Biology and Fertility of Soils

, Volume 41, Issue 6, pp 379–388

Contributions of nitrification and denitrification to N2O emissions from soils at different water-filled pore space

Authors

  • E. J. Bateman
    • Department of Agricultural Sciences, Wye CampusImperial College London
    • Department of Biology 3University of York
    • School of Biological SciencesUniversity of Aberdeen
Original Paper

DOI: 10.1007/s00374-005-0858-3

Cite this article as:
Bateman, E.J. & Baggs, E.M. Biol Fertil Soils (2005) 41: 379. doi:10.1007/s00374-005-0858-3

Abstract

A combination of stable isotope and acetylene (0.01% v/v) inhibition techniques were used for the first time to determine N2O production during denitrification, autotrophic nitrification and heterotrophic nitrification in a fertilised (200 kg N ha−1) silt loam soil at contrasting (20–70%) water-filled pore space (WFPS). 15N-N2O emissions from 14NH415NO3 replicates were attributed to denitrification and 15N-N2O from 15NH415NO3 minus that from 14NH415NO3 replicates was attributed to nitrification and heterotrophic nitrification in the presence of acetylene, as there was no dissimilatory nitrate reduction to ammonium or immobilisation and remineralisation of 15N-NO3. All of the N2O emitted at 70% WFPS (31.6 mg N2O-N m−2 over 24 days; 1.12 μg N2O-N g dry soil−1; 0.16% of N applied) was produced during denitrification, but at 35–60% WFPS nitrification was the main process producing N2O, accounting for 81% of 15N-N2O emitted at 60% WFPS, and 7.9 μg 15N-N2O m−2 (0.28 ng 15N-N2O g dry soil−1) was estimated to be emitted over 7 days during heterotrophic nitrification in the 50% WFPS treatment and accounted for 20% of 15N-N2O from this treatment. Denitrification was the predominant N2O-producing process at 20% WFPS (2.6 μg 15N-N2O m−2 over 7 days; 0.09 ng 15N-N2O g dry soil−1; 85% of 15N-N2O from this treatment) and may have been due to the occurrence of aerobic denitrification at this WFPS. Our results demonstrate the usefulness of a combined stable isotope and acetylene approach to quantify N2O emissions from different processes and to show that several processes may contribute to N2O emission from agricultural soils depending on soil WFPS.

Keywords

DenitrificationNitrificationNitrous oxideSoil water-filled pore space

Introduction

The atmospheric concentration of N2O (314 ppb) is increasing at a rate of 0.25% per year (Houghton et al. 2001). This is of concern because N2O is a greenhouse gas and is involved in the destruction of stratospheric ozone (Badr and Probert 1993). Direct and indirect emissions of N2O from agricultural systems are now thought to contribute 6.2 Tg N2O-N year−1 to a total global source strength of 17.7 Tg N2O-N year−1 (Kroeze et al. 1999). N2O is produced biologically in soils during denitrification (Smith and Arah 1990), nitrification (Blackmer and Bremner 1978) and nitrifier denitrification (Poth and Focht 1985; Wrage et al. 2001). These processes may occur simultaneously in different microsites of the same soil (Stevens et al. 1997) but there is often uncertainty associated with which process is predominantly contributing to emissions from a particular soil.

A wide range of heterotrophic bacteria and fungi are able to reduce NO3 and NO2 to N2O or N2 during denitrification under anaerobic conditions (Knowles 1982; Shoun et al. 1992). This process is often considered to be the main N2O-producing process in soils inasmuch as many studies have shown N2O emissions after N application to increase with increasing soil water content and most rapidly above 60% water-filled pore space (WFPS) (e.g. Dobbie et al. 1999; Abbasi and Adams 2000; Skiba and Ball 2002). However, there is increasing evidence that aerobic denitrification may be significant in environments where O2 is not limiting (Patureau et al. 2000) because many bacteria isolated from soils and sediment (including Pseudomonas, Aeromonas and Moraxella genera) are capable of nitrate respiration in the presence of O2 (Carter et al. 1995; Patureau et al. 2000).

N2O production in soil during autotrophic nitrification is traditionally considered to be minor in comparison with denitrification. The oxidation of NH3 to NO2 by ammonia oxidisers and oxidation of NO2 to NO3 by nitrite oxidisers is optimal in aerobic conditions since this oxidation requires O2 as the terminal electron acceptor. However, significant N2O production during nitrification has been measured in cultures of nitrifiers under reduced O2 potential (Goreau et al. 1980) and in soils up to 60% WFPS (Linn and Doran 1984; Abbasi and Adams 2000). There is also evidence that ammonia oxidisers are able to reduce NO2 to N2O or N2 under short-term O2 limitation (nitrifier denitrification) (Poth and Focht 1985; Wrage et al. 2001). Thus, the contribution of autotrophic nitrification to N2O emissions from soils is not confined to strictly aerobic conditions and its contribution may have previously been underestimated. Furthermore, it is now known that the ability to nitrify extends beyond the closely related (monophyletic) autotrophic nitrifiers. The ability of some heterotrophic microorganisms, such as Alcaligenes faecalis, to nitrify organic and inorganic N compounds and produce N2O has been demonstrated in culture (Papen et al. 1989; Anderson et al. 1993). To date, evidence for heterotrophic nitrification in soils is limited to acidic and organic rich forest soils where autotrophic nitrification can be inhibited (Robertson and Tiedje 1987; Pedersen et al. 1999), and the contribution of heterotrophic nitrification to N2O emissions from fertilised arable soils where inorganic NH4+ is the most abundant source of N is unknown.

Acetylene (C2H2) inhibits ammonia oxidation by autotrophic nitrifiers at low (10 Pa) concentrations, and inhibits N2O reductase during denitrification at high (10 kPa) concentrations (Berg et al. 1982). Its application has provided evidence that both processes may proceed simultaneously in soil (Webster and Hopkins 1996; Abbasi and Adams 2000; Garrido et al. 2002). C2H2 is not reported to inhibit the oxidation of NH4+ by heterotrophic nitrifiers at the low concentrations used to inhibit autotrophs (Hynes and Knowles 1982; Moir et al. 1996; Daum et al. 1998). Recent advances in stable isotope techniques facilitating direct measurement of 15N-N2O allow a more accurate determination of the source of N2O where several processes are contributing to emissions (Stevens et al. 1997; Baggs et al. 2003) and can be coupled with C2H2 (10 Pa; 0.01% v/v) inhibition to determine the respective contributions of denitrification, autotrophic nitrification and heterotrophic nitrification.

Here we report an experiment in which we used a combination of C2H2 (0.01% v/v) to inhibit autotrophic nitrification (Berg et al. 1982), and stable isotope (15N) techniques (Baggs et al. 2003) to determine the respective contributions of denitrification, autotrophic nitrification and heterotrophic nitrification to N2O emissions after application of inorganic fertiliser to soil at different water contents (WFPS). Our aim was to identify the effect of soil WFPS on N2O production by each of these processes. We hypothesised that denitrification would be the predominant process contributing to N2O emissions above 60% WFPS and nitrification would be the predominant source of N2O below this WFPS.

Materials and methods

Soil

Soil (0–15 cm depth) was sampled from an arable field on the Imperial College London Estate at Wye in January 2002. The soil was a brown earth silt loam [17% sand, 68% silt, 15% clay, total carbon 1.9%, total N 0.2%, pH (H2O) 7.1, bulk density 1.23 g cm−3] of the Coombe series classified as a Cambisol (FAO classification). Soil was air-dried, sieved <2 mm and stored at 4°C until establishment of the experiment in March 2002.

Experimental set-up

The experiment was established in 500-ml Kilner jars with gas-tight lids fitted with a gas sampling port. Soil (200 g, 2% gravimetric water content) was weighed into each jar and soil water content amended to achieve the target WFPS of 20, 35, 50, 60 and 70%. Soils were conditioned at 20% below their target WFPS for 4 days before the experiment to initiate microbial activity and to minimise changes in soil water content at the start of the experiment.

On day 0, 15N-labelled NH4NO3 (710 μg N g−1, equivalent to 200 kg N ha−1 on a surface area basis when added to 200 g soil with a surface area of 71 cm2) was added in solution to soil at each WFPS as either (treatment a) 14NH415NO3 or (treatment b) 15NH415NO3 (10 at.% excess 15N). Unfertilised control treatments were also established. Additional water was added to achieve the target WFPS. Each treatment was replicated four times for gas analyses and an additional four times for soil analyses. The lids of the Kilner jars were then closed to create gas-tight incubations. C2H2 gas [5 ml of 1% (v/v in air)] was added to the headspace of a further four replicates of the 15NH415NO3 fertiliser treatments (treatment c) (Table 1) to give a final concentration of 0.01% (v/v) sufficient to inhibit ammonia oxidation. After 17 h, the jars were reopened. Preliminary tests (data not shown) showed that total inhibition of autotrophic nitrification continued for at least 3 days after addition of 0.01% (v/v) C2H2, and that there was no effect of either periodic closure of incubations on trace gas production, or of such low concentrations of C2H2 on denitrification in this soil. Thus, incubation of soil with 0.01% (v/v) C2H2 for 17 h was repeated every 3 days to maintain inhibition of autotrophic nitrification. All treatments were incubated at 21°C in the dark for 24 days after fertiliser application, and soil WFPS was maintained on a weight basis.
Table 1

Outline of the 15N fertiliser and C2H2 inhibition treatments used to estimate the contributions of denitrification and autotrophic and heterotrophic nitrification to 15N-N2O emissions

Treatment

Source of 15N-N2O

(a) 14NH415NO3

Denitrification

(b) 15NH415NO3

Denitrification and nitrification (autotrophic and heterotrophic)

(c) 15NH415NO3+ C2H2(0.01% v/v)

Denitrification and heterotrophic nitrification

(c) minus (a)

Heterotrophic nitrification

Gas sampling and analysis

Gas samples for N2O and CO2 determination were taken from the Kilner jars on days 1, 2, 3, 5, 7, 10, 14 and 24 after fertiliser application using gas-tight syringes. Jars were closed for 1 h during gas sampling. Samples were taken at 20, 40 and 60 min after jar closure and the measured flux determined by linear interpolation between these samples. Both 12- and 125-ml gas samples were taken from each jar headspace. The 12-ml samples were stored in pre-evacuated 12-ml gas vials (Labco), and 1 ml of this gas analysed for N2O and CO2 on an Agilent 6890 gas chromatograph fitted with an electron capture detector and flame ionisation detector with methaniser. The 125-ml samples for analysis of the fertiliser-derived 15N-N2O and 15N-N2 were taken on days 1, 3, 7, 14 and 24 using gas-tight syringes, stored in helium-flushed and pre-evacuated 125-ml gas-tight glass bottles (Supelco), and 125-ml samples analysed on a Europa 20:20 isotope ratio mass spectrometer after cryofocusing in an ANCA TGII gas preparation module (PDZ/Europa).

Soil sampling and analysis

Soil was destructively sampled from additional replicates (four per treatment) on days 1, 3, 7, 14 and 24. NH4+-N and NO3-N were extracted from these soils with 1 M KCl (25 g soil to 100 ml KCl solution) and determined by colorometric analysis on a Burkard SFA2 continuous-flow analyser. The 15N enrichment of NH4+ and NO3 was determined by diffusion methodology (Brookes et al. 1989) and analysis on the mass spectrometer. Gross nitrification rates were calculated for the 14NH415NO3 replicates according to Davidson et al. (1991).

Determining the source of N2O production

Application of 15N-labelled NH4NO3 and 0.01% (v/v) C2H2 enabled the source of fertiliser-derived 15N-N2O to be determined (Table 1). N2O emitted from C2H2 treatments could be attributed to either denitrification or heterotrophic nitrification (treatment c), and 15N-N2O production from 14NH415NO3 replicates (treatment a) could be attributed to denitrification, having confirmed in preliminary tests (data not shown) that there was no 15N enrichment of NH4+ or NO2 in this treatment, and thus negligible dissimilatory reduction of 15N-NO3, or immobilisation of 15N-NO3 and remineralisation as 15N-NH4+. The difference in emission of 15N-N2O between the 14NH415NO3 (treatment a) and 15NH415NO3 (treatment b) replicates was attributed to nitrification (Baggs et al. 2003), and the difference between 14NH415NO3 (treatment a) and 15NH415NO3+C2H2 (treatment c) was attributed to heterotrophic nitrification. Thus, autotrophic nitrification accounted for the nitrified-N2O emission that could not be attributed to heterotrophic nitrification.

Statistical analysis

All data were analysed using the SPSS statistical package (v11.0). Data were tested for normality and log-transformed where appropriate, before means comparisons using independent t tests and tests of association using Spearman’s rank correlation.

Results

Emissions of 14+15N-N2O and CO2

Nitrous oxide emitted from the fertilised 70% WFPS treatment was higher (P<0.05) throughout the experiment than from the other treatments, with 31.6 mg 14+15N m−2 (1.1 μg N2O-N g dry soil−1) emitted over 24 days from this treatment; 6 and 16 times greater than that emitted at 60 and 20% WFPS, respectively (Table 2; Fig. 1). There was no significant difference in total 14+15N-N2O emitted over 24 days from the 35 and 50% WFPS treatments. Up to 47% of the total N2O emitted over the 24-day experimental period was emitted in the first 3 days.
Table 2

Total emissions of 14+15N-N2O (mg N2O-N m−2) ± 1 SEM, and percentage contributions of autotrophic nitrification and denitrification or heterotrophic nitrification to total N2O-N emissions over 3, 7 and 24 days after application of NH4+NO3, as indicated by the acetylene inhibition technique

 

20% WFPS

35% WFPS

50% WFPS

60% WFPS

70% WFPS

Total N2O-N mg m−2 (3 days) (μg N2O-N g dry soil−1)

0.2±0.02 (0.01)

1.2±0.1 (0.04)

1.1±0.3 (0.04)

2.4±1.2 (0.09)

14.9±5.0 (0.53)

% Denitrification and/or heterotrophic nitrification

48.5

19.8

19.5

10.08

100.0

% Autotrophic nitrification

51.5

80.2

80.5

89.91

0

Total N2O-N mg m−2 (7 days) (μg N2O-N g dry soil−1)

0.5±0.03 (0.02)

2.0±0.10 (0.07)

1.5±0.14 (0.05)

3.5±1.68 (0.12)

22.4±8.05 (0.80)

% Denitrification and/or heterotrophic nitrification

48.5

22.3

23.5

12.9

100.0

% Autotrophic nitrification

51.5

77.7

76.5

87.1

0

Total N2O-N mg m−2 (24 days) (μg N2O-N g dry soil−1)

2.0±0.2 (0.07)

3.3±0.1 (0.12)

3.1±0.1 (0.11)

5.2±1.7 (0.19)

31.6±10.7 (1.12)

% Denitrification and/or heterotrophic nitrification

52.3

22.1

32.4

19.0

100.0

% Autotrophic nitrification

47.7

77.9

67.6

81.0

0

N2O emitted from C2H2 treatments was attributed to either denitrification or heterotrophic nitrification (treatment c) and the contribution of nitrification was taken as the difference in N2O emission with (treatment c) and without C2H2. Mean emissions of 14+15N-N2O are given in parentheses on a per gram dry soil basis (μg N2O-N g dry soil−1)

Fig. 1

Total N2O-N emissions over 24 days after application of NH4+NO3 to soils at different WFPS. Black bars represent the contribution from denitrification or heterotrophic nitrification in C2H2 treatments and open bars represent the contribution from autotrophic nitrification determined from the difference in emission from C2H2 and non-C2H2 treatments. Error bars are 1 SE of the difference (n=4) for the total N2O emission

Daily fluxes of N2O from the fertilised 35–70% WFPS treatments decreased throughout the experiment but were higher (P<0.01) than from the unfertilised controls (Fig. 2). The greatest fluxes (P<0.05) were measured from the 70% WFPS treatment with a flux of 5.7 mg 14+15N m−2 day−1 (0.2 μg N2O-N g dry soil−1 day−1) measured on day 1. A flux of 0.5 mg 14+15N m−2 day−1 was measured from both the 35 and 50% WFPS treatments on day 1 and was higher (P<0.001) than the flux of 0.1 mg 14+15N m−2 day−1 (0.04 μg N2O-N g dry soil−1 day−1) measured from the 20% WFPS treatment on this day. Although the mean N2O emission rate from soils at 60% WFPS was consistently greater (P<0.05) over the first 5 days than from soils at 35 and 50% WFPS the variability between replicates was greater at 60% WFPS. By day 24 emissions were low from all treatments. The percentage of N applied emitted as 14+15N2O-N over 24 days increased with increasing WFPS from 0.01% at 20% WFPS to 0.16% at 70% WFPS. Daily fluxes of 14+15N-N2O were positively correlated with available NH4+ and most strongly at 35–60% WFPS (r=0.74–0.88; P<0.05), but negatively correlated with available NO3 (r=−0.57 to −0.83; P<0.05).
Fig. 2

Daily N2O-N fluxes after application of NH4+NO3 to soils at (a) 70% WFPS, (b) 20–60% WFPS, and (c) from unfertilised control soils (20–70% WFPS). Error bars represent +1 SEM (n=8)

A total of 9.6 g C m−2 (0.34 mg CO2-C g dry soil−1) was emitted as CO2 over 24 days from the 70% WFPS treatment and the total CO2 emitted over the first 7 days from this treatment was significantly higher (P<0.05) than from all other treatments. A maximum flux of 1.5 g C m−2 day−1 (0.05 mg CO2-C g dry soil−1 day−1; P<0.001) was measured on day 1 from the 70% WFPS treatment and fluxes from the 35 and 60% WFPS treatments were higher (P<0.001) than from the 50 and 20% WFPS treatments on this day (Fig. 3). The CO2 flux from the 70% WFPS treatment on day 24 was lower (P<0.05) than fluxes measured from other treatments on this day. Only at 35 and 50% WFPS were CO2 emissions from fertilised treatments greater than from the controls (P<0.05). 14+15N-N2O and CO2 fluxes were positively correlated, and most strongly in the 35, 60 and 70% WFPS treatments and controls (r=0.91–0.95; P<0.05).
Fig. 3

Daily CO2-C fluxes after application of NH4+NO3 to soils and from unfertilised control soils at different WFPS. Error bars are ±1 SEM (n=8)

Emission of 14+15N-N2O from autotrophic nitrification as determined by acetylene inhibition

Total emissions of N2O over 24 days were significantly lowered (P<0.05) from 20 to 60% WFPS treatments in the presence of C2H2, indicating the contribution of autotrophic nitrification to N2O production (Table 2; Fig. 1). N2O emitted in the presence of C2H2 was attributed to heterotrophic nitrification and/or denitrification. Autotrophic nitrification did not contribute to N2O production at 70% WFPS. The highest N2O emission from autotrophic nitrification of 4.2 mg N m−2 (0.15 μg N g dry soil−1) over 24 days (81% of total N2O emitted and 0.02% of N applied) was from the 60% WFPS treatment. Between 68 and 81% of N2O emitted from the 35, 50 and 60% WPFS treatments was from autotrophic nitrification, and this proportion was slightly higher (80–90%) over the first 3 days. At 20% WFPS, autotrophic nitrification only contributed to 48% of total N2O emitted.

Emission of 15N-N2O

A total 15N-N2O emission of 1.3 mg 15N m−2 (45 ng 15N g dry soil−1) was measured over 7 days from the 70% WFPS treatment (Table 3) and accounted for 0.06% of 15N applied. There was no significant difference between total 15N-N2O emitted from 14NH415NO3 or 15NH415NO3 replicates of this WFPS treatment, indicating that this N2O was produced during denitrification and confirming the lack of inhibition by 0.01% v/v C2H2. At all water contents, 15N-N2O fluxes followed the same trend as fluxes of 14+15N-N2O. Maximum fluxes of 306.3, 87.1, 20.3, 17.0 and 0.4 μg N m−2 were measured from the 70, 60, 50, 35 and 20% WFPS treatments, respectively, on day 1. Autotrophic nitrification was the predominant source of 15N-N2O from 35 to 60% WFPS treatments and accounted for 81% (0.01% 15N applied) of 15N-N2O emitted from the 60% WFPS treatment (Fig. 4). 15N-N2O emitted from the 15NH415NO3 C2H2 50% WFPS replicate was higher than from the corresponding 14NH415NO3 replicate, indicating a contribution from heterotrophic nitrification of 7.9 μg 15N m−2 (0.28 ng 15N g dry soil−1) over 7 days (20% of the total 15N-N2O emission) in this treatment (Fig. 4). Of the 15N-N2O emitted over 7 days from this treatment, 81% was lost in the first 3 days. 15N-N2O emitted from the 20% WFPS treatment was low (3.1 μg 15N m−2 over 7 days; 0.11 ng 15N g dry soil−1) and denitrification was the predominant source of 15N-N2O in this treatment accounting for 2.6 μg 15N m−2 over 7 days (0.09 ng 15N g dry soil−1), or 85% of the emission.
Table 3

Total emissions of 15N-N2O (μg 15N2O-N m−2) ± 1 SEM, and percentage contributions of autotrophic and heterotrophic nitrification and denitrification to total 15N-N2O emissions over 3 and 7 days after application of 15NH415NO3 or 14NH415NO3

 

20% WFPS

35% WFPS

50% WFPS

60% WFPS

70% WFPS

Total 15N-N2O μg m−2 (3 days) (ng 15N-N2O g dry soil−1)

0.7±0.1 (0.03)

29.2±6.4 (1.04)

31.9±10.6 (1.13)

149.0±125.8 (5.29)

824.7±295.0 (29.28)

% Denitrification

83.1

37.1

22.1

24.5

100.0

% Autotrophic nitrification

16.9

62.9

64.7

75.5

0

% Heterotrophic nitrification

0

0

13.2

0

0

Total 15N-N2O μg m−2 (7 days) (ng 15N-N2O g dry soil−1)

3.1±0.4 (0.11)

46.5±12.8 (1.65)

39.5±10.6 (1.40)

227.6±195.6 (8.08)

1280.1±494.6 (45.44)

% Denitrification

84.8

36.7

23.9

18.5

100.0

% Autotrophic nitrification

15.3

63.3

56.2

81.5

0

% Heterotrophic nitrification

0

0

20.0

0

0

Mean emissions of 15N-N2O are given in parentheses on a per gram dry soil basis (ng 15N-N2O g dry soil−1)

Fig. 4

The contribution of nitrification and denitrification to 15N-N2O emissions over 7 days after fertiliser application to soils at different WFPS. Total bar height represents 15NH415NO3 (all processes), black bar represents 14NH415NO3 (denitrification), dark grey bar represents the difference between emissions from 15NH415NO3 C2H2 and 14NH415NO3 (heterotrophic nitrification), light grey bar represents the difference in 15N-N2O between 15NH415NO3 and 14NH415NO3 minus estimated heterotrophic nitrification contributions (autotrophic nitrification). Error bars are 1 SE of the difference (n=4)

Soil mineral N

Concentrations of NH4+ decreased (P<0.05) between days 1–24 in the 35–60% WFPS treatments and the rate of decrease was more rapid with higher WFPS in this range (Fig. 5). Concentrations in the C2H2 treatments varied little throughout the experiment indicating effective inhibition of NH4+ oxidation to NO3 by C2H2 at all soil WFPS. Available NO3 increased in the 35–60% WFPS treatments between days 1–24 (P<0.01) and this increase was more gradual in the C2H2 replicates, indicating that autotrophic nitrification was the main source of NO3.
Fig. 5

Soil mineral N after application of NH4+NO3 to soils at different WFPS. (a) NH4+-N, (b) NH4+-N in acetylene treated soils, (c) NO3-N, (d) NO3-N in acetylene-treated soils. Error bars are ±1 SEM or are smaller than the symbols (n=3)

The 15N enrichment of NO3 from the 14NH415NO3 treatment declined between days 1–7 in the 35, 50 and 60% WFPS treatments indicating 15N-NO3 pool dilution as a result of nitrification of 14NH4. Gross nitrification was highest (P<0.05) in the 60% WFPS treatment with a rate of 6.9 μg N g dry soil−1 between days 1–7 (Table 4). N2O production attributable to nitrification in this treatment was 0.17% of the estimated gross nitrification during days 1–7. The highest percentage production of N2O-N relative to NO3-N during nitrification (days 1–7) was in the 35% WFPS treatment because whilst gross nitrification was low, 15N-N2O production during nitrification was comparable to that in the 50% WFPS treatment.
Table 4

Soil 15N-NO3-N (μg 15N-NO3-N g dry soil−1) after fertiliser application (14NH415NO3) and gross nitrification rates

 

Soil 15NO3-N (μg 15N-NO3-N g dry soil−1)

35% WFPS

50% WFPS

60% WFPS

15N-NO3-N day 1

24.5 (0.5)

25.4 (1.0)

23.0 (1.7)

15N-NO3-N day 3

25.8 (0.7)

24.4 (0.5)

23.7 (0.4)

15N-NO3-N day 7

24.7 (1.2)

25.5 (0.3)

25.7 (0.5)

Gross nitrification days 1–7 (μg N g−1 dry soil day−1)

1.1

2.4

6.9

% N2O-N:NO3-N during nitrification days 1–7

0.53

0.18

0.17

Values in parentheses are 1 SEM

Discussion

The effect of soil WFPS on N2O emissions

The magnitude of N2O emissions increased with increasing WFPS, in agreement with other studies (e.g. Dobbie et al. 1999; Abbasi and Adams 2000; Skiba and Ball 2002); however, this increase was not linear. The significant increase in emissions between 20 and 35% WFPS suggests that N2O production at 20% WFPS was limited by substrate diffusion and water availability for microbial activity (Stark and Firestone 1995) and that this limitation did not occur at 35% WFPS. This was confirmed by higher CO2 emissions at 35% than at 20% WFPS, particularly over the first 3 days. With increasing soil WFPS O2 diffusion into the soil is restricted and the proportion of soil volume that is anaerobic increases (Smith 1980; Sexstone et al. 1988; Renault and Sierra 1994). Soil microsites were assumed to be predominantly anaerobic at 70% WFPS, as the high emission from this treatment was solely produced during denitrification.

The significant increase in N2O emissions between 60 and 70% WFPS is in agreement with field measurements from temperate soils. Clayton et al. (1997) found 65% WFPS to be a critical threshold above which emissions from a grassland soil increased significantly. Here we report a tenfold increase in emissions between 60 and 70% WFPS, whilst increases of up to 12 and 30 times have been measured from temperate grassland and arable soils with an increase in WFPS from 60 to 80% (Dobbie and Smith 2001).

Use of both stable isotopes and acetylene (0.01% v/v) enabled determination of the respective contributions of denitrification, autotrophic nitrification and heterotrophic nitrification to N2O emissions that were occurring simultaneously in the same soil (Fig. 6). Our model shows that the contribution of these processes to N2O emissions from soil changes with different WFPS. The absolute contributions varied depending on method used, with a greater contribution of autotrophic nitrification estimated using acetylene inhibition alone (51–90% contribution at 20–60% WFPS), than when used in combination with the stable isotope technique (17–76% contribution at 20–60% WFPS), although the only major discrepancy above 15% was at 20% WFPS. This may indicate an advantage of direct measurement of 15N-N2O rather than reliance on inhibitors that may also affect denitrification. However, the trends with different WFPS and conclusions drawn were the same irrespective of method adopted.
Fig. 6

Summary model of the contributions of autotrophic and heterotrophic nitrification and denitrification to N2O emissions in response to soil WFPS

N2O emission from nitrification

Autotrophic nitrification was the predominant process contributing to N2O emissions from the 35–60% WFPS treatments and accounted for 81% of N2O emitted at 60% WFPS. A WFPS around 60% offers optimal conditions for nitrification because neither the diffusion of substrates nor the diffusion of O2 is restricted (Parton et al. 1996). N2O production during autotrophic nitrification declined slightly over time, most likely due to the progressive development of anaerobic microsites resulting from microbial respiration (Wolf and Russow 2000). However, there is evidence that autotrophic nitrification may proceed under short-term O2 limitation in the process of nitrifier denitrification (Goreau et al. 1980; Poth and Focht 1985; Kester et al. 1997; Bollmann and Conrad 1998; Wrage et al. 2001). The greater contribution of autotrophic nitrification to N2O emissions, particularly 15N-N2O, at 60% than at 35% WFPS, may indicate a propensity for nitrifier denitrification at this higher WFPS, where short-term O2 limitation may have occurred (Bollmann and Conrad 1998). Thus, the threshold at which our soil became too anaerobic for either nitrifier denitrification or aerobic nitrification was above 60% WFPS and below 70% WFPS where nitrification did not contribute to N2O production. A high N2O-to-NO3product ratio during nitrification has previously been shown at low O2 availability (Kester et al. 1997), but here we found a higher ratio at 35% than at 50 or 60% WFPS due to low gross nitrification at 35% WFPS but N2O production comparable to that at 50% WFPS.

Our results indicate that heterotrophic nitrification accounted for 20% of 15N-N2O emitted over 7 days at 50% WFPS, but this should be treated with caution due to the high variability associated with this treatment. However, the possibility that some of the heterotrophic nitrification may be inhibited by C2H2 means it is more likely to be underestimated than overestimated. The possible occurrence of heterotrophic nitrification only in this WFPS treatment suggests that O2 and water diffusion or C availability were limiting for this process, or that N2O production during this process was too low to detect at other WFPS. N2O production by the heterotroph A. faecalis has been shown to be more sensitive to reduced O2 availability in culture than N2O production by the autotroph Nitrosomonas europaea (Anderson et al. 1993). To our knowledge, our results provide the first indication for N2O production during heterotrophic nitrification in arable soils where the main source of available N is NH4+. Heterotrophic nitrification of background organic N in the soil before fertiliser application is considered negligible given the excess of inorganic N that was applied. The proportion of N2O produced during heterotrophic nitrification would be expected to be greater in acidic grassland or woodland soils where organic N favours heterotrophs that are capable of nitrifying organic sources, where autotrophic nitrification is often inhibited, and where fungal biomass predominates (Killham 1986; Pedersen et al. 1999; Laverman et al. 2000).

N2O emission from denitrification

All N2O emitted at 70% WFPS was produced during denitrification, in agreement with the model of Davidson (1991), and was attributed to the likely predominance of anaerobic microsites at this WFPS (Sexstone et al. 1988). The main difference between our results and the model of Davidson (1991) was that we measured high N2O release by denitrification at 70% WFPS, whereas the model assumes the maximum emission to occur at 60% WFPS where both nitrification and denitrification significantly contribute to N2O production, and above which most of the N2O is reduced to N2. This discrepancy could be due to differences in experimental design and soil types. Alternatively, application of NH4NO3 at 200 kg N ha−1 may have temporarily inhibited the conversion of N2O to N2 during denitrification because NO3 is preferred over N2O as an electron acceptor at concentrations of >10 μg g−1 (Blackmer and Bremner 1978). Thus, it is possible that concentrations of up to 400 μg g−1 in our experiment were inhibitive for reduction of N2O to N2 at 70% WFPS. Despite this possibility only up to 0.16% of N applied was lost as N2O in the first 24 days after fertiliser application, which is below the range of Bouwman’s (1996) estimate of 1.25±1% of N applied.

Although N2O emissions were low at 20% WFPS, approximately 50% of 14+15N-N2O and 85% of 15N-N2O emitted from this treatment was produced during denitrification. This suggests that either denitrification was proceeding in anaerobic microsites in this treatment or that aerobic denitrification was occurring (Carter et al. 1995; Patureau et al. 2000). The periplasmic nitrate reductase involved in the first reduction step during aerobic denitrification is not inhibited by O2, unlike the membrane-bound nitrate reductase of anaerobic denitrification (Carter et al. 1995), and so it is possible that aerobic denitrification was occurring at 20% WFPS. Although aerobic denitrifiers are thought to be present in high numbers in soils (Patureau et al. 2000), to our knowledge this is the first reported indication of the contribution of this process to N2O emissions from arable soils. Some heterotrophic nitrifiers, such as Paracoccus denitrificans, are able to couple heterotrophic nitrification with aerobic denitrification (Robertson and Kuenen 1990). However, with our methodology, any coupled heterotrophic nitrification–denitrification would have been attributed to heterotrophic nitrification, and thus it is possible that this occurred in the 50% WFPS treatment.

Conclusions

Our study demonstrates that a combined stable isotope and acetylene inhibition (0.01% v/v) approach enables determination of the relative contributions of autotrophic and heterotrophic nitrification and denitrification to N2O emission from soils at varying WFPS. Nitrification was the main source of N2O in soils at 35–60% WFPS, indicating the significance of this process for global warming, despite its role having often been underplayed compared to that of denitrification. Our results indicate N2O production during heterotrophic nitrification in our soil at 50% WFPS and the possibility of aerobic denitrification. Thus, several processes may simultaneously produce N2O in soil at 60% WFPS and below, and these processes need to be considered when predicting atmospheric loading of N2O and proposing strategies to mitigate N2O emissions from agricultural soils.

Acknowledgements

This work was funded by a Research Committee studentship awarded by the Biotechnology and Biological Sciences Research Council, UK. We thank Jon Fear for the stable isotope analyses and Trudi Krol for assisting with the soil mineral N analyses.

Copyright information

© Springer-Verlag 2005