A new mechanistic model of δ18O-N2O formation by denitrification

Available online at www.sciencedirect.com

Geochimica et Cosmochimica Acta 112 (2013) 102–115 www.elsevier.com/locate/gca

A new mechanistic model of d18O-N2O formation by denitrification David M. Snider a,⇑, Jason J. Venkiteswaran a, Sherry L. Schiff a, John Spoelstra a,b b

a University of Waterloo, Department of Earth and Environmental Sciences, Waterloo, Ontario, Canada N2L 3G1 Environment Canada, National Water Research Institute, Canada Centre for Inland Waters, Burlington, Ontario, Canada L7R 4A6

Received 19 October 2012; accepted in revised form 3 March 2013; available online 13 March 2013

Abstract Anaerobic incubations of flooded and non-flooded agricultural soils and stream sediment were conducted with 18Olabelled water to investigate the stable isotope ratios (d15N and d18O) of nitrous oxide (N2O) produced from denitrification. The rates of N2O production and d15N- and d18O-N2O values were measured. The amount of oxygen exchange (O-exchange) with water, and the nitrogen and oxygen isotope effects (e) were calculated. The net 15N isotope effect (NO3 ! N2O) for denitrification in this study varied from 30& to 9&. The net 18O isotope effect ranged between +32& and +60& and was negatively correlated to the total fraction of O-exchange, which varied between 0.40 and 0.94. This manuscript describes a new, comprehensive set of mathematical expressions that can be used to model d18O values of N2O formed by denitrification and calculate the magnitude of O-exchange and the net 18O isotope effect for denitrification. This mathematical approach is compared to another method of approximating O-exchange (Snider et al., 2009), and we show that this older method provides a minimum estimate of O-exchange. Using this mechanistic model, we discuss how N2O consumption, open/closed systems, and variations in the N2O:N2 ratio can influence the observed d18O-N2O. The net 18O isotope effect for denitrification in this study was partially controlled by the fractions of O-exchange and N2O reduction, which were likely influenced by the actively denitrifying microbial community and the soil moisture. We conclude that d18O-N2O values are, in many cases, useful tracers of N2O production because they are often higher than the majority of nitrification-derived d18O-N2O values. The usefulness of d18O values to apportion sources must be determined on a case-bycase basis. Ó 2013 Elsevier Ltd. All rights reserved.

1. INTRODUCTION The global agricultural sector has great potential to lower its climate change impact by reducing emissions of the greenhouse gas nitrous oxide (N2O). Much of the recent rise in atmospheric N2O concentrations is due to an increased usage of synthetic nitrogen (N) fertilizers (Park et al., 2012). Cultivated soils emit the largest amounts of N2O ⇑ Corresponding author. Present address: University of Guelph,

School of Environmental Sciences, Guelph, Ontario, Canada N1G 2W1. Tel.: +1 519 824 4120x58329; fax: +1 519 837 0756. E-mail addresses: [email protected] (D.M. Snider), [email protected] (J.J. Venkiteswaran), sschiff@uwaterloo.ca (S.L. Schiff), [email protected] (J. Spoelstra). 0016-7037/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.gca.2013.03.003

(US-EPA, 2006), and there is great interest in understanding the soil N and carbon (C) processes responsible for GHG production so that effective policies and management practises can be implemented to successfully mitigate emissions. N2O is produced in soils through microbially-mediated redox reactions that occur in impacted and pristine environments under a variety of geochemical settings. In oxygenated environments, when ammonium (NH4+) is oxidized to nitrate (NO3) (nitrification), N2O is an alternate end-product formed by the oxidation of hydroxylamine (NH2OH), or by the reduction of nitrite (NO2) (nitrifierdenitrification). In oxygen-depleted zones, N2O is an obligatory intermediate compound of denitrification. There are many knowledge gaps that still need to be addressed if we

D.M. Snider et al. / Geochimica et Cosmochimica Acta 112 (2013) 102–115

are to successfully implement mitigation strategies. In many environments multiple N2O producing processes may occur simultaneously. Our current inability to unequivocally separate nitrifier-produced N2O from denitrifier-derived N2O is of paramount importance (Chen et al., 2008), because management strategies may affect each N2O-producing process differently. The stable isotope analysis of N and oxygen (O) (d15N and d18O) can help constrain N2O production pathways because the magnitude of isotopic fractionation associated with each pathway may differ. The 15N isotope effect of denitrification (enitrous oxidenitrate = 10& to 45&; Snider et al., 2009) is smaller than the 15N isotope effect of nitrifierdenitrification. The latter has been measured using cultures of single organisms and estimates of enitrous oxideammonium range from 47& to 55& (Sutka et al., 2006; Mandernack et al., 2009). Pe´rez et al. (2006) calculated much larger (negative) values of the 15N isotope effect for nitrifier-denitrification in soils (e = 111 ± 12&), however, such large e values have not yet been corroborated by any other study. The 15Ne for N2O produced by NH2OH oxidation is poorly constrained (+8& to 32&, Sutka et al., 2003, 2004, 2006). These initial estimates would suggest that the d15N-N2O formed from NH2OH oxidation may be indistinguishable from the d15N-N2O formed from denitrification if the endmembers have similar d15N values. Other parameters such as d18O values or isotopomeric nitrogen ratios (d15Na and d15Nb) are tools that might prove useful in separating N2O produced by these two pathways. The use of d18O-N2O to partition N2O sources is complicated by several O endmembers (NO3, H2O, and O2), oxygen exchange (O-exchange) between H2O and N2O precursors (Kool et al., 2009; Snider et al., 2009; Casciotti et al., 2010), and multiple isotope effects (e); many of which are not measured or robustly defined (Fig. 1). Despite this, it appears that d18O values of N2O formed by nitrification (NH2OH oxidation and nitrifier-denitrification) ranges between +13& and +35& regardless of variations in d18OH2O, d18O-O2, isotope effects, or O-exchange (Snider et al., 2012).

103

In contrast, almost all of the published d18O-N2O data reported to be produced by denitrification in soils is higher than +30& and ranges between approximately +35& and +45& (e.g., Toyoda et al., 2005, 2009; Opdyke et al., 2009; Ostrom et al., 2010). Denitrifier-produced N2O in saturated environments (e.g., oceans, rivers, groundwaters) is even more enriched in 18O with d18O values that typically range between +45& and +55& (e.g., Boontanon et al., 2000; Ostrom et al., 2000). Well et al. (2005) have even measured groundwater d18O-N2O with an astonishingly high value of +90&. To date, there is no clear understanding why denitrifier-N2O produced in these differing environments has distinct d18O-values. Although N2O reduction increases both the N and O isotope ratios of N2O, the effects of consumption (alone) do not adequately explain these differences. Earlier experiments by Snider et al. (2009) showed that O-exchange in temperate forest upland and wetland soils was confined to narrow ranges that differed between soil types (90–91% and 65–70%, respectively). The magnitude of O-exchange was independent of temperature, soil moisture, or the rate of net N2O production. With the exception of one experiment, the net 18O isotope effect in these experiments ranged from +37& to +43&. The objectives of this study are to improve our understanding of the systematics of O-exchange and O isotope fractionation in denitrification using incubations of agricultural soil. Additionally, we evaluate a new approach to calculate O-exchange and isotope effects and compare this to previously developed methods. 2. METHODS 2.1. Soil collection, processing, and characterization Agricultural soils and stream sediment were collected in bulk from the Strawberry Creek Watershed (Petrone et al., 2006), located near Maryhill, Ontario, Canada on July 5, 2007. One mineral soil with low organic matter (OM) was collected from an upland site, and another mineral soil with

Fig. 1. Proposed systematics of oxygen isotope enrichment and exchange during denitrification. The net O isotope effect measured between N2O and NO3 (18Oenet) is a function of all the individual isotope effects (e1–e6), a variable fraction of O-exchange that occurs during NO2 and NO reduction, and the relative rates of N2O production and N2O consumption. In this conceptual model of denitrification it is assumed that NO3 and N2O are the only compounds that accumulate in the intra- or extracellular environments. If any other intermediate compounds accumulate and egress from the cell, then additional isotope effects could result if the steps are rate-limiting. If complete denitrification occurs and 100% of the N2O is reduced to N2 (sequentially, within the same organism), then the observed 18O/16O discrimination is zero because no N2O remains. If an organism is ‘leaky’ (as described in the ‘hole-in-the-pipe’ model by Firestone and Davidson, 1989), then a portion of the N2O is reduced to N2 (sequentially, within the same organism) and another portion is ‘leaked’ out of the cell. In this scenario e4 > zero, and the accurate modelling of 18O/16O requires knowledge of the relative rates of N2O production and N2O consumption (the N2O:N2 ratio). Similarly, if extracellular N2O is taken up by N2O consuming organisms, an intermolecular isotope effect occurs (e4) because the rate of N216O reduction to N2 is faster than the reduction of N218O to N2.

104

D.M. Snider et al. / Geochimica et Cosmochimica Acta 112 (2013) 102–115

high OM was collected from a wetter site. At both locations, late vegetative stage corn was being grown at the time of collection. Crop residue and stones were removed from the surface and the soils were collected from the ground surface down to the top of a dense, till layer (25 cm deep). No distinguishable soil-horizons were present in this shallow soil zone because cropping methods employed mechanical tillage. In addition to the mineral soils, sediment was collected from a nearby creek, which was not flowing (the low summer water table was below the sediment surface). Riparian grasses and an upper layer of medium-textured sand (15 cm thick) were removed, and the underlying sediment was collected in bulk. The sediment layer was wet, but not saturated. All soils were air-dried, homogenized, sieved to 2 mm, and stored in resealable freezer bags in darkness at room temperature. Total carbon (TC), total nitrogen (TN), d13C-TC, and d15N-TN were analyzed on an elemental analyzer coupled to an isotope ratio mass spectrometer (IRMS). Carbon:nitrogen ratios are reported on a molar basis for a non-acidified sample. Organic matter content (%) was quantified by loss on ignition (LOI). Soils were extracted with 2 M potassium chloride (KCl) and Nanopure deionized water (DI) (Barnstead International, Dubuque, IA) for analysis of extractable NH4+ and NO3 concentrations, respectively. Soil pH was determined in calcium chloride (CaCl2) and water following ASTM methods (ASTM International, 2007). A particle size analysis (%sand, %silt, %clay) was determined for each soil after oxidizing the organic matter with hydrogen peroxide. 2.2. Incubation protocols Five separate anaerobic incubation experiments were conducted in this study (experiments 1–5), with 9 replicate incubations per experiment. In each experiment 22.5 grams of oven-dry soil [dry weight (dw.)] was wet up to either 100% water holding capacity (WHC; field capacity) or flooded to 200% WHC, and incubated in the dark at laboratory temperature (20–22 °C). The non-flooded soils were suspended in the incubation jars using sacs made of nylon screening to facilitate gas diffusion (Snider et al., 2009). The well-drained, low OM soil (experiment 1) was only incubated at 100% WHC because this site does not experience flooding in situ. Three different levels of d18O-H2O were used in each experiment, which corresponded to low, medium, and high 18 O/16O enrichments (Table 1). The incubation waters used in this study were prepared following methods outlined in Snider et al. (2009). Any water remaining in the air-dried soils was extracted by azeotropic distillation, and the d18O-H2O was analyzed. The final d18O values of the soil water (after soils were moistened to their target moisture content) was calculated and the results are shown in Table 1. All soils were fertilized with potassium nitrate (KNO3) (1.47 mg N/g-soildw., d15N = +13.8 ± 0.3& and d18O = +28.0 ± 0.8&) after an anaerobic pre-incubation period that lasted 5 days. Incubations were sampled for

N2O concentration, d15N-N2O, and d18O-N2O five times (non-flooded) or four times (flooded) over the 8 h experimental period. Anaerobiosis was maintained throughout the study by flushing incubation jars with He for 25 min at 600 mL/min. Flushing occurred once a day during the pre-incubation period, after fertilization, and after each sampling event. The flooded soils (experiments 3 and 5) were slowly swirled on a reciprocal shaker throughout the pre-incubation and experimental periods to minimize diffusion effects and increase the rate of equilibration between the headspace and the flooded soil. To assess the amount of abiotic N2O production, controls were performed with sterilized soils. Three replicate jars of each soil were sterilized, fertilized, then flushed with He, and the concentration of N2O that accumulated over 5 days was monitored. A detailed description of N2O sample storage, and concentration and isotopic analysis is provided in Snider et al. (2009). Nitrogen and oxygen isotope ratios (R) of N2O are expressed in delta (d) notation (Rsample  Rstandard  1) and reported in units of per mill (&), where the international standards for N and O are atmospheric N2 (AIR) and Vienna Standard Mean Ocean Water (VSMOW), respectively. Instantaneous nitrogen isotope effects (e) are calculated as (a  1) and reported in & units, where a = 15N/14N product  15N/14N substrate. 2.3. Determination of oxygen isotope exchange and fractionation (Method I) The 18O/16O ratios of N2O and H2O were normalized by expressing these values in delta (d) notation relative to the 18 O/16O ratio of the NO3 endmember, where the 18 O/16O of the NO3 fertilizer was 0.0020613 (Eqs. (1a) and (1b); Snider et al., 2009). As a result, the linear regression of d18 O-N2 O ðrel: d18 O-NO3  Þ versus d18 O-H2 O ðrel: d18 O-NO3  Þ has a slope (m) that approximates the mean fraction of O-exchange that occurred between H2O and N2O precursors, and an intercept (b) that approximates the net 18O isotope effect (18Oenet) for denitrification (NO3 ! N2O). The results of this regression analysis are shown in Table 1. d18 O-N2 O ðrel: d18 O-NO3  Þ ¼ ðRN2 O  RNO3   1Þ  1000 18

18



d O-H2 O ðrel: d O-NO3 Þ ¼ ðRH2 O  RNO3   1Þ  1000

ð1aÞ ð1bÞ

2.4. Determination of oxygen isotope exchange and fractionation (Method II) In this study a new approach to calculate O-exchange and fractionation during denitrification is evaluated (Method II). This approach is similar to that used by Snider et al. (2010, 2012), and defines O-exchange and the net 18O isotope effect as mathematical expressions (Eqs. (2)–(9)) that describe the step-wise reduction of NO3 to N2 (Fig. 1). The following equations are different than those developed for the ‘bacterial denitrifier method’ (Casciotti et al., 2002). Most of the parameters in the following equations are unknown or not robustly defined in the literature. In this

D.M. Snider et al. / Geochimica et Cosmochimica Acta 112 (2013) 102–115

105

Table 1 The geochemical characteristics of the agricultural soils incubated in this study, the mean N2O production rate, d15N-N2O, 15N isotope effect, d18O-H2O, d18O-N2O, and the corresponding results of simple linear regression (Model I) of d18O-N2O versus d18O-H2O for experiments 1–5. Experiment

1

2

Description

Well-drained, low OM soil at field capacity Silt loam (43, 41, 16)

Poorly-drained, high OM soil at field capacity and flooded

Stream sediment at field capacity and flooded

Loam (19, 64, 17)

Fine sandy loam (55, 33, 12)

100 0.52

100 0.79

200 1.58

100 0.69

200 1.38

1.9 21.8 0.2 6.7 9.7 4.2 0.2 17.2

8.7 23.0 0.6 6.1 12.7 14.9 1.1 4.8

8.7 23.0 0.6 6.1 12.7 14.9 1.1 4.8

6.4 13.2 0.3 6.7 17.0 7.1 1.2 1.2

6.4 13.2 0.3 6.7 17.0 7.1 1.2 1.2

7.3, 7.0 10.0 (1.8)

7.5, 7.2 8.9 (2.5)

7.5, 7.2 6.1 (5.0)

7.5, 7.3 29.0 (7.7)

7.5, 7.3 10.8 (8.1)

6.9 (1.9)

2.9 (1.1)

4.5 (2.7)

16.1 (1.0)

10.4 (3.5)

20.5 (1.9)

16.4 (1.0)

9.2 (2.7)

29.5 (1.0)

23.9 (3.4)

Texture (%sand, %silt, %clay) Water holding capacity (%)a Gravimetric soil water content (g H2O/g-soildw.) TC (%) d13C-TC (& rel. to VPBD) TN (%) d15N-TN (& rel. to AIR) C:N ratio (mol:mol) LOI (%) Extractable NH4+ (lg N/g-soildw.) Extractable NO3 (lg N/gsoildw.)b Soil pH (H2O, 0.1 M CaCl2) Meanc production rate (nmol N2O/h/g-soildw.) of last three samplings (±1r) Meanc d15N-N2Orel. AIR (&) of last three samplings (±1r) Meanc 15N isotope effect (e) (&) of last three samplings (±1r) d18O-H2Orel. VSMOW (&)d Meane d18O-N2Orel. VSMOW (&) of last three samplings (±1r)

3

4

5

–8.2 45.6 105.9 –9.5 45.6 106.5 –10.1 48.1 113.3 –9.5 47.6 111.5 –10.1 49.3 115.9 25.5 80.7 133.5 53.4 87.5 128.0 54.2 93.5 134.5 52.6 80.2 113.5 73.7 99.2 124.2 (1.0) (2.1) (1.3) (1.8) (1.0) (1.7) (8.5) (0.7) (4.4) (2.0) (2.0) (2.0) (2.2) (2.1) (3.1)

Linear regression analysis of d18O-N2O (rel. VSMOW) versus d18O-H2O (rel. VSMOW) of last three samplings Slope (m) (±SE) 0.94 (0.01) 0.64 (0.01) 0.65 (0.02) 0.50 (0.01) Intercept (b) (±SE) 32.8 (0.7) 59.0 (0.4) 61.3 (1.5) 57.0 (0.6) Coefficient of determination (r2) 0.9968 0.9973 0.9748 0.9938 Linear regression analysis of d18O-N2O (rel. d18O-NO3) versus d18O-H2O (rel. d18O-NO3) of last three samplings Slope (m) (± SE) [total fraction of 0.94 (0.01) 0.64 (0.01) 0.65 (0.02) 0.50 (0.01) oxygen exchange]f Intercept (b) (±SE) [net 18O 32.4 (0.5) 47.7 (0.3) 50.1 (1.1) 41.9 (0.4) isotope effect]f Coefficient of determination (r2) 0.9968 0.9973 0.9748 0.9938

0.40 (0.01) 78.3 (0.7) 0.9857 0.40 (0.01) 59.9 (0.5) 0.9857

a

Soil moisture at field capacity (100% WHC) or flooded (200% WHC). b Although antecedent NO3 was present in the field-collected soils, it was removed prior to fertilization (by soil denitrifiers during the preincubation period). c n = 27 measurements. d Soils were incubated with three different waters: low, medium, and high d18O-H2O. e n = 9 measurements. f Determined using Method I, which is described in Sections 3.4 and 3.5.

study, the only measured parameters are d18O-NO3, d18ON2O, and d18O-H2O. d18 O-NO2  ¼ d18 O-NO3  þ e1 18

ð2Þ

18



d O-NO ¼ ð1  fexch:-nir Þ  ðd O-NO2 þ e2 Þ þ fexch:-nir  ðd18 O-H2 O þ e5 Þ 18

ð3Þ

18

d O-N2 O ¼ ð1  fexch:-nor Þ  ðd O-NO þ e3 Þ þ fexch:-nor  ðd18 O-H2 O þ e6 Þ

ð4Þ

where, isotope effects (e) are defined as productsubstrate, e = (a  1)  1000, and a = 18O/16O of product  18O/16O

of substrate; and, e1 is a net O isotope effect of two processes: (i) the inter-molecular 18O/16O fractionation associated with NO3 reduction to NO2 in the inner bacterial membrane (Granger et al., 2008), where the rate of N16O16O16O reduction occurs faster than the rate of N16O16O18O reduction (e.g., Bo¨ttcher et al., 1990, 8&; Mengis et al., 1999, 18&) [negative e]; and, (ii) the intra-molecular 18O/16O fractionation associated with NO3 reduction, where the rate of   N  16O bond breakage occurs faster than the rate of   N  18O bond breakage. This yields 18O-depleted H2O and 18O-enriched NO2 [positive e]; and, e2 is the intra-molecular O isotope effect associated

106

D.M. Snider et al. / Geochimica et Cosmochimica Acta 112 (2013) 102–115

with NO2 reduction (similar to that described above for NO3 reduction) [positive e]; and, e3 is the inter-molecular O isotope effect resulting from the reduction of 2NO ! N2O, where the rate of N  16O bond breakage occurs faster than the rate of N  18O bond breakage. This yields 18O-depleted H2O and 18O-enriched N2O [positive e]; and, f exch.-nir and f exch.-nor are the fractions (f ) of O-exchange that occur at nitrite reductase (nir) and nitric oxide reductase (nor), respectively; and, e5 and e6 are the hypothesized O isotope effects associated with enzymatic O-exchange at nir and nor, respectively. This inter-molecular fractionation may occur in a manner that is similar to the 18O/16O fractionation that occurs during the H2O uptake reactions of nitrification [described by Casciotti et al. (2010), Buchwald and Casciotti (2010); and Buchwald et al. (2012)]. If the rate of H2  16O bond breakage occurs faster than the rate of H2  18O bond breakage, then 16O is more readily incorporated into NO2 and NO during O-exchange [negative e]. When there is no N2O consumption the d18O-N2O can be modelled by Eq. (5), which is derived by substituting Eqs. (2) and (3) into Eq. (4):

For illustration purposes, it is helpful to rewrite Eq. (7) by substituting 0% and 100% O-exchange. When f exch.-nir = 0 and f exch.-nor = 0 (i.e., no O-exchange), then Eq. (7) reduces to Eq. (8): d18 O-N2 Ored: ¼ ðd18 O-NO3  þ e1 þ e2 þ e3 þ 1000Þ  ð1  fN2 O red: Þe4 =1000  1000

ð8Þ

When f exch.-nir = 1 and f exch.-nor = 1 (i.e., complete O-exchange), then Eq. (7) reduces to Eq. (9): d18 O-N2 Ored: ¼ hð1  fN2 O red: Þe4=1000 i  d18 O-H2 O þ hðe6 þ 1000Þ  ð1  fN2 O red: Þe4 =1000  1000i ð9Þ Algebraic rearrangements of Eqs. (2)–(9) were completed with MuPADÒ from the Symbolic Math Toolbox for MATLABÒ. To interpret the data measured in this study and obtain estimates of O-exchange and the net 18O isotope effect, the d18O-N2O values are regressed against the d18O-H2O values of the soil water. The linear regression results are shown in Table 1, and are discussed using Methods I and II in Sections 3.4–3.5 and 3.8–3.9, respectively.

d18 O-N2 O ¼ hfexch:-nir þ fexch:-nor  ðfexch:-nir  fexch:-nor Þi 3. RESULTS AND DISCUSSION

 d18 O-H2 O þ h½e6  fexch:-nor  þ ½1  fexch:-nor   ½e3 þ ðe5  fexch:-nir Þ þ ð1  fexch:-nir Þ  ðe1 þ e2 þ d18 O-NO3  Þi ð5Þ Note that Eq. (5) is presented here in the linear form (y = mx + b), where y = d18O-N2O; x = d18O-H2O; and m and b are the slope and the intercept of the linear regression of d18O-N2O versus d18O-H2O, respectively. When N2O reduction does occur, the d18O-N2O after partial reduction to N2 (d18O-N2Ored.) can be approximated by a Rayleigh distillation (Eq. (6), Mariotti et al., 1981) shown here using delta and epsilon units rather than ratios and alphas. d18 O-N2 Ored: =1000 þ 1 ¼ ðd18 O-N2 O=1000 þ 1Þ  ð1  fN2 O

red: Þ

ðe4 =1000Þ

ð6Þ

where, d18O-N2O is the initial value prior to reduction (Eq. (5)), and d18O-N2Ored. is the final value after some fraction of N2O has been consumed (f N2O red.); and, (1  f N2O red.) is the fraction of N2O that remains; and, e4 is the inter-molecular O isotope effect associated with N2O reduction, where the rate of N  16O bond breakage occurs faster than the rate of N  18O bond breakage. This yields 18O-depleted H2O and 18O-enriched N2O [positive e]. The substitution of Eq. (5) into Eq. (6) and the rearrangement into a linear form yields Eq. (7): d18 O-N2 Ored: ¼ h½fexch:-nir þ fexch:-nor  ðfexch:-nir  fexch:-nor Þ  ½1  fN2 O red: e4 =1000 i  d18 O-H2 O þ hh½e6  fexch:-nor  þ ½1  fexch:-nor   ½e3 þ ðe5  fexch:-nir Þ þ ð1  fexch:-nir Þ  ðe1 þ e2 þ d18 O-NO3  Þ þ 1000i  h1  fN2 O red: ie4 =1000  1000i

ð7Þ

3.1. Soil parameters The mineral soils incubated in this study were low in carbon (1.9–8.7%) and nitrogen (0.2–0.6%), and the soil pH was typical (7.0–7.5) for soils of the region, which are heavily influenced by the carbonate-rich bedrock (Table 1). Soil collected from the upland site contained a very low OM content (4.2%). The wetter soil and the stream sediment contained higher amounts of OM (14.9% and 7.1%, respectively; Table 1). The total amount of inorganic N extracted from all soils was relatively low (2.4–17.4 lg N/g-soildw.). 3.2. Net N2O production rates The 5-day accumulation of N2O in the sterilized control incubations was small and headspace concentrations were lower than the detection limit of the gas chromatograph (100 ppb N2O). This indicates that abiotic N2O production was a negligible source of N2O in these experiments. The rates of N2O produced in all experiments varied from almost zero (poorly drained soil, flooded) to 35 nmol N2O/h/g-soildw. (stream sediment, field capacity) (Table 1 and Fig. 2). This range is very similar to the production rates observed in incubations of temperate forest soils (1–39 nmol N2O/h/g-soildw.; Snider et al., 2009). In this study, however, divergent production rates could not be explained by variations in temperature because all experiments were conducted at room temperature. Instead, the variable N2O production rates observed in this study can be explained by differences in soil type (and its geochemical properties such as organic carbon), soil moisture, and the fraction of N2O reduction (f N2O red.). The mean rates of N2O production in the two soils incubated at field capacity (9–10 nmol N2O/h/g-soildw.) were not statistically different (2-tailed t-test, unequal variance,

D.M. Snider et al. / Geochimica et Cosmochimica Acta 112 (2013) 102–115

p = 0.052). This suggests that both soils have a similar potential to emit N2O in situ. In contrast, the stream sediment had a much higher N2O production rate at field capacity (29 nmol N2O/h/g-soildw.), which was statistically different from the cultivated soils (2-tailed t-test, unequal variance, p < 1012). The stream sediment is an aquatic soil that is flooded for most of the year by nutrient-rich agricultural drainage waters. As such, characteristics of the microbial community (such as species composition and density of denitrifiers) in this soil could be very different than those of the terrestrial soils. The flooded incubations had lower net N2O production rates than their non-flooded counterparts (Table 1). In this study net N2O production is balanced by gross N2O production and N2O consumption. Lower rates of net N2O production in the flooded soils suggest that either: (i) the overall denitrification rate was lower in the flooded experiments; (ii) the N2O:N2 end-product ratio was different; or, (iii) the rate of consumption of extracellular N2O was elevated in the flooded soils (Fig. 1). Given the initial rate of NO3 fertilization (105 lmol/gsoildw.), the denitrifying community was not N-limited. Simulating an extreme scenario, if there was a constant N2O:N2 ratio of 0.01, and a net N2O production rate of 25 nmol N2O/h/g-soildw., only 20% of the supplied NO3 would be consumed after 8 h of incubation. Additionally, if the NO3 pool was limiting, the d15N-N2O and d18ON2O values would increase over the course of the incubations. This was not observed to any appreciable degree in any of the experiments (Fig. 2). 3.3. d15N-N2O and

15

N isotope effects (e)

The mean d15N-N2O produced from all experiments ranged from 16.1& (stream sediment, field capacity) to +4.5& (poorly-drained soil, flooded; Table 1). The d15NN2O of the last three samplings in all experiments remained relatively constant (quasi steady-state), and these data were used to calculate isotope effects. The instantaneous isotope effects (15Nenitrous oxidenitrate) varied between 29.5& and 9.2&, and are similar to literature estimates of soil denitrifiers (e.g., Mariotti et al., 1981, 1982; Menyailo and Hungate, 2006; Pe´rez et al., 2006). The 15N isotope effects reported here for agricultural soils are slightly smaller than those measured in the temperate forest soils studied by Snider et al. (2009), which ranged between 35& and 19&). There is a negative exponential relationship between the d15N-N2O (and 15Ne) and the rate of N2O production in this study (Fig. 3) that was not observed in Snider et al. (2009). This relationship is mainly driven by the data from experiments 3 and 4, and the standard deviations of the mean N2O production rates are outside the 95% confidence intervals of the regression, which suggests the relationship is weak (Fig. 3). Further, a negative correlation is not expected. If there were any relationship at all, a positive correlation might exist because as the rate of N2O production increases the 15Ne might decrease due to limited N supply. It is important to remember that N2O is an obligatory intermediate of denitrification. Any changes in net N2O production may be due, in part, to differing rates of N2O

107

reduction. This N2O-consuming process also affects the d15N-N2O by enriching the residual pool of N2O in 15N. Therefore, changing rates of N2O reduction could provide a partial explanation for the negative correlation shown in Fig. 3. In fact, the datum with the highest 15Ne value and the lowest N2O production rate was measured in the flooded, poorly-drained soil (experiment 3) which may have experienced significant N2O reduction that changed with time. On the other hand, the flooded stream sediment had one of the lowest values of 15Ne, and significant N2O reduction did occur in this experiment (see Section 3.6). Therefore the observed relationship in Fig. 3 may be fortuitous, and the high 15Ne datum from experiment 3 and the low 15 Ne datum from experiment 4 may be caused by non-steady-state rates of N2O production (Fig. 2). 3.4. d18O-N2O and oxygen isotope exchange (Method I) Linear regression analysis of d18O-N2O versus d18OH2O indicated that d18O-N2O values were highly influenced by the d18O values of the soil water (r2 = 0.97–1.00; Table 1). Determinations of O-exchange and 18Oenet were made using Method I following the procedures outlined in Section 2.3 and in Snider et al. (2009). The total fraction of O-exchange (f exch.-nir + f exch.-nor) between H2O and N2O precursors varied widely across all experiments. Very high amounts of O-exchange occurred in well-drained soil at field capacity (94%). Moderate and consistent amounts of O-exchange occurred in the poorly-drained soil (64%–field capacity; 65%–flooded), despite the differences in soil moisture and N2O production rate between these experiments. Finally, lower amounts of O-exchange occurred in the stream sediment (50%-field capacity and 40%-flooded). There is an interesting trend in the magnitude of O-exchange during denitrification in these soils and in the forested soils incubated in Snider et al. (2009). Very high amounts of O-exchange occurred in all the well-drained soils (87–94%) regardless of soil type (organic versus mineral) or soil history (forested versus continuous agriculture). Although these soils are very different, neither of them experiences in situ saturated conditions for any considerable length of time. In contrast, the poorly-drained agricultural soil of this study and the forested wetland soil of Snider et al. (2009) do flood during snowmelt and high precipitation events throughout the year. These wetter soils exhibited lower amounts of O-exchange (64–70%). Finally, the agricultural stream sediment, which is flooded for most of the year exhibited the lowest amounts of O-exchange (non-flooded, 50%; flooded, 40%). This trend suggests that O-exchange during soil denitrification is controlled by the microbial composition of the denitrifier community, which in turn, is controlled by the soil’s landscape position and its antecedent hydrological conditions. This is a hypothesis that needs to be tested and verified with future research. 3.5. d18O-N2O and oxygen isotope fractionation (Method I) The net 18O/16O fractionation that occurs during denitrification (18Oenet) is a cumulative isotope effect composed of

108

D.M. Snider et al. / Geochimica et Cosmochimica Acta 112 (2013) 102–115

Fig. 2. (i–v) N2O production, d15N-N2O (rel. AIR), and d18O-N2O (rel. VSMOW) versus time for flooded and non-flooded anaerobic incubations of agricultural soil and stream sediment. Each N2O production rate (bar plot) and d15N-N2O symbol (triangle) represents a mean value with n = 9. Each d18O-N2O symbol represents a mean value with n = 3, and corresponds to soils incubated with water at low (white circle), medium (grey circle), and high (black circle) 18O-enrichment. Error bars represent 1 sigma from the mean, and, where they are absent, the error is smaller than the size of the symbol. All scales are identical on every plot. The timing of all isotope data has been artificially shifted on these figures to reduce symbol overlap and improve visibility.

D.M. Snider et al. / Geochimica et Cosmochimica Acta 112 (2013) 102–115

65 60

5

3

-10

net O isotope effect (N2O-NO3 ) (‰)

y = -27.9x (± 7.2) + 9.2 (± 7.7); 2 r = 0.83 (for circles)

-

55

15

-

N isotope effect (N2O-NO3 ) (‰)

-5

109

3

-15

50 2

2

45 B

1 C 5

-25

C

F

40 35

1 30

B

A

-30

DE

18

-20

D4 E

25

4

20 A

-35 0.0

0.5

1.0

1.5

2.0

2.5

Log [Production Rate (nmol N2O/hr/g soil-dw.)]

0.0

15 0.5

1.0

1.5

2.0

2.5

Log [Production Rate (nmol N2O/hr/g soil-dw.)]

Fig. 3. The relationship between the 15N isotope effect (left side) and the 18O isotope effect (right side) and the rate of net N2O production by denitrification. Data collected in this study (circles) are labelled by their experiment number, and data published in Snider et al. (2009) (squares) are labelled by experiment letter. There is a significant, negative exponential relationship (slope is not equal to zero) between the 15 Ne and the production rate in the agricultural soils (grey regression line, 95% confidence interval, and equation with standard error shown); yet the relation is weak and largely controlled by experiments 3 (flooded poorly-drained soil) and 4 (stream sediment at field capacity). There are no other apparent relationships between the isotope effects and the N2O production rate. The production rate data was log10 transformed to facilitate linear regression.

NO2  þ 2Hþ þ e NO þ H2 O

ð10Þ

If O-exchange is a fractionating process, the isotope effect might be similar to the 18O/16O fractionations that occur during O2 and H2O uptake in nitrification. These have recently been determined using marine nitrifiers by Casciotti et al. (2010) and Buchwald and Casciotti (2010). As an example, the isotope effect for H2O uptake during the final oxidation of NO2 to NO3 ranges from 27& to 1& (Buchwald and Casciotti, 2010; Buchwald et al.,

65 60

5

55 3 2

50 45 4

40

B

D C

E

35 1

18

net O isotope effect (‰)

several fractionating steps that are described in Section 2.4 (e1e6; Fig. 1). In this study the values of 18Oenet varied from +32& to +60& (Table 1). In Snider et al. (2009) the 18Oenet was constrained to +40 ± 3&, with the exception of one experiment that had a lower value equal to +17&. In both studies there was no correlation between 18Oenet and the N2O production rate (Fig. 3). There was, however, a negative correlation between 18Oenet and O-exchange (Fig. 4). This relationship is expected because when the fraction of O-exchange decreases, a larger proportion of the early isotope effects (e.g., e1 + e2, Fig. 1) are expressed in the d18O-N2O (Eq. (8)). When the amount of O-exchange increases, the early isotope effects are diminished by the incorporation of d18O-H2O (Eq. (9)). It is unknown if O-exchange is a fractionating process and this cannot be independently determined with the data generated from this study. Although the exact mechanism of O-exchange is not well-defined, it is known to occur during the NO2 and NO reduction steps (Ye et al., 1991), and the exchange may be a result of reversible hydrolysis reactions (Kool et al., 2007). Presumably, as the rate of the backward reaction increases (e.g., Eq. (10)), so does the amount of O-exchange.

30 2

agricultural (expt. 4 excluded): y = -51.7x + 81.5; r = 0.98

25

2

agricultural (expt. 4 included): y = -55.5x + 81.1; r = 0.67 2

temperate forest (A excluded): y = -18.2x + 54.5; r = 0.98

20

A

15 0.4

0.5 0.6 0.7 0.8 0.9 O-exchange (ƒexch.-nir + ƒexch.-nor)

1.0

Fig. 4. A regression of the net 18O isotope effect versus Oexchange. Two lines are shown through the agricultural soils data (black circles with corresponding experiment numbers) representing regression with and without experiment 4 as an outlier. One regression line is shown through the forested soils data (Snider et al., 2009; white squares with corresponding experiment letters) with experiment A excluded as an outlier. Model II linear regression (ranged major axis) is used because both 18Oenet and O-exchange are subject to natural variation.

2012). If e5 and e6 were large and negative values, then 18Oenet for denitrification would decrease with increasing amounts of O-exchange. This provides a further explanation for the negative correlation that was observed in this study between 18Oenet and O-exchange (Fig. 4).

D.M. Snider et al. / Geochimica et Cosmochimica Acta 112 (2013) 102–115 65

The consumption of N2O to N2 causes the remaining portion of the N2O pool to become enriched in 18O and 15 N (Menyailo and Hungate, 2006; Ostrom et al., 2007; Vieten et al., 2007; Jinuntuya-Nortman et al., 2008; Well and Flessa, 2009). The ratio of the isotope effects for N2O consumption (18O:15N) ranges between 2.3 and 3.0 and N2O isotope samples that have undergone consumption may plot along a trajectory with a slope close to 2.5 (d18ON2O:d15N-N2O). Demonstrations of this distinctive relationship in soil-N2O studies are provided by JinuntuyaNortman et al. (2008) and Well and Flessa (2009). Most of the experiments in this study had d15N-N2O and 18 d O-N2O values that did not change appreciably with time (Fig. 2). These incubations were designed to reach a quasi steady-state with continuous net N2O production (i.e., open system). As such, cross-plots of d18O-N2O versus d15N-N2O for individual experiments cannot be used in this study to look for evidence of N2O consumption. N2O reduction likely occurred in the experiments of this study; especially in the flooded soils. Saturated environments present ideal conditions for N2O reduction to N2 because the rate of N2O escape to the atmosphere is greatly reduced. For illustrative purposes only, if we were to assume that e1–e3 and e5–e6 were identical in the non-flooded and flooded experiments of each soil, then any differences in net isotope effects ought to be a result of differences in N2O consumption (e4). The d18O-N2O and d15N-N2O values in the flooded stream sediment were higher than the values measured in the non-flooded sediment, and a line drawn through the 18Oe:15Ne ratios of these two experiments had a slope of 3.2 (Fig. 5). However, the variability of the 15Ne estimates was large and the slope varied from 1.8 up to 15. Given the mean slope was close to 2.5, and the net rate of N2O production in the flooded sediment was 1=3 of the rate measured in the non-flooded sediment (Table 1), it is probable that N2O reduction occurred to a greater extent in the flooded experiment than in the non-flooded experiment. This distinctive trajectory was not observed in the N2O isotope ratios of the flooded and non-flooded poorlydrained soil incubations (experiments 2 and 3; Fig. 5). Although the 15Ne of the flooded soil was 7.2& higher than the non-flooded soil, there was very little difference (2&) in 18Oenet between the two experiments. Additionally, the net rates of N2O production in these incubations were not significantly different at a = 0.01 (2-tailed t-test, unequal variance, p < 0.014). This suggests that regardless of the soil moisture differences, both experiments experienced similar amounts of N2O reduction.

60

3.7. Predicting N2O isotope ratios in open and closed systems The net production of N2O is the difference between gross production (NO3 ! N2O) and gross consumption (N2O ! N2). When N2O reduction is absent (the gross consumption rate = zero) the d18O-N2O can be modelled by Eq. (5). When N2O reduction does occur Eqs. (6) and (7) can be used to predict d18O-N2O values. These equations

5

= 3.

2

3.6. Evidence of N2O reduction

t. 4 ) -exp e (ex p t. 5

50

3 2

t. (exp slope

slop

-

net O isotope effect (N2O-NO3 ) (‰)

55

45

B

4

) pt. 2 3-ex

= 0 .3

D

40 E

2 .5

C

1

e (N 2 O re du c

18

tion

)=

35

30

25

slop

110

20 A

15 -35

-30

-25 15

-20

-15

-10

-5

N isotope effect (N2O-NO3-) (‰)

Fig. 5. The oxygen and nitrogen isotope effects of the agricultural soils incubated in this study (black circles, labelled by experiment number) and the temperate forest soils of Snider et al. (2009) (white squares, labelled by experiment letter). The ratio of the oxygen and nitrogen isotope effects for N2O reduction is approximately 2.5:1. This figure would suggest that in the stream sediment, more N2O reduction occurred in experiment 5 (flooded) than in experiment 4 (non-flooded).

require knowledge of the fraction of N2O reduced (f N2O red.). When a system is truly closed there is no N2O production, and the f N2O red. is entirely dependent upon the initial size of the N2O pool and the rate of N2O consumption. In this case, the classic Rayleigh distillation applies (Mariotti et al., 1981; Eq. (6)) and the d18O-N2Ored. increases exponentially as the finite pool of N2O diminishes (Fig. 6i). When a system is open, such as the soil incubation experiments of this study, N2O is an intermediate compound that is continuously produced and consumed. In this case the f N2O red. is dependent upon the relative rates of gross production and consumption. This can be expressed as the N2O:N2 ratio (Eq. (11)) or as the N2O yield (%): N2 O : N2 ratio ¼ N2 O  ½N2 O þ N2 

ð11Þ

It is difficult to accurately measure the N2O:N2 ratio because the existing methods are fraught with errors (e.g., the acetylene block technique, Bollman and Conrad, 1997a,b; Wrage et al., 2004; Groffman et al., 2006) and the direct measurement of N2 at low concentrations is prone to contamination by atmospheric air-N2. A few studies have accurately quantified N2O:N2 ratios in soils using 13N- or 15N-labelling techniques, or by carefully measuring the N2 production in soils incubated in artificial atmospheres (helium or argon, e.g., Scholefield et al.,

D.M. Snider et al. / Geochimica et Cosmochimica Acta 112 (2013) 102–115

111

Fig. 6. Predicting the d18O-N2Ored. value in open and closed systems. (i) A closed system Rayleigh distillation of d18O-N2Ored. with different values of the isotope effect of N2O consumption. (ii) In an open system with net N2O production and N2O consumption the fraction of N2O remaining in the system increases with time and approaches 1. The rate of increase is governed by the N2O:N2 ratio. (iii) As the remaining fraction of N2O approaches 1, the observed 18O/16O discrimination decreases, and the d18O-N2Ored. approaches the initial d18O value.

1997; Delaune et al., 1998; Butterbach-Bahl et al., 2002; Bol et al., 2003; Meijide et al., 2010; Bergstermann et al., 2011). The N2O:N2 ratio in soils often increases following soil disturbance or N fertilization, and it can be highly variable over short time-scales (hours–days). When the N2O:N2 ratio > 0 (i.e. net N2O production > N2O consumption) the size of the N2O pool will increase with time. The rate of increase of the N2O pool is directly proportional to the N2O:N2 ratio (Fig. 6ii). This has important implications for the modelling of d18O-N2Ored. because when the N2O:N2 ratio is high (0.5 or greater), the fraction of N2O remaining (1  f N2O red.)

quickly approaches 1. As a result, very little 18O/16O discrimination is observed because the isotopic ‘signal’ of N2O consumption is quickly inundated by the newly produced N2O. The fraction of N2O remaining will also approach 1 when the N2O:N2 ratio is lower (0.01–0.1), albeit, it takes longer for the N2O pool to build in size (Fig. 6iii). Of course none of the model simulations illustrated in Fig. 6i–iii account for any physical loss of N2O. Diffusion is an important mechanism of N2O loss in soils, and advection and gas exchange may be important in aqueous systems. The relative rate of N2O loss out of the system will

112

D.M. Snider et al. / Geochimica et Cosmochimica Acta 112 (2013) 102–115

govern the upper size limit of the N2O pool and may prevent the fraction of N2O remaining from reaching 1. The imposition of physical N2O losses will cause the slope of the trajectories in Figs. 6ii–iii to decrease. This is important because the magnitude of the observed isotopic separation increases as the fraction of N2O remaining decreases (Fig. 6i).

When the remaining fraction of N2O is small (such as in a closed system), the effect of N2O reduction causes the slope (m) to increase slightly because consumption affects the d18O-N2O (but not the d18O-H2O). As a result, as N2O consumption increases, the difference between the slope (m) and the true O-exchange decreases (Table 2). In many cases, estimates of O-exchange derived using Method I are a good approximation of the true O-exchange (f exch.-nir + f exch.-nor; derived using Method II). It is only when f exch.-nir and f exch.-nor are both large that Method I fails to accurately estimate the amount of O-exchange (with or without N2O reduction).

3.8. Oxygen isotope exchange (Method II) The total O-exchange (f exch.-nir + f exch.-nor) is related to the slope of the linear regression of d18O-N2O(rel. VSMOW) versus d18O-H2O(rel. VSMOW) (Table 1) and is described in Eq. (7). An expression of the slope term is given below in Eq. (12).

3.9. Oxygen isotope fractionation (Method II) The net 18O isotope effect (18Oenet) is related to the intercept of the linear regression of d18O-N2O(rel. VSMOW) versus d18O-H2O(rel. VSMOW) (Table 1) and is described in Eq. (7). An expression of the intercept term is given below in Eq. (13).

Slope ðmÞ ¼ ½fexch:-nir þ fexch:-nor  ðfexch:-nir  fexch:-nor Þ  ½1  fN2 O

e4 =1000 red: 

ð12Þ

Eq. (12) is non-determinate (there is no unique solution) and f exch.-nir + f exch.-nor cannot be solved explicitly because there are too many unknown parameters. In an open system with a high N2O:N2 ratio (large fraction of N2O remaining) the total O-exchange (f exch.-nir + f exch.-nor) is approximately equal to the slope (m) of the linear regression (Table 1) plus the product of the individual O-exchange terms (f exch.-nir  f exch.-nor). Therefore, the slope (m) of the linear regression (as it is used in Method I) is a minimum estimate of O-exchange. If the majority of the O-exchange occurs at only one enzyme, and f exch.-nir  f exch.-nor (or vice versa) then f exch.-nir  f exch.-nor becomes very small, and the slope (m) f exch.-nir + f exch.-nor (Table 2). However, if there is a large amount of O-exchange and it occurs at both enzymes (f exch.-nir f exch.-nor), then f exch.-nir  f exch.-nor is a large and significant term. In this case the slope (m) < f exch.-nir + f exch.-nor and it is a poor approximation of the true magnitude of O-exchange (Table 2).

Intercept ðbÞ ¼ h½e6  fexch:-nor  þ ½1  fexch:-nor   ½e3 þ ðe5  fexch:-nir Þ þ ð1  fexch:-nir Þ  ðe1 þ e2 þ d18 O-NO3  Þ þ 1000i  h1  fN2 O red: ie4 =1000  1000

ð13Þ

The mathematical approach (Method II) used here to define the net 18O isotope effect for denitrification demonstrates the complex nature of 18Oenet, which depends on the d18O-NO3 value, the individual isotope effects (e1–e6), and the fractions of O-exchange and N2O reduction. If the system is open and the N2O:N2 ratio is high (or there is a large fraction of N2O remaining), then the isotope effect associated with N2O consumption (e4) approaches zero. In this case, Eq. (13) can be simplified to Eq. (14):

Table 2 The slope (m) of the linear regression of d18O-N2O(rel. VSMOW) versus d18O-H2O(rel. VSMOW) is a minimum estimate of O-exchange during denitrification. This table demonstrates the difference between the approximated O-exchange (column A) and the true O-exchange (column B) across the range of O-exchange measured in this study (0.40–0.94) when (i) f exch.-nir f exch.-nor, (ii) f exch.-nir – f exch.-nor, (iii) f N2O red. = 0, and (iv) f N2O red. > 0. Slope ðmÞ ¼ ½fexch:-nir þ fexch:-nor  ðfexch:-nir  fexch:-nor Þ  ½1  fN2 O

red: 

e4 =1000

(12)

A B C Approx. O-exchange (derived from True O-exchange Error slope of regression analysis) (f exch.-nir + f exch.-nor,derived from |A  B| equations)

D E F G H Suitability of f exch.-nir f exch.-nor f exch.-nir  f exch.-nor fN2 O slope to predict O-exchange

0.94 0.94 0.97 0.98 0.40 0.40 0.41 0.42

Poor Good Poor Good Good Good Good Good

1.50 0.98 1.50 0.98 0.45 0.41 0.45 0.41

0.56 0.04 0.53 0.00 0.05 0.01 0.04 0.01

0.75 0.94 0.75 0.94 0.22 0.39 0.22 0.39

0.75 0.04 0.75 0.04 0.23 0.02 0.23 0.02

0.56 0.04 0.56 0.04 0.05 0.01 0.05 0.01

red:

0 0 0.9 0.9 0 0 0.9 0.9

n.b. The value of e4 (within the range of literature estimates of 18Oe = 26& to 5&) has very little effect on the error (column C). As the strength of the isotope effect increases (e4 becomes more negative), the difference between the true and approximate O-exchange increases slightly. In these evaluations of Eq. (12), e4 was held constant at 15&.

D.M. Snider et al. / Geochimica et Cosmochimica Acta 112 (2013) 102–115

have even greater d18O-N2O values and be more clearly distinct from nitrifier-N2O.

Intercept ðbÞ ¼ ½e6  fexch:-nor  þ ½1  fexch:-nor   ½e3 þ ðe5  fexch:-nir Þ þ ð1  fexch:-nir Þ  ðe1 þ e2 þ d18 O  NO3  Þ

113

ð14Þ

It is not possible to isolate any further useful expressions of the 18O isotope effects from Eqs. (13) and (14) because there are too many unmeasured or unknown parameters in this study. However, the mathematical approach of defining 18Oenet (and O-exchange) that is presented here is useful for future studies that model the O isotope ratios of N2O produced by denitrification. 4. CONCLUSIONS The instantaneous 15N isotope effect (enitrous oxidenitrate) for denitrification measured for three agricultural soils ranged from 30& to 9&, which is consistent with other measurements described in the literature. The net 18O isotope effect calculated for all experiments in this study varied from +32& to +60&, and the approximate fraction of Oexchange (f exch.-nir + f exch.-nor) ranged from 0.40 in the flooded stream sediment to 0.94 in the non-flooded low OM soil. Moderate amounts of O-exchange occurred in the non-flooded and flooded high OM soil (0.64–0.65). A new mathematical approach to calculate O-exchange shows that the existing method derives approximations of O-exchange that are minimum estimates. These approximations become less accurate as O-exchange increases and when significant proportions of the total O-exchange occur at both nitrite and nitric oxide reductase. This new mathematical approach also demonstrated the complex nature of the net 18O isotope effect for denitrification, which is dependent upon the d18O-NO3 value, multiple isotope effects, and the fractions of O-exchange and N2O reduction. The 18Oenet was negatively correlated to O-exchange, and the well-drained soil had the highest amount of O-exchange and the lowest apparent oxygen isotope enrichment. Correspondingly, the stream sediment had the lowest amounts of O-exchange and the highest apparent oxygen isotope enrichment calculated for the agricultural and forest soils studied in this study and in Snider et al. (2009). This trend helps explain why the d18O-N2O values measured in oceans, rivers and groundwaters tends to be higher than the d18O-N2O values measured in soils. O-exchange is likely controlled by the microbial composition of the denitrifying community, which is influenced by the soil’s landscape position and its associated hydrology (well-drained versus saturated). The dominant denitrifying communities in well-drained soils may be very different than the communities found in saturated soils, groundwaters, surface waters, and oceans. Large amounts of O-exchange during denitrification can, in some cases, make d18O-N2O a less useful tracer of production pathways. When O-exchange is high the d18ON2O value is lower and may overlap with d18O values of nitrifier-N2O. Despite this, it appears that most of the in situ literature d18O values of denitrifier-N2O are >+35&, which is greater than the range of nitrification d18O–N2O (+13& to +35&; Snider et al., 2012). Furthermore, denitrifier-N2O that is affected by consumption will

ROLE OF THE FUNDING SOURCES Financial support for this research was provided to S. Schiff from Strategic and Discovery Grants from the Natural Sciences and Engineering Research Council of Canada (NSERC) and the Canadian Foundation for Climate and Atmospheric Sciences (CFCAS), and to D. Snider from a NSERC Postgraduate Scholarship and a Banting Postdoctoral Fellowship. The sponsors had no involvement in the study design, collection, analysis and interpretation of data, nor in the writing of the report. ACKNOWLEDGEMENTS We are grateful to Richard Elgood and Bill Mark for their skillful technical assistance in the laboratory.

REFERENCES ASTM International. ASTM Standard D4972-01. (2007) Standard Test Method for pH of Soils. ASTM Int., West Conshohocken, PA. http://dx.doi.org/10.1520/D4972-01R07. Available at: http://www.astm.org. Bergstermann A., Ca´rdenas L., Bol R., Gilliam L., Goulding K., Meijide A., Scholefield D., Vallejo A. and Well R. (2011) Effect of antecedent soil moisture conditions on emissions and isotopologue distribution of N2O during denitrification. Soil Biol. Biochem. 43, 240–250. Bol R., Toyoda S., Yamulki S., Hawkins J. M. B., Cardenas L. M. and Yoshida N. (2003) Dual isotope and isotopomer ratios of N2O emitted from a temperate grassland soil after fertiliser application. Rapid Commun. Mass Spectrom. 17, 2550–2556. Bollman A. and Conrad R. (1997a) Enhancement by acetylene of the decomposition of nitric oxide in soil. Soil Biol. Biochem. 29, 1057–1066. Bollman A. and Conrad R. (1997b) Acetylene blockage technique leads to underestimation of denitrification rates in oxic soils due to scavenging of intermediate nitric oxide. Soil Biol. Biochem. 29, 1067–1077. Boontanon N., Ueda S., Kanatharana P. and Wada E. (2000) Intramolecular stable isotope ratios of N2O in the tropical swamp forest in Thailand. Naturwissenschaften 87, 188–192. Bo¨ttcher J., Strebel O., Voerkelius S. and Schmidt H.-L. (1990) Using isotope fractionation of nitrate–nitrogen and nitrate– oxygen for evaluation of microbial denitrification in a sandy aquifer. J. Hydrol. 114, 413–424. Buchwald C. and Casciotti K. L. (2010) Oxygen isotopic fractionation and exchange during bacterial nitrite oxidation. Limnol. Oceanogr. 55, 1064–1074. Buchwald C., Santoro A. E., McIlvin M. R. and Casciotti K. L. (2012) Oxygen isotopic composition of nitrate and nitrite produced by nitrifying cocultures and natural marine assemblages. Limnol. Oceanogr. 57, 1361–1375. Butterbach-Bahl K., Willibald G. and Papen H. (2002) Soil core method for direct simultaneous determination of N2 and N2O emissions from forest soils. Plant Soil 240, 105–116. Casciotti K. L., McIlvin M. and Buchwald C. (2010) Oxygen isotopic exchange and fractionation during bacterial ammonia oxidation. Limnol. Oceanogr. 55, 753–762.

114

D.M. Snider et al. / Geochimica et Cosmochimica Acta 112 (2013) 102–115

Casciotti K. L., Sigman D. M., Galanter Hastings M., Bo¨hlke J. K. and Hilkert A. (2002) Measurement of the oxygen isotopic composition of nitrate in seawater and freshwater using the denitrifier method. Anal. Chem. 74, 4905–4912. Chen D., Li Y., Grace P. and Mosier A. R. (2008) N2O emissions from agricultural lands: a synthesis of simulation approaches. Plant Soil 309, 169–189. Delaune R. D., Lindau C. W., Sulaeman E. and Jugsujinda A. (1998) Nitrification and denitrification estimates in a Louisiana swamp forest soil as assessed by 15N isotope dilution and direct gaseous measurements. Water Air Soil Pollut. 106, 149–161. Firestone M. K. and Davidson E. A. (1989) Microbial basis of NO and N2O production and consumption in soils. In Exchange of Trace Gases Between Terrestrial Ecosystems and the Atmosphere (eds. M. O. Andreae and D. S. Schimel). John Wiley and Sons, New York, pp. 7–21. Granger J., Sigman D. M., Lehmann M. F. and Tortell P. D. (2008) Nitrogen and oxygen isotope fractionation during dissimilatory nitrate reduction by denitrifying bacteria. Limnol. Oceanogr. 53, 2533–2545. Groffman P. M., Altabet M. A., Bo¨hlke J. K., Butterback-Bahl K., David M. B., Firestone M. K., Giblin A. E., Kana T. M., Nielsen L. P. and Voytek M. A. (2006) Methods for measuring denitrification: diverse approaches to a difficult problem. Ecol. Appl. 16, 2091–2122. Jinuntuya-Nortman M., Sutka R. L., Ostrom P. H., Gandhi H. and Ostrom N. E. (2008) Isotopologue fractionation during microbial reduction of N2O within soil mesocosms as a function of water-filled pore space. Soil Biol. Biochem. 40, 2273–2280. Kool D. M., Wrage N., Oenema O., Dolfing J. and Van Groenigen J. W. (2007) Oxygen exchange between (de)nitrification intermediates and H2O and its implications for source determination of NO3 and N2O: a review. Rapid Commun. Mass Spectrom. 21, 3569–3578. Kool D. M., Wrage N., Oenema O., Harris D. and Van Groenigen J. W. (2009) The 18O signature of biogenic nitrous oxide is determined by O exchange with water. Rapid Commun. Mass Spectrom. 23, 104–108. Mandernack K. W., Mills C. T., Johnson C. A., Rahn T. and Kinney C. (2009) The d15N and d18O values of N2O produced during the co-oxidation of ammonia by methanotrophic bacteria. Chem. Geol. 267, 96–107. Mariotti A., Germon J. C., Hubert P., Kaiser P., Letolle R., Tardieux A. and Tardieux P. (1981) Experimental determination of nitrogen kinetic isotope fractionation: some principles; illustrations for the denitrification and nitrification processes. Plant Soil 62, 413–430. Mariotti A., Germon J. C. and Leclerc A. (1982) Nitrogen isotope fractionation associated with the NO2 ! N2O step of denitrification in soils. Can. J. Soil Sci. 62, 227–241. Meijide A., Cardenas L. M., Bol R., Bergstermann A., Goulding K., Well R., Vallejo A. and Scholefield D. (2010) Dual isotope and isotopomer measurements for the understanding of N2O production and consumption during denitrification in an arable soil. Eur. J. Soil Sci. 61, 364–374. Mengis M., Schiff S., Harris M., English M., Aravena R., Elgood R. and MacLean A. (1999) Multiple geochemical and isotopic approaches for assessing ground water NO3 elimination in a riparian zone. Ground Water 37, 448–457. Menyailo O. V. and Hungate B. A. (2006) Stable isotope discrimination during soil denitrification: production and consumption of nitrous oxide. Global Biogeochem. Cycles 20. http://dx.doi.org/10.1029/2005GB002527. Opdyke M. R., Ostrom N. E. and Ostrom P. H. (2009) Evidence for the predominance of denitrification as a source of N2O in temperate agricultural soils based on isotopologue measure-

ments. Global Biogeochem. Cycles 23. http://dx.doi.org/ 10.1029/2009GB003523. Ostrom N. E., Russ M. E., Popp B., Rust T. M. and Karl D. M. (2000) Mechanisms of nitrous oxide production in the subtropical North Pacific based on determinations of the isotopic abundances of nitrous oxide and di-oxygen. Chemosphere Global Change Sci. 2, 281–290. Ostrom N. E., Pitt A., Sutka R., Ostrom P. H., Grandy A. S., Huizinga K. M. and Robertson G. P. (2007) Isotopologue effects during N2O reduction in soils and in pure cultures of denitrifiers. J. Geophys. Res. 112. http://dx.doi.org/10.1029/ 2006JG000287. Ostrom N. E., Sutka R., Ostrom P. H., Grandy A. S., Huizinga K. M., Gandhi H., vonFischer J. C. and Robertson G. P. (2010) Isotopologue data reveal bacterial denitrification as the primary source of N2O during a high flux event following cultivation of a native temperate grassland. Soil Biol. Biochem. 42, 499–506. Park S., Croteau P., Boering K. A., Etheridge D. M., Ferretti D., Fraser P. J., Kim K.-R., Krummel P. B., Langenfelds R. L., van Ommen T. D., Steele L. P. and Trudinger C. M. (2012) Trends and seasonal cycles in the isotopic composition of nitrous oxide since 1940. Nat. Geosci. 5, 261–265. Pe´rez T., Garcia-Montiel D., Trumbore S., Tyler S., De Camargo P., Moreira M., Piccolo M. and Cerri C. (2006) Nitrous oxide nitrification and denitrification 15N enrichment factors from Amazon forest soils. Ecol. Appl. 16, 2153–2167. Petrone R., Smith C., Macrae M. and English M. (2006) Riparian zone equilibrium and actual evapotranspiration in a first order agricultural catchment in Southern Ontario, Canada. Agr. Water Manage. 86, 240–248. Scholefield D., Hawkins J. M. B. and Jackson S. M. (1997) Development of a helium atmosphere soil incubation technique for direct measurement of nitrous oxide and dinitrogen fluxes during denitrification. Soil Biol. Biochem. 29, 1345–1352. Snider D., Schiff S. and Spoelstra J. (2009) 15N/14N and 18O/16O stable isotope ratios of nitrous oxide produced during denitrification in temperate forest soils. Geochim. Cosmochim. Acta 73, 877–888. Snider D. M., Spoelstra J., Schiff S. L. and Venkiteswaran J. J. (2010) Stable oxygen isotope ratios of nitrate produced from nitrification: 18O-labeled water incubations of agricultural and temperate forest soils. Environ. Sci. Technol. 44, 5358–5364. Snider D. M., Venkiteswaran J. J., Schiff S. L. and Spoelstra J. (2012) Deciphering the oxygen isotope composition of nitrous oxide produced by nitrification. Global Change Biol. 18, 356– 370. Sutka R. L., Ostrom N. E., Ostrom P. H., Breznak J. A., Gandhi H., Pitt A. J. and Li F. (2006) Distinguishing nitrous oxide production from nitrification and denitrification on the basis of isotopomer abundances. Appl. Environ. Microbiol. 72, 638–644. Sutka R. L., Ostrom N. E., Ostrom P. H., Gandhi H. and Breznak J. A. (2003) Nitrogen isotopomer site preference of N2O produced by Nitrosomonas europaea and Methylococcus capsulatus Bath. Rapid Commun. Mass Spectrom. 17, 738–745. Sutka R. L., Ostrom N. E., Ostrom P. H., Gandhi H. and Breznak J. A. (2004) Erratum. Nitrogen isotopomer site preference of N2O produced by Nitrosomonas europaea and Methylococcus capsulatus Bath. Rapid Commun. Mass Spectrom. 18, 1411– 1412. Toyoda S., Iwai H., Koba K. and Yoshida N. (2009) Isotopomeric analysis of N2O dissolved in a river in the Tokyo metropolitan area. Rapid Commun. Mass Spectrom. 23, 809–821. Toyoda S., Mutobe H., Yamagishi H., Yoshida N. and Tanji Y. (2005) Fractionation of N2O isotopomers during production by denitrifier. Soil Biol. Biochem. 37, 1535–1545.

D.M. Snider et al. / Geochimica et Cosmochimica Acta 112 (2013) 102–115 US-EPA. (2006) Global anthropogenic non-CO2 greenhouse gas emissions: 1990–2020. United States Environmental Protection Agency, EPA 430-R-06-003, June 2006. Washington, DC, Available from: . Vieten B., Blunier T., Neftel A., Alewell C. and Conen F. (2007) Fractionation factors for stable isotopes of N and O during N2O reduction in soil depend on reaction rate constant. Rapid Commun. Mass Spectrom. 21, 846–850. Well R., Flessa H., Jaradat F., Toyoda S. and Yoshida N. (2005) Measurement of isotopomer signatures of N2O in ground water. J. Geophys. Res. 110. http://dx.doi.org/10.1029/ 2005JG000044.

115

Well R. and Flessa H. (2009) Isotopologue enrichment factors of N2O reduction in soils. Rapid Commun. Mass Spectrom. 23, 2996–3002. Wrage N., Velthof G. L., Oenema O. and Laanbroek H. J. (2004) Acetylene and oxygen as inhibitors of nitrous oxide production in Nitrosomonas europaea and Nitrosospira briensis: a cautionary tale. FEMS Microbiol. Ecol. 47, 13–18. Ye R. W., Toro-Suarez I., Tiedje J. M. and Averill B. A. (1991) H218O isotope exchange studies on the mechanism of reduction of nitric oxide and nitrite to nitrous oxide by denitrifying bacteria. J. Biol. Chem. 266, 12848–12851. Associate editor: Martin Novak