The field-plot experiment was carried out in 2022–2024 at Anyang (36°11’51’’N, 114°20’56’’E), Henan Province, China ( Figure 1 ). The soil type is classified as a Fluvisols. The study site has an average elevation of 84.3 m above sea level, and the mean annual temperature and rainfall is 14°C and 557 mm, respectively. The soil pH was 7.57, and the contents of soil organic carbon (SOC), total N (TN), available phosphorus, and available potassium were 11.22 g kg⁻¹, 1.09 g kg⁻¹, 12.9 mg kg⁻¹, and 89.2 mg kg⁻¹, respectively.

Six treatments were applied in the delayed N application experiment during the 2022-2023 winter wheat growing season: a control group without N fertilizer application (CK); two rates of N fertilizer application, namely, 180 kg N ha⁻¹ (optimal N application; ON) and 270 kg N ha⁻¹ (farmer conventional N application; FN); ON + biochar at the rate of 15 t ha⁻¹ (ONB); ON + DMPP (OND); and ON + biochar + DMPP (ONDB). Five treatments were applied in the normal N application experiment during the 2023-2024 winter wheat growing season. The experimental treatments included CK, ON, ONB, OND, and ONDB. It is well established that elevated N application rates lead to increased N₂O emissions and higher emission factors, the experimental results from the first wheat-growing season fully support this conclusion. This study focuses on the environmental effects resulting from the integration of optimal N application (ON) with biochar or DMPP. Therefore, the FN treatment was excluded from the normal N application regime.

Urea was utilized as the N fertilizer applied in two distinct phases: 60% of the N fertilizer was applied during the first fertilization, and the remaining 40% was applied as topdressing ( Table 1 ). In addition, phosphate and potassium fertilizers, together with biochar, were applied on October 20, 2022 and October 16, 2023. Biochar was prepared from corn stalks at 450°C. The biochar C and N contents were 507 g kg⁻¹ and 2.1 g kg⁻¹, respectively, and the pH was 9.7. Each experimental plot has an area of 2 m × 2 m, with three replicates per treatment. Wheat seeds were sown on October 25, 2022, and October 16, 2023, while the mature grains were harvested on June 10, 2023 and June 4, 2024, respectively.

Nitrous oxide gas was collected in a sealed chamber, following the methodology described by Wu et al. (2024). Eighteen static opaque chamber bottoms were inserted into the soil within the study plot at 8 cm depth until the harvest of winter wheat. The static chambers were equipped with an electric fan and a thermometer on top. Gas samples were extracted from the chamber at four time points (0, 15, 30, and 45 min) using a 50 ml plastic syringe following its closure. Simultaneously, the temperature inside the static chambers was recorded. The electric fan operated continuously throughout the sampling process to maintain air homogeneity within the enclosed space. A total of 72 gas samples were collected each sampling day over a continuous 7-day period following N application; thereafter, the sampling was conducted at 7- to 10-day intervals.

N₂O flux was determined using a GC-2010 Plus gas chromatograph. The emission flux of N₂O (f) was calculated with the following formula:

where ρ (kg m⁻³) is the N₂O density, V is the volume of the sealing chamber (m³), A is the bottom area of the chamber (m²), ΔC/ΔT (μL L⁻¹ h⁻¹) is the temporal variation in N₂O concentration in the sealed chamber, and T (C) is the mean temperature inside the chamber. The cumulative emission of nitrous oxide (CE-N₂O) was estimated by employing linear interpolation of N₂O flux and time (Allen et al., 2010).

The N₂O emission factor was calculated as follows:

where CE_{N fertilizer} and CE_{no N fertilizer} represents CE-N₂O from treatments with N application and without, respectively. N_input represents the amount of N fertilizer applied.

Fresh soil samples were collected subsequent to gas sampling for determination of the ammonia-N (NH₄ ⁺-N) and Nitrate-N (NO₃ ⁻-N) contents, which were extracted using 2 M L⁻¹ KCl solution and subsequently determined with a flow analyzer. Soil samples collected at harvest were used to determine physicochemical parameters. Dissolved organic N (DON) was measured by subtracting NH₄ ⁺-N and NO₃ ⁻-N from the total soluble N (TSN) content. Soil bulk density, pH, TSN, SOC), and TN were determined following soil agricultural chemistry analysis (Lu, 2000). Each sample is analyzed in duplicate, and the relative deviation between the duplicate samples must not exceed 5%. The soil water-filled pore space (WFPS) was determined as follows.

Where, θv represents the volumetric water content, ρ represents for soil bulk density.

Total DNA from 0.5 g soil samples was extract by using the E.Z.N.A.^® Soil DNA Kit (Omega Bio-Tek, USA). The purified DNA was stored at −80°C until analysis. Gene copy numbers were determined using fluorescent qPCR (ABI 7500, USA) following protocols described by Huang. Information on the primers used and the parameters for the qPCR reactions are presented in Supplementary Table S1 .

The qPCR products were identified, purified, and quantified using 2% agarose gel electrophoresis, the AxyPrep DNA Gel Extraction Kit, and a Quantus™ Fluorometer, respectively. The NEXTFLEX Rapid DNA-Seq Kit was used to construct a DNA library, which was sequenced using an Illumina platform (NovaSeq PE250). The initial sequences were subsequently refined and concatenated to yield high-quality sequences. The UPARSE software was employed for clustering of operational taxonomic units (OTUs), following the clustering protocols and methodologies outlined by Bi et al. (2023). The RDP Classifier was used to annotate the species classification of the sequences, and the classification information for each OTU was derived by comparison with the Silva 16S rRNA database. The UCLUST algorithm was used for further taxonomic analysis of the representative OTU sequences.

The data were analyzed statistically with SPSS version 25.0. ANOVA was employed to assess the significance of differences among the indexes, post hoc tests were executed using the LSD, with a significance threshold of P < 0.05. Redundancy analysis was conducted using canoco 5.0 to examine the relationships between soil indicators and microbial communities. Correlation analysis was conducted using Origin 2021 software. A random forest analysis was conducted to identify microbial genera that significantly influence N₂O emission using the ‘rfPermute’ package for R.

The temporal dynamics of N₂O emission showed discernible fluctuations. N fertilizer application greatly stimulated soil N₂O emission. A distinct difference in soil N₂O emissions between the first and second N applications was observed ( Figure 2 ). The N₂O flux increased markedly following the second N application. The N application treatments exhibited significantly higher N₂O emissions compared with those of the CK. Under the delayed N application regime, the highest CE-N₂O (1.96 kg ha⁻¹) was observed under the FN treatment. The ONB treatment (1.47 kg ha⁻¹) led to a slightly higher CE-N₂O than in the ON treatment (1.42 kg ha⁻¹) ( Figure 2A ). Addition of DMPP resulted in significant reduction of N₂O emissions; the OND and ONDB treatments exhibited reductions in CE-N₂O of 32% and 38%, respectively, compared with the CE-N₂O of the ON treatment. The N₂O emission factors for different treatments under the delayed N application regime ranged from 0.23% to 0.56%, which were substantially lower than the factors ranging from 0.60% to 1.03% under the normal N application regime ( Figure 2B ). These results indicate that implementing a delayed N application strategy can effectively mitigate N₂O emissions.

Soil NO₃ ⁻-N and NH₄ ⁺-N contents were significantly increased following application of N. The NH₄ ⁺-N content under the ONB treatment was lower than the ON treatment following N fertilization. The NO₃ ⁻-N content in response to DMPP application (the OND and ONDB treatments) was comparatively low ( Figures 3A, B ). The WFPS ranged between 16.58% and 71.46%, exhibiting similar tendencies under the different treatments ( Figure 3C ). According to the average inorganic-N content of the soil from the initial N application until the wheat harvesting period, the FN treatment resulted in the highest NH₄ ⁺-N and NO₃ ⁻-N contents ( Table 2 ). The ONB treatment decreased NH₄ ⁺-N content and significantly increased NO₃ ⁻-N content, whereas treatment with DMPP (OND and ONDB) led to a significant increase in NH₄ ⁺-N content and a significant decrease in NO₃ ⁻-N content. The NO₃ ⁻-N content differed significantly among the treatments, except for OND and ONDB. Biochar amendment significantly enhanced the soil pH and SOC content, compared with the ON treatment; the ONB and ONDB treatments increased pH by 3% and 5%, respectively, and SOC by 7% and 8%, respectively. Application of DMPP led to slight, but non-significant, increases in soil SOC and pH. The soil DON concentration increased markedly with increase in the N application rate. The ONB and ONDB treatments slightly enhanced the soil DON concentration, whereas the OND treatment had the opposite effect.

Ammonia-oxidizing and denitrifying bacteria exhibited distinct variation in abundance among the treatments ( Table 3 ), as indicated by the copy numbers of microbial functional genes. The CK, ON, and FN treatments exhibited significant elevation in AOB gene copy numbers with increasing N input, whereas no notable differences in AOA were observed. In comparison with the ON treatment, the ONB treatment markedly enhanced the abundance of AOB gene copies, whereas ONDB treatment had the most pronounced effect in reducing AOB gene copy numbers. The ONB treatment significantly enhanced the AOA abundance, whereas the OND and ONDB treatments had no significant effect. The ratio of AOA to AOB gene copy numbers was smallest under the ON treatment (0.35) and largest under the ONB treatment (0.69). The quantity of nirS, nirK, and nosZ gene copies declined with increase in the N application rate. Relative to the ON treatment, the ONB treatment substantially increased the nirS abundance, whereas the OND treatment significantly increased the abundance of nirK ( Table 3 ). The highest number of gene copies was observed for nirK, whereas the lowest number of copies detected was for nirZ. The nirK/nosZ ratio varied between 1.27 and 1.53, with no statistically significant differences among treatments detected.

To graphically illustrate the effects of different treatments on the N cycling microbial community, hierarchical clustering analysis of OTUs was conducted across treatments ( Figure 4 ). Based on the OTU clustering results for AOB, OND and ONDB were initially grouped before being clustered with CK, whereas ON and ONB were preferentially grouped and then clustered with FN. The CK, ON, and FN treatments were grouped into distinct clusters, indicating that the different N fertilizer rates significantly influenced the soil AOB community. It is noteworthy that the cluster heatmaps for the AOA, AOB, and nirK-containing communities revealed preferential combination of the OND and ONDB treatments ( Figures 4A–C ), indicating that DMPP application altered the composition of the amoA- and nirK-containing communities. The ON and OND treatments were initially grouped, suggesting that DMPP application alone had a minimal impact on the nirS- and nosZ-containing communities ( Figures 4D, E ). In contrast, the different N application treatments (CK, ON, FN) were grouped into separate clusters, suggesting that N application significantly influenced the nirS- and nosZ-containing communities.

Based on the microbial communities at the genus level ( Figure 5 ), the AOA community structure was relatively simple, with Candidatus Nitrosocosmicus identified as the dominant genus ( Figure 5A ). The AOB community was primarily composed of the genera Nitrosospira and Nitrospira at the genus level ( Figure 5B ), with Nitrosospira exhibiting the highest relative abundance. Application of N or biochar decreased the relative abundance of Nitrospira, whereas DMPP had a stimulatory effect on Nitrospira abundance. Notably, combined application of DMPP and biochar led to a more pronounced stimulation of Nitrospira abundance. The composition of the nirK-, nirS-, and nosZ-containing communities exhibited greater diversity at the genus level ( Figures 5C–E ). A high rate of N input or biochar application resulted in a shift of the most dominant nirS-containing genus from Bradyrhizobium to Pseudomonas. The most dominant genus for the nirK- and nosZ-containing communities was Bradyrhizobium. Biochar amendment markedly enhanced the relative abundance of Ramlibacter in the nosZ-containing community, resulting in a shift of the most dominant genus to Ramlibacter under the ONB treatment.

Redundancy analysis was conducted to examine the inter-relationships among N₂O emission, soil physicochemical properties, and microbial gene abundance. Axes 1 and 2 accounted for 67.99% of the total variance ( Figure 6 ). The ON, ONB, and DMPP addition treatments were resolved as distinct on axis 1 (49.80%). The contents of NO₃ ⁻-N and inorganic N, DOC, pH, and TN were critical indicators that influenced the experimental system. Correlation analysis indicated that NO₃ ⁻-N, inorganic N, AOB, nirK, DON, and NH₄ ⁺-N were key indicators that influenced soil N₂O emission ( Figure 7 ). Random forest analysis further revealed that Nitrospira was the genus within the AOB community that most significantly affected N₂O emission, Cronobacter was the dominant genus responsible for N₂O emission in the nirK-containing community, whereas Ramlibater and Methylobacillus in the nosZ-containing community were the significantly predictors of N₂O emission ( Figure 8 ).

The N₂O released from agricultural soils is a byproduct of nitrification and denitrification. Carbon and N play crucial roles influencing the emission of N₂O (Cayuela et al., 2014; Li et al., 2022). Our findings indicate that the co-application of biochar and DMPP caused the most effective inhibition of N₂O emission. Compared with the ON treatment, the ONDB treatment led to a 38% reduction in CE-N₂O. This effect is attributed to the substantial decrease in content of the denitrification substrate (NO₃ ⁻-N) under the ONDB treatment, which consequently inhibited the activity of denitrifying bacteria. The regression analysis revealed a significant negative correlation between pH and CE-N₂O (P < 0.01, Supplementary Figure S1 ). The ONDB treatment significantly increased the soil pH, which may be an additional factor that contributes to the synergistic effects of biochar and DMPP in mitigating N₂O emission. Therefore, we proposed that DMPP’s inhibition of AOB combined with biochar’s pH modulation jointly reduced N₂O. Recent studies have demonstrated that biochar has a markedly superior capacity for N₂O adsorption compared with soil and its mineral constituents (Xiao et al., 2018). Consequently, the ONDB treatment may enhance the adsorption and stabilization of specific N₂O molecules within the biochar matrix. However, Li et al. (2023) reported that combination of biochar and DMPP did not lead to a significant reduction in N₂O emissions in agricultural systems, which may be attributable to regional soil characteristics and the intrinsic properties of biochar.

Biochar application alone resulted in an increase in soil CE-N₂O compared with that of the ON treatment. The ONB treatment significantly enhanced the soil NO₃ ⁻-N content ( Table 2 ), indicating that biochar incorporation significantly enhanced soil nitrification, consistent with the findings of Chen et al. (2019). Previous research has demonstrated that biochar is abundant in various volatile compounds and serves as an organic C source for denitrifying bacteria (Fu et al., 2022), thereby stimulating N₂O emission. The present study revealed that the ONB treatment significantly increased the DOC compared with the ON treatment, thereby confirming that the incorporation of biochar (an exogenous source of organic C) enhanced soil denitrification (Weldon et al., 2019). Furthermore, biochar application significantly enhanced the SOM, thereby increasing the availability of C and N within the soil (Chagas et al., 2022). This enhancement fosters elevated diversity and activity of soil microorganisms, leading to increased oxygen consumption (Xu W. et al., 2023), and as a result, localized anoxic conditions are more conducive to denitrification.

Li et al. (2023) reported DMPP decreases N₂O emissions by disrupting the N conversion processes within the soil, ultimately causing reduced availability of N for nitrification and denitrification. A experiment conducted by Zhao on a wheat–maize rotation system demonstrated that DMPP significantly mitigated soil N₂O emissions, consistent with the present findings. The current study demonstrated that DMPP application resulted in a significant 32% reduction in CE-N₂O compared with that of the ON treatment, which was largely consistent with the results of a meta-analysis of agricultural systems conducted by Ekwunife et al. (2022). Nevertheless, this reduction was markedly less than in lab experiments (Fan et al., 2019).

Fluctuations in air temperature and precipitation affect soil aeration and oxygen concentrations, which subsequently impact on N₂O production (Yang Y. et al., 2021); in addition, the redox environment of the soil plays a critical role (Xu P. et al., 2023). In the present study, a notable increase in soil N₂O flux emissions was observed following the second N application. Fluctuations in soil moisture, in combination with optimal surface-soil temperatures ranging from 19 to 27°C, resulted in frequent cycles of drying and wetting within the soil environment ( Supplementary Figure S2 ). This dynamic created alternating conditions of oxidation and reduction, which further enhanced the denitrification process facilitated by both nitrifying and denitrifying bacteria, thereby increasing N₂O emissions. Theodorakopoulos et al. (2017) reported that N₂O emissions were predominantly attributable to nitrification at WFPS < 60%, and by denitrification at WFPS > 60%. The soil WFPS ranged between 17% and 71% in the present study. Consequently, it is probable that N₂O emissions from the soil primarily originated from the nitrification pathway. Significant correlations between NO₃ ⁻-N, inorganic N, NH₄ ⁺-N, and CE-N₂O were observed, which is inconsistent with the findings of Huang et al. (2019). We propose that the N₂O emissions observed in the present study primarily originated from oxidation of soil ammonia, particularly through hydroxylamine decomposition. Furthermore, the notable positive correlation between soil DON and CE-N₂O reinforces that DON-mediated heterotrophic ammoxidation may serve as a pivotal contributor to N₂O production. However, the findings of this study were derived from two wheat growing seasons, and the long-term efficacy of combined biochar and DMPP application in mitigating N₂O emissions remains to be confirmed.

The activity of the nitrifying bacterial community is significantly influenced by the soil environment (Zheng et al., 2019). AOA and AOB display distinct adaptations to soil NH₄ ⁺ environments, with AOB predominant under elevated NH₄ ⁺ conditions, whereas AOA exhibits the opposite trend (Fang et al., 2023). Previous investigations have revealed that fertilizer application markedly increases nitrifying microbial activity in the soil, leading to elevated N₂O emissions (Li et al., 2020). The microcosmic examination of ammonia-oxidizing processes under different N fertilization regimes is rather complex, owing to the participation of a diverse array of ammonia-oxidizing microorganisms (Yang L. et al., 2021).

The present study detected a positive correlation between N₂O emission and the abundance of AOB, which in turn increases with elevation of the N application rate. This result accords with a meta-analysis of 157 field observation datasets conducted by Ouyang et al. (2018). However, their study indicated that the abundance of AOA increases in response to N application. The present findings showed that biochar application significantly enhanced the abundance of both AOA and AOB, consistent with previous research demonstrating that biochar stimulates nitrification activity and fosters the proliferation of ammonia-oxidizing microorganisms (Xu W. et al., 2023). Application of DMPP mitigated the impact of N fertilization on the abundance of AOB. Furthermore, the synergistic effect of biochar and DMPP significantly decreased the abundance of AOA and AOB. The copy number of AOA genes was lower than that for AOB genes (AOA/AOB=0.52) at wheat harvesting ( Table 3 ). These findings indicate that AOB may exhibit greater abundance and demonstrate enhanced ammoxidation activity in agricultural soils with elevated N contents (Yan et al., 2018). With regard to OTU clustering within AOB, the OND and ONDB treatments were clustered and subsequently linked with CK to form a single cluster. The ONB and ON treatments were preferentially linked before being grouped with FN to establish a distinct cluster. This finding elucidates the variation in CE-N₂O under the different treatments. Redundancy analysis indicated that AOB was the most significant positive factor that influenced N₂O emissions, while correlation analysis revealed that AOB contributes substantially more to N₂O emissions than AOA.

The impact of AOB on N₂O emissions is significantly greater than that of AOA. Further identification of specific microorganisms within AOB that modulate N₂O emissions is warranted. Cytryn et al. (2012) amplified amoA gene fragment and revealed that the AOB community in paddy soil is predominantly composed of Nitrosomonas. The relative abundance of Nitrosospira decreased compared with Nitrosomonas as N application increased., as the addition of N promotes a shift from a less nutrient-rich bacterial community to a more symbiotic community. However, high-throughput sequencing revealed that the predominant genus of AOB was Nitrosospira in this study and that the proportion of Nitrosospira rose with increase in the N application rate, whereas Nitrospira showed an inverse relationship. Similarity, Bi et al. (2023) reported that Nitrosospira is the predominant ammonia-oxidizing genus in agricultural soils. Liu et al. (2021) identified Nitrosospira as the dominant genus in environments with high NH₄ ⁺ concentrations, demonstrating an enhanced capacity for ammonia oxidation. This genus plays a crucial role in N₂O emissions from soils characterized by high concentrations of NH₄ ⁺. In the current study, the application of biochar alone (ONB) significantly enhanced the proportion of Nitrosospira compared with the ON treatment, while concurrently reducing Nitrospira abundance. However, Lin et al. (2017) demonstrated that exogenous organic C altered the AOB community, shifting from Nitrosospira to Nitrosomonas.

We propose that the primary factor contributing to this discrepancy is soil pH. In acidic conditions, the nitrification activity of Nitrosomonas surpasses that of Spirulina Nitrosomonas; however, the adaptability of Nitrosomonas to alkaline environmental stress is significantly lower than that of Spirulina Nitrosomonas. Nitrospira can oxidize nitrite and convert urea into ammonia, promoting the growth of nitrifying bacteria. Although its abundance remains relatively stable with increasing N application rates, it exhibits a significant correlation with N₂O emissions (Liu et al., 2024). Likewise, in the present study, treatments that included DMPP were observed to enhance the relative abundance of Nitrospira, whereas the ONB and FN treatments led to a decrease in its relative abundance. Random forest analysis further demonstrated that Nitrospira exhibit a significant predictive capacity for N₂O emissions. Additionally, in comparison with the ON treatment, the combination of biochar and DMPP significantly reduced the α-diversity indices ( Supplementary Table S2 ), indicating that ONDB treatment significantly reduced AOB richness and diversity. Consequently, we propose that DMPP exerts an inhibitory effect on N₂O emissions by diminishing the abundance and α-diversity of AOB, as well as by increasing the relative proportion of Nitrospira within the AOB community.

Denitrification, mediated by heterotrophic bacteria and fungi, primarily occurs in anaerobic environments, where nitrate undergoes a series of transformations (e. g. NO₂ ⁻, NO, N₂O) that ultimately yield N₂. Previous studies have established that denitrifying bacteria harboring nirS and nirK are the primary contributors to denitrification-mediated N₂O production, primarily because of inadequate genetic capacity for the reduction of N₂O (Huang et al., 2019). The nirK and nosZ copy numbers under the ONB treatment were lower than those detected under the ON treatment ( Table 3 ). Biochar application can foster a soil environment with an elevated C/N ratio, thereby promoting N assimilation within the soil and diminishing the availability of N substrates for denitrification (Liu Z. et al., 2021). The OTU cluster heatmaps revealed that the nirS-, nirK-, and nosZ-containing communities exhibited distinct responses to the various treatments. Our findings indicated that DMPP treatment (OND and ONDB) significantly altered the composition of nirK-containing communities ( Figure 4C ); however, DMPP application alone exhibited limited effects on the nirS- and nosZ-containing communities ( Figures 4D, E ). Conversely, biochar application alone had a pronounced impact on the nirS- and nosZ-containing communities ( Figures 4D, E ). Xiao et al. (2021) reported that application of exogenous N or C influences the soil microbial community, while the simultaneous addition may alter the dominant genera within the denitrification gene community. Similarly, in the present study, we observed a shift in the most dominant genus of nirS-containing community from Bradyrhizobium under the ON treatment to Pseudomonas under the FN and ONB treatments. The most dominant genus of nosZ-containing community transitioned from Bradyrhizobium to Ramlibacter under the ONB treatment. Bradyrhizobium is an aerobic azotobacter within the rhizobia order (Geddes et al., 2020), and was the most dominant genus in the nosZ-containing community in all treatments except for the ONB treatment. This finding emphasizes the possibility for denitrification within an aerobic environment. Both nirK and nosZ were significantly correlated with DON. The correlation coefficient between SOC and nirS exceeded that of SOC with the other denitrification genes (nirK, nosZ) ( Figure 7 ), indicating that nirK- or nosZ-containing bacteria showed heightened sensitivity to exogenous N, whereas nirS-containing bacteria exhibited greater sensitivity to exogenous C compared with that of nirK- or nosZ-containing bacteria.

In this study, CE-N₂O was significantly negatively correlated with the nirK gene copy number (P < 0.01). This finding is consistent with the conclusions by Wu et al. (2023), who established that the presence of nirK-containing denitrifying bacteria are critical determinants in both N₂O consumption and production. Random forest analysis identified Cronobacter as a critical genus driving N₂O emissions in nirK-containing communities. Similarly, Xie et al. (2024) reported that substitution with organic fertilizers affected the relative abundance of Cronobacter in the nirK-containing community, thereby mitigating N₂O emissions. Interestingly, our findings demonstrate that despite the lack of a markedly correlation between CE-N₂O emissions and nosZ gene copy numbers, while a significant negative correlation was observed with the nirK/nosZ ratio (Shi et al., 2019). Furthermore, the present findings revealed that Ramlibater and Methylobacillus in the nosZ-containing community exhibited a significant predictive capacity for N₂O emissions. Consequently, the nosZ gene may still be among the key contributors to N₂O emissions mediated by denitrification. Tang et al. (2024) reported that niche variations in the denitrification genes nirK and nirS resulted in differential N₂O emissions. However, a weak correlation was observed between CE-N₂O and nirS gene copy number in the present study, indicating that nirS was not the most critical factor influencing N₂O emissions. Nevertheless, correlation analysis revealed a significant correlation between AOB and nirK copy number with N₂O emissions; however, this does not provide direct evidence for microbiome-mediated N₂O emission. Therefore, it is essential to conduct a more in-depth analysis in future to elucidate the contributions of various microorganisms to N₂O emissions using microbial molecular ecology and isotope tracer methodologies.