Year-round N 2 O production by benthic NO x reduction in a monomictic south-alpine lake

Nitrous oxide (N2O) is a potent greenhouse gas, generated through microbial nitrogen (N) turnover processes, such as nitrification, nitrifier denitrification, and denitrification. Previous studies quantifying natural sources have mainly focused on soils and the ocean, but the potential role of terrestrial water bodies in the global N 2O budget has been widely neglected. Furthermore, the biogeochemical controls on the production rates and the microbial pathways that produce benthic N 2O in lakes are essentially unknown. In this study, benthic N 2O fluxes and the contributions of the microbial pathways that produce N 2O were assessed using 15N label flow-through sediment incubations in the eutrophic, monomictic south basin of Lake Lugano in Switzerland. The sediments were a significant source of N 2O throughout the year, with production rates ranging between 140 and 2605 nmol N 2O h−1 m−2, and the highest observed rates coinciding with periods of water column stratification and stably anoxic conditions in the overlying bottom water. Nitrate (NO−3 ) reduction via denitrification was found to be the major N2O production pathway in the sediments under both oxygen-depleted and oxygen-replete conditions in the overlying water, while ammonium oxidation did not contribute significantly to the benthic N 2O flux. A marked portion (up to 15 %) of the total NO−3 consumed by denitrification was reduced only to N 2O, without complete denitrification to N2. These fluxes were highest when the bottom water had stabilized to a low-oxygen state, in contrast with the notion that stable anoxia is particularly conducive to complete denitrification without accumulation of N 2O. This study provides evidence that lake sediments are a significant source of N2O to the overlying water and may produce large N 2O fluxes to the atmosphere during seasonal mixing events.


Introduction
Nitrous oxide (N 2 O) is a potent greenhouse gas with a global warming potential that is ∼ 300 times higher than that of CO 2 over a 100 yr time horizon (Forster et al., 2007).Furthermore, N 2 O is the most important stratospheric ozonedepleting substance currently being emitted to the atmosphere (Ravishankara et al., 2009).The atmospheric concentration has increased from 270 ppb in 1750 (Forster et al., 2007) to 323 ppb in 2011 (AGAGE, 2012), but there are still large uncertainties with regard to the relative contributions of the major sources and sinks of N 2 O (Forster et al., 2007).Microbiological processes in soils and the ocean are the most important natural N 2 O sources (Forster et al., 2007).However, the recent increase in atmospheric N 2 O concentration is largely due to human intervention in the nitrogen (N) cycle, in particular through the agricultural use of synthetic N-based fertilizers (Codispoti et al., 2001;Bouwman et al., 2002;Mosier et al., 1998).In aquatic systems, anthropogenic fixed nitrogen loading can have multiple detrimental environmental effects, such as eutrophication, acidification, and the reduction of biodiversity (Galloway et al., 2003).In these environments, microbial processes taking place in redox transition zones play an important role in removing fixed N. Denitrification, for example, can be an important mechanism for removing fixed N along the land-ocean continuum by reducing it back to N 2 .N 2 O is a free intermediate in this process that may be released to the environment under certain conditions.N 2 O is also produced during other N transformation reactions (Galloway et al., 2003) such as nitrification (specifically, ammonia oxidation) and nitrifier denitrification (Fig. 1).

Published by Copernicus Publications on behalf of the European Geosciences Union.
Distinguishing the relative contributions of each of these major N 2 O production and consumption pathways to the total N 2 O flux is often challenging in aquatic systems because different types of microorganisms perform these pathways under overlapping environmental conditions.Furthermore, certain microbes carry out more than one pathway in response to changes in biogeochemical conditions.N 2 O is produced during the aerobic oxidation of ammonium (NH + 4 ) to nitrite (NO − 2 ) when hydroxylamine (NH 2 OH), an intermediate in the reaction, decomposes (Stein, 2011).Rates of ammonia oxidation depend primarily on substrate (NH + 4 ) and oxygen (O 2 ) availability (Ward, 2008).However, in sediments, aerobic NH + 4 oxidation and NO − 2 oxidation to nitrate (NO − 3 ) can be closely coupled to anaerobic NO − 3 reduction in the redox transition zone (coupled nitrificationdenitrification; Ward, 2008).Ammonia oxidizers also produce N 2 O through a second mechanism known as nitrifier denitrification, an enzymatic pathway that sequentially reduces NO − 2 to nitric oxide (NO), and then N 2 O (Wrage et al., 2001).The importance of nitrifier denitrification as a N 2 O source appears to be higher under low-O 2 conditions (Ritchie and Nicholas, 1972;Poth and Focht, 1985).However, unlike denitrification, O 2 does not seem to inhibit nitrifier denitrification to the same extent or through the same mechanisms (Kool et al., 2011).Denitrification is the reduction of NO − 3 to N 2 via the gaseous intermediates NO and N 2 O (Knowles, 1982) under anoxic or suboxic conditions (i.e., [O 2 ] < 2-5 µmol L −1 ; Devol, 2008;Codispoti et al., 2001).Denitrification can both produce and consume dissolved N 2 O, releasing N 2 O under conditions that suppress the activity of the N 2 O reductase enzyme, for example, under low O 2 concentrations (Firestone et al., 1979;Otte et al., 1996).Furthermore, rapid transitions between oxic and suboxic conditions may cause "stop-and-go" denitrification, which causes N 2 O accumulation in aquatic environments (Naqvi et al., 2000;Codispoti et al., 2001).
In lacustrine sediments, microbial activity consumes O 2 rapidly in the topmost millimeters, leading to suboxic or anoxic conditions in deeper sediment horizons, where denitrification becomes an important redox process (Hunting and van der Geest, 2011).The O 2 penetration depth is closely related to the O 2 concentration in the overlying water and the sediment reactivity (Lehmann et al., 2009;Thibodeau et al., 2010).A decrease in bottom water O 2 concentration is reflected in a narrower oxygenated zone in the sediment (Rasmussen and Jørgensen, 1992).Narrow redox zonation leads thus to an equally narrow succession of microbial processes (Stockdale et al., 2009).Changes in the redox zonation may have profound consequences on N 2 O production (Otte et al., 1996).Seasonal cycles of water column mixing and stagnation can influence the oxidation state of surface sediments and modulate the penetration of redox boundaries into the sediments, potentially changing the redox environments of nitrifiers and denitrifiers (Rasmussen and Jørgensen, 1992).
The few studies that have quantified N 2 O fluxes from freshwater sediments indicate that lake sediments can be a significant source of N 2 O.They also highlight that factors influencing N 2 O production pathways in the benthic environment are still not clearly identified, particularly with regard to the relative importance of nitrification, nitrifier denitrification, and denitrification (Mengis et al., 1996;Liikanen et al., 2003b;Liikanen and Martikainen, 2003;McCrackin and Elser, 2010).
In this study, N 2 O production pathways in lacustrine sediments were studied using ex situ steady state flow-through incubations with intact sediment cores (Lavrentyev et al., 2000;McCarthy et al., 2007;Liikanen et al., 2002aLiikanen et al., -c, 2003a, b;, b;Liikanen and Martikainen, 2003) in combination with substrate 15 N labeling to assess benthic N 2 O production rates and pathways.The experiments were conducted with sediments from a eutrophic, monomictic lake in southern Switzerland, the south basin of Lake Lugano.Monomixis and the resulting intermittent anoxia and suboxia of the bottom waters makes the south basin an ideal study site for testing the effects of variable bottom water oxygenation on the benthic N 2 O production in a lake.Furthermore, previous measurements (Wenk, 2013) indicate high bottom water N 2 O accumulation in the deep hypolimnion during thermal stratification in summer and fall, begging the question as to what causes N 2 O accumulation in near-bottom waters.Upon water column overturn in winter, N 2 O-laden bottom waters may be advected to the surface, enhancing N 2 O fluxes into the atmosphere.
The study's objectives were (1) to estimate N 2 O fluxes from the sediments to the overlying water column and to assess seasonal variations in these fluxes, (2) to identify the dominant benthic N 2 O-producing processes in the lake, and (3) to study the possible impact of variable redox conditions of bottom waters during the seasonal cycle on N 2 O production rates and pathways.

Site description
Lake Lugano is located in southern Switzerland/northern Italy at an altitude of 271 m above sea level (Fig. 2).It is divided by a natural dam into two main basins: the northern and southern basin (Barbieri and Polli, 1992;Lehmann et al., 2004a).Due to the limited water exchange, the basins are characterized by a distinct limnology, so that the northern and southern basins can be regarded as two separate lake systems that are connected by a narrow opening at Melide.A detailed overview of the lake's limnology can be found in Barbieri and Polli (1992) and Barbieri and Simona (2001).
This study focuses on the southern basin.It has an area of 20.3 km 2 , a volume of 1.14 km 3 and a maximum depth of 95 m (Barbieri and Polli, 1992).Lazzaretti and Hanselmann (1992) and Lehmann et al. (2004a, b) described in detail the changes in seasonal redox conditions in the southern basin.During the mixing period (January/February to April), the whole water column becomes oxygenated and oxic conditions are found at the sediment/water interface until late spring.With the onset of thermo-stratification, generally in April, together with the increased phytoplankton production in surface waters and organic matter export to the hypolimnion, oxygen concentrations in the deep hypolimnion decrease, and by June/July, the redox transition zone has migrated from within the sediments into the water column by several meters.Complete anoxia prevails in the bottom waters until the water column turns over again in winter.Previous work on phosphorus (P) accumulation in the sediments of the south basin has shown spatial changes of the mineral content of the sediments within the south basin, and in turn the distribution of adsorbed P, as a result of the heterogeneity of the catchment geology (Veronesi, 1999).As for other sediment characteristics (porosity, organic N, organic C) the different sites within the south basin are comparable (Veronesi, 1999).The sampling site in this study is located west of the village of Figino (45 • 57 N, 8 • 54 E; Fig. 2) close to the deepest spot in the south basin (95 m).With regard to the abovementioned sediment parameters, the chosen site can be considered most representative for the studied lake basin.

Sampling
Six ∼ 50 cm sediment cores with 20 cm overlying bottom water were taken with a 5.7 cm diameter gravity corer in 2010 (April, August, October) and 2011 (January, May).The cores were stored upright and in the dark during transport to the home laboratory on the day of sampling.In addition, 3 × 20 L of bottom water were sampled using 10 L Niskin  Barbieri and Polli, 1992).
to minimize the risk of O 2 contamination, only the lower three quarters of the Niskin bottle contents were used, and in October 2010 and January 2011, the bags were additionally stored underwater.In situ oxygen concentration of the bottom water was measured with a CTD profiler (Idronaut Ocean Seven 316Plus, Idronaut).

Steady state flow-through experiments
Steady state flow-through experiments were set up according to Gardner et al. (1991) andLavretyev et al. (2000) (Fig. 3) within approx.6 h after sampling in a cold room at near-in situ temperature (6.5 • C).The top caps on the liners were removed and replaced with gastight, O-ring sealed PVC plungers containing two holes.The plungers were lowered into the liners until all headspace air was released through the holes.Subsequently, the inlet water reservoir was connected to the core with gastight tubes (FEP, 0.8 mm inner diameter).A second tube connected the core with the sampling vial.A constant flow of ∼ 1 mL min −1 was established with a peristaltic pump.For each sampling campaign, three duplicate flow-through experiments were set up, where two core incubations were supplied with water from one inlet water reservoir, respectively.One of the three inlet water reservoirs was amended with 15 N-NH + 4 (ammonium chloride, > 99 % 15 N atom, Spectra Stable Isotopes), one with 15 N-NO − 3 (potassium nitrate, > 99 % 15 N atom, Spectra Stable Isotopes) and one was left unamended as a control.The labeled substrates were added so that the in situ concentrations were doubled (final 15  et al., 2013), the ammonia oxidation half-saturation constant was less than 0.1 µmol L −1 (i.e., significantly lower than the ambient [NH +  4 ] in this study).After a conditioning period of > 24 h (Gardner and McCarthy, 2009), the in-and outflows were sampled daily.Results are presented as the average of two (August 2010) or three incubation days (October 2010, January 2011, May 2011).In April 2010, samples for N 2 O analysis were only taken on the last day of the experiment.Oxygen concentrations in the inlet water reservoirs and the outflow were measured daily with an optical sensor system (PreSens dipping probe; detection limit 0.5 µM, analytical error at suboxic oxygen levels 0.2 µM).
For N 2 O analyses, glass vials (21 mL) were filled from bottom to top, and allowed to overflow for at least two bottle volumes to minimize N 2 O exchange with air.The vials were capped with aluminium crimp caps with silicone septa (CS-Chromatographie Service GmbH, art.no.300227).Subsequently, a 10 mL He headspace was added in exchange with water.The samples were sterilized with 0.2 mL of 10 mol L −1 NaOH to prevent further microbial activity (Sigman et al., 2001).Sample treatment was done within 1 h after sampling.After NaOH addition, the samples were analyzed within 3 days.

Determination of N 2 O concentrations and benthic fluxes
N 2 O concentrations and stable isotope ratios were determined using an isotope ratio mass spectrometer (IRMS, Thermo Finnigan Deltaplus XP), coupled to an automated purge and trap system (Thermo Finnigan GasBench II).N 2 O concentration standards were produced using the denitrifier method (Sigman et al., 2001) to reduce NO − 3 to N 2 O. Six KNO 3 solutions were produced to yield the following N 2 O standard concentrations: 0.03, 0.1, 0.2, 0.5, 1.0, and 1.5 nmol N 2 O mL −1 .Detector-sensitivity corrected IRMS peak areas of N 2 O standards were compared to the respective concentrations in a regression analysis, and the resulting transfer function (Fig. 4 Fluxes of N 2 O of masses 44, 45 and 46 [nmol h −1 m −2 ] from the sediment to the water column were then calculated from the concentration changes in the in-and out-flowing water:   (Wenk, 2013).

Water column characteristics
The physical and chemical bottom water parameters at the five sampling dates are presented in Table 1.In April 2010, the water column was well mixed so that bottom waters were fully oxygenated, the NH + 4 concentration was low (0.3 µmol L −1 ) and the NO − 3 concentration was comparatively high (83.7 µmol L −1 ).By August 2010, the lake was stratified, a 2 m thick anoxic near-bottom layer had developed, and the NH + 4 concentration in the bottom water had increased (7.1 µmol L −1 ).In October 2010, with ongoing stratification and organic matter decomposition, the anoxic bottom layer expanded by another 6 to 7 m into the water column.NO − 2 (2.4 µmol L −1 ) as well as NH + 4 (44.5 µmol L −1 ) accumulated in the bottom water, but NO − 3 concentrations (50.3 µmol L −1 ) decreased.In January 2011, samples were collected right at the beginning of the winter overturn.The O 2 gradient started to collapse, and suboxic to hypoxic concentrations of O 2 (11 µmol L −1 ) were measured in bottom waters.In May 2011, the water column was fully oxygenated again.

Benthic N 2 O fluxes
Net benthic N 2 O fluxes were calculated for the total N 2 O as well as the unlabeled ( 14 N 14 N 16 O), the singly labeled ( 14 N 15 N 16 O, 15 N 14 N 16 O), and the doubly labeled N 2 O ( 15 N 15 N 16 O) according to Eq. (1) (Fig. 5, Table 2).In all incubations, except January 2011, total N 2 O fluxes were positive, indicating that sediments released N 2 O to the overlying water column.from unamended cores showed a strong variation between duplicate cores.In April and August 2010 the difference between the N 2 O fluxes of the replicate cores was most pronounced (difference ∼ 700 and ∼ 850 nmol N 2 O h −1 m −2 , respectively) reflecting the relatively high heterogeneity of the sediments (Fig. 5).Independent of the variance between duplicate cores, inter-seasonal differences between oxic and anoxic conditions were significant (see below).

15 N-N 2 O recovery from the 15 N labeling experiments
15 N label was detected as 15 N-N 2 O in all 15 N-NO − 3 -labeled sediment core incubations (Fig. 5).Generally, 15   Crackin and Elser, 2010;Liikanen and Martikainen, 2003;Mengis et al., 1996).The N 2 O fluxes reported here ranged between 140 nmol N 2 O h −1 m −2 (April 2010) and 1115 nmol N 2 O h −1 m −2 (August 2010), and were comparable to measurements reported previously for other eutrophic lakes.However, the October 2010 sampling stands out in this regard, with an exceptionally high production rate of 2605 nmol N 2 O h −1 m −2 (Fig. 5c).Mengis et al. (1996) In general, comparison of the total N 2 O fluxes from unamended cores (Fig. 5) to the NH + 4 -or NO − 3 -amended cores shows no significant stimulation of N 2 O production.In January 2011 the N 2 O fluxes in the 15 N-NO − 3 -labeled cores were even lower than in the unamended ones.This suggests that the in situ microbial processes were not substrate-limited so that the rates presented here are real rather than potential rates.The expectation that net N 2 O fluxes vary seasonally, with potential links to water column stratification and oxygenation, was supported by our data.Incubations during fully established anoxic bottom-water conditions (August and October 2010) showed significantly (p = 0.004) higher N 2 O fluxes than during oxic incubations (April 2010 and May 2011).Total N 2 O fluxes seemed to increase during the stagnation period with increasing anoxia from August to October 2010, when maximal values were reached.Such high sedimentary N 2 O production during stratification can lead to N 2 O accumulation and oversaturation in the bottom water, as measured in 2009 (maximum of > 900 nmol L −1 ; Wenk, 2013).At the beginning of the water column overturn in January 2011, together with the breakdown of the O 2 gradient, small amounts of O 2 (11 µmol L −1 O 2 ) were measured in bottom waters, and again, significantly (p = 0.007) lower benthic N 2 O fluxes than during fully anoxic conditions were observed.The fact that more reduced conditions during stratification foster N 2 O production, and that O 2 seems to hinder benthic N 2 O production, implies that denitrification is the dominant N 2 O production pathway (see below).It has been suggested that the rapid injection of small quantities of O 2 into O 2 -deficient environments may be conducive to N 2 O release by denitrifiers (Codispoti et al., 2001).However, our results indicate that total N 2 O production rates were in fact very low during the initial phase of the destratification period (January 2011), the period when pulses of O 2 -laden bottom waters would have been expected to cause this type of perturbation of the denitrifiers' redox environment.Rather, the reduced N 2 O production in January 2011 suggests that the introduction of low concentrations of O 2 may, in fact, inhibit N 2 O release.
We cannot rule out the possibility that other factors besides O 2 also impacted the activity of the microbial community, and thus N 2 O production.Sediment heterogeneity on small spatial scales, and the availability of reducing substrates like organic carbon or sulfide (Wenk et al., 2013a), and inorganic nitrogen compounds are, along with oxygen availability, factors that influence microbial turnover within the nitrogen cycle.Yet, the concentrations of these substrates, as well as the composition of the microbial community itself, often also depend on the seasonally changing redox conditions, so that it is difficult to fully resolve the causalities (direct or indirect) between benthic N 2 O production and redox conditions.

Denitrification is the main sedimentary source of N 2 O
Significant 15 N-N 2 O recovery was only observed in the overlying water of the 15 N-NO − 3 -labeled cores, but not in the 15 N-NH + 4 -labeled cores (Fig. 5), suggesting that denitrification is the predominant N 2 O production pathway in the Lake Lugano south basin sediments.Even in April 2010 and May 2011, when the sediment/water interface was fully oxygenated and nitrification was likely to occur (Ward, 2008), N 2 O derived from 15 N-NH + 4 was not observed.Therefore, it is very likely that canonical anoxic denitrification is the main source of N incorporated into the N 2 O produced during the incubations.Rapid oxygen consumption supported by high rates of organic matter input to the sediments can reduce the oxygen penetration depth into the sediments, producing conditions favorable to denitrification close to the sediment/water interface (Li et al., 2007).It is still not clear, however, whether canonical denitrifiers carry out all of the steps involved in reducing NO − 3 to N 2 O.The process of nitrifier denitrification may also produce N 2 O in the presence of NO − 2 and small amounts of O 2 (as low as 5 µmol L −1 ; Frame and Casciotti, 2010).Although NO − 3 cannot serve as an electron acceptor instead of NO − 2 during nitrifier denitrification, we cannot rule out the possibility that denitrification in the ambient environment supplies NO − 2 to ammonia oxidizers that then convert it to N 2 O during nitrifier denitrification.On the other hand, with active N 2 O production through nitrifier denitrification, at least some 15 N-N 2 O production in the 15 N-NH + 4 -labeled cores would also be expected, but was not observed.
In general, N 2 O production through NH + 4 oxidation was not observed at the oxygenated sediment/water interface in April 2010 and May 2011 (Fig. 5a, e), even though conditions were seemingly conducive to benthic ammonium oxidation.Obviously, there was always N 2 O mass 44 production from ambient substrates, which could in theory originate from the nitrification of ambient pore water NH + 4 .This would imply a dilution of the 15 N-NH + 4 tracer by rapid NH + 4 regeneration from organic N in the sediments, as shown for soil incubations by Norton and Stark (2011).However, isotope pairing predicts that even in the case of high dilution of the labeled ammonium with ambient ammonium from N remineralization within the sediments, a small contribution of NH + 4 oxidation to N 2 O production would be indicated by a measurable efflux of 15 N-N 2 O in the 15 N-NH + 4 -labeled experiments.It is also possible that the added 15 N-NH + 4 was assimilated by the microbial community before it reached the nitrification zone of the sediment, and thus before nitrifiers had a chance to oxidize it.Only in January 2011 did NH + 4 oxidation produce N 2 O, when O 2 was low but not absent, an observation that agrees with previous work showing that N 2 O production during nitrification is enhanced under low O 2 conditions (Stein, 2011;Goreau et al., 1980).Finally, distinguishing between N 2 O produced solely by denitrification and N 2 O produced by nitrification coupled to denitrification is not possible using the 15 N tracer approach adopted here, although, as with nitrification, some 15 N-N 2 O production in the 15 N-NH + 4 -labeled cores would be expected in the case of active coupling of nitrification to denitrification to N 2 O.
Denitrification rates determined in a parallel study were relatively high under the truly anoxic conditions observed in October 2010 (28.2 ± 23.7 µmol N h −1 m −2 ; Wenk et al., 2013b), once redox gradients and, presumably, the microbial community had stabilized.At that time and also under anoxic conditions in August 2010, up to 15 % of total NO − 3 that was reduced was released as N 2 O without being completely reduced to N 2 .The comparatively high net N 2 O fluxes measured during these incubations indicate that N 2 O production was much more efficient than N 2 O consumption.Although methodological limitations prevent us from completely ruling out N 2 O production during ammonia oxidation, we conclude that N 2 O released from the sediment is mainly produced through nitrate reduction, and accumulates in bottom waters particularly when denitrification rates are high.

Conclusions
This study demonstrates that the sediments of the Lake Lugano south basin are a net source of N 2 O to the water column.NO − 3 reduction by denitrification was found to be the primary source of this N 2 O, while N 2 O production during ammonia oxidation was minimal or not observed at all, even when oxic conditions prevailed at the sediment/water interface.Net N 2 O fluxes displayed significant temporal variations, and 1-15 % of the reduced NO − 3 -N accumulated as N 2 O in the water column.Surprisingly, the highest fraction of NO − 3 reduced to N 2 O was observed after water column stratification and the development of anoxia in the deep hypolimnion, conditions that are generally thought to favor complete denitrification to N 2 .
If we extrapolate the average value of total N 2 O fluxes presented in this study (879 nmol N 2 O h −1 m −2 ) in time and over the area of the Lake Lugano south basin (20.3 km 2 ), we can estimate that the sediments of the Lake Lugano south basin release ∼ 6.9 t N 2 O yr −1 to the overlying water.We are aware of the fact that benthic conditions are highly variable in time and space, and that this extrapolation is likely accompanied by a large uncertainty.Capturing all spatio-temporal variability for a more robust quantitative extrapolation would require a sampling regime with higher temporal resolution and additional sampling sites.Nevertheless, our study demonstrates that considerable amounts of N 2 O may be produced in the sediments of the Lake Lugano south basin, with benthic N 2 O fluxes that episodically exceed 2500 nmol N 2 O h −1 m −2 .The fate of N 2 O in the water column remains uncertain.N 2 O fluxes from the surface waters to the atmosphere were not addressed in this study, and the capacity of microbial processes in the water column of the lake basin to remove the N 2 O before it escapes to the atmosphere has yet to be determined.We speculate, however, that significant, year-round benthic N 2 O production, as observed for Lake Lugano, can lead not only to the accumulation of N 2 O in the bottom waters, but ultimately also to its episodic evasion to the atmosphere during periods of destratification and water column overturn.The relevance of such lacustrine N 2 O degassing events for regional or global N 2 O budgets awaits further investigation.

Fig. 2 .
Fig. 2. Location and map of Lake Lugano.The sampling station (red triangle) is located in the south basin, west of the village of Figino, close to the point of maximum depth (modified fromBarbieri and Polli, 1992).

Fig. 4 .
Fig. 4. Calibration curve and transfer function produced using the denitrifier method to measure dissolved N 2 O concentrations.Data from January 2011.
where [N 2 O out/in ] are the measured N 2 O concentrations (nmol L −1 ), Q the average flow rate (L h −1 ) and A the sediment core surface area (m 2 ).Positive flux values indicate a net increase in N 2 O concentration between the in-and outflowing water, and thus fluxes out of the sediments.Statistical analysis was done with the program Prism 6. Non-parametric tests (Kruskal-Wallis test and Mann-Whitney test, confidence level 95 %) were used to test for any statistical difference in N 2 O fluxes between different experimental treatments, and between experiments under oxic vs. anoxic conditions.
N 2 O flux measurements in unamended cores (core 1 and core 2), which represent N 2 O fluxes under in situ NO − 3 and NH + 4 concentration conditions, changed from 831 and 140 nmol N 2 O h −1 m −2 in April 2010, to 1115 and 259 nmol N 2 O h −1 m −2 in August 2010, to 2426 and 2605 nmol N 2 O h −1 m −2 in October 2010.In January and May 2011, average N 2 O fluxes were again relatively low (202 and 195 nmol N 2 O h −1 m −2 in January 2011, and 178 and 189 nmol N 2 O h −1 m −2 in May 2011 for core 1 and core 2, respectively).Overall, total N 2 O fluxes calculated N-N 2 O fluxes in the 15 N-NH + 4 -labeled cores did not exceed natural abundance levels.Only in January 2011 a slightly elevated 15 N-N 2 O flux (16 nmol N 2 O h −1 m −2 ) was measured in one of the duplicate cores.Total N 2 O fluxes in cores with 15 N-NH + 4 and 15 N-NO − 3 additions were not significantly different from N 2 O fluxes of unamended cores in April, August, October 2010 and May 2011 despite the two-fold increase in NO − 3 and the addition of NH + 4 .In January 2011, total N 2 O fluxes in the 15 N-NO − 3 cores were significantly lower than in the unamended and the 15 N-NH + 4 -labeled cores (p = 0.002 and p = 0.015, respectively).

Fig. 5 .
Fig. 5. N 2 O fluxes from the sediment to the overlying water.Each column represents the average value of the N 2 O fluxes over the experiment run time.Error bars show the standard error of the fluctuations over incubation time (April: single measurements; August: duplicate measurements; October-May: triplicate measurements).White: 14 N 14 NO (mass 44).Light grey: 14 N 15 NO and 15 N 14 NO (mass 45).Dark grey: 15 N 15 NO (mass 46).Black: total N 2 O flux.Note the different scale in (c).

Table 1 .
In situ temperature and concentrations of dissolved O 2 , NH + 4 , NO − 3 and NO − 2 in bottom waters.Dissolved inorganic N concentrations were measured in the inlet water reservoirs of the control experiments

Table 2 .
Average benthic N 2 O fluxes of the masses 44 ( 14 N 14 NO), 45 ( 14 N 15 NO and 15 N 14 NO) and 46 ( 15 N 15 NO) and total fluxes measured in the flow-through experiments in April, August and October 2010, and January and May 2011 (SE denotes standard error; no replicate time points for April 2010).