emissions, but we know of no published measurements of N₂O production via nitrification in streams. Several studies have shown that nitrification rates can be equal to or greater than denitrification rates in streams (26–28) and rivers (29), and the IPCC assumes nitrification rates exceed denitrification by twofold (30). Measurements of the nitrification N₂O yield (i.e., the fraction of nitrified N escaping as N₂O) are sparse, but it appears to be within the same range as the denitrification N₂O yield (9). Therefore, the IPCC assumes that nitrification produces twice as much N₂O emission as denitrification in streams and rivers. Given that N₂O_DN produced within the stream contributes an average of 26% of the N₂O emitted by headwater streams (Fig. 1E), nitrification could account for as much as an additional 52%, with groundwater inputs and indirect denitrification composing the remainder (Fig. 3). This budget highlights the potential importance of nitrification and indirect denitrification to stream N₂O production, but these processes remain poorly understood and therefore represent critical research gaps. Nevertheless, our research demonstrates that headwater streams are not only conduits for the emission of groundwater-derived N₂O to the atmosphere, but also active sites of in situ N₂O production, particularly where NO₃⁻ concentrations are elevated by anthropogenic N loading.

Fig. 3.

Average N₂O fluxes estimated in this study (all units are μg N·m⁻²·h⁻¹). Black arrows represent fluxes that were directly measured and the white arrow with dashed boundaries represents fluxes that were estimated by mass balance. (A) N₂O produced in the stream, or imported to the stream via groundwater, temporarily resides in a pool of dissolved N₂O before being emitted to the atmosphere. (B) Direct denitrification is the conversion of stream water nitrate to N₂ and N₂O. Less than 1% of stream water nitrate subject to direct denitrification is converted to N₂O, but this is the source of 26% of the N₂O emitted by the stream. (C) The balance of N₂O emission in excess of that produced via direct denitrification (e.g., 37 − 9.6 = 27.4) must have entered the stream via another mechanism. Likely mechanisms include indirect denitrification (e.g., the denitrification of nitrate generated within the sediments), nitrification, and inputs of N₂O-supersaturated groundwater.

The IPCC and others have estimated global anthropogenic N₂O emissions from streams and rivers by assuming all anthropogenic N that enters a river network is nitrified to NO₃⁻ and half of this NO₃⁻ is then denitrified; the N₂O yield is assumed to range from 0.3% to 3.0% in each transformation (9, 30, 31). This approach has shown that streams and rivers may be the source of 15% of global anthropogenic emissions, but this estimate is poorly constrained due to uncertainty in the N₂O yield and proportion of anthropogenic N inputs denitrified in river networks. We improved the estimate of global anthropogenic N₂O emissions from lotic systems by modifying an existing global river network model (32) to include spatially explicit N loading in the contemporary era, an empirically derived relationship between denitrification and NO₃⁻ concentrations based on the LINX II ¹⁵N tracer studies (22), and the mean N₂O yield of 0.9% reported here. The model estimates the percentage of dissolved inorganic nitrogen (DIN) delivered to the world's streams and rivers that is converted to N₂O via direct denitrification as water flows through the river network, including lakes and reservoirs. However, the model does not include indirect denitrification or denitrification associated with off-channel features (e.g., floodplains, riparian zones) and therefore provides a conservative estimate of anthropogenic N₂O emissions (Fig. 3).

The percentage of DIN inputs converted to N₂O via direct denitrification of water column NO₃⁻ in river networks across the globe ranges from 0% to 0.9% (Fig. 4). The percentage of N inputs converted to N₂O decreases with increasing N inputs because denitrification becomes less efficient as a NO₃⁻ sink at higher NO₃⁻ concentrations (22). We expected that the longer water-residence time in large rivers would result in a larger percentage of N inputs being denitrified compared with smaller river networks. However, we found no effect of catchment area (a surrogate for river network length), likely because the size of the network is confounded by other factors including variation in the distribution of N inputs, temperature, runoff conditions, and the presence of lakes and reservoirs within river networks (32).

Fig. 4.

Relationship between the percentage of dissolved inorganic nitrogen (DIN) inputs to river networks that are converted to N₂O via denitrification and the amount of DIN delivered to the river network from the catchment. Data are from 866 river networks included in a global river network model. Data points are split into rivers draining basins >100,000 km² (solid black circles) and those draining between 10,000 and 100,000 km² (open red circles).

We estimate that at the global scale, 0.75% of DIN inputs to river networks are converted to N₂O via direct denitrification and nitrification, threefold greater than the IPCC's estimate. This N₂O is likely to be emitted to the atmosphere from the turbulent water columns of streams and rivers. Using the IPCC's modeling framework and the results of our work, we estimate that nitrification and denitrification in river networks convert 0.68 Tg·y⁻¹ of anthropogenic DIN inputs to N₂O globally, equivalent to 10% of the global anthropogenic N₂O emissions of 6.7 Tg N·y⁻¹ (2) (for calculation details see Global N₂O Budget in SI Materials and Methods). This estimate of anthropogenic N₂O emissions from river networks is conservative because our model does not include several potentially large sources of N₂O (e.g., indirect denitrification and groundwater inputs). We also caution that our estimate of N₂O emissions attributed to nitrification is supported by few data (see above and Fig. 3).

We found that the combination of high denitrification rates and large anthropogenic DIN inputs results in substantial anthropogenic N₂O emissions from river networks, even though <1% of denitrified NO₃⁻ was converted to N₂O, a much lower percentage than has been reported for upland or flooded soils. Management efforts to enhance stream denitrification will reduce the delivery of N to sensitive coastal waters with less concomitant N₂O emissions than the enhancement of a comparable amount of denitrification in soils. Unfortunately, river networks have a limited capacity to remove NO₃⁻ from the water column and anthropogenic N inputs have already overwhelmed this capacity in many river systems (33, 34). Whereas the trade-off between desirable N removal and undesirable N₂O production may be smaller in streams than in soils, the best way to reduce N export to coastal waters without enhancing N₂O emissions is to reduce N inputs to watersheds.

Materials and Methods

LINX II consisted of ¹⁵NO₃⁻ additions to 72 small streams distributed across three land-use categories and eight regions to provide in situ measurements of N₂ and N₂O production via denitrification at the whole-stream scale. We used a standardized set of experimental protocols to measure biogeochemical process rates including denitrification, ecosystem respiration, and gross primary production (35). We also measured a broad suite of physicochemical characteristics including organic matter standing stocks, water column nutrient concentrations, effective stream depth, stream width, stream discharge, and water velocity. The experiments were conducted as previously described (22, 35, 36) and as reported online in the project protocols (http://www.biol.vt.edu/faculty/webster/linx/). A detailed description of the experimental protocols, study site locations, and characteristics can be found in SI Materials and Methods, Fig. S1, and Table S1.

Site Selection.

Study sites were selected to encompass a broad range of conditions across three land-use categories and eight regions. Within each region headwater streams (discharge ranged from 0.2–268 L·s⁻¹) were selected draining basins dominated by native vegetation (reference), urban land use, or agricultural land use, with three sites in each land use for a total of nine sites per region (Table S1, Fig. S1). We selected stream reaches that had minimal groundwater or surface water inputs and were long enough to allow for a measureable amount of in-stream N processing (105–1,830 m).

Isotope Addition and Sampling.

The production of N₂ and N₂O via denitrification was measured by continuously adding a solution of sodium bromide (NaBr, conservative tracer) and ¹⁵N-enriched potassium nitrate (K¹⁵NO₃⁻, 98+% ¹⁵N) to each stream for 24 h beginning at ≈13:00 hours using a small pump. The pump rate and injectate concentration were chosen to increase the stream water δ¹⁵NO₃⁻ by 20,000‰ and the Br⁻ concentration by 100 μg·L⁻¹. The conservative tracer was used to account for ground and surface water inputs to the reach and to measure channel hydraulic properties. The K¹⁵NO₃⁻ addition resulted in a relatively small (∼7.5%) increase in stream water NO₃⁻ concentration.

Ten sampling stations were selected along the study reach and water samples were taken for NO₃⁻ (concentration and δ¹⁵N), N₂ (δ¹⁵N), and N₂O (concentration and δ¹⁵N) several hours before, 12 h after, and 23 h after the K¹⁵NO₃⁻ addition began. Nitrate samples were filtered (GFF; 0.7-μm pore size; Whatman) in the field and stored on ice or frozen before analysis. Our protocol for dissolved gas sampling is described in detail elsewhere (37) and is briefly outlined here. Gas samples were taken by slowly withdrawing 40 or 120 mL of stream water into a 60- or 140-mL polypropylene syringe (BD Falcon and Harvard Apparatus) equipped with a polycarbonate stopcock. Twenty mL of ultrahigh purity helium was then added to the syringes, which were gently shaken for 5 min to equilibrate the dissolved N₂ and N₂O between the aqueous and gas phases. The headspace gas was then transferred to a preevacuated 12-mL Exetainer (Labco) and stored underwater before analysis. All gas transfers were done underwater to minimize contamination from atmospheric N₂ and N₂O.

Air–Water Gas Exchange Rate Measurement.

Within 1 d of the ¹⁵NO₃⁻ addition the air–water gas exchange rate (k₂, units of time⁻¹) was measured using the steady-state tracer gas injection method with either propane or SF₆ as tracers (Table S1) (38, 39). The tracer gas and a conservative tracer [e.g., sodium chloride (NaCl) or rhodamine] were added to the stream at a constant rate. Water samples were collected from each of the 10 downstream sampling stations after the tracer concentrations reached a plateau throughout the reach as indicated by conductivity or fluorescence at the most downstream sampling station. Gas tracer samples were collected in 5-mL polypropylene syringes and injected into preevacuated glass storage vials. The air–water gas exchange rate was calculated from the dilution-corrected decline in tracer gas concentration across the experimental reach (Table S1).

Analytical Methods.

¹⁵N-NO₃ was determined on filtered water samples following Sigman et al. (40). Samples were analyzed for ¹⁵N on a Finnigan Delta-S or a Europa 20/20 mass spectrometer in the Mass Spectrometer Laboratory of the Marine Biological Laboratory in Woods Hole, MA (http://ecosystems.mbl.edu/SILAB/about.html); a Europa Integra mass spectrometer in the Stable Isotope Laboratory of the University of California, Davis, CA (http://stableisotopefacility.ucdavis.edu/); or a ThermoFinnigan DeltaPlus mass spectrometer in the Stable Isotope Laboratory at Kansas State University, Manhattan, KS (http://www.k-state.edu/simsl).

Gas samples were analyzed for δ¹⁵N₂, δ¹⁵N₂O, and N₂O concentration by mass spectrometry using either a Europa Hydra Model 20/20 mass spectrometer at the Stable Isotope Laboratory of the University of California, Davis, CA, or a GV Instruments Prism Series II mass spectrometer in the Biogeochemistry Laboratory, Department of Zoology, Michigan State University, East Lansing, MI. The original stream water N₂O concentration was calculated using temperature-corrected Bunsen solubility coefficients, a mass balance of the gas–liquid system in the sample syringe, and an atmospheric N₂O partial pressure of 315 ppbv (41). Propane and SF₆ concentrations were measured using gas chromatography with flame ionization and electron capture detectors, respectively.

¹⁵N content of all samples was reported in δ¹⁵N notation, where δ¹⁵N = [(R_SA/R_ST) – 1] × 1,000, R = ¹⁵N/¹⁴N, and the results are expressed as per mil deviation of the sample (SA) from the standard (ST), N₂ in air (δ¹⁵N = 0‰). All δ¹⁵N values were converted to ¹⁵N mole fractions (MF = ¹⁵N/¹⁴N + ¹⁵N), and tracer ¹⁵N fluxes were calculated for each sample by multiplying the ¹⁵N MF, corrected for natural abundances of ¹⁵N by subtracting the average ¹⁵N MF for samples collected before the ¹⁵N addition, by the concentrations of NO₃⁻, N₂, or N₂O in stream water (concentrations of NO₃⁻ and N₂O were measured, whereas N₂ was taken as the concentration in equilibrium with air at the ambient stream temperature) and stream discharge derived from the measured conservative solute tracer concentrations.

Gas Production and Emission Calculations.

Rates of N₂ and N₂O production were calculated as best-fit model parameters from a two-compartment model of denitrification linking ¹⁵N₂, ¹⁵N₂O, and ¹⁵NO₃⁻ over the study reach described in SI Materials and Methods. Nitrous oxide emission rates via diffusive evasion (F, μg N₂O-N·m⁻²·h⁻¹) were calculated as

where h is the stream depth, N₂O_obs is the measured concentration of dissolved N₂O in the water (average across all sampling stations), and N₂O_equil is the N₂O concentration expected if the stream were in equilibrium with the atmosphere.

Global River Network N₂O Emission Model.

Global anthropogenic N₂O emissions from river networks were estimated using a river network model. The model was run under mean annual conditions and accounts for the spatial distribution of DIN loading, temperature, hydrology, and denitrification efficiency loss. Model details and prediction errors can be found in Fig. S2 and SI Materials and Methods.

Acknowledgments

We are grateful to N. E. Ostrom for assistance with stable isotope measurements of N₂ and N₂O and G. P. Robertson for comments on the manuscript. We thank the US Forest Service, National Park Service, and many private landowners for permission to conduct experiments on their lands. We also acknowledge the many workers who helped with the Lotic Intersite Nitrogen experiments. Funding for this research was provided by the National Science Foundation (DEB-0111410). The National Science Foundation’s Long Term Ecological Research (NSF-LTER) network hosted many of the study sites included in this research and partially supported several of the authors during the project. We specifically acknowledge Andrews, Central Arizona-Phoenix, Coweeta, Kellogg Biological Station, Konza, Luquillo, Plum Island, and Sevilleta NSF-LTERs for support.

Footnotes

²To whom correspondence should be addressed. E-mail: beaulieu.jake{at}epa.gov.
Author contributions: J.L.T., S.K.H., R.O.H., P.J.M., B.J.P., L.R.A., L.W.C., C.N.D., W.K.D., N.B.G., S.L.J., W.H.M., G.C.P., and H.M.V. designed research; J.J.B., J.L.T., S.K.H., W.M.W., R.O.H., P.J.M., B.J.P., L.R.A., W.K.D., N.B.G., S.L.J., W.H.M., H.M.V., C.P.A., M.J.B., A.B., C.C., A.M.H., L.T.J., J.M.O., J.D.P., R.W.S., D.J.S., and S.M.T. performed research; J.J.B. analyzed data; J.J.B., J.L.T., and S.K.H. wrote the paper; and W.M.W. performed the global river N₂O production modeling.
The authors declare no conflict of interest.
↵*This Direct Submission article had a prearranged editor.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1011464108/-/DCSupplemental.

Freely available online through the PNAS open access option.

View Abstract

References

↵
1. Robertson GP,
2. Vitousek PM
(2009) Nitrogen in Agriculture: Balancing the cost of an essential resource. Annu Rev Environ Resour 34:97–125.
OpenUrl CrossRef
↵
1. Solomon S,
2. et al.
1. Forster P,
2. et al.
(2007) in Climate Change 2007: The Physical Science Basis. Contributions of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Changes in atmospheric constituents and in radiative forcing, ed Solomon S, et al. (Cambridge Univ Press, New York).
↵
1. Ravishankara AR,
2. Daniel JS,
3. Portmann RW
(2009) Nitrous oxide (N₂O): The dominant ozone-depleting substance emitted in the 21st century. Science 326:123–125.
OpenUrl Abstract/FREE Full Text
↵
1. Stehfest E,
2. Bouwman L
(2006) N₂O and NO emission from agricultural fields and soils under natural vegetation: Summarizing available measurement data and modeling of global annual emissions. Nutr Cycl Agroecosyst 74:207–228.
OpenUrl CrossRef
↵
1. Andreae MO,
2. Schimel DS
1. Firestone MK,
2. Davidson EA
(1989) in Exchange of Trace Gases Between Terrestrial Ecosystems and the Atmosphere, Microbiological basis of NO and N₂O production and consumption in soil, eds Andreae MO, Schimel DS (Wiley, Chichester, UK), pp 7–21.
↵
1. Codispoti LA
(2010) Oceans. Interesting times for marine N2O. Science 327:1339–1340.
OpenUrl Abstract/FREE Full Text
↵
1. Cole JJ,
2. Caraco NF
(2001) Emissions of nitrous oxide (N₂O) from a tidal, freshwater river, the Hudson River, New York. Environ Sci Technol 35:991–996.
OpenUrl PubMed
↵
1. Beaulieu JJ,
2. Arango CP,
3. Hamilton SK,
4. Tank JL
(2008) The production and emission of nitrous oxide from headwater streams in the midwestern USA. Glob Chang Biol 14:878–894.
OpenUrl CrossRef
↵
1. Seitzinger SP,
2. Kroeze C
(1998) Global distribution of nitrous oxide production and N inputs in freshwater and coastal marine ecosystems. Global Biogeochem Cycles 12:93–113.
OpenUrl CrossRef
↵
1. Stevens RJ,
2. Laughlin RJ,
3. Malone JP
(1998) Measuring the mole fraction and source of nitrous oxide in the field. Soil Biol Biochem 30:541–543.
OpenUrl CrossRef
↵
1. Stevens RJ,
2. Laughlin RJ
(1998) Measurement of nitrous oxide and di-nitrogen emissions from agricultural soils. Nutr Cycl Agroecosyst 52:131–139.
OpenUrl CrossRef
↵
1. Alexander RB,
2. Boyer EW,
3. Smith RA,
4. Schwarz GE,
5. Moore RB
(2007) The role of headwater streams in downstream water quality. J Am Water Resour Assoc 43:41–59.
OpenUrl CrossRef
↵
1. Groffman PM,
2. et al.
(2006) Methods for measuring denitrification: Diverse approaches to a difficult problem. Ecol Appl 16:2091–2122.
OpenUrl CrossRef
↵
1. Burgin A,
2. Hamilton SK
(2007) Have we overemphasized the role of denitrification in aquatic ecosystems? A review of nitrate removal pathways. Front Ecol Environ 5:89–96.
OpenUrl CrossRef
↵
1. Firestone MK,
2. Firestone RB,
3. Tiedje JM
(1980) Nitrous oxide from soil denitrification: Factors controlling its biological production. Science 208:749–751.
OpenUrl Abstract/FREE Full Text
↵
1. Firestone MK,
2. Smith MS,
3. Firestone RB,
4. Tiedje JM
(1979) The influence of nitrate, nitrite, and oxygen on the composition of the gaseous products of denitrification in soil. Soil Sci Soc Am J 43:1140–1144.
OpenUrl
↵
1. Harper LA,
2. Mosier AR,
3. Duxbury JM,
4. Rolston DE
1. Hutchinson GL,
2. Davidson EA
(1993) in Agricultural Ecosystem Effects on Trace Gases and Global Climate Change, Processes for production and consumption of gaseous nitrous oxide in soil, eds Harper LA, Mosier AR, Duxbury JM, Rolston DE (Am Soc Agron, Madison, WI), pp 79–94.
↵
1. Swerts M,
2. Merckx R,
3. Vlassak K
(1996) Denitrification, N₂-fixation and fermentation during anaerobic incubation of soils amended with glucose and nitrate. Biol Fertil Soils 23:229–235.
OpenUrl CrossRef
↵
1. Miller MN,
2. et al.
(2008) Crop residue influence on denitrification, N2O emissions and denitrifier community abundance in soil. Soil Biol Biochem 40:2553–2562.
OpenUrl CrossRef
↵
1. Zehnder AJB
1. Tiedje JM
(1988) in Biology of Anaerobic Microorganisms, Ecology of denitrification and dissimilatory nitrate reduction to ammonium, ed Zehnder AJB (Wiley, New York), pp 179–244.
↵
1. Craig LS,
2. et al.
(2008) Stream restoration strategies for reducing river nitrogen loads. Front Ecol Environ 6:529–538.
OpenUrl CrossRef
↵
1. Mulholland PJ,
2. et al.
(2008) Stream denitrification across biomes and its response to anthropogenic nitrate loading. Nature 452:202–205.
OpenUrl CrossRef PubMed
↵
1. Schlesinger WH,
2. Reckhow KH,
3. Bernhardt ES
(2006) Global change: The nitrogen cycle and rivers. Water Resour Res 42:W03S06.
OpenUrl CrossRef
↵
1. Laursen AE,
2. Seitzinger SP
(2004) Diurnal patterns of denitrification, oxygen consumption and nitrous oxide production in rivers measured at the whole-reach scale. Freshw Biol 49:1448–1458.
OpenUrl CrossRef
↵
1. Seitzinger S,
2. et al.
(2006) Denitrification across landscapes and waterscapes: A synthesis. Ecol Appl 16:2064–2090.
OpenUrl CrossRef PubMed
↵
1. Holmes RM,
2. Jones JB,
3. Fisher SG,
4. Grimm NB
(1996) Denitrification in a nitrogen-limited stream ecosystem. Biogeochemistry 33:125–146.
OpenUrl
↵
1. Webster JR,
2. et al.
(2003) Factors affecting ammonium uptake in streams—an inter-biome perspective. Freshw Biol 48:1329–1352.
OpenUrl CrossRef
↵
1. Arango CP,
2. Tank JL
(2008) Land use influences the spatiotemporal controls of nitrification and denitrification in headwater streams. J North Am Benthol Soc 27:90–107.
OpenUrl CrossRef
↵
1. Richardson WB,
2. et al.
(2004) Denitrification in the Upper Mississippi River: Rates, controls, and contribution to nitrate flux. Can J Fish Aquat Sci 61:1102–1112.
OpenUrl CrossRef
↵
1. Mosier A,
2. et al.
(1998) Closing the global N₂O budget: Nitrous oxide emissions through the agricultural nitrogen cycle—OECD/IPCC/IEA phase II development of IPCC guidelines for national greenhouse gas inventory methodology. Nutr Cycl Agroecosyst 52:225–248.
OpenUrl
↵
1. Seitzinger SP,
2. Kroeze C,
3. Styles RV
(2000) Global distribution and N₂O emissions from aquatic systems: Natural emissions and anthropogenic effects. Chemosphere Glob Chang Sci 2:267–279.
OpenUrl CrossRef
↵
1. Wollheim WM,
2. et al.
(2008) Global N removal by freshwater aquatic systems using a spatially distributed, within-basin approach. Global Biogeochem Cycles 22:GB2026.
OpenUrl CrossRef
↵
1. Caraco NF,
2. Cole JJ
(1999) Human impact on nitrate export: An analysis using major world rivers. Ambio 28:167–170.
OpenUrl
↵
1. Bernot MJ,
2. Dodds WK
(2005) Nitrogen retention, removal, and saturation in lotic ecosystems. Ecosystems (N Y) 8:442–453.
OpenUrl CrossRef
↵
1. Bernot MJ,
2. et al.
(2010) Inter-regional comparison of land-use effects on stream metabolism. Freshw Biol 55:1874–1890.
OpenUrl CrossRef
↵
1. Mulholland PJ,
2. et al.
(2009) Nitrate removal in stream ecosystems measured by ¹⁵N addition experiments: Denitrification. Limnol Oceanogr 54:666–680.
OpenUrl
↵
1. Hamilton SK,
2. Ostrom NE
(2007) Measurement of the stable isotope ratio of dissolved N₂ in ¹⁵N tracer experiments. Limnol Oceanogr Methods 5:233–240.
OpenUrl
↵
1. Genereux DP,
2. Hemond HF
(1992) Determination of gas-exchange rate constants for a small stream on Walker Branch Watershed, Tennessee. Water Resour Res 28:2365–2374.
OpenUrl CrossRef
↵
1. Wanninkhof R,
2. Mulholland PJ,
3. Elwood JW
(1990) Gas-exchange rates for a 1st-order stream determined with deliberate and natural tracers. Water Resour Res 26:1621–1630.
OpenUrl
↵
1. Sigman DM,
2. et al.
(1997) Natural abundance-level measurement of the nitrogen isotopic composition of oceanic nitrate: An adaptation of the ammonia diffusion method. Mar Chem 57:227–242.
OpenUrl CrossRef
↵
1. Weiss RF,
2. Price BA
(1980) Nitrous oxide solubility in water and seawater. Mar Chem 8:347–359.
OpenUrl CrossRef

Article Alerts

Email Article

Citation Tools

Request Permissions

Proceedings of the National Academy of Sciences: 108 (1)

Table of Contents

Submit

Article Classifications

Biological Sciences
Environmental Sciences

[1] ↵
Robertson GP,
Vitousek PM
(2009) Nitrogen in Agriculture: Balancing the cost of an essential resource. Annu Rev Environ Resour 34:97–125.
OpenUrl CrossRef

[2] Robertson GP,

[3] Vitousek PM

[4] ↵
Solomon S,
et al.
Forster P,
et al.
(2007) in Climate Change 2007: The Physical Science Basis. Contributions of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Changes in atmospheric constituents and in radiative forcing, ed Solomon S, et al. (Cambridge Univ Press, New York).

[5] Solomon S,

[6] et al.

[7] Forster P,

[8] et al.

[9] ↵
Ravishankara AR,
Daniel JS,
Portmann RW
(2009) Nitrous oxide (N₂O): The dominant ozone-depleting substance emitted in the 21st century. Science 326:123–125.
OpenUrl Abstract/FREE Full Text

[10] Ravishankara AR,

[11] Daniel JS,

[12] Portmann RW

[13] ↵
Stehfest E,
Bouwman L
(2006) N₂O and NO emission from agricultural fields and soils under natural vegetation: Summarizing available measurement data and modeling of global annual emissions. Nutr Cycl Agroecosyst 74:207–228.
OpenUrl CrossRef

[14] Stehfest E,

[15] Bouwman L

[16] ↵
Andreae MO,
Schimel DS
Firestone MK,
Davidson EA
(1989) in Exchange of Trace Gases Between Terrestrial Ecosystems and the Atmosphere, Microbiological basis of NO and N₂O production and consumption in soil, eds Andreae MO, Schimel DS (Wiley, Chichester, UK), pp 7–21.

[17] Andreae MO,

[18] Schimel DS

[19] Firestone MK,

[20] Davidson EA

[21] ↵
Codispoti LA
(2010) Oceans. Interesting times for marine N2O. Science 327:1339–1340.
OpenUrl Abstract/FREE Full Text

[22] Codispoti LA

[23] ↵
Cole JJ,
Caraco NF
(2001) Emissions of nitrous oxide (N₂O) from a tidal, freshwater river, the Hudson River, New York. Environ Sci Technol 35:991–996.
OpenUrl PubMed

[24] Cole JJ,

[25] Caraco NF

[26] ↵
Beaulieu JJ,
Arango CP,
Hamilton SK,
Tank JL
(2008) The production and emission of nitrous oxide from headwater streams in the midwestern USA. Glob Chang Biol 14:878–894.
OpenUrl CrossRef

[27] Beaulieu JJ,

[28] Arango CP,

[29] Hamilton SK,

[30] Tank JL

[31] ↵
Seitzinger SP,
Kroeze C
(1998) Global distribution of nitrous oxide production and N inputs in freshwater and coastal marine ecosystems. Global Biogeochem Cycles 12:93–113.
OpenUrl CrossRef

[32] Seitzinger SP,

[33] Kroeze C

[34] ↵
Stevens RJ,
Laughlin RJ,
Malone JP
(1998) Measuring the mole fraction and source of nitrous oxide in the field. Soil Biol Biochem 30:541–543.
OpenUrl CrossRef

[35] Stevens RJ,

[36] Laughlin RJ,

[37] Malone JP

[38] ↵
Stevens RJ,
Laughlin RJ
(1998) Measurement of nitrous oxide and di-nitrogen emissions from agricultural soils. Nutr Cycl Agroecosyst 52:131–139.
OpenUrl CrossRef

[39] Stevens RJ,

[40] Laughlin RJ

[41] ↵
Alexander RB,
Boyer EW,
Smith RA,
Schwarz GE,
Moore RB
(2007) The role of headwater streams in downstream water quality. J Am Water Resour Assoc 43:41–59.
OpenUrl CrossRef

[42] Alexander RB,

[43] Boyer EW,

[44] Smith RA,

[45] Schwarz GE,

[46] Moore RB

[47] ↵
Groffman PM,
et al.
(2006) Methods for measuring denitrification: Diverse approaches to a difficult problem. Ecol Appl 16:2091–2122.
OpenUrl CrossRef

[48] Groffman PM,

[49] et al.

[50] ↵
Burgin A,
Hamilton SK
(2007) Have we overemphasized the role of denitrification in aquatic ecosystems? A review of nitrate removal pathways. Front Ecol Environ 5:89–96.
OpenUrl CrossRef

[51] Burgin A,

[52] Hamilton SK

[53] ↵
Firestone MK,
Firestone RB,
Tiedje JM
(1980) Nitrous oxide from soil denitrification: Factors controlling its biological production. Science 208:749–751.
OpenUrl Abstract/FREE Full Text

[54] Firestone MK,

[55] Firestone RB,

[56] Tiedje JM

[57] ↵
Firestone MK,
Smith MS,
Firestone RB,
Tiedje JM
(1979) The influence of nitrate, nitrite, and oxygen on the composition of the gaseous products of denitrification in soil. Soil Sci Soc Am J 43:1140–1144.
OpenUrl

[58] Firestone MK,

[59] Smith MS,

[60] Firestone RB,

[61] Tiedje JM

[62] ↵
Harper LA,
Mosier AR,
Duxbury JM,
Rolston DE
Hutchinson GL,
Davidson EA
(1993) in Agricultural Ecosystem Effects on Trace Gases and Global Climate Change, Processes for production and consumption of gaseous nitrous oxide in soil, eds Harper LA, Mosier AR, Duxbury JM, Rolston DE (Am Soc Agron, Madison, WI), pp 79–94.

[63] Harper LA,

[64] Mosier AR,

[65] Duxbury JM,

[66] Rolston DE

[67] Hutchinson GL,

[68] Davidson EA

[69] ↵
Swerts M,
Merckx R,
Vlassak K
(1996) Denitrification, N₂-fixation and fermentation during anaerobic incubation of soils amended with glucose and nitrate. Biol Fertil Soils 23:229–235.
OpenUrl CrossRef

[70] Swerts M,

[71] Merckx R,

[72] Vlassak K

[73] ↵
Miller MN,
et al.
(2008) Crop residue influence on denitrification, N2O emissions and denitrifier community abundance in soil. Soil Biol Biochem 40:2553–2562.
OpenUrl CrossRef

[74] Miller MN,

[75] et al.

[76] ↵
Zehnder AJB
Tiedje JM
(1988) in Biology of Anaerobic Microorganisms, Ecology of denitrification and dissimilatory nitrate reduction to ammonium, ed Zehnder AJB (Wiley, New York), pp 179–244.

[77] Zehnder AJB

[78] Tiedje JM

[79] ↵
Craig LS,
et al.
(2008) Stream restoration strategies for reducing river nitrogen loads. Front Ecol Environ 6:529–538.
OpenUrl CrossRef

[80] Craig LS,

[81] et al.

[82] ↵
Mulholland PJ,
et al.
(2008) Stream denitrification across biomes and its response to anthropogenic nitrate loading. Nature 452:202–205.
OpenUrl CrossRef PubMed

[83] Mulholland PJ,

[84] et al.

[85] ↵
Schlesinger WH,
Reckhow KH,
Bernhardt ES
(2006) Global change: The nitrogen cycle and rivers. Water Resour Res 42:W03S06.
OpenUrl CrossRef

[86] Schlesinger WH,

[87] Reckhow KH,

[88] Bernhardt ES

[89] ↵
Laursen AE,
Seitzinger SP
(2004) Diurnal patterns of denitrification, oxygen consumption and nitrous oxide production in rivers measured at the whole-reach scale. Freshw Biol 49:1448–1458.
OpenUrl CrossRef

[90] Laursen AE,

[91] Seitzinger SP

[92] ↵
Seitzinger S,
et al.
(2006) Denitrification across landscapes and waterscapes: A synthesis. Ecol Appl 16:2064–2090.
OpenUrl CrossRef PubMed

[93] Seitzinger S,

[94] et al.

[95] ↵
Holmes RM,
Jones JB,
Fisher SG,
Grimm NB
(1996) Denitrification in a nitrogen-limited stream ecosystem. Biogeochemistry 33:125–146.
OpenUrl

[96] Holmes RM,

[97] Jones JB,

[98] Fisher SG,

[99] Grimm NB

[100] ↵
Webster JR,
et al.
(2003) Factors affecting ammonium uptake in streams—an inter-biome perspective. Freshw Biol 48:1329–1352.
OpenUrl CrossRef

[101] Webster JR,

[102] et al.

[103] ↵
Arango CP,
Tank JL
(2008) Land use influences the spatiotemporal controls of nitrification and denitrification in headwater streams. J North Am Benthol Soc 27:90–107.
OpenUrl CrossRef

[104] Arango CP,

[105] Tank JL

[106] ↵
Richardson WB,
et al.
(2004) Denitrification in the Upper Mississippi River: Rates, controls, and contribution to nitrate flux. Can J Fish Aquat Sci 61:1102–1112.
OpenUrl CrossRef

[107] Richardson WB,

[108] et al.

[109] ↵
Mosier A,
et al.
(1998) Closing the global N₂O budget: Nitrous oxide emissions through the agricultural nitrogen cycle—OECD/IPCC/IEA phase II development of IPCC guidelines for national greenhouse gas inventory methodology. Nutr Cycl Agroecosyst 52:225–248.
OpenUrl

[110] Mosier A,

[111] et al.

[112] ↵
Seitzinger SP,
Kroeze C,
Styles RV
(2000) Global distribution and N₂O emissions from aquatic systems: Natural emissions and anthropogenic effects. Chemosphere Glob Chang Sci 2:267–279.
OpenUrl CrossRef

[113] Seitzinger SP,

[114] Kroeze C,

[115] Styles RV

[116] ↵
Wollheim WM,
et al.
(2008) Global N removal by freshwater aquatic systems using a spatially distributed, within-basin approach. Global Biogeochem Cycles 22:GB2026.
OpenUrl CrossRef

[117] Wollheim WM,

[118] et al.

[119] ↵
Caraco NF,
Cole JJ
(1999) Human impact on nitrate export: An analysis using major world rivers. Ambio 28:167–170.
OpenUrl

[120] Caraco NF,

[121] Cole JJ

[122] ↵
Bernot MJ,
Dodds WK
(2005) Nitrogen retention, removal, and saturation in lotic ecosystems. Ecosystems (N Y) 8:442–453.
OpenUrl CrossRef

[123] Bernot MJ,

[124] Dodds WK

[125] ↵
Bernot MJ,
et al.
(2010) Inter-regional comparison of land-use effects on stream metabolism. Freshw Biol 55:1874–1890.
OpenUrl CrossRef

[126] Bernot MJ,

[127] et al.

[128] ↵
Mulholland PJ,
et al.
(2009) Nitrate removal in stream ecosystems measured by ¹⁵N addition experiments: Denitrification. Limnol Oceanogr 54:666–680.
OpenUrl

[129] Mulholland PJ,

[130] et al.

[131] ↵
Hamilton SK,
Ostrom NE
(2007) Measurement of the stable isotope ratio of dissolved N₂ in ¹⁵N tracer experiments. Limnol Oceanogr Methods 5:233–240.
OpenUrl

[132] Hamilton SK,

[133] Ostrom NE

[134] ↵
Genereux DP,
Hemond HF
(1992) Determination of gas-exchange rate constants for a small stream on Walker Branch Watershed, Tennessee. Water Resour Res 28:2365–2374.
OpenUrl CrossRef

[135] Genereux DP,

[136] Hemond HF

[137] ↵
Wanninkhof R,
Mulholland PJ,
Elwood JW
(1990) Gas-exchange rates for a 1st-order stream determined with deliberate and natural tracers. Water Resour Res 26:1621–1630.
OpenUrl

[138] Wanninkhof R,

[139] Mulholland PJ,

[140] Elwood JW

[141] ↵
Sigman DM,
et al.
(1997) Natural abundance-level measurement of the nitrogen isotopic composition of oceanic nitrate: An adaptation of the ammonia diffusion method. Mar Chem 57:227–242.
OpenUrl CrossRef

[142] Sigman DM,

[143] et al.

[144] ↵
Weiss RF,
Price BA
(1980) Nitrous oxide solubility in water and seawater. Mar Chem 8:347–359.
OpenUrl CrossRef

[145] Weiss RF,

[146] Price BA

Main menu

User menu

Search

New Research In

Nitrous oxide emission from denitrification in stream and river networks

Abstract

Results and Discussion