Methane and Nitrous Oxide Emissions from a Temperate Peatland under Simulated Enhanced Nitrogen Deposition

Abstract

Nitrogen (N) deposition has increased in recent years and is significantly affected by global change and human activities. Wetlands are atmospheric CH₄ and N₂O sources and may be affected by changes in N deposition. To reveal the effects of increased N deposition on peatland greenhouse gas exchange, we observed the CH₄ and N₂O emissions from controlled microcosms collected from a temperate peatland in the Xiaoxing’an mountains, Northeast China. We found that the moss biomass did not change, but the total herb biomass increased by 94% and 181% with 5 and 10-times-higher N deposition, respectively. However, there were no significant changes in CH₄ emissions from the microcosms with N addition. The unchanged CH₄ emissions were mainly caused by the opposite effect of increased nitrate and ammonium concentrations on soil CH₄ production and the increased plant biomass on CH₄ emission. We also found that the manipulated microcosms with 5 and 10-times-higher N deposition had 8 and 20-times-higher seasonal average N₂O emissions than the control microcosms, respectively. The increased N₂O emissions were mainly caused by short-term (≤7 d) pulse emissions after N addition. The pulse N₂O emission peaks were up to 1879.7 and 3836.5 μg m⁻² h⁻¹ from the microcosms with 5 and 10-times-higher N deposition, respectively. Nitrate and ammonium concentrations increasing in the soil pore water were the reason for the N₂O emissions enhanced by N addition. Our results indicate that the increase in N deposition had no effects on the CH₄ emissions but increased the N₂O emissions of the temperate peatland. Moreover, pulse emissions are very important for evaluating the effect of N addition on N₂O emissions.

1. Introduction

The rate of N deposition in the atmosphere has rapidly increased since the 19th century due to the impact of human activities such as fertilizer application and the combustion of fossil fuels [1]. It is predicted that by 2050 the global N deposition will be nearly double that of the early 1990s [2]. Consistent with the trend of global N deposition, the average annual N deposition in mainland China has increased by about 60% over the past 30 years (from 13.2 to 21.1 kg N ha⁻¹ yr⁻¹) [3]. Meanwhile, the N deposition rates in the northeast of China have also significantly increased, reaching 0.11 kg N ha⁻¹ yr⁻¹ from precipitation alone [3,4].

Wetlands are one of the major carbon sinks on the Earth. The carbon pool stored in the form of peat in northern wetlands is 500 ± 100 Gt [5]. Meanwhile, wetlands emit a large amount of CH₄ and N₂O from soils into the atmosphere every year [6,7]. Since most of the plants in wetlands are N-limited, an increase in N deposition will greatly affect the vegetation composition and net primary productivity of plants in wetland ecosystems [8,9]. An increase in N deposition will further affect the decomposition rates of litter and soil organic matter [10] accompanying the changes in nutrient availability in soil [11]. These are important factors which control CH₄ and N₂O emissions from wetland soil [12,13,14]. Therefore, CH₄ and N₂O emissions may be greatly affected by an increase in N deposition.

CH₄ is produced by methanogens under anaerobic conditions during the decomposition of wetland soil organic matter [15]. Wetland CH₄ fluxes may be affected by several factors, such as soil anaerobic conditions [16,17], soil temperature [18,19], plant biomass [12,20], etc. CH₄ exchange between wetland soil and the atmosphere involves three complex processes: CH₄ production in anaerobic soil, CH₄ transport and oxidation in the soil or water, and CH₄ release to the atmosphere [21]. An increase in N deposition will affect one or all of these three processes. Therefore, the impacts of increased N deposition on CH₄ emissions from wetlands are different in the findings of different studies: some studies found an increase [22,23,24], no significant impacts [25,26], and some found a decrease in N deposition [13,27]. A study on CH₄ emissions from boreal forested peatlands showed that only a few ecosystems emitted more CH₄ after N addition, and there was no change in most of the rest ecosystems in the same region [28]. Even in the same wetland ecosystem, the application of different forms of nitrogen or different observation years had different impacts on CH₄ emissions [29,30,31]. Among the environmental factors affected by an increase in N deposition mentioned above, an increase in plant coverage and productivity may play a key role in the changes in CH₄ emissions [28,32]. More coverage and a higher productivity of plants may provide sufficient substrates for methanogens and transport pathways for CH₄ emissions, thus enhancing CH₄ emissions.

N₂O is an obligatory intermediate in the denitrification pathway of wetland soil [33], and N₂O production is greatly affected by the nitrogen content in the soil [34,35]. Differing from the uncertainty of the effects of increasing N deposition on CH₄ emissions, most studies conclude that an increase in N deposition promotes N₂O emissions from wetlands [22,30,36,37,38]. An increase in N deposition promotes N₂O emissions mainly because of the increased N availability in soil, which enhances nitrification or denitrification [14,25,34]. However, a few studies show that an increase in N deposition has no significant effect on N₂O emissions [26,28]. Different results indicate that the effects of increased N deposition on N₂O emissions may be related to the trophic classes of the studied wetlands. Ombrotrophic peatlands seem to be more susceptible than minerotrophic peatland to changes in N deposition [39]. The dose of the N deposition may also play an important role. A larger external N supply may have greater effects [40,41]. Meanwhile, different forms of N may also have different effects on N₂O emissions from wetlands [14].

The Xing’an mountains are located in Northeast China. It is a cold region as it is affected by the Siberian climate. The wetlands in this area are well developed, and peat accumulates in this region due to its flat valleys and the existence of underground permafrost. It is one of the major areas of peatlands in China [42]. Peatlands are sensitive to global change due to the relatively high labile fractions in their soil that can easily be converted into greenhouse gases [43]. However, there are few studies on the effects of increasing N deposition on CH₄ and N₂O fluxes from temperate peatlands in this region [13,14]. Will N deposition increase GHG (CH₄ and N₂O) fluxes from temperate peatland? What is the main factor controlling the CH₄ and N₂O changes with increasing N deposition? For the peatlands of the Xiaoxing’an mountains, the answers to these questions remain a mystery. In this study, a typical temperate peatland in the Xiaoxing’an mountains was selected to control N deposition using microcosms. We observed the CH₄ and N₂O emissions from these microcosms. The objectives of this study were to (1) investigate the influence of N addition on CH₄ and N₂O emissions from temperate peatlands and (2) reveal the factors controlling the changes in CH₄ and N₂O emissions. These results are helpful for revealing the response and mechanism of peatland in the second largest permafrost region of China to the increase in N deposition. At the same time, because the peatland in this region is located at the southern edge of the Eurasian permafrost zone, these results can act as a reference for global permafrost peatlands.

2. Materials and Methods

2.1. Site Description and Microcosms Sampling

We used microcosms collected from a temperate peatland in Xiaoxing’an Mountains Northeast China to control nitrogen(N) addition and measured CH₄ and N₂O emissions. The sampling site was Yongqing forest farm of Youhao Forestry Bureau of Yichun City (48°03′53″–48°17′11″ N, 128°30′36″–128°45′00″ E; 260–500 m a.s.l.). The average annual temperature is 0.4 °C, and the average annual precipitation is 630 mm. Peat has accumulated in the wetlands in this region due to the cold and hydrologic conditions caused by underground permafrost. Peat depths are no more than 1 m in most of the wetlands in the Xiaoxing’an mountains.

Microcosms were sampled from a forested peatland in this area. The dominant tree in the peatland is Larix gmelinii, and trees are short and sparse due to the poor nutrients and low temperatures in the soil. Ledum palustre var. angustum and Vaccinium uliginosum are dominant under the trees. The height of these shrubs is between 0.2 m and 0.5 m. The herb species of Calamagrostis angustifolia and Eriophorum vaginatum were scattered among the shrubs. Sphagnum spp. and mosses cover 70% to 90% of the ground surface. The peat soil is acid with low pH values of 4.75 ± 0.05 at the depth of 0–20 cm and 4.89 ± 0.02 at the depth of 20–40 cm. Soil organic carbon and total nitrogen concentrations at the depth of 0–20 cm are 238.91 ± 3.57 g kg⁻¹ and 13.33 ± 0.41 g kg⁻¹, respectively. The corresponding concentrations at a depth of 20–40 cm are 290.96 ± 64.76 g kg⁻¹ and 13.28 ± 0.42 g kg⁻¹, respectively.

During sampling, trees and high shrubs were avoided, and only short shrubs and herbs were chosen. Samples were taken at the end of autumn before the soil was frozen. The microcosms were excavated as a whole and then moved into small plastic buckets. These small buckets had a bottom diameter of 22 cm, a top diameter of 28 cm, and depths of 30 cm. A total of 12 microcosms were collected. These microcosms were transported to Sanjiang Plain Wetland Ecological Experimental Station of Chinese Academy of Sciences. This wetland station is located in Honghe farm (47°35′ N, 133°31′ E) of Sanjiang Agriculture Administration Bureau of Heilongjiang Province. The annual average temperature is 2.5 °C, and the annual average precipitation is 558 mm. After arrived at the wetland station, we buried the small buckets in 30 cm deep holes to avoid too much disturbance to soil microbial activity due to the very low temperature in winter. In the following spring, the small buckets and the microcosms were taken out after the soil thawed, and then the N addition experiment was carried out. Twelve microcosms were divided into three groups with four replications in each group. Two groups were treated with N addition and one group as the control. There was a gap of approximately 0.5 m between groups and between single microcosms in the same group.

2.2. Simulation of N Deposition

Before the N addition experiment, one month of incubation was carried out in June when all the plants in the buckets started to grow again. Incubation aimed to recover the soil microbial activity. During one month of incubation, the water table level of the microcosms was maintained at 5 cm below the soil surface. After the incubation time in early July, a simulation of increased N deposition was carried out, and the microcosms received water only by precipitation. The average N deposition in Northeast China is approximately 1.5 g N m⁻² yr⁻¹ according to a nitrogen deposition map [4]. Therefore, 7.5 and 15.0 g N m⁻² yr⁻¹ were added to the manipulated microcosms, respectively, to simulate 5 and 10 times the level of natural N deposition. N was directly poured into the microcosms via a NH₄NO₃ solution every month from July to September. For the time interval of N addition, the setting time varies greatly among studies. N fertilization is usually conducted at intervals of between two weeks and one month in the growing season [11,22,44]. The longer N application interval was two or three times a year in boreal peatlands [25,45]. The time interval of N addition used in our study was once a month in the growing season. This fertilization interval was within a reasonable range according to the reported studies. To reduce the influence of N addition on soil moisture between manipulated and control microcosms, the same volume of deionized water was applied by the same method to the control microcosms when fertilizing.

2.3. CH₄ and N₂O Flux Measurements

Before N addition, CH₄ and N₂O emissions were observed three times at the later stage of one month incubation. The sampling frequency was once every two days. According to the results of the sampling on three occasions, there was no significant difference between CH₄ and N₂O emissions in the treatment and control microcosms (see Table 1 below). After one month of incubation, two groups of the microcosms had NH₄NO₃ solution added, and one group was kept as a control. CH₄ and N₂O emissions were measured every 5 to 10 days during the experimental period. We sampled 1–2 times after N addition to capture the possible pulse greenhouse gases fluxes. The sampling lasted to the end of the growing season. Sampling was carried out for at least 30 days after the last addition of N. During the whole control experiment period from June to October, we sampled greenhouse gases a total of 20 times.

We used static chambers to collect gas samples. The chambers were made of big opaque plastic buckets. These big buckets had a bottom diameter of 31 cm, a top diameter of 39 cm, and a depth of 40 cm; they covered the entire microcosms. There was a rubber stopper and a fan on the top of the chambers for headspace gas sampling and air mixing in the chamber. We used 50 mL syringes to draw four gas samples from the chamber through the rubber stopper at 10 min intervals. Samples were analyzed within 12 h on a gas chromatograph (GC, Agilent 7890, Agilent Co., Santa Clara, CA, USA). CH₄ and N₂O concentration analysis and calculation methods were conducted according to Song et al. [7]. The height of the chamber in the equation was obtained by dividing the volume difference between the big bucket and the small bucket by the top surface area of the small bucket. Data were accepted when R² of the linear regression between gas concentrations and time was ≥0.90 for CH₄ or ≥0.80 for N₂O.

2.4. Environmental Factors Measurement

Environmental factors were recorded each time gas was sampled in order to explain the variations in CH₄ and N₂O emissions. A digital thermometer (JM624, China) was used to record the 10 cm soil temperature. The water table level relative to the soil surface was measured by inserting a fine steel rod with a scale into the soil. Soil pore water at 20 cm depth was sampled with a syringe connected with a 2 mm diameter fine steel pipe. The concentration of nitrate (NO₃⁻ -N) and ammonium (NH₄⁺ -N) of the soil pore water was analyzed. NO₃⁻ -N concentration was analyzed using ultraviolet spectrophotometry at 220 nm and 275 nm. NH₄⁺ -N concentration was analyzed using the indophenol blue spectrophotometric method. By the end of the experiment, the aboveground and belowground biomass of herb and all biomass of moss were measured by digging and harvesting all of the whole plants in the microcosms. The shoots and roots of the herbs were separated and then separately dried to a constant weight in an oven at 60 °C. The whole moss was dried under the same conditions.

2.5. Data Analysis

CH₄ or N₂O emissions and environmental factors were the average values of the four replicate microcosms. We used the one-way ANOVA (Duncan comparison) to test the differences in CH₄ or N₂O emissions among groups with different N treatments and the controls. We used the correlation analysis to test the relationship between CH₄ or N₂O emissions and the environmental factors. We used a log 10 transformation before we analyzed the correlation between environmental factors and log 10 transformation of N₂O fluxes due to the great fluctuations in emissions after N addition. To make sure all of the data were positive, we added a value of 30 to the flux data before transformation. When the p value was less than 0.05, the test was considered significant. The statistical analysis was performed using SPSS version 18.0 software (SPSS, Inc. Chicago, IL, USA).

3. Results

3.1. Variations of CH₄ and N₂O Emissions before N Addition

Three times sampling before N addition showed that there were no significant differences in CH₄ and N₂O emissions between the treated and control microcosms (p > 0.05, Table 1).

Table 1. CH₄ and N₂O fluxes of different microcosms during the incubation period, n = 12.

Treatment	CH4 Flux (mg m−2 h−1)		N2O Flux (μg m−2 h−1)		n
	Mean	SE	Mean	SE
CK	2.50	0.40	55.3	7.5	12
N5	1.79	0.66	55.1	6.9	12
N10	2.57	0.95	55.5	6.5	12
p value (ANOVA)	0.471		0.959

CK represents the control microcosms; N5 and N10 are the treated microcosms of 5 and 10-times natural N deposition, respectively.

Table 1. CH₄ and N₂O fluxes of different microcosms during the incubation period, n = 12.

Treatment	CH4 Flux (mg m−2 h−1)		N2O Flux (μg m−2 h−1)		n
	Mean	SE	Mean	SE
CK	2.50	0.40	55.3	7.5	12
N5	1.79	0.66	55.1	6.9	12
N10	2.57	0.95	55.5	6.5	12
p value (ANOVA)	0.471		0.959

CK represents the control microcosms; N5 and N10 are the treated microcosms of 5 and 10-times natural N deposition, respectively.

3.2. Effects of N Addition on Vegetation and Environmental Factors

Nitrogen addition significantly increased (p < 0.001) the aboveground, belowground and total biomass of herbs in the microcosms. The aboveground biomass of herbs in N5- and N10-treated microcosms was 135% and 291% higher (p < 0.05) than that of the control, respectively; the belowground biomass values were 82% and 148% higher (p < 0.05) than those of the control, respectively; and the total biomass values were 94% and 181% higher (p < 0.05) than those of the control microcosms, respectively. However, there was no significant difference in moss biomass among the N5- and N10-treated and control microcosms (p > 0.05, Figure 1).

There was no significant difference in soil temperature at a 10 cm depth and the water table between the treated and control microcosms (p > 0.05, Figure 2a,b). NO₃⁻ -N in soil pore water in the N10-treated microcosms was significantly higher than the control (p < 0.05). NO₃⁻ -N concentration of the N5-treated microcosms was also higher than that of the control microcosms; however, it was not significantly different from both the N10-treated and control microcosms (p > 0.05, Figure 2c). NH₄⁺ -N in soil pore water increased with N addition; however, this difference was not significant (p > 0.05, Figure 2d).

3.3. Effects of N Addition on Seasonal Variation of CH₄ Emissions

The same variations in CH₄ emissions were observed in the treated and control microcosms. CH₄ emissions were low in spring and autumn but high in summer in all the microcosms. Three emission peaks were observed, and the largest peaks were 39.74, 24.27 and 26.04 mg m⁻² h⁻¹ for the control, N5- and N10-treated microcosms, respectively. No CH₄ absorption events were detected in all microcosms during the experimental period (Figure 3).

CH₄ emissions had positive exponential correlations with soil temperature at 10 cm depth (p < 0.05,

Table 2. Correlations between CH₄ fluxes and soil temperature (°C) or water table (cm), n = 20.

Treatment	Variable	Equation	R2	p
CK	T	F = 0.252 e0.145T	0.288	0.015 *
	WT	F = 4.545 − 0.875 WT	0.492	0.001 **
N5	T	F = 0.118e0.16T	0.34	0.007 **
	WT	F = 2.031 − 0.738 WT	0.786	0.000 **
N10	T	F = 0.1 e0.173T	0.343	0.007 **
	WT	F = 3.35 − 0.62 WT	0.543	0.000 **

F indicates CH₄ fluxes, T indicates soil temperature at 10 cm depth, and WT indicates water table. * and ** indicate significant level at 0.05 and 0.01 level.

) and had negative linear correlations with the water table (p < 0.01, Table 2) in all the treated and control microcosms. However, there was no significant correlation between CH₄ emissions and NO₃⁻ -N or NH₄⁺ -N concentrations in soil pore water (p > 0.05).

3.4. Effects of N Addition on Seasonal Average CH₄ Emissions

Seasonal average CH₄ emissions of N5- and N10-treated microcosms were slightly lower than those of the control microcosms. However, the differences were not significant (p > 0.05, n = 20, Figure 4).

3.5. Effects of N Addition on Seasonal Variation of N₂O Emissions

Seasonal variations in N₂O emission from the microcosms were changed by N addition. Emission rates from the control microcosms were high in spring and ranged from 47.8 to 64.7 μg m⁻² h⁻¹. The emissions decreased in summer, and most emission values did not exceed 30 μg m⁻² h⁻¹. The emissions further decreased in autumn, and most values were negative, indicating that the microcosms started to absorb the N₂O in the atmosphere. N₂O emissions ranged from −21.4 to 64.7 μg m⁻² h⁻¹ during the observed growing season (Figure 5).

Similar emission rates were found between the treated and control microcosms in spring and autumn. However, the differences were large when N was added in summer. Pulse N₂O emissions (≤7 d) were observed from the treated microcosms after each N addition. N₂O peaks of the pulse emissions were 1879.7 and 3836.5 μg m⁻² h⁻¹ for the N5- and N10-treated microcosms, respectively. These values were 29 and 59 times higher compared to the N₂O emission peak of the control microcosms (Figure 5).

N₂O emissions from the control microcosms had positive linear correlations with soil temperature at a 10 cm depth (p < 0.01), and had positive logarithmic correlations with NO₃⁻ -N or NH₄⁺ -N concentrations in soil pore water (p < 0.001,

Table 3. Correlations between N₂O fluxes and soil temperature (°C), concentrations of NO₃⁻ -N and NH₄⁺ -N in soil pore water (mg L⁻¹), n = 20.

Treatment	Variable	Equation	R2	p
CK	T	F = −71.108 + 4.35T	0.447	0.001 **
	NO3− -N	F = −85.483 + 46.561ln(NO3− -N)	0.651	0.000 **
	NH4+ -N	F = −107.521 + 46.275ln(NH4+ -N)	0.528	0.000 **
N5	T	Log10(F + 30) = 0.365e0.068T	0.211	0.03 *
	NO3− -N	Log10(F + 30) = −0.617 + 0.926ln(NO3− -N)	0.47	0.001 **
	NH4+ -N	Log10(F + 30) = −1.051 + 0.945ln(NH4+ -N)	0.285	0.015 *
N10	T	Log10(F + 30) = 1.409e0.013T	0.03	0.464
	NO3− -N	${\mathrm{Log}}_{10}(\mathrm{F}+30)=1.258{\mathrm{e}}^{0.025\left({\mathrm{NO}}_3^{-} \text{-}\mathrm{N}\right)}$	0.354	0.006 **
	NH4+ -N	${\mathrm{Log}}_{10}(\mathrm{F}+30)=0.962{\mathrm{e}}^{0.032\left({\mathrm{NH}}_4^{+} \text{-}\mathrm{N}\right)}$	0.265	0.020 *

F indicates N₂O fluxes, and T indicates soil temperature at 10 cm depth. * and ** indicate significant level at 0.05 and 0.01 level.

). Transformed N₂O emissions from the N5-treated microcosms had positive exponential correlations with soil temperature (p < 0.05) and positive logarithmic correlations with NO₃⁻ -N or NH₄⁺ -N concentrations in soil pore water (p < 0.01 or p < 0.05, respectively, Table 3). Transformed N₂O emissions from the N10-treated microcosms had positive exponential correlations with NO₃⁻ -N or NH₄⁺ -N concentrations in soil pore water (p < 0.01 or p < 0.05, respectively), but had no significant correlation with the soil temperature (p > 0.05, Table 3). There were no significant correlations between N₂O emissions and the water table in any one of the treated and control microcosms (p > 0.05).

3.6. Effects of N Addition on Seasonal Average N₂O Emissions

Nitrogen addition increased seasonal average N₂O emissions from microcosms. The seasonal average N₂O emission from the control microcosms was 16.9 μg m⁻² h⁻¹, whereas the seasonal average emission rates from the N5- and N10-treated microcosms were 136.4 and 329.1 μg m⁻² h⁻¹, respectively. These values were about 8 and 20 times higher compared to that in the control microcosms, respectively. However, ANOVA analysis showed that there was no significant difference in N₂O emissions between the control and treated microcosms (p > 0.05, n = 20, Figure 6).

4. Discussion

4.1. Effects of N Addition on CH₄ Emissions

Our results show that there was no significant effect of N5 or N10 addition on CH₄ emissions from the peatland microcosms. This is because soil temperature and the water table, the two key factors controlling CH₄ emissions, did not significantly change after N addition; therefore, there were only minor changes in CH₄ emissions. However, plant biomass, as another indicator of CH₄ emissions from wetlands, significantly changed in this study. Most results show that vascular plants can promote CH₄ emissions during the growing season [12,20]. Meanwhile, moss can promote CH₄ uptake, and thus reduce CH₄ emissions from wetlands [13,46]. We found that the aboveground, belowground and total biomass of herbs increased, while the biomass of moss did not change with N addition (Figure 1). However, the increase in herb biomass did not promote CH₄ emissions. This may be because N addition increased nitrate and ammonium concentrations in peat (Figure 2). Nitrate and ammonium provide alternative electron acceptors for microbial metabolism, thus inhibiting CH₄ production in soil [47]. The inhibition of nitrate and ammonium on methane emission offsets the promotion of increased plant biomass on methane emissions. Therefore, CH₄ emissions did not change with the increased herb biomass. Granberg et al. [48] also found that N addition promoted plant growth but inhibited CH₄ emissions from a boreal mire. They believe that this inhibition was mainly caused by the shift in root allocation from deep to shallow roots, which reduced the supply of root exudates to methanogens and the transport of CH₄. Although we did not measure the change in plant root depth in this study, we still believe that this speculation might be one of the explanations for the unchanged CH₄ emissions after N addition.

CH₄ emissions had positive exponential correlations with soil temperature in all treated and control microcosms (p < 0.05, Table 2). This is consistent with previous results because a high temperature promotes the decomposition of soil organic matter and the growth of plants, which can both provide more substrates for methanogens [18]. Moreover, with the increase in temperature, the activity of methanogens also increases. Therefore, a high temperature promotes CH₄ production [18,19] and further accelerates CH₄ emission.

CH₄ emissions have a negative linear correlations with the water table in all the treated and control microcosms (p < 0.05, Table 2). This is different from most previous studies, which showed that a higher water table aggravated the anaerobic conditions of the soil, which is conducive to the production of CH₄ [16,49,50]. Therefore, CH₄ emissions usually positively correlate with the water table [17,51]. If CH₄ emission rates are low or their variability is large, CH₄ emissions may not be related to the water level [52]. However, a few studies reported that CH₄ emissions are negatively correlated with the water table [53,54], these were in accordance with our results. The negative correlations between CH₄ emissions and the water table in this study may be caused by two factors. Firstly, the severe fluctuations of the water table led to passive high CH₄ emissions from the soil pore water. We observed three CH₄ emission peaks in this study (Figure 3). The first peak (18–34 mg m⁻² h⁻¹) occurred in late July due to the decrease in water table affected by low precipitation. The water table decreased from 1 cm above the soil surface to 12 cm below the surface (Figure S1). The second peak (8–14 mg m⁻² h⁻¹) occurred in early August when the water table decreased from 2 cm above the soil surface to 5 cm below the surface (Figure S1). The third peak (24–40 mg m⁻² h⁻¹) occurred in mid-August when the water table decreased from 6 cm to 30 cm below the surface (Figure S1). The largest decrease in water table resulted in the highest methane emission peaks. Similar results were found in previous studies: the decrease in water table leads to the rapid release of CH₄ dissolved in pore water to the atmosphere [55,56], thus inducing a peak in CH₄ emissions.

The second factor that caused negative correlations between CH₄ emissions and the water table may be temperature interference. The water table decrease between mid-July and late August was caused by local high temperatures and low precipitation. During this period, the water table value decreased, but soil temperatures remained over 20 °C for most of the observed time and were highest in the growing season (Figure S2). High temperatures promoted CH₄ production [50]. Therefore, the CH₄ emission peaks appeared when the soil temperature was highest, but the water table was low. The CH₄ emission peaks caused by the soil temperature increase and three-fold decease in the water table value between July and August led to the negative correlations between CH₄ emissions and the water table.

4.2. Effects of N Addition on N₂O Emissions

Our results show that N addition not only changes the seasonal variation and the total amount of N₂O emissions, but also has a great impact on the relationship between N₂O emissions and environmental factors. N₂O emissions from the 5 and 10-times-higher N deposition microcosms had 8 and 20-times-higher seasonal average N₂O emissions than the control microcosms, respectively. This is in accordance with most previous results. Increases in N₂O emissions were caused by pulse emissions and continued no longer than seven days after N addition (Figure 5). The differences in N₂O emissions were significant between the treated and control microcosms within seven days (p < 0.001). After the high pulse emission, N₂O emission rates decreased to the same level as the control microcosms. High pulse emission peaks after N addition resulted in large but insignificant differences in seasonal average N₂O emissions between treated and control microcosms (Figure 6). The rapid increase in N₂O emissions in a short period (one to several days) was also observed in boreal peatland [25], estuarine marsh [57] and paddy fields [58]. Anthony and Silver [59] reported that ‘hot moments’ of N₂O emissions contributed to 45% of mean annual N₂O fluxes. This indicates the importance of the observation frequency to the accuracy of estimating accumulated N₂O. Missing the first few days after fertilizing may underestimate the total N₂O emissions caused by the addition of N.

The increase in N addition to N₂O emissions from wetlands may be due to the following two reasons. Firstly, the concentrations of nitrate and ammonium in soil increases (Figure 2), which increases the available substrates of nitrifiers and denitrifiers. Sufficient substrates can promote nitrification and denitrification, and thus increase N₂O emissions in wetlands [25,34,35,58]. Secondly, the increase in plant biomass (Figure 1) can also increase the available substrates of bacteria and provide a pathway for N₂O transport from soil to the atmosphere [60]. As a result, N₂O emissions come from sites with plant/high plant biomass, rather than those without plant/low plant biomass [61]. However, the increased N₂O emissions found in this study mainly derive from high-pulse emissions after N application, rather than the continuous high emissions. Therefore, we speculate that the increase in N availability is the main reason for the increase in N₂O emissions from the treated microcosms, while the increase in plant biomass plays a minor role. Cui et al. [14] also found that N addition promoted N₂O emissions from peatland in the Great Xing’an mountains of Northeast China, which is close to our research region. However, there might be different control mechanisms regarding the promotion between the two studies. These authors found that N₂O emissions was positively correlated to NO₃⁻ -N rather than NH₄⁺ -N concentrations. Therefore, they found that the denitrification was the main mechanism influencing N₂O emissions [14]. Though there were similar soil chemical characteristics between our study and theirs, we deduced that both nitrification and denitrification controlled N₂O emissions. This is due to the different hydrological conditions used in each study. The water table value decreased to below the surface twice in this study, and thus created aerobic conditions in soil. The alternation of aerobic and anaerobic conditions in the soil facilitated both nitrification and denitrification, and both conditions controlled the N₂O emissions.

Due to the high-pulse N₂O emissions after N addition in the treated microcosms, the correlations between N₂O emissions and environmental factors were different between the treated and control microcosms. N₂O emissions from the control microcosms had a positive correlations with soil temperature and NO₃⁻ -N or NH₄⁺ -N concentrations in soil pore water (p < 0.001, Table 3). The increase in soil temperature promoted N₂O production by increasing microbial activity [6,44], while the increase in N availability provides more substrates for nitrifiers and denitrifiers [62,63]. Thus, both of them could promote N₂O emissions. This is consistent with previous results for different wetlands [64,65]. Transformed N₂O emissions from the N-treated microcosms had a positive correlations with NO₃⁻ -N or NH₄⁺ -N concentrations in soil pore water (p < 0.01 or p < 0.05, respectively), but the correlations of N5 (R² = 0.285 − 0.47) and N10 (R² = 0.265 − 0.354) were lower than those in control microcosms (R² = 0.528 − 0.651). Transformed N₂O emissions from the N5-treated microcosms had positive correlations with soil temperature (p < 0.05). However, there were no significant correlations between the N₂O emissions and soil temperature in N10-treated microcosms. The reduced and unrelated correlations between N₂O emissions and environmental factors in the treated microcosms were induced by N addition. The high-pulse N₂O emissions stimulated by the N addition confused the relationships between N₂O emission and environmental factors. Therefore, a higher dose of N addition caused lower correlations between N₂O emissions and the environmental factors. Due to the limitation of the microcosms experiment, we cannot collect soil samples each time to determine the NO₃⁻ -N or NH₄⁺ -N concentrations. From the reported studies, we can determine the relationship between N₂O emission and mineral N content in soil. Whether it was mineral soil [35,66] or peat soil [65,67], researchers found that N₂O emissions increased with a higher content of mineral N in soil. N₂O emission from wetlands is mainly caused by denitrification of N in soil [68]. Therefore, based on previous research results, we speculate that the NO₃⁻ -N or NH₄⁺ -N concentrations in soil may have higher correlations with N₂O emissions in this study. N addition can promote N₂O emission by increasing the NO₃⁻ -N or NH₄⁺ -N concentrations in soil and water. There were no significant correlations between N₂O emissions and the water table in all of the treated and control microcosms. This indicates that the water table may not be a good predictor for N₂O emissions from peatlands in this region.

5. Conclusions

There were no significant effects of N5 and N10 addition on soil temperature and water table of the microcosms. Although the plant biomass of herbs was increased by N addition, CH₄ emissions did not change significantly. This is because the increased nitrate and ammonium by N addition inhibited CH₄ production in the soil and further affected CH₄ emissions. CH₄ emissions had positive correlations with soil temperature but had negative correlations with the water table in all the treated and control microcosms. These relationships were due to the temporary high CH₄ emissions caused by the high temperature and rapid decrease in the water table.

N addition not only changed the seasonal variations and the total amount of N₂O emissions, but also had a great influence on the relationships between N₂O emissions and environmental factors. There were significant positive correlations between N₂O emissions and the soil temperature or N availability in the control microcosms. However, these relationships were lower or disappeared after N addition. Seasonal average N₂O emissions from the N5 and N10 addition microcosms were 8 and 20 times higher than the control microcosms, respectively. The increased N₂O emissions mainly came from high-pulse emissions after N addition, which lasted for no more than seven days. The increased nitrate and ammonium concentrations in soil pore water were the reason for the N₂O emissions promoted by N addition.