Introduction

Nitrous oxide (N2O) is a greenhouse gas that is mainly produced during microbial nitrogen transformation processes in natural and agricultural soils. In 2011, atmospheric N2O concentration was at 324 ppb, which is an increase of 20% in comparison to 17501. Since N-fertilization provides the substrates for N2O production in soils atmospheric N2O concentrations have increased together with the intensification of agricultural practices. The rising demand for food and the reliance of agriculture on nitrogen fertilizers will likely further increase N2O emissions in the future2. In order to counteract this trend, strategies to mitigate N2O emissions arising from current agricultural practice are urgently required.

Among other strategies, biochar application to soils has been discussed as a potential option to mitigate N2O emissions from agricultural soils3,4. Biochar is a carbon-rich solid, produced with the intent to be used as soil amendment. It is produced by thermal decomposition of organic material under low oxygen conditions. Recently biochar has gained a lot of attention due to its potential to not only improve soil quality but also serve as a potent climate change mitigation strategy5,6,7. For example, it has been demonstrated that due to its physicochemical properties biochar can improve plant growth while simultaneously sequestering atmospheric carbon dioxide8,9. Although biochar’s physicochemical properties vary strongly with the type of biomass used as feedstock and the charring conditions during thermal decomposition, most biochars share common characteristics. The majority of biochars have a highly aromatic carbon structure, a neutral to alkaline pH, and a large surface area6,10. It has been shown in a recent meta-analysis that soil biochar amendment can significantly mitigate soil-derived N2O emissions11,12. In addition, various studies revealed biochar effects on functional microbial guilds involved in nitrogen transformation processes such as nitrification and denitrification13,14,15,16,17.

Microbial nitrification and denitrification are thought to be the main pathways for N2O production in natural and agricultural soils18. Even though both processes are considered to account for the majority of soil-derived N2O emissions, other microbial processes such as dissimilatory nitrate reduction to ammonium (DNRA) or co-denitrification have been assumed to contribute significantly to total N2O emissions under specific environmental conditions19,20. Oxygen partial pressure, which is largely controlled by the soil water content and aerobic respiration, and the availability of nitrogen compounds represent key factors controlling N2O production and release18. The highest N2O emissions often occur in situations when oxygen is limiting and mineral nitrogen contents are high, for example after heavy rainfall on fertilized soils. Although high water contents (up to total saturation) in most soils in temperate regions are usually temporary, it has been shown that the amount of N2O emitted under these conditions accounts for a high proportion of the total annual N2O emissions21,22. Low oxygen partial pressure promotes anaerobic nitrogen transformation processes such as denitrification, which is considered as the prevailing N2O producing process when the water-filled pore space (WFPS) reaches 60–90%18. Denitrification is performed by many facultative and strict anaerobic chemoorganotrophic bacteria and describes the stepwise enzymatic reduction of nitrate (NO3) to dinitrogen (N2)18,23. Obligate intermediates are nitrite (NO2), nitric oxide (NO) and N2O. The enzymes catalysing the reduction reactions are encoded by the functional genes narG and napA (nitrate reductases), nirK and nirS (nitrite reductases), norB (nitric oxide reductase) and nosZ (nitrous oxide reductase)23,24. As some denitrifiers lack a functional nosZ gene and nitrous oxide reductases are highly oxygen and pH sensitive, the last step of denitrification, N2O reduction to N2, is often impaired and N2O is released as the denitrification end product25,26,27. Although N2O can be produced by several nitrogen transformation processes, the only known dominant sink for N2O is the microbial reduction to N2 via nosZ-encoded nitrous oxide reductases28. Hence, the quantity of N2O emitted from soils depends substantially on the activity of N2O-reducing microorganisms as the enzymatic reduction of N2O to N2 directly controls the N2O/(N2O + N2) emission ratio. For a long time N2O reduction was thought to be a functional trait restricted to ‘classical’ denitrifiers. In addition to the nosZ gene ‘classical’ denitrifiers usually also harbour other denitrification genes such as nirK and nirS and most of them are affiliated with the Proteobacteria23,29. However, recent studies have provided evidence for the existence of a new clade of N2O reducers containing phylogenetically and physiologically more diverse microorganisms30,31. These microbes carry an atypical form of the nosZ gene and about half of them do not carry a nitrite reductase gene. Thus they are not ‘classical’ denitrifiers and many of them only have the genetic capability to perform the last step of denitrification, i.e. the reduction of N2O to N230,31,32,33. Independently of the dominating microbial N2O production pathway, a high amount of water-saturated soil pores also directly affects diffusion of N2O molecules from the site of production to the soil surface34. When compared to air, the diffusion coefficient of N2O is approximately 4 orders of magnitude lower in water35. Consequently, N2O can be retained in the water-filled pores for several days while slowly diffusing through the network of soil pores towards the soil surface which can lead to enhanced N2O reduction activities due to the long residence time of N2O in soil34,36,37. This retardation process has been described as gas “entrapment” in previous studies and occurs frequently in soils with elevated water content34.

According to recent studies that reported lower N2O emissions from biochar-amended soil likely explanations for the suppression of N2O emission are (i) a biochar-induced decrease in denitrification activity resulting in lower N2O production rates38,39, (ii) an increase in microbial N2O reduction (e.g. by a pH increase) leading to a decreased N2O/(N2O + N2) emission ratio11,15,16,40, or (iii) the contribution of abiotic processes such as sorption of N2O onto biochar particles and abiotic reduction of N2O to N2 by redox active compounds41,42. However, systematic studies focusing on the underlying geochemical and microbial processes causing N2O emission mitigation are rare and the contributing mechanisms are not fully understood yet.

In order to further improve our understanding of the underlying processes, we set up soil microcosm experiments with wet soil (90% WFPS) and wood-derived biochar, which were fertilized with a 15NO3-labelled NH415NO3 solution. We quantified important soil chemical properties, determined the isotopic signatures of different nitrogen pools and quantified denitrification-specific functional marker genes and transcripts. The main objectives of this study were to determine biochar’s impact (i) on nitrification and denitrification rates, (ii) the contribution of source processes to N2O emissions, (iii) NO3-derived N2O and N2 emissions, (iv) N2O and N2 remaining entrapped in soil, and (v) the abundance and activity of functional microbial guilds contributing to soil denitrification.

Material and Methods

Soil and biochar properties

Soil was sampled from the top 10 cm at an urban gardening site of the University of Tuebingen, Germany (48°32′3″N, 9°4′16″E). The soil was classified as an Anthrosol (World Reference Base for Soil Resources, 2014) with 49.8% sand, 25.8% silt, and 24.4% clay (USDA soil texture class: sandy clay loam). Biochar was obtained from Swiss Biochar (Belmont-sur-Lausanne, Switzerland). It was produced by slow pyrolysis (620 °C) of wood residues from a wood chip production plant in accordance with the standards of the European Biochar Certificate (EBC, http://www.european-biochar.org). Important physical and chemical properties of soil and biochar are summarized in Table 1.

Table 1 Properties of soil and biochar.

Soil microcosm setup and sampling

In total three similar soil microcosm experiments with slightly different sampling strategies were set up (i.e. experiment 1 to 3) to quantify N2O and N2 fluxes, soil nitrogen transformation rates and functional denitrification genes and mRNA transcripts.

Prior to microcosm setup, the field moist soil (water-filled pore space (WFPS): 40.6%) sampled from the top 10 cm and the biochar were passed through a 2 mm sieve and thoroughly homogenized. Soil microcosms were set up in 250 ml glass beakers (experiment 1 and 2) or 250 ml wide neck glass bottles (experiment 3). Control microcosms consisted of the field moist equivalent of 100 g dry soil. Biochar microcosms contained the field moist equivalent of 95 g dry soil and 5 g dry biochar, which corresponded to a biochar content of 5% (w/w). Both, control microcosms and biochar microcosms were set up in triplicates. After preparation, all soil microcosms were homogenized with a spatula and compacted by exposing them to 10 blows of a 125 g hammer falling 20 cm. The surface area of the hammer was identical to the soil surface area to ensure homogenous compaction. At beginning of the experiments the soil microcosms were fertilized with a 15NO3-labelled NH415NO3 solution (experiment 1: 50 at% 15N, experiment 2 and 3: 66 at% 15N) at a rate of 332 mg N kg−1 (equivalent to 100 kg N ha−1 estimated based on the soil surface area in the microcosms). To uniformly distribute the 15N label in the soil microcosms the NH415NO3 solution was spread evenly over the soil surface to allow homogenous infiltration. The amount of water added with the NH415NO3 solution (control: 36 ml, biochar: 49 ml) was adapted to adjust WFPS to 90%. To allow gas exchange with ambient atmosphere but decrease evaporation rates soil microcosms were covered with a perforated aluminium foil. During incubation WFPS was controlled gravimetrically and held constant by periodically replenishing the evaporated water with deionized water. Soil microcosms were incubated at a constant temperature of 20 °C.

Experiment 1: Total and NO3-derived N2O emissions, soil chemical analyses and qPCR analyses

In experiment 1 sampling took place right after microcosm setup (day 0) and after 1, 2, 4, 7 and 10 days of incubation. We set up 36 soil microcosms to allow destructive soil sampling of a set of 6 microcosms (3 control and 3 biochar microcosms) at each sampling time point. At each time point of sampling we collected soil samples for soil chemical and qPCR analyses as well as gas samples for N2O and NO3-derived N2O quantification from the same set of microcosms.

For the determination of N2O and NO3-derived N2O emission rates (ERp), soil microcosms (250 ml glass beakers) were placed in gas-tight enrichment containers consisting of modified 1.7 l preserving glass jars equipped with luer-lock stopcocks for gas sampling. Gas-tightness was confirmed experimentally prior to experiment start. Enrichment containers were closed and after 1 h of enrichment headspace gas samples were taken and transferred immediately to evacuated 20 ml (N2O) and 120 ml (15N2O) glass vials, capped with butyl rubber septa, using a syringe. In addition, lab air gas samples were taken directly before the containers were closed for gas enrichment, to obtain starting values for N2O and NO3-derived N2O.

Directly after the gas sample was taken (1 h) soil microcosms were removed from the enrichment containers and the entire soil was transferred into clean vessels, homogenized using a spatula and sub sampled for nucleic acid extraction and soil chemical analyses. Sub samples for nucleic acid extraction consisted of the equivalent of 2 g dry soil and were directly frozen at −80 °C.

Experiment 2: NO3-derived N2O and N2 emissions

In experiment 2 gas samples were collected repeatedly from six soil microcosms at the same time points as in experiment 1 (days 0, 1, 2, 4, 7 and 10).

For NO3-derived N2O and N2 emission rate (ERp) determination, soil microcosms (250 ml glass beakers) were placed inside enrichment containers as described above. To decrease the high atmospheric N2 background (78%) and thus improve 15N2 quantification43,44, gas enrichment was performed in an artificial atmosphere. After closure, enrichment containers were purged with an artificial gas mixture consisting of 21% O2, 2% N2 and 77% He. To remove remaining gases from the headspace and the soil we first applied a vacuum using a pump. When an under-pressure of −700 mbar was reached, containers were filled with the artificial gas mixture until an over-pressure of 300 mbar was achieved. This procedure was repeated 5 times. The last filling was brought to ambient atmospheric pressure by exhausting excess gas through a water-filled vessel.

Headspace samples to analyse 15N enrichment of N2O and N2 by isotope ratio mass spectroscopy were taken 1 and 3 h after purging the enrichment container. For gas sample collection, a 12 ml screw-cap exetainer (Labco Limited, Lampeter, UK), filled with the artificial gas mixture, was connected to the gas enrichment container using a needle. In addition, a syringe (70 ml) was connected to the exetainer with a second needle. Headspace gas was repeatedly pumped through the exetainer into the syringe and back into the headspace of the gas enrichment container by drawing and expelling the syringe completely (70 ml) for 5 times. Preliminary tests confirmed that this procedure ensures equal concentrations in the 12 ml screw-cap exetainer and the enrichment container. All equipment such as tubes, syringes, needles and stopcocks were purged with the artificial gas mixture before sampling and kept closed (tubes, stopcocks) or were sealed using butyl rubber stoppers.

Experiment 3: NO3-derived N2O and N2 entrapment in soil

In experiment 3 six soil microcosms (250 ml wide neck glass bottles) were incubated for two days.

In addition to NO3-derived N2O and N2 emission rate (ERp) determination, we also determined the soil entrapment rate (SERp) i.e. the fraction of NO3-derived N2O and N2 produced within 3 h of gas enrichment that remained entrapped in the soil. It has been shown by Holt et al.45 that destroying the soil matrix by creating and shaking soil slurries and subsequent sampling of the headspace provides an appropriate method to quantify entrapped gases45. To allow the creation of a shakeable soil slurry after 3 h of gas enrichment we used 250 ml wide neck glass bottles which served as soil microcosms and enrichment containers instead of 250 ml glass beakers and the enrichment containers described above.

After closure of the bottles using butyl rubber stoppers purging with the artificial gas mixture as well as the gas sampling after 1 and 3 h was performed exactly as described for experiment 2. After the second gas sample was taken (3 h after purging) a defined volume of deionized water was added to create a soil slurry. The resulting over-pressure was set off by exhausting excess gas through a water-filled vessel. The slurry was horizontally shaken on a shaker at 150 rpm for 2 min. Afterwards another gas sample was collected from the headspace as described for experiment 2.

Soil chemical analyses and gross nitrogen transformation rates

Soil and biochar characterization

Properties of the initial soil (Table 1) were determined using standard laboratory methods commonly used in soil sciences. For details please refer to the Supplementary Information. All biochar properties were determined by Eurofins (Eurofins Umwelt Ost, Freiberg, Germany) according to the guidelines given by the European Biochar Certificate (http://www.european-biochar.org).

Soil chemical and stable isotope analyses

Soil extracts for chemical analyses were prepared using a 2 M KCL solution directly after soil sub sampling at each time point of sampling (0, 1, 2, 4, 7, 10 days). The equivalent of 5 g dry soil was mixed with 25 mL of 2 M KCL and shaken horizontally for 1 h at 150 rpm. After sedimentation of larger particles (15 min) the supernatant was filtered through a 0.45 μm pore size syringe filter. The obtained filtrate was stored at 4 °C and analysed for NH4+, 15NH4+, NO3, 15NO3 within 4 weeks. The absence of nitrogen transformation reactions in the filtrates under these storage conditions was experimentally confirmed in preliminary tests.

NH4+ and NO3 contents were determined spectrophotometrically using a segmented flow analyser (AA3 AutoAnalyzer 3 HR, Seal Analytical, Southampton, UK). 15N enrichment of NH4+ and NO3 was quantified directly from the filtrate using the SPINMAS technique, as described in detail in Stange et al.46.

Gross nitrification and NO3 consumption rates

Gross nitrification and NO3 consumption rates were determined using 15N pool dilution. Individual rates were determined for all time frames (0–1, 1–2, 2–4, 4–7 and 7–10 days) based on the approach first described by Kirkham and Bartholomew47 using the equations from Davidson et al.48. The equations (1–2) can be found in the Supplementary Information.

Quantitative polymerase chain reaction (qPCR) analyses

Nucleic acid extraction

For quantitative polymerase chain reaction (qPCR) analyses nucleic acids (RNA and DNA) were extracted from frozen (−80 °C) soil sub samples taken at each time point of sampling (0, 1, 2, 4, 7, 10 days) of experiment 1. Total RNA and DNA were extracted simultaneously from the moist equivalent of 2 g dry soil using the RNA PowerSoil Total RNA Isolation Kit in combination with the RNA PowerSoil DNA Elution Accessory Kit according to the manufacturer protocol (MO BIO Laboratories, Carlsbad, CA, USA). Concentration and quality of RNA and DNA isolates were determined spectrophotometrically (NanoDrop 1000, Thermo Scientific, Waltham, MA, USA), fluorometrically (Qubit 2.0 Fluorometer, Life Technologies, Carlsbad, CA, USA), and by gel electrophoresis (Experion Automated Electrophoresis Station, Bio-Rad Laboratories, Hercules, CA, USA).

Reverse transcription

Digestion of residual DNA in RNA extracts was performed using the Ambion TURBO DNA-free Kit (Life Technologies, Carlsbad, CA, USA) according to the instructions given by the manufacturer. Successful DNA removal was confirmed by PCR using primers 27F (5′-AGAGTTTGATCMTGGCTCAG-3′)49 and PC5 (5′-TACCTTGTTACGACTT-3′)50 with the following conditions: hot start at 70 °C, 5 min at 95 °C, 35 cycles with 1 min at 95 °C, 1 min at 44 °C and 3 min at 72 °C followed by a final elongation step of 10 min at 72 °C. If no PCR products were obtained, RNA extracts were used for cDNA synthesis via reverse transcription. Reverse transcription of DNA-free RNA extracts was performed with random primers (Life Technologies, Carlsbad, CA, USA) using the SuperScript III Reverse Transcriptase (Life Technologies, Carlsbad, CA, USA) according to the manufacturer’ s protocol.

Nucleic acid extraction efficiencies

To allow the determination of DNA and RNA extraction efficiencies of each sample we added prior to extraction 4 × 1010 DNA and 7 × 1010 RNA copies of a linearized plasmid carrying a fragment of a cassava mosaic virus (accession number: AJ427910) as internal standard. This fragment is not thought to naturally occur in our soil and biochar and has been used as internal standard to monitor DNA recovery before51. In order to determine DNA and RNA extraction efficiencies the number of internal standard copies in the final DNA and cDNA extracts was quantified using qPCR. Extraction efficiencies were calculated according to equation (3) in the Supplementary Information.

Quantitative polymerase chain reaction (qPCR)

The quantification of internal standard copies (DNA and cDNA) and functional marker genes (DNA) and transcripts (cDNA) (napA, narG, nirK, nirS, typical nosZ, and atypical nosZ) was carried out by qPCR using the iCycler iQ Real-Time PCR Detection System and the iQ 5 Optical System software (Bio-Rad Laboratories, Hercules, CA, USA). qPCRs were setup in 96 well plates with a reaction volume of 10 μl. Reaction mixtures consisted of SsoAdvanced Universal SYBR Green Supermix or IQ SYBR Green Supermix (Bio-Rad Laboratories, Hercules, CA. USA), gene-specific primers and template DNA or cDNA. As standards, dilution series with plasmids containing fragments of the target genes were used. In all qPCRs, standards, samples and negative controls were analysed in triplicates and amplicon specificity was confirmed by performing melt curve analyses and 2% agarose gels. To account for potential differences in extraction efficiency, gene and transcript copy numbers were divided by the corresponding average control and biochar extraction efficiencies. Details about the qPCRs are listed in Tables S1 and S2 in the Supplementary Information.

Gas analyses

Determination of concentrations and isotopic signatures of N2O in experiment 1

N2O concentrations in the 20 ml septum-capped glass vials sampled from the gas enrichment containers were measured using a gas chromatograph (GC) equipped with an electron capture detector (63Ni-ECD) (5890 Series II, HP, Palo Alto, CA, USA). The GC setup and configuration has been described in detail in Loftfield et al.52. The precision of N2O determination, expressed as the coefficient of variation for repeated standard gas quantifications, was below 3%.

The abundance of 15N-substituted isotopologues of N2O i.e. (14N15N16O + 15N14N16O) and 15N15N16O was quantified in the gas samples collected in the 120 ml septum-capped glass vials from the same gas enrichment containers. Isotopic signatures of N2O molecules were determined by analysing m/z 44, 45 and 46 using an isotope ratio mass spectrometer (IRMS, DeltaPlus XP IRMS, Thermo Finnigan, Waltham, MA, USA). The IRMS was connected to a pre-concentration device, equipped with a chemical CO2-trap and an automated cryo-trap (PreCon, Thermo Finnigan, Waltham, MA, USA), and a Gasbench II (Thermo Finnigan, Waltham, MA, USA). A CTC PAL autosampler (CTC Analytics AG, Zwingen, Switzerland) was used to collect gas from the sample vials by flushing the vials with a continuous flow of He carrier gas (25 ml/min). The precision for 45R and 46R determination, expressed as the coefficient of variation for repeated standard gas quantifications at different N2O concentrations was below 1.6% and 1.1%, respectively.

Determination of isotopic signatures of N2O and N2 in experiment 2 and 3

In experiment 2 and 3 abundances of 15N substituted isotopologues of N2 (15N14N, 15N15N) and N2O ([14N15N16O + 15N14N16O] and 15N15N16O) were determined in the gas samples collected in the 12 ml screw-cap exetainers during gas enrichment. Isotopologues of N2O and N2 were analysed using a modified GasBench II preparation system coupled to a MAT 253 isotope ratio mass spectrometer (Thermo Finnigan, Waltham, MA, USA) as described in Lewicka-Szczebak et al.53. In this setup N2O is converted to N2 prior to analysis enabling the simultaneous quantification of 29R (29N2/28N2) and 30R (30N2/29N2) of N2, N2O + N2 and N2O. The precision of 29R and 30R determination of this setup is below 0.15% and 1%, respectively53.

N2O source partitioning and rates of NO3-derived N2O and N2

Apart from slight modifications at some steps the general analysis procedure was performed according to Buchen et al.54. We used the non-equilibrium approach55,56 to calculate the 15N enrichment of the N2O-producing NO3 pool (ap) and the fractions of pool-derived (NO3) N2O and N2 (fp) using the equations described in Spott et al.57. Details about the used equations (4–9) can be found in the Supplementary Information.

Concentrations of NO3-derived N2O and N2 (cp) were determined by multiplication of fp values with the headspace concentration of total N2O and N2 (equation (10) in the Supplementary Information). cp values were corrected for dilution effects (ccp) when needed and subsequently used to calculate NO3-derived N2O and N2 emission rates (ERp) using equations (11–13) and (14–15) respectively, as described in the Supplementary Information. NO3-derived N2O and N2 soil entrapment rates (SERp) and total production rates (TPRp) were determined based on the soil entrapment/emission ratio (SEp/Ep) using equations (16–18) in the Supplementary Information.

The contribution of N2O from the labelled NO3 pool to total soil-derived N2O emissions (fnitrate) was determined using equation (19), as described in the Supplementary Information.

Statistical analysis

The effects of biochar addition and time (main effects) as well as their interaction (biochar*time) on NO3-derived N2O and N2 emission rates, NO3-derived N2O/(N2O + N2) emission ratios and the numbers of gene and transcript copies were evaluated using two-way ANOVA. Two-way ANOVAs were carried out in SAS (SAS 9.2, SAS Institute, Cary, NC, USA) using PROC MIXED. The model is given by:

where yijk is the k th replicate at the i th biochar treatment at the j th time point. μ is the total effect, ai is the main effect of the i th biochar treatment (control or biochar), βj is the main effect of the j th time point (days 0, 1, 2, 4, 7, 10) and (αβ)ij is the interaction of the i th biochar treatment and the j th time point. eijk is the error of yijk.

In case it increased the model fit, group specific error variances were considered. Selection of the best fitting model was based on the Akaike Information Criterion (AIC). If necessary, a logarithmic transformation was used to reach normally distributed residuals with homogeneous variances. For data presentation log transformed means were back transformed. The corresponding standard errors were back transformed using the delta method. When significant interaction effects were determined simple mean comparisons were performed using the SLICE statement in PROC MIXED.

Temporal averages of the sources of N2O emissions, individual and mean nitrogen transformation rates, cumulative NO3-derived N2O and N2 emissions, as well as NO3-derived N2O and N2 emission, soil entrapment and total production rates determined only at day 2 in experiment 3 were analysed for differences in control and biochar microcosms using t-tests. In addition, t-tests were applied to evaluate differences between NO3-derived N2O and N2 emission and soil entrapment rates within treatments (control or biochar). T-tests were carried out in SAS using PROC TTEST.

Results

Gross nitrification and NO3 consumption rates

Individual gross nitrification rates (experiment 1; days 0–1, 1–2, 2–4, 4–7, 7–10) ranged from 0.07 ± 0.02 to 0.28 ± 0.03 mg N kg−1 dry soil h−1 (Table 2). Comparison of control and biochar microcosms indicated slightly higher mean nitrification rates in control compared to biochar-amended microcosms. However, individual gross nitrification rates were only significantly higher in control microcosms between days 2–4 and between days 7–10. For all other time periods, rates were either equal (days 4–7) or higher in biochar compared to control microcosms (days 0–1 and 1–2).

Table 2 Gross nitrification and NO3 consumption rates in control and biochar microcosms (experiment 1).

Individual gross NO3 consumption rates (experiment 1; days 0–1, 1–2, 2–4, 4–7, and 7–10) ranged from 0.07 ± 0.02 to 0.61 ± 0.04 mg N kg−1 dry soil h−1 (Table 2). Mean NO3 consumption rates were higher in biochar-amended than in control microcosms and also individual NO3 consumption rates were higher in biochar microcosms for 4 out of 5 time periods. Significantly higher individual rates in biochar microcosms were determined between day 0 and 1 and between day 7 and 10.

In control microcosms mean nitrification and NO3 consumption rates were very similar, 0.19 ± 0.02 and 0.20 ± 0.05 mg N kg−1 dry soil h−1, respectively. In biochar-amended microcosms, however, the mean NO3 consumption rate was significantly (p = 0.017) higher (0.3 ± 0.05 mg N kg−1 dry soil h−1) than the mean nitrification rate (0.16 ± 0.02 mg N kg−1 dry soil−1).

Sources of N2O emissions

N2O source partitioning (experiment 1) indicated that the predominant source of N2O emissions was the 15N-labelled NO3 pool. On average 100.9 ± 1.7% (control) and 100.1 ± 2.3% (biochar-amended microcosms) of total soil-derived N2O emissions originated from the 15N-labelled NO3 pool. Statistically significant differences between control and biochar microcosms were not detected. The contribution of NO3-derived N2O to total soil-derived N2O emissions over time is shown in Figure S1 in the Supplementary Information.

NO3-derived N2O and N2 emissions

Time courses of NO3-derived N2O emission rates (experiments 1) differed strongly in control and biochar-amended microcosms (Fig. 1A). In general, emission rates were much higher in control compared to biochar microcosms. Furthermore, the addition of biochar also changed the temporal dynamics of NO3-derived N2O emission rates. In control microcosms the highest rates were quantified after 4 days of incubation (0.160 mg N2O-N kg−1 dry soil h−1). In the presence of biochar, highest NO3-derived N2O emission rates were determined at day 10 (0.067 mg N2O-N kg−1 dry soil h−1). Between day 1 and 10 emission rates were always higher in control microcosms. Two-way ANOVA revealed that biochar amendment and time as well as their interaction (biochar*time) had a significant effect (Table 3). Significantly higher emission rates in control compared to biochar microcosms were determined at day 1, 2 and 4 (p < 0.001 for each time point). Cumulative (day 0–10) NO3-derived N2O emissions were significantly (p = 0.004) higher in control (27.2 ± 1.9 mg N2O-N kg−1 dry soil) than in biochar (12.6 ± 0.2 mg N2O-N kg−1 dry soil) microcosms resulting in biochar-induced N2O emission mitigation of 54 ± 16%.

Table 3 Results from two-way ANOVAs for gas emission, transcript, and transcript ratio data (experiments 1 and 2).
Figure 1
figure 1

NO3-derived N2O emission rates (A), NO3-derived N2 emission rates (B), and the NO3-derived N2O/(N2O + N2) emission ratio (C) in control (white circles) and biochar (black circles) microcosms over time (experiments 1 and 2).

Data points and error bars represent means and standard errors (n = 3), respectively.

NO3-derived N2 emission rates (experiment 2) were substantially lower than the corresponding NO3-derived N2O emission rates and showed only minor differences between control and biochar-amended microcosms (Fig. 1B). Highest rates, with values of 0.03 mg N2-N kg−1 dry soil h−1 in control and 0.02 mg N2-N kg−1 dry soil h−1 in biochar microcosms, were observed at day 10. Between day 1 and 10, emission rates were slightly higher in control microcosms. According to the two-way ANOVA, NO3-derived N2 emission rates were not significantly affected by biochar addition, time or their interaction (Table 3). Cumulative (day 0–10) NO3-derived N2 emissions were slightly (control: 5.3 ± 1.4 mg N2-N kg−1 dry soil, biochar: 3.3 ± 0.2 mg N2-N kg−1 dry soil), but not statistically significantly higher in control microcosms.

The NO3-derived N2O/(N2O + N2) emission ratio (experiment 2) ranged from 0.71 to 0.91 (Fig. 1C). The ratio was always lower in biochar microcosms at each time point of sampling from day 1 to day 10. Nonetheless, two-way ANOVA revealed that in contrast to time, which had a significant effect on the N2O/(N2O + N2) emission ratio, the effect of biochar addition as well as the interaction of time and biochar did not significantly affect the N2O/(N2O + N2) ratio (Table 3).

Gene and transcript copies of functional denitrification genes

napA, narG, nirK, nirS, typical nosZ and atypical nosZ transcript copy numbers (experiment 1) were strongly affected by biochar addition and changed as a function of time (Fig. 2). For napA, narG and nirS transcript copy numbers two-way ANOVA revealed a significant interaction (biochar*time), indicating that biochar effects were correlated to the time of sampling (Table 3).

Figure 2
figure 2

Transcript copy numbers of functional marker genes of denitrification in control (white circles) and biochar (black circles) microcosms over time (experiment 1).

The different panels show: napA (A), narG (B), nirK (C), nirS (D), typical nosZ (E), and atypical nosZ (F). Panels G and H show the typical nosZ/nirS and the atypical nosZ/nirS transcript ratio, respectively. Data points and error bars represent means and standard errors (n = 3), respectively.

napA transcript levels were relatively similar in control and biochar microcosms and ranged from 9.1 × 105 to 4.6 × 106 transcripts g−1 dry soil (Fig. 2A). Significantly more napA transcripts in biochar microcosms were quantified at day 7 (p = 0.022). narG ranged from 1.6 × 106 to 6.3 × 106 transcripts g−1 dry soil and showed pronounced differences between control and biochar microcosms with higher values in biochar microcosms at day 1, 2 and 4, and lower transcript levels in the biochar setups at day 10 (Fig. 2B). Significantly higher narG transcript copy numbers in biochar microcosms were quantified at day 4 (p < 0.001) and significantly lower values were determined at day 10 (p = 0.004). nirS transcript copy numbers ranged from 8.6 × 105 to 2.2 × 106 transcripts g−1 dry soil (Fig. 2D). Transcript levels were higher in biochar microcosms at day 1, 2, 4, and 7 with significantly higher values at day 1 (p = 0.026) and 4 (p = 0.018). Significantly lower values were quantified after 10 days of incubation (p = 0.027).

Transcript copy numbers of nirK, typical nosZ and atypical nosZ were significantly affected by both main effects, biochar addition and time (Table 3). Transcript copy numbers of nirK, typical nosZ and atypical nosZ ranged from 6.3 × 105 to 1.8 × 107, 3.5 × 105 to 2.7 × 106, and 3.3 × 106 to 6.6 × 108 transcript copies g−1 dry soil, respectively. All three were higher in biochar microcosms at days 1, 2, 4 and 7 (Fig. 2C). According to two-way ANOVA nirK, typical nosZ and atypical nosZ transcript copy numbers were significantly higher in biochar microcosms compared to the control microcosms (Table 3).

While some of the transcript copy ratios were affected by biochar addition and time (typical nosZ/nirS and atypical nosZ/nirS; Fig. 2G,H), others showed similar values in control and biochar microcosms and hardly changed over time (typical nosZ/nirK and atypical nosZ/nirK; time course data not shown). Time courses for the typical nosZ/nirS transcript ratio ranged from values of 0.36 to 1.29 (Fig. 2G). The typical nosZ/nirS transcript ratio was higher in the biochar microcosms at days 1, 4, 7, and 10 and two-way ANOVA revealed significant time (p < 0.001) and biochar effects (p = 0.031) (Table 3). The atypical nosZ/nirS transcript ratio ranged from values of 179.1 to 404.2 and was hardly affected by biochar addition. In contrast to the typical nosZ/nirS transcript ratio, time was the only significant effect on the atypical nosZ/nirS transcript ratio (Table 3). According to two-way ANOVA, no significant effects (biochar, time, biochar*time) could be determined for typical nosZ/nirK and atypical nosZ/nirK transcript ratios (Table 3).

Gene copy numbers of napA, narG, nirK, nirS, typical nosZ and atypical nosZ showed similar dynamics as the corresponding transcripts (Figure S2). However, differences between control and biochar microcosms were less prominent. Two-way ANOVA revealed a significant biochar effect on napA gene abundance, a significant biochar*time effect on typical nosZ, and significant time and biochar*time effects on nirS gene copy numbers (Table S3).

Soil entrapment of NO3-derived N2O and N2

Quantification of NO3-derived N2O and N2 remaining in the soil at day 2 (experiment 3) revealed that the extent of soil gas content significantly differed in control and biochar-amended microcosms and strongly depended on the type of nitrogen gas (N2O, N2) (Fig. 3).

Figure 3
figure 3

Emission, soil entrapment and total production rates of NO3-derived N2O (A) and N2 (B) in control (white bars) and biochar (black bars) microcosms at day 2 (experiment 3).

Panel C shows the N2O/(N2O + N2) product ratio derived from (A) and (B). Bars and error indicators represent means and standard errors (n = 3), respectively.

In control microcosms soil entrapment/emission ratios (SEp/Ep) for NO3-derived N2O and N2 were 0.5 and 2.7, respectively. Accordingly, emission rates were significantly higher than soil entrapment rates for NO3-derived N2O (p = 0.018) and significantly lower for NO3-derived N2 (p < 0.001) (Fig. 3A,B). In biochar microcosms the soil entrapment/emission ratios (SEp/Ep) reached values of 7.5 (NO3-derived N2O) and 42.9 (NO3-derived N2) resulting in significantly higher soil entrapment than emission rates for both NO3-derived N2O (p < 0.001) and N2 (p = 0.021) (Fig. 3A,B). Comparison of NO3-derived N2O emission, soil entrapment, and total production rates in control and biochar microcosms revealed that while biochar addition significantly decreased the emission rate (p = 0.003), soil entrapment (p < 0.001) and total production rates (p = 0.009) were significantly higher in biochar-amended compared to control microcosms (Fig. 3A). Similar results were found for NO3-derived N2. Although biochar addition did not significantly affect the NO3-derived N2 emission rate, it significantly increased NO3-derived N2 soil entrapment (p = 0.027) and total production (p = 0.030) rates (Fig. 3B).

Independent of biochar addition, NO3-derived N2O/(N2O + N2) ratios were significantly higher for emitted than for entrapped gases (control: p < 0.001, biochar: p < 0.001) (Fig. 3C). Comparison of NO3-derived N2O/(N2O + N2) ratios in control and biochar microcosms revealed a significant biochar effect. While soil biochar amendment did not significantly change the ratio of NO3-derived N2O and N2 that was emitted, it significantly decreased the ratio of NO3-derived N2O and N2 that remained in the soil (p = 0.003) and the ratio of total NO3-derived N2O and N2 that was produced (p < 0.001) (Fig. 3C).

Nitrogen balance

Nitrogen balances for the first 2 days of incubation (experiments 1, 2, and 3) were estimated based on the initial (day 0) and residual (day 2) NO3 and NH4+ contents and the cumulative (day 0–2) NO3-derived N2O and N2 losses (Table 4). Scenario I considered only emitted cumulative N2O and N2. Scenario II considered emitted and soil entrapped cumulative N2O and N2. In scenario I, 66.9 and 94.5% of the lost N from the mineral nitrogen pool (NO3 and NH4+) remained unaccounted in control and biochar-amended microcosms, respectively. Cumulative emissions of NO3-derived N2O and N2 accounted for only 33.1 (control) and 5.5% (biochar), respectively. In scenario II, which considered emitted and soil entrapped gases, cumulative NO3-derived N2O and N2 accounted for 58.6 (control) and 79.1% (biochar) of the lost N. Thus, only 41.4 and 20.9% of the lost N remained unaccounted in control and biochar microcosms, respectively.

Table 4 Nitrogen balance after 2 days of incubation (experiments 1, 2, and 3).

Discussion

In fertilized soil systems, highest N2O emission rates frequently occur after heavy rainfall when water-filled pore spaces (WFPS) are high (i.e. O2 availability in the soil is low) and anaerobic nitrogen transformation processes such as denitrification are prevailing21,22. In order to simulate such situations, we set up a short-term soil microcosm experiment with a WFPS of 90% to which we added a NH415NO3 solution to prevent NH4+ and NO3 limitation for microbial nitrogen transformation processes.

N2O source partitioning revealed that soil biochar amendment did not significantly affect the sources of N2O. Independent of biochar addition, N2O emissions solely originated from the NO3 pool suggesting that denitrification was the main N2O-producing pathway. This finding is in good agreement with other studies which were carried out under similar environmental conditions58,59,60.

Individual gross NO3 consumption rates were higher in biochar microcosms for 4 out of 5 time periods when compared to the control microcosms. These results are in line with the findings of Prommer et al.61 and Hu et al.62 who also documented a biochar-induced increase in NO3 consumption. According to the 15N pool dilution approach used in this study, gross NO3 consumption covers microbial NO3 assimilation, dissimilatory nitrate reduction to ammonium (DNRA) and denitrification48. However, the environmental conditions in the soil microcosms suggest that the majority of NO3 was consumed by microbial denitrification. This is in contrast to other studies which assumed that a decrease in the denitrification rate would provide a potential explanation for decreased N2O emissions from biochar-amended soil38,39. However, Xu et al.15 and Jones et al.63 quantified higher denitrification enzyme activities (DEA) in the presence of biochar. Also Obia et al.40 quantified higher denitrification rates in biochar-amended soil slurries.

Although gross NO3 consumption rates were higher, NO3-derived N2O emission rates were significantly lower in biochar-amended microcosms. Furthermore, the calculation of cumulative NO3-derived N2O emissions showed that soil biochar amendment significantly decreased N2O emissions by 54%. These results confirmed what has been shown in several lab and field studies before11,64,65,66,67,68,69, that biochar amendment can mitigate soil N2O emission. In addition to N2O emission reduction, Cayuela and colleagues showed that soil biochar amendment substantially decreased the N2O/(N2O + N2) emission ratio64. Similar effects were also found in the soil slurry experiments by Obia et al.40 as well as in this study. Although two-way ANOVA revealed no significant biochar, time, or interaction effects, biochar addition decreased the NO3-derived N2O/(N2O + N2) emission ratio at all sampling points. In contrast to Obia et al.40 who reported an increase in N2 emissions following biochar addition, NO3-derived N2 emission rates and cumulative NO3-derived N2 emissions were almost equal in control and biochar microcosms in our study. Thus, according to the NO3-derived N2O and N2 emission data, biochar addition decreased total denitrification-related nitrogen loss (NO3-derived N2O + N2).

Several studies showed that in flooded and wet soils significant quantities of N2O and N2 can remain entrapped in the soil matrix45,70,71. As diffusion coefficients are significantly lower (approximately 4 orders of magnitude) in water compared to air, diffusion of microbially-released N2O and N2 from the site of production to the soil surface takes considerably longer in soils with elevated water-filled pore space compared to well-aerated soils34,35. Furthermore, N2O and N2 emission can be suppressed when N2O and N2 accumulate in gas bubbles entrapped within the water-saturated pore network34,72. For flooded soils it has been reported that gas sampling after physically disturbing the soil (e.g. by shaking the soil microcosm), resulted in up to 12 times more N2O and N2 than could be quantified in the headspace of undisturbed microcosms45,70,71. We also showed that physical destruction of the soil matrix by shaking and subsequent gas sampling led to significantly higher quantities of NO3-derived N2O and N2 in the microcosm headspace. Interestingly, the nitrogen gas entrapment in the soil (both NO3-derived N2O and N2) was significantly higher in the presence of biochar. It remains to be shown if this increase in soil gas entrapment is a direct biochar effect e.g. by sorption of N2O and N2 onto biochar particles41,42 or related to biochar-induced alterations of the soil pore structure and soil aggregation that resulted in changes of the soil hydraulic properties73,74.

The high relevance of N2O and N2 soil entrapment was also supported by nitrogen balance calculations. When only gas emission data from undisturbed microcosms was taken into account, a relatively high fraction of nitrogen could not be accounted for in the nitrogen balance calculations. However, this fraction decreased substantially when headspace gas and soil entrapped nitrogen gases (i.e. total production) were both considered. This indicates that NO3-derived N2O and N2 represented the major sinks of mineral nitrogen, even though a substantial fraction was not accounted for by headspace sampling from undisturbed microcosms. Due to the high entrapment rate and the higher NO3 consumption rates, biochar significantly increased total production rates of NO3-derived N2O and N2 (sum of entrapped and emitted gases). Additionally, NO3-derived N2O/(N2O + N2) ratios calculated based on total production rates were significantly lower in biochar-amended microcosms. These findings suggest that when total production rates and not only emission rates were considered, biochar actually increased denitrification-related nitrogen losses. However, as indicated by the lower product ratio, biochar addition significantly decreased the relative contribution of N2O to the total nitrogen gas loss.

Although in the present study biochar slightly increased napA, narG, nirK, nirS, typical nosZ and atypical nosZ gene copy numbers, the observed differences were mostly not statistically significant. This could be explained by the short incubation time and the fact that no carbon source was added. Thus no significant changes in the abundance of the denitrifying microbial community were observed in response to soil biochar amendment. No significant effect of biochar on gene copy numbers of functional denitrification genes has also been reported in other studies75,76. In contrast to our results on gene copy numbers, soil biochar amendment significantly increased nirK, typical nosZ, and atypical nosZ transcript copy numbers. Furthermore, narG and nirS transcript levels were elevated in biochar microcosms at most time points with significantly higher values only at certain dates. Higher numbers of nitrate, nitrite and nitrous oxide reductase transcripts suggest a higher denitrification activity in biochar-amended microcosms. This is in accordance with the higher gross NO3 consumption rates. Interestingly biochar addition most significantly increased typical and atypical nosZ transcript copy numbers indicating a biochar-induced promotion of N2O-reducing microorganisms. In consequence, the typical nosZ/nirS transcript ratio, i.e. the number of expressed nitrous oxide reductase genes per expressed nitrite reductase genes, was significantly higher in biochar microcosms. These results are in good agreement with other studies in which functional denitrification genes14,16,77 and their mRNA15,16 have been quantified and suggest that biochar addition enhances microbial N2O reduction. As reported in several studies nitrous oxide reductases are highly pH sensitive and are impaired due to post-transcriptional effects at pH values below 6.125,78,79. Obia et al.40 showed in soil slurry experiments with acidic soil that the addition of alkaline biochar significantly increased the pH of the slurries and decreased the N2O/(N2O + N2) product ratio. Based on their findings the authors concluded that due to the higher pH in biochar-amended slurries impairment of nitrous oxide reductases was less prominent, which resulted in a higher N2O reduction activity40. In the present study, however, we used a slightly alkaline soil. Measured pH values never dropped below 7.6 at any sampling time point of our microcosm experiment (data not shown). Although we cannot exclude that pH values were lower at specific microsites of the soil, for example due to the oxidation of NH4+ during nitrification80 it is unlikely that pH values dropped below 6.1 frequently. Therefore, we think that pH-driven impairment of nitrous oxide reductases did not represent a major controlling factor for N2O reduction activity in the present study.

In the present soil microcosm study, which was conducted with a slightly alkaline sandy clay loam soil and a wood-derived biochar, the observed biochar-induced N2O emission mitigation and lower N2O/(N2O + N2) ratio can potentially be explained by an interrelation of soil physical and microbiological parameters. Our results suggest that under the environmental conditions (N fertilization, high WFPS) of the present study, N2O soil entrapment is the main mechanism causing N2O emission mitigation. In biochar-amended microcosms we quantified 4.1 times more soil matrix entrapped NO3-derived N2O compared to the control microcosms. In addition, we determined an overall decrease of the N2O/(N2O + N2) ratio which was correlated to significantly increased typical and atypical nosZ transcript copy numbers and an increased typical nosZ/nirS transcript ratio, suggesting that soil biochar amendment enhanced N2O reduction activity of “classical” denitrifiers and atypical N2O reducers. Recent studies revealed that about half of the so far detected atypical nosZ gene carrying N2O reducers lack other functional denitrification genes and thus depend on the supply of N2O by other organisms32,33. These microorganisms as well as typical nosZ containing N2O reducers that specialized on N2O reduction81 might benefit from the entrapped N2O released by denitrifiers not capable of N2O reduction26, and further reduce it to N2. Support for this hypothesis comes from a recent sequencing study that revealed a biochar-induced increase in relative sequence abundance of microbial species carrying atypical nosZ genes that lack other denitrification genes and classical denitrifier species efficiently performing complete denitrification82. According to Clough et al.34 the potential for microbial reduction of N2O to N2 increases when N2O remains entrapped in the soil pore space. It is conceivable that an increase in N2O availability as consequence of the increased retention time of N2O in water-filled pores, stimulated microbial N2O reduction and thus decreased the N2O/(N2O + N2) ratio. We cannot exclude, however, that abiotic N2O reduction reactions with redox active organic compounds or organo-mineral phases high in Fe also contributed to the lower N2O/(N2O + N2) ratio in soil microcosms amended with biochar42.

In conclusion, this study confirmed that biochar amendment can significantly decrease N2O emissions from a slightly alkaline sandy clay loam soil under denitrifying conditions. Interestingly, N2O emission mitigation occurred although biochar addition stimulated denitrification gene expression, increased denitrification rates, and elevated total N2O and N2 production rates. Our data suggest that the lower N2O emissions from biochar-amended soil in the present study are caused by N2O entrapment in water-saturated soil pores and consequent stimulation of microbial N2O reduction by classical denitrifiers and atypical N2O reducers leading to a lower N2O/(N2O + N2) product ratio. Our findings emphasize the importance of considering soil entrapped nitrogen gases in biochar studies, especially under conditions of an elevated WFPS. However, the data shown here is based on a plant-free, short-term microcosm experiment with one specific sandy clay loam soil and a wood-derived biochar. The relevance of N2O entrapment for other soil-biochar combinations and the conditions under which any soil entrapped N2O might be emitted (e.g. extreme events such as drought or tillage) should be systematically evaluated in future lab and field studies.

Additional Information

How to cite this article: Harter, J. et al. Gas entrapment and microbial N2O reduction reduce N2O emissions from a biochar-amended sandy clay loam soil. Sci. Rep. 6, 39574; doi: 10.1038/srep39574 (2016).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.