Composite samples of SF and SC (photos of them at one site are shown in Supplementary Figure S1) were collected from a plateau of a permafrost mound of a palsa mire near Kilpisjärvi (68°43′; 21°25′), northwestern Finland (Figure 1A). Each sample of SC/SF was formed from three random sampling sites with three replicates in August–September, 2014. SC and SF were collected from the same patch (within 50–100 cm) and the sampling sites were 50 to 100 m away from each other. From each sampling site, random 533 to 565 individual plants of either SC or SF were collected and mixed for each sample in order to guarantee the sample's representation. Both SC and SF were collected from large homogenous stands with a 40 cm thaw layer above the permafrost surface. The region has a low annual mean temperature (−2.3°C) and moderate mean annual precipitation (487 mm). The growing season is one of the shortest in continental Europe (~100 d when the mean daily temperature is ≥5°C). The Sphagnum samples stored in Ziploc® bags at 4°C were used for further culture-based N₂O emission measurements.

To evaluate the potential for N₂O emission of the two Sphagnum mosses under an experimental nitrogen load, we took three Sphagnum mosses plants (~0.1 g in dry weight) randomly from the respectively, composite sample of SC and SF using sterilized tweezers. At the same time, we standardized the dry weight for the N₂O assay. Either 100 μL of Sphagnum moss leaf extract (100 mg/10 ml) or 3 plants were added to N₂O assay medium [10 ml of modified Winogradsky's Gellan (MWG) medium containing 0.005% yeast extract and solidified with 3% gellan gum with 22.6 ml of headspace in each vial (30 ml gas-chromatographic vial with a butyl rubber plug) (Nichiden-Rika Glass Co., Kobe, Japan)] with 0.05% sucrose diluted with sterilized Milli-Q water (the solution was adjusted to pH = 5.0 with 2 M H₂SO₄) (three replicates in each case) (Hashidoko et al., 2008). After incubation at 15°C (according to the mean value of summer temperature of Finland) for 7 days in the dark, an N₂O assay was carried out by using an electron capture detector(ECD)-gas chromatograph (Shimadzu GC-14B, 125 Kyoto, Japan) connected to a Porapak N column (1 m long, Waters, Milford, MS, USA). In another treatment, 0.1 g L⁻¹ of caffeic acid instead of 0.05% sucrose was added as the carbon source to the vials with three plants (~0.1 g in dry weight) randomly taken from the above composite samples (pH 5). A control for the assay, without any carbon source, was also performed simultaneously (three replicates in each case). After incubation at 15°C in the dark for 4, 8, and 15 days, an assay of N₂O was performed as mentioned above.

Polymerase chain reaction-denatured gradient gel electrophoresis (PCR-DGGE) was performed to observe the culture-based bacterial communities on the leaves of the two Sphagnum mosses. First, genomic DNA was extracted from the medium after the N₂O assay using an Isoplant II DNA Extraction kit (Nippon Gene, Toyama, Japan). The PCR steps and conditions were as follows: PCR denaturation for 5 min at 95°C, and 30 cycles of amplification (15 s at 95°C, 30 s at 55°C, 30 s at 72°C), and 10 min elongation at 72°C. Then PCR products for DGGE were obtained by using the common 16S rRNA primers GC-341F (CGC CCG CCG CGC CCC GCG GGG GTC CCG CCG CCC CCG CCC GCC T AC GGG AGG CAG CAG) and 907R (CCG TCA ATT CCT TTR AGT TT) (Ferris et al., 1996) and run on a 30–70% denatured gradient gel (6% w/v). The sequences of DGGE-cutting bands were obtained using an ABI prism^TM 310 Genetic Analyzer and retained in the NCBI (BioProject No. PRJNA681491).

100 μl of medium with three Sphagnum mosses (after incubation for 7 days) was diluted 1× 10⁴- and 10⁶-Fold and inoculated onto MWG plates to screen N₂O emitters. After incubation for 5 days at 20°C in the dark, 13 distinguishable bacterial colonies characterized by colony characteristics were selected for streak cultivation on MWG plates and transferred to potato dextrose agar (PDA) plates until purified. Each Pure strain [a total of 108 isolates (13 bacterial colonies with 8 replicates), with 100 μl of each bacterial cell suspension (OD_660nm = 0.9–1.0)] was inoculated into an N₂O assay vial with 10 ml of modified MWG medium to test their N₂O emission ability. The three pure strains SC-K1, SC-L1, and SC-H2 (from SC) showed relatively higher N₂O production and were active N₂O emitters (Supplementary Table S2, data collected from six top active N₂O emission-bacterial colonies). The genomic DNA of each strain was extracted, and the 16S RRNA gene was amplified through PCR by using a series of primers 27F, 338R, 341F, 907R, 1080R, 1380R, 1492R, 1112F, and 1525R. Sequencing was performed with an ABI Prism^TM 310 Genetic Analyzer (Applied Biosystems, USA) (Nie et al., 2015). All the resulting 16S RRNA gene sequencing datasets were deposited in the NCBI database (accession nos. MW301596–MW301598) and compared with sequences in the nucleotide basic local alignment search tool (BLASTN) database program provided by NCBI (National Center of Biotechnology Information, Bethesda, MD, USA; http://Blast.Ncbi.nlm.nih.gov/Blast.cgi).

The pure isolates (SC-K1, SC-L1, and SC-H2) pre-cultured on PDA for 4 days at 15°C were separately scraped with a nichrome wire loop and suspended into 1.5 ml Milli-Q water (equal amounts of each pure strain was guaranteed). A 20 μl portion of the inoculant that showed an optical density of OD_660nm 0.9–1.0 was added to the N₂O assay vial and then was thoroughly vortexed for 30s. 1 mM NH₄NO₃, KNO₃, and NH₄Cl were tested and incubated at 15°C for 5 days with 0.05% sucrose (pH = 5.0) to determine the optimal nitrogen substrates for pure N₂O emitters. The pH was adjusted with 1 M H₂SO₄ and 1 M KOH solutions to 4.6, 5.0, 5.7, 6.8, and 7.3 before autoclaving and incubated at 15°C for 5 days with 0.05% sucrose to determine the optimal pH for N₂O emitters. Different temperatures (4, 10, 15, 20, 25, and 30°C) were set in separate incubators and incubated for 5 days with 0.05% sucrose to find the appropriate temperature. All experiments were performed with three replicates.

Sucrose and E-caffeic acid were applied as carbon sources and secondary metabolites (polyphenols), respectively, for the microbiota inhabiting Sphagnum moss (Nie et al., 2015). The inoculants were prepared as described in Nie et al. (2015). To observe the responses of the N₂O emitters (SC-K1, SC-L1, SC-H2) to sucrose, 0 (control), 0.05, and 0.5% sucrose were used for the separated/cultivated bacterial strains. To determine the optimal concentrations of E-caffeic acid for N₂O emitters (SC-K1, SC-L1, SC-H2), 0 (control), 0.005, 0.01, 0.05, 0.1, 0.5, and 1 g L⁻¹ E-caffeic acid were used. Each treatment contained three analytical replicates incubated at 15°C for 5 days with inoculants for N₂O assays. Their N₂O emissions were separately measured.

We applied the acetylene inhibition assay, which is widely used to measure denitrification rates (Sørensen, 1978). The activity of N₂O reductase was inhibited by adding acetylene (C₂H₂) at pH 5.0 and 7.0, and 10% C₂H₂ gas was injected into the headspace of vials inoculated with N₂O emitters (the same with above inoculation method) (Bollmann and Conrad, 1997). At the same time, treatments without injected C₂H₂ gas were carried out as controls to compare the N₂O reductase activity (three replicates in each case). Incubation conditions were the same as described above.

Functional genes of nitrogen cycling, including narG, nirK, nirS, and nosZ (Supplementary Figure S4), were detected by using the PCR method. The target genes were amplified by using the primers narGF (TCG GGC AAG GGC CAT GAG TAC) and narGR (TTT CGT ACC AGG TGG CGG TCG), nirSCd3Af (AAC GYS AAG GAR ACS GG) (Nie et al., 2015) and nirSR3cd (GAS TTC GGR TGS GTC T) (Throbäck et al., 2004), nirK-1F (GGM ATG GTK CCS TGG CA) and nirK-5R (GCC TCG ATC AGR TTR TGG) (Braker et al., 1998), nosZ-1111F (STA CAA CWC GGA RAA SG), nosZ-661F (CGG CTG GGG GCT GAC CAA), nosZ-1527R (CTG RCT GTC GAD GAA CAG), and nosZ-1773R (ATR TCG ATC ARC TGB TCG TT) (Scala and Kerkhof, 1998). The exact reaction conditions of the PCR amplifications are presented in Supplementary Table S1.

The data were expressed as mean with standard error (SE). The data were examined for normality and homoscedasticity using the Shapiro-Wilk's and Levene's tests, respectively (SPSS, version 23.0). All data was found to fit the normal distribution and homogeneity of variances. Comparisons were made using a one-way analysis of variance (ANOVA) among two or more groups. One-way ANOVA was used to compare differences in N₂O emission with different inoculants (Sphagnum mosses or their leaves washing), physicochemical factors [pH, temperature, sucrose, nitrogen types, and secondary metabolite (E-caffeic acid)], and C₂H₂ inhibition assay. Using the Fisher's Least Significant Difference(LSD) method, multiple comparisons were carried out using IBM SPSS 23.0 software (Chicago, Illinois, USA).

After incubation for 7 days, we found that the average N₂O emissions of SF were 1.9 ng vial⁻¹ d⁻¹ in the leaf extract and 69.9 ng vial⁻¹ d⁻¹ in the leaf samples. The SC sample showed N₂O emissions of 9.1 in the leaf extract and 956.2 ng vial⁻¹ d⁻¹ in the leaf samples (Figures 1B,C).

The PCR-DGGE profile showed that the major culture-based bacterial communities in these Sphagnum mosses were similar. However, the SC sample harbored the family Enterobacteriaceae (Figure 1D, Supplementary Figure S2), while the SF sample contained the genus Dyella of Gammaproteobacteria (Supplementary Figure S2). N₂O production increased with 0.1 g L⁻¹ caffeic acid addition in both samples, and the effect was significantly larger in the SC sample than in the SF sample (p < 0.05) (Figures 2A,B).

Compared to PCR-DGGE, the culture-based approach revealed distinctive profiles of N₂O emitters (Supplementary Figure S2). Two Burkholderia spp. were isolated from the SF sample, while three Gammaproteobacteria (one Pseudomonas sp., one Serratia sp., and an unidentified Enterobacteriaceae bacterium) and one Burkholderia sp. were isolated from the SC sample. Among them, Serratia sp. SC-K1, Enterobacteriaceae bacterium SC-L1, and Pseudomonas sp. SC-H2 showed the most efficient N₂O emissions, and the activity of N₂O emissions was the greatest in Pseudomonas sp. SC-H2, then Enterobacteriaceae bacterium SC-L1, and then Serratia sp. SC-K1 (pH 5) (Table 1, Supplementary Table S2).

According to the N₂O production responses to different nitrogen sources, KNO₃ was the most efficient substrate for N₂O emission, followed by NH₄NO₃, while almost no N₂O emissions were found with NH₄Cl as the substrate. Active N₂O emissions from KNO₃ indicated that the three N₂O emitters were nitrate reducers (Figure 3). N₂O emissions increased as the pH increased from 4.6 to 7.3. Enterobacteriaceae bacterium SC-L1 and Serratia sp. SC-K1 showed a temporary increase at a pH value of 5 but no drastic increase in N₂O emissions, indicating adaptation to acidic environments (Figures 4A,B). At pH values over 6, Pseudomonas sp. SC-H2 emissions increased sharply, making it the most likely N₂O emitter (Figure 4C). For the three strains used, N₂O emissions also increased with increasing temperature from 4 to 30°C (Figures 4D–F).

The three microbial strains exhibited disparate responses to sucrose and E-caffeic acid (Figure 5). In the absence of added sucrose (control treatment), Serratia sp. SC-K1 emitted more N₂O than Enterobacteriaceae bacterium SC-L1 and Pseudomonas sp. SC-H2, while these last two strains emitted N₂O at higher levels with 0.05% sucrose supplementation (Figures 5A,B). Notably, the response of Pseudomonas sp. SC-H2 to 0.05% sucrose was very drastic, resulting in emission ~2x10³ times higher than without sucrose (Figure 5C). This result demonstrated that Serratia sp. SC-K1 is an oligotrophic bacterium, whereas Enterobacteriaceae bacterium SC-L1 and Pseudomonas sp. SC-H2 are eutrophic bacteria.

For the pure strains of Enterobacteriaceae bacterium SC-L1 and SC-K1, a relatively lower concentration of E-caffeic acid ( ≤ 0.1 g L⁻¹) increased N₂O emissions of these two strains, and the optimum concentration was 0.1 g L⁻¹ (Figures 5D,E). Among them, Serratia sp. SC-K1 was very sensitive to 0.1 g L⁻¹, and 13-fold higher N₂O production was found than without E-caffeic acid (Figure 5E). For Pseudomonas sp. SC-H2, when the concentration of E-caffeic acid was above 0.01 g L⁻¹, N₂O emissions decreased significantly (p < 0.01) (Figure 5F).

There was no detectable difference between the 10% C₂H₂ and control treatment emissions at a pH value of 5.0. However, in Pseudomonas sp. SC-H2 cultured at a pH value of 7.0, N₂O emissions upon exposure to C₂H₂ were drastically increased to four-fold higher than that of the control. Without 10% C₂H₂, the production level of N₂O at a pH value of 7.0 was higher than that at a pH value of 5.0 (Figure 6). This result suggested that the peat ecosystem was highly disturbed at a pH value of 7.0, denitrification was greatly accelerated, and the final denitrification step to reduce N₂O to N₂ was driven by N₂O reductase.

PCR assays detected the narG gene in the three N₂O emitter strains, but only Pseudomonas sp. SC-H2 contained nirS and nosZ genes (Table 2; Supplementary Figure S3). In combination with the results of the C₂H₂ assay, these results suggested that Pseudomonas sp. SC-H2 is a complete denitrifier. The nirK gene was not detected within Enterobacteriaceae bacterium SC-L1 and Serratia sp. SC-K1.

Increased atmospheric N deposition can reduce the growth of some Sphagnum species, such as Sphagnum magellanicum (Aerts et al., 2001; Limpens and Berendse, 2003). In contrast, the production of SF increased with elevated N deposition but decreased as N deposition reached 14.0 kg ha⁻¹ yr⁻¹ as reported by Vitt et al. (2003). SC can also tolerate a high N supply (Bonnett et al., 2010). Our study offered evidence that individual samples of the latter two Sphagnum species had N₂O emission potential reasonably associated with their bacterial communities. In particular, the SC sample harbored specific bacterial communities associated with high N₂O emission. Surprisingly, the N₂O emission of the SC sample was significantly greater than that of the SF sample (Figure 1B) (p < 0.01). Such a large difference in N₂O emission between the SF and SC species gives precedence to the hypothesis of potential N₂O emission differences in different Sphagnum species.

Based on the analysis of bacterial communities using culture-based PCR-DGGE and isolation of N₂O emitters, the major Sphagnum-associated bacterial communities of our samples were consistent with boreal mire and tropical peat forest and included Burkholderia, Mucilaginibacter, Rhodanobacter, and Janthinobacterium but their N₂O emission activity was different in varied sites due to differences in climate and habitat environments (Hashidoko et al., 2008; Sun et al., 2014). Janthinobacterium spp. did not show high N₂O emission potential in subarctic palsa bog unlike in the tropical peatland soil, which suggested that the N₂O emission functions of N₂O emitters were changing in different climate zones. Previous experimentation has shown that the Sphagnum microbiota supported the host plant and the entire ecosystem under environmental changes (Bragina et al., 2014). Burkholderia spp. were N₂O emitters, but their N₂O emission functions were significantly lower than the acid-tolerant Janthinobacterium sp. in a deforested tropical peatland soil, which was previously determined by soil pH (Hashidoko et al., 2010). The Burkholderia spp. isolates in SF were similar to another climate zone in Finland, showing the same species of Sphagnum although in a different climate zone (Nie et al., 2015). Within this study, some unique bacterial strains were found in the leaves of SC, including a Pseudomonas sp. and two Enterobacteriaceae family members. In numerous previous studies, Pseudomonas species (P. denitrificans, P. perfectomarinus, P. fluorescens, P. stutzeri, P. aeruginosa, and P. nautica) were found performing denitrification (Delwiche, 1959; Payne et al., 1971; Balderston et al., 1976; Sørensen et al., 1980; Dooley et al., 1987; Viebrock and Zumft, 1988; SooHoo and Hollocher, 1991; Prudêncio et al., 2000). The isolated Pseudomonas sp. was not found in the bands of PCR-DGGE, possibly due to relatively low abundance under acidic conditions (pH 5) (Figure 4C). Anderson and Levine (1986) offered evidence that Enterobacteriaceae and Serratia sp.'s nitrate respiration produces N₂O, which was also found in our SC sample. Enterobacter sp. was also found as dissimilatory nitrate reduction to ammonium (DNRA) bacteria in agricultural soils (Heo et al., 2020). Pseudomonas sp. SC-H2, Enterobacteriaceae bacterium SC-L1, and Serratia sp. SC-K1 were responsible for N₂O emissions in our Sphagnum samples (SC). These findings suggest that the variation in the N₂O emission potential of Sphagnum found in peatlands is associated with species-specific bacterial communities, which are variable under different species and environments.

The top three active N₂O emitters (Pseudomonas sp. SC-H2, Enterobacteriaceae bacterium SC-L1, and Serratia sp. SC-K1) from SC increased N₂O production with increasing temperature up to 30°C (Figures 4D–F), illustrating a potential rise in N₂O emissions following global warming (Pfenning and McMahon, 1997; Voigt et al., 2017a; Chen et al., 2020). For the three active N₂O emitters, N₂O production was relatively high at a pH value of 7.0 (Figures 4A–C), which is much higher than the naturally low pH of Sphagnum microhabitats (Tahvanainen and Tuomaala, 2003). Although N₂O reduction to N₂ by Pseudomonas sp. SC-H2 was obvious, the N₂O production was still high after 5 days of incubation (Figure 6). This result indicated that N₂O emission hotspots are inclined to be in neutral peatlands, as supported by Palmer and Horn (2015). Combining these results with acetylene inhibition assays at pH value of 5.0 and 7.0 showed that N₂O reduction to N₂ was almost negligible at a pH value of 5 for these three active N₂O emitters. This result is consistent with a previous study of the lack of N₂O reductase (nos) function at low pH (Liu et al., 2014). This result also suggested that N₂O reduction was inhibited in the acidic environment in the peat bogs. Since the Sphagnum microhabitats are very acidic, N₂O reductase activity is repressed, supporting that N₂O reduction is not a pathway decreasing N₂O emissions in the pristine Sphagnum bog system. Under low-pH conditions, N₂O production by Pseudomonas sp. SC-H2 was small, but N₂O could be accumulated. However, the palsa mounds are formed due to the ice core under the Sphagnum peat layer in the subarctic climate, and once they collapse after permafrost thawing, the peat acidity will be neutralized to some extent by mixing with mineral material and minerogenic water flow (Seppälä, 2011; Takatsu et al., 2022).

Sphagnum mosses are important for peat accumulation and form a carbon pool of global significance. Increasing atmospheric N deposition can activate phenol oxidase in peat bogs and destabilize peat carbon (Bragazza et al., 2006). Phenol oxidase requires bimolecular oxygen for its activity (Freeman et al., 2004), and drying increases aerobic conditions in peatlands (Swindles et al., 2019) and can degrade recalcitrant phenolic materials. Tahvanainen and Haraguchi (2013) showed that this phenolic mechanism is affected by pH. Such changes may reduce the generally high C:N ratio, which increases net N mineralization, nitrification, and denitrification rates, while subsequently increasing the potential of N₂O production in peat bogs, while lower C:N ratios ( ≤ 25–30) stimulate N₂O emissions (Huang et al., 2004; Klemedtsson et al., 2005; Maljanen et al., 2012). Connected mechanisms and the release of ice-trapped N₂O are further impacted by thawing permafrost (Voigt et al., 2017b). Our findings indicate that N₂O emissions are not exceptionally high under the naturally cold temperatures and low pH of Sphagnum habitats; rather, substantially high pH and temperatures, and perhaps a connected imbalance of microbial communities in such conditions, induced the highest N₂O emissions. The results warrant caution in interpretation and against unexpected emission potential under rapidly changing conditions. It also calls for a need to monitor the in situ N₂O emissions from different permafrost Sphagnum species in the permafrost in future studies.

Without sucrose, the N₂O emitters Enterobacteriaceae bacterium SC-L1 and Pseudomonas sp. SC-H2 could not emit N₂O because of their low growth. This result indicated that these two strains were heterotrophic microorganisms that needed to gain C sources from Sphagnum moss and form plant-microbial symbionts between plants and microbes. Interestingly, Serratia sp. SC-K1 grew well without sucrose and emitted much more N₂O; meanwhile, it could be significantly inhibited by adding a low concentration of sucrose (0.05%). This result indicated that this strain is an autotrophic microorganism adapted to nutrient-poor environments, using carbon dioxide (CO₂) as a C source. These autotrophic microorganisms contribute to CO₂ uptake and carbon sequestration. Drained peatland ecosystems have an immense potential for C sinks to maintain the C balance, even though droughts are occasionally caused by decreasing photosynthesis (Minkkinen et al., 2018).

Our study showed that N₂O emitters (Serratia sp. SC-K1 and Enterobacteriaceae bacterium SC-L1) could resist relatively higher concentrations of caffeic acid ( ≤ 0.1 g L⁻¹), while the N₂O emitter (Pseudomonas sp. SC-H2) had low resistance to caffeic acid ( ≤ 0.005 g L⁻¹) (Figures 5D–F). These results could explain why we could not find the Pseudomonas spp. using DGGE band sequencing. Polyphenol (caffeic acid) from Sphagnum moss inhibits growth and results in a low relative abundance of Pseudomonas spp. The more abundant Serratia sp. SC-K1 and Enterobacteriaceae bacterium SC-L1 were the dominant N₂O emitters due to their higher resistance to polyphenolic compounds. The stimulated N₂O production in the Sphagnum moss-microbe vial with 0.1 g L⁻¹ caffeic acid confirmed Serratia sp. SC-K1 and Enterobacteriaceae bacterium SC-L1 were the dominant N₂O emitters. Serratia spp. are gram-negative bacilli and belong to the family Enterobacteriaceae. The interaction of polyphenolic compounds and Enterobacteriaceae bacteria might directly influence N₂O emissions in peatland ecosystems. High concentrations of polyphenols are likely to lower N₂O emissions. The response of phenol oxidase to N deposition differs by ecosystem type. In peat bogs, elevated N deposition decreased polyphenols' contents and decreased the polyphenol ratio to N, which may increase N₂O production due to an inverse relationship between N₂O emissions and the polyphenol to nitrogen ratio (Pimentel et al., 2015).

The three N₂O emitters preferred KNO₃ as a substrate over NH₄Cl. This result suggested that these three isolates mainly use DNRA or denitrification to produce N₂O gas. For the Enterobacteriaceae bacterium SC-L1 and Serratia sp. SC-K1, the nirS, nirK, and nosZ genes were not detected, but the narG gene was, suggesting that they do not have nitrite reductase and are non-denitrifiers consistent with other Enterobacteriaceae bacteria emitting N₂O as a final product (Arkenberg et al., 2011). Enterobacter species are often reported as producing N₂O by DNRA (Smith and Zimmerman, 1981). This result indicated that they are also important sources for N₂O emissions in SC dominant bogs. Pseudomonas sp. SC-H2 harbored nosZ, nirS, and narG. Therefore, Pseudomonas sp. SC-H2 was a typical denitrifier. Microbial heterotrophic denitrification and DNRA compete for shared resources (Jia et al., 2020).

Although the N₂O potential was relatively high in the SC sample, the N₂O emissions in situ in the peat bogs were generally low in northern Finland, which might be impacted by the complexity of environmental conditions (Dinsmore et al., 2017). The potential N₂O emissions in the field (Repo et al., 2009; Voigt et al., 2017b) and laboratory incubations (Elberling et al., 2010) increase with increasing mineral N availability, permafrost thawing, and drainage. A previous study suggested that drainage of bogs alters nutrient cycling and microbial communities to increase N₂O emissions (Frolking et al., 2011). Unvegetated (free of vascular plants) peat surfaces resulting from wind erosion and frost action were hot spots for N₂O emission in subarctic permafrost peatlands due to the absence of plant nitrogen uptake, a low C:N ratio, and sufficient drainage (Marushchak et al., 2011; Voigt et al., 2017b). Pseudomonas sp. SC-H2 had negligible N₂O emissions at low pH (<4.5), while the other two N₂O-emitting Enterobacteriaceae bacteria from SC exhibited contrasting patterns in the Sphagnum bogs. Therefore, the contribution of denitrification and DNRA to N₂O emissions in boreal peat bogs should be considered in future studies.