Five-year study of the effects of simulated nitrogen deposition levels and forms on soil nitrous oxide emissions from a temperate forest in northern China

Few studies have quantified the effects of different levels and forms of nitrogen (N) deposition on soil nitrous oxide (N2O) emissions from temperate forest soils. A 5-year field experiment was conducted to investigate the effects of multiple forms and levels of N additions on soil N2O emissions, by using the static closed chamber method at Xi Mountain Experimental Forest Station in northern China. The experiment included a control (no N added), and additions of NH4NO3, NaNO3, and (NH4)2SO4 that each had two levels: 50 kg N ha−1 yr−1 and 150 kg N ha−1 yr−1. All plots were treated to simulate increased N deposition on a monthly schedule during the annual growing season (March to October) and soil N2O emissions were measured monthly from March 2011 to February 2016. Simultaneously, the temperature, moisture, and inorganic N contents of soil were also measured to explore how the main factors may have affected soil N2O emission. The results showed that the types and levels of N addition significantly increased soil inorganic N contents, and the accumulation of soil NO3–-N was significantly higher than that of soil NH4+–N due to N addition. The three N forms significantly increased the average N2O emissions (P < 0.05) in the order of NH4NO3 > (NH4)2SO4 > NaNO3 by 355.95%, 266.35%, and 187.71%, respectively, compared with control. The promotion of N2O emission via the NH4+–N addition was significantly more than that via the NO3––N addition, while N addition at a high level exerted a stronger effect than at the low-level. N addition exerted significantly stronger effects on cumulative N2O emissions in the initial years, especially the third year when the increased cumulative N2O emission reached their maximum. In the later years, the increases persisted but were weakened. Increasing inorganic N concentration could change soil from being N-limited to N-rich, and then N-saturated, and so the promotion on soil available N effect increased and then decreased. Moreover, the soil NH4+–N, NO3–-N, temperature, and water-filled pore space were all positively correlated with soil N2O emissions. These findings suggest that atmospheric N deposition can significantly promote soil N2O emission, and that exogenous NH4+–N and NO3–-N inputs into temperate forests can have synergic effects on soil N2O emission. In future research, both aspects should be better distinguished in the N cycle and balance of terrestrial ecosystems by using 15N tracer methods.

Few studies have quantified the effects of different levels and forms of nitrogen (N) deposition on soil nitrous oxide (N 2 O) emissions from temperate forest soils. A 5-year field experiment was conducted to investigate the effects of multiple forms and levels of N additions on soil N 2 O emissions, by using the static closed chamber method at Xi Mountain Experimental Forest Station in northern China. The experiment included a control (no N added), and additions of NH 4 NO 3 , NaNO 3 , and (NH 4 ) 2 SO 4 that each had two levels: 50 kg N ha −1 yr −1 and 150 kg N ha −1 yr −1 . All plots were treated to simulate increased N deposition on a monthly schedule during the annual growing season (March to October) and soil N 2 O emissions were measured monthly from March 2011 to February 2016. Simultaneously, the temperature, moisture, and inorganic N contents of soil were also measured to explore how the main factors may have affected soil N 2 O emission. The results showed that the types and levels of N addition significantly increased soil inorganic N contents, and the accumulation of soil NO 3 --N was significantly higher than that of soil NH 4 + -N due to N addition. The three N forms significantly increased the average N 2 O emissions (P < 0.05) in the order of NH 4 NO 3 > (NH 4 ) 2 SO 4 > NaNO 3 by 355.95%, 266.35%, and 187.71%, respectively, compared with control. The promotion of N 2 O emission via the NH 4 + -N addition was significantly more than that via the NO 3 --N addition, while N addition at a high level exerted a stronger effect than at the low-level. N addition exerted significantly stronger effects on cumulative N 2 O emissions in the initial years, especially the third year when the increased cumulative N 2 O emission reached their maximum. In the later years, the increases persisted but were weakened. Increasing inorganic N concentration could change soil from being N-limited to N-rich, and then N-saturated, and so the promotion on soil available N effect increased and then decreased. Moreover, the soil NH 4 + -N, NO 3 --N, temperature, and water-filled pore space were all positively correlated with soil N 2 O emissions. These findings suggest that atmospheric N deposition can significantly promote soil N 2

Introduction
Nitrous oxide (N 2 O) is not only a potent greenhouse gas whose global warming potential is 298-and 21-fold that of CO 2 and CH 4 , but it also contributes to stratospheric ozone depletion [1]. Emissions of N 2 O from soil have been identified as the primary source (57%) of total global N 2 O emissions [2]. Nitrification and denitrification are the two main processes that produce N 2 O in soils and both can occur simultaneously (Fig 1). N 2 O is produced by denitrifying bacteria during the reduction of NO 3 - [3]. These two processes may be affected by soil water content, temperature, N availability and pH, as well as other particular biotic or abiotic properties [4][5][6]. Inorganic N is a key factor regulating soil N 2 O emission [4][5][7][8] (Fig  1). In general, increasing available mineral N in soils leads to enhanced N 2 O formation and emission via increased nitrification and denitrification rates [9]. Soil N 2 O emission is also driven by soil temperature and water content [10]. Some previous studies indicated soil N 2 O emissions were increased under conditions of higher soil water content and soil temperature [10]. The latter may regulate soil N 2 O emission by influencing N 2 O-producing microorganisms, such as nitrifying and denitrifying bacteria [11]. Furthermore, low soil moisture can reduce the temperature sensitivity of soil microbes, so that the diffusion of extracellular enzymes in the substrate are lowered [12].  China is now ranked third behind Europe and North America in terms of the scale of anthropogenic reactive N emissions, and has been experiencing a dramatic increase in anthropogenic reactive N due to its rapid economic development [10]. The average N deposition in our study area was 13.2 kg N ha −1 yr −1 in the 1980s and 21.1 kg N ha −1 yr −1 in the 2000s [13]. Alongside increases in N deposition there have been decreases in the ratio of NH 4 + -N to NO 3 − -N deposition, from approximately 5 to 2, from the 1980s to the 2000s, although NH 4 + -N remains the dominant form of N deposition [13]. Nationally, N deposition is a more serious issue in the north compared with the other regions of China [14].
Increasing N deposition could influence the production and emission of N 2 O by disturbing the balance between microbial N mineralization and immobilization, with the consequences for the relative availability of soil NH 4 + -N and NO 3 − -N [15] (Fig 1). Most studies report that raising N addition levels could linearly stimulate soil N 2 O emissions [4][5][16][17]. A metaanalysis of global N addition experiments showed that N additions increased soil N 2 O emissions by an average of 134% in terrestrial ecosystems [18]. Some plausible mechanisms have been proposed to clarify the promotion effect of N addition for soil N 2 O emissions: (1) Without additional N, the N retention in soil is mainly used by plants and microorganisms to maintain biomass and growth, so less N becomes lost as gaseous N [19]; (2) The amount of additional N greatly exceeds the atmospheric N deposition, thus leading to N accumulation in forest soil, which can benefit nitrifying and denitrifying bacteria [20] which would stimulate the nitrification rate and N 2 O emissions [21]. However, some studies indicated that N addition has no significant effect on soil N 2 O emission, which might be attributed to particular N addition threshold level for increased N 2 O emissions [7,22]. Thornton and Valente [23] found that the increased rate of soil N 2 O emissions was low at high N-addition levels; this may have occurred because the high level N addition to soil drove other limitations, such as carbon availability, thereby decreasing the C/N ratios that regulate the status of N saturation, which likely had a strong influence on N 2 O emission [24]. Furthermore, some studies have shown denitrification to be the main source of soil N 2 O emissions [25][26], whereas other studies reported that nitrification were primarily responsible for soil N 2 O emissions [7,[27][28]. Clearly then, how soil N 2 O emissions respond to additional N appears to be inconsistent.
The NH 4 + -N/NO 3 --N ratio showed a decreasing trend in our study area [29], and so clarifying the response of soil N 2 O emission to different forms and levels of N addition now is necessary. However, several previous studies that stimulated N deposition only considered NH 4 NO 3 [2,4,6], while others that did examine N deposition in varied N forms only reported their short-term effects on N 2 O emission [30]. In addition, some studies have focused on soil core incubations in the laboratory [31], which are conditions that differ greatly from those in the field. Therefore, from both a scientific and management perspective, further examination of the characteristics of different levels and forms of N addition is critically important for better understanding how N deposition affects soil N 2 O emissions in temperate forest soils.
In our study, we report the results of continuous measurements of soil N 2 O emissions over a 5-year period from a temperate forest in northern China. Based on the above analysis, we hypothesized that (1) N addition could increase soil N 2 O emission and that this promotion effect likely increased with the N addition level; (2) Applying NO 3 − and NH 4 + -N in combination could promote soil N 2 O emission more than would their respective single applications.

Study area
The study was conducted in a temperate forest of the Xi Mountain Experimental Forest Station (31˚54 0 32@ N, 110˚68 0 08@ E, 133 m a.s.l.) in Beijing, northern China. The station belongs to Beijing Forestry University. The study area is characterized by a temperate continental monsoon climate with a maximum air temperature of 31˚C in July and a minimum of -9˚C in January. Mean annual temperature is 11.6˚C and the average annual precipitation is 630 mm. During the 5-year experimental period, the yearly maximum and minimum temperatures were, respectively, 31, 31, 32, 33, 31˚C and -9, -8, -8, -5, -5˚C, while the total precipitation received annually was 721, 759, 508, 500, and 459 mm. At this research station, Quercus liaotungensis is the zonal vegetation with an average age of 62 years. The diameter at breast height, canopy closure, average height, and density were 9.7 cm, 69%, 8.4 m, and 2963 trees ha −1 . The soil here is classified as Chromic Luvisols (WRB Soil Classification) composed of 51% sand, 40% silt, and 9% clay. The thickness of the soil humus horizon (A horizon) is approximately 3−5 cm, and the O horizon thickness < 3 cm. Before starting the experiment, soil samples from the upper 10 cm of soil in each plot (with three replicates) were collected by using corers in March 2011. Initial soil properties were measured and showed no significant differences among the plots (Table 1).

Experimental design
The experiments were performed from March 2011 to February 2016. Seven 10 m × 10 m N addition plots, with three replicates each (n = 21 plots in total), were randomly established and distributed on a flat ground dominated by the Quercus liaotungensis community at the research station. To ensure plot independence, 1.5-m buffer strips were set up between adjacent plots. As deposition of NH 4 + -N and NO 3 --N showed great variation from month to month in the study area [30], three N-addition forms, namely NaNO 3 , (NH 4 ) 2 SO 4 , and NH 4 NO 3 , were used to simulate the effects of deposited NH 4 + -N, NO 3 --N, and their combination. According to the current level of atmospheric N deposition (30.6 kg N ha -1 yr -1 ) at the experimental site [30], two N-addition levels referred to as low N (L: 50 kg N ha -1 yr -1 ) and high N (H: 150 kg N ha -1 yr -1 ) were used to simulate a future increase in atmospheric N deposition by 1.5-fold and 5-fold. A control (0 kg N ha -1 yr -1 ) was used to calculate the net effect of naturally occurring N addition to the soil. From 2011 to 2015, additional N was evenly sprayed on the soil surface in plots by using sprayers, with eight equal applications made from March to October (i.e., the growing season). If it rained, the scheduled N addition was postponed to 1 day after the rain day. To reduce the effect of additional water on the experiment, control plots received an equivalent deionized water treatment.

Gas sampling and measurement
Soil N 2 O emission measurements were performed three times in the first week of each month, from March 2011 to February 2016. Soil N 2 O emissions were measured using a static closed opaque chamber and gas chromatography method [32]. The chamber was made of stainless steel and consisted of a fixed base and a removable top (without bottom, length × width × height = 50 cm × 50 cm × 50 cm). Before measurement, the base, which supported the sampling chamber, was installed into the soil at a depth of 20 cm for the entire experiment to avoid soil disturbance. Soil temperatures were measured in each plot at a depth of 5 cm nearby the chamber before and after collecting gas samples. And the average temperature value was used for emission calculation. The fixed base frame was free of vegetation. When collecting the gases, we inserted the removable top into the fixed base. The chamber was covered with thermal insulation cotton to reduce the impact of direct radiative heating in the chamber and a digital thermometer in the chamber was used to record its air temperature. Two fans were used to increase mixing and uniformity of air in the chamber. Gas samples were collected three times, from a sampling outlet at the top of the chamber, from 09:00 to 11:00 AM. local time on the first, fourth, and seventh day after N addition in each month from March 2011 to February 2016. If unpredicted extreme weather occurred, such as heavy rain or snow, this gas sampling was rescheduled. Gas samples were taken using 100 mL plastic syringes at intervals of 0, 10, 20, and 30 minutes after closing the chamber and inserting polyethylene-coated aluminum bags for soil N 2 O concentration analysis. Gas samples were analyzed within 6 h in a gas chromatograph (Agilent 7890A, Agilent Technologies Inc., Palo Alto, CA, USA) [33].
Soil N 2 O emissions were calculated as follows [33]: where, F N 2 O refers to N 2 O emission (μg m -2 h -1 ); D refers to the gas density of the chamber (mol m -3 ); D = WP/RT; W refers to the molar mass of N 2 O (g mol -1 ); P refers to air pressure (Pa); T refers to the air temperature inside the chamber (K); R refers to the gas constant (J mol -1 K -1 ); H refers to the height of the sampling chamber (m); and Δc/Δt denotes the linear slope of the concentration change over the measurement period.
where, F is the N 2 O emissions (μg m -2 h -1 ); i is the sampling number, i.e., samples collected in March had a value of 1 and those collected next February had a value of 12; and t is the sampling time based on the Julian day.

Soil sampling and measurement
Considering that N 2 O release mainly occurred in the mineral horizon, litter was first removed from the soil surface (O horizon < 3 cm) when sampling the soil. Soil samples at 0-10 cm depth were collected from near the static chambers monthly. Soil samples were passed through a 2-mm sieve to remove roots, gravel, and stones for soil analyses. Part of the fresh soil was used for soil NH 4 + -N and NO 3 --N content analyses, while the remaining portion was airdried for pH measurement. Soil NO 3 --N and NH 4 + -N concentrations were determined by the KCl extraction method [5]. Soil water content (WC) was measured using the standard oven-drying method at 105˚C for 8 h. Bulk density (BD) was determined by the core method.
Water-filled pore space (WFPS) (%) was calculated based on the equation: where 2.65 (g cm -3 ) refers to the assumed soil particle density.

Statistical analysis
All statistical analyses were conducted by SPSS v22.0 (IBM Corp., Armonk, USA) and the significance level for all statistical tests was set at P = 0.05. The differences in initial soil properties between different N-addition plots were examined using one-way analysis of variance

Soil N 2 O emissions under N addition
During the 5-year experimental period, the temperate forest soil was a net source of N 2 O. Soil N 2 O emissions were higher between May and September, but the values were lower and leveled off in other times of each year. Meanwhile, the peak of soil N 2 O emissions was concentrated in August of each year (Fig 2). Soil N 2 O emissions were significantly influenced by N forms, N levels, and the sampling time (P < 0.01), but the interaction effect of N forms and levels, months and N levels or months, N forms and N levels, did not significantly influence the soil N 2 O emissions (P > 0.05, Table 2).

Promotion effects of different N forms and levels
Different levels and forms of N addition and experimental time all significantly influenced the soil N 2 O emissions (P < 0.01, Table 2). As for the two N-level addition treatments, the N-addition treatments significantly increased soil N 2 O emissions, and this promotion effect was enhanced as the N-addition levels increased (Table 2, Fig 2). Soil N 2 O emissions ranged from 1.30 μg m -2 h -1 to 34.44 μg m -2 h -1 , with an average value of 11.55 μg m -2 h -1 in the control plots (Fig 2). Compared to the control, the average N 2 O emissions in the low-and high-level N addition plots significantly increased by 186.02% and 353.98%, respectively. The maximal emissions were obtained in August 2013 for the low and high nitrogen addition serials, which were 163.23 and 276.33 μg m -2 h -1 in the L-NH 4 NO 3 and H-NH 4 NO 3 addition plots respectively, for all the three added nitrogen forms (Fig 2). As for the N-addition treatments using the different forms of N, soil N 2 O emissions were significantly increased by NH 4 NO 3 , (NH 4 ) 2 SO 4 , and NaNO 3 additions in the order of NH 4 NO 3 > (NH 4 ) 2 SO 4 > NaNO 3 > control for the same level of N addition (Fig 2, Table 2). Compared to the control, the average N 2 O emissions in the NH 4 NO 3 , (NH 4 ) 2 SO 4 , and NaNO 3 addition plots significantly increased by 355.95%, 266.35%, and 187.71%, respectively (Fig 2).
There was no significant interaction between N form and N level on soil N 2 O emissions (P > 0.05, Table 2).

Interannual soil cumulative N 2 O emissions under N addition
Except for the interaction between N form and N level, year, N form and N level as well as all their interactions exerted significant effects on cumulative N 2 O emissions ( Table 2). In the Nitrogen deposition on soil N2O emissions control plot, cumulative N 2 O emissions showed no significant differences among the 5 years (Table 3). However, in the N addition plots, the promotion effect of additional N on soil N 2 O emissions increased over time in the initial years, but then it decreased. As for the three N-form additions, ANOVA showed that the annual emissions were basically elevated by NH 4 NO 3 , (NH 4 ) 2 SO 4 , and NaNO 3 additions in the order of NH 4 NO 3 > (NH 4 ) 2 SO 4 > NaNO 3 > control (P < 0.05), but no significant differences were found between NaNO 3 and control plots in the first year (Table 3, P > 0.05).

Environmental variables and their correlation with N 2 O emissions
During the 5-year period, air temperature had a clear seasonal pattern with higher temperatures in wet seasons (May to September) and lower in dry seasons (November to February). Soil temperature (ST) at the 5-cm depth fluctuated greatly, following changes in air temperature. The highest ST was 29.9˚C and lowest was -7.2˚C. WFPS ranged from 10.20% to 69.84% and fluctuated greatly (Fig 3). There were no significant differences among different N-addition plots on ST and WFPS (P > 0.05, Table 4). Soil NH 4 + -N and NO 3 --N concentrations exhibited significant seasonal variation, with a single peak value in the N addition plots. The maximum value appeared between June and August, while the minimum was observed from November to March (Fig 4, Fig 5).

Table 2. Summary of repeated measures ANOVA results (F values) indicating the effects of different forms and levels of N addition and experimental time on temporal variation of soil N 2 O emissions and annual cumulative
Soil NO 3 --N accumulated significantly in N addition plots and its concentration ranged from 15.24 to 53.33 mg kg -1 . N level had a significant promotion effect on soil NO 3 --N concentrations, with those under the and high level N addition had a significantly greater promotion on it compared with that of low level (P < 0.05, Table 4). The concentrations from lowand high-level N addition plots were, respectively, 142.14% and 172.90% greater than those from the control (12.03 mg kg -1 ). Soil NH 4 + -N significantly accumulated in the N addition plots and its concentration ranged from 2.32 to 6.74 mg kg -1 . The accumulation of NH 4 + -N caused by N addition was less than that of NO 3 --N in soil. Soil NH 4 + -N concentration was significantly influenced by the   three N forms, which increased NH 4 + -N in the order of (NH 4 ) 2 SO 4 > NH 4 NO 3 > NaNO 3 and by 57.40%, 36.27%, and 31.84% when compared with the control (2.98 mg kg -1 ), respectively (Fig 4, Fig 5). The correlation analysis showed that soil N 2 O emissions were positively correlated with ST at 5 cm depth, WFPS at a 10-cm depth, and soil inorganic nitrogen concentration (Fig 6). In addition, a linear equation showed that soil N 2 O emissions were extremely significantly (P < 0.01) correlated with ST, WFPS, and soil NH 4 + -N and NO 3 --N (Fig 6). Nitrogen deposition on soil N2O emissions

Promotion effects of N addition on soil N 2 O emissions
Our results showed that the temperature plantation in northern China was a source of atmospheric N 2 O under natural conditions. The mean N 2 O emissions value in the control was 11.55 μg N 2 O-N m -2 h -1 ; this rate is comparable to that reported by Butterbach-Bahl et al. [16] who found that the N 2 O emissions in soils of spruce forests in Germany and Ireland ranged from 3.5 to 16.4 μg N m -2 h -1 . In our study, NaNO 3 , (NH 4 ) 2 SO 4 , and NH 4 NO 3 addition at levels of 50 and 150 kg N ha -1 yr -1 significantly increased soil N 2 O emissions by an average of 115.26% to 260.15%, 182.92% to 349.77%, 259.89% to 452.02%, respectively. The rate of increase was lower than that for a subtropical forest of the Qianyanzhou Ecological Station, where it was increased by 403% to 762% [5]. Except for L-NaNO 3 addition, the increase in soil N 2 O emission was higher than the global average (134%) [18]. On one hand, these results indicate that the temperature plantation had high turnover rates of soil N and responded to the increased N deposition. On the other hand, to measure the peak N 2 O emissions in our study, the gas samples were collected in the first, fourth and seventh days after N was added; hence, the cumulative N 2 O emissions might have been overestimated since the N 2 O emissions should have been measured weekly during growing season (to properly reflect an average impact over time). In our previous work, N addition significantly increased the amount of soil microbes and changed the soil microbial community structure in our study area [36]. Soil urease activities were significantly increased by N additions, which promoted soil N 2 O emission [37]. Therefore, the weakened N limitation brought about by a higher litter decomposition rate and greater microbial activity could explain the increased N 2 O emissions we found here [38].

N 2 O emissions under different N addition forms and levels
Based on our observations over 5 years, the results supported our hypothesis that soil annual cumulative N 2 O emissions increased under elevated N-addition levels. Positive correlations between N-addition levels and soil N 2 O emissions have been found in many previous studies [4,6,[39][40]. However, at our site, NaNO 3 addition at a rate of 50 kg N ha -1 yr -1 did not stimulate a significant increase in the cumulative N 2 O emissions in the first year. Perhaps this is because of a threshold response of soil N 2 O emissions to the N additions [4,7,41]. Specifically, such a response is determined by the competition between plants and soil microbes for available N, and thus emissions will not significantly increase until the plant N demands have been satisfied [4,7,42].
Considering sions from ammonium sources than from nitrate sources [7,26]. Two potential mechanisms may be responsible for this phenomenon: (1) high immobilization of NO 3 --N and nitrification rates, coupled to a low denitrification potential, led more NO 3 --N to accumulate in soil [44]; (2) poor mobility of NH 4 + created depletion zones around the plant roots, leaving more N input exposed to microorganisms in soils. However, most research to date suggests that denitrification is the main process driving N 2 O production [25,45]. Yet when WFPS is in the range of 30 to 70%, nitrification can become the main process driving N 2 O production, as denitrification rates increase rapidly when WFPS exceeds 60% [7]. WFPS in our research plots was at a low level for most of the 5-yr monitoring period, only exceeding 60% for a few months, which likely provided less than optimal conditions for the denitrification process [46]. Given this trend in WFPS, we indirectly conclude that NH 4 + -N had higher conversion efficiency to N 2 O than NO 3 --N at our forest site. . In our study area, N was clearly a limiting factor in the initial years based on amount and stimulating effect of N addition upon tree biomass.
Continuing the N addition could shift the soil from being N-limited to N-rich, and then becoming N-saturated, such that soil N 2 O emissions may appear to reach a steady state at high N levels [23]. In addition, Liu and Song [51] found that soil microbial activities may be limited by carbon availability when N is abundant. The suppression of soil N 2 O emissions by longterm N additions was possibly due to a lack of readily available organic carbon [52] and/or adverse effects on mineralization of organic carbon under conditions of high N addition [53]. Therefore, our field experiment highlights the importance carrying out long-term studies to avoid possibly overestimating the N addition effects on N 2 O emissions from short-term observations.

Relationships between soil N 2 O emissions and soil properties
In our study area, the soil concentration of NO 3 --N was higher than that of NH 4 + -N, and the accumulation of NO 3 --N caused by N addition was more than soil NH 4 + -N concentration. On the one hand, although the soil NO 3 --N concentration was directly increased by NO 3 --N addition, the NH 4 + -N addition could have enhanced the activity of soil nitrifiers and led to the NO 3 --N accumulation in soil we found. This finding and interpretation is consistent with some previous studies carried out in tropical and subtropical forests [54][55]. On the other hand, several studies using the 15 N tracing method suggest that plants in temperate forest at our site preferred NH 4 + -N, which led to more NH 4 + -N becoming assimilated, such that the accumulation of NH 4 + -N in the soil was relatively little and brief [56]. We found that the soil N 2 O emissions were significantly correlated with concentrations of soil NH 4 + -N and NO 3 --N, suggesting soil N 2 O emission was dominated by both nitrification and denitrification processes. Since atmospheric N deposition can significantly promote soil N 2 O emission, and exogenous NH 4 + -N and NO 3 --N inputs into temperate forests may have synergic effects on soil N 2 O emission, in the future both of these aspects ought to be distinguished in the dynamics of the N cycle and balance in terrestrial ecosystems by using 15 N tracer methods. High ST, together with a relatively high WFPS, tend to promote both nitrification and denitrification processes [57] and consequently, high N 2 O emissions, an interpretation that is consistent with many previous findings [58][59]. In particular, high WFPS may promote microbial movement and the expansion of the soil anaerobic microbial community [43]. Warm temperatures benefit soil nitrifying and denitrifying bacteria activities [11], which may explain the seasonal variation in the relatively high N 2 O emissions that occurred from May to September that we observed in this study. Many other complex factors may have a played a role in determining our results, such as soil pH, soil C availability, and the microbial community structure, since they jointly influence the two key processes of nitrification and denitrification that are involved in soil N 2 O production [36,60].

Conclusions
This study emphasizes the effects of different N forms and levels on N 2 O emissions from a temperate forest over 5-year experimental period. We found that the accumulation of soil NO 3 --N was significantly higher than that of soil NH 4 + -N due to N addition. N addition initially promoted soil N 2 O emission yet this promoting effect, although it existed, weakened in the following years. High level N addition had a stronger promotion effect upon soil N 2 O emission than did the low level N addition. Meanwhile, the combined application of NH 4 + -N and NO 3 --N promotes N 2 O emissions more than their single applications, and NH 4 + -N addition had a stronger promotion effect for soil N 2 O emission than did the NO 3 --N addition. In addition, WFPS, ST, soil NH 4 + -N, and NO 3 --N were all positively related to the N 2 O emissions. In the future, the long-term observation of soil N 2 O emissions, and the measurement of microbial functional groups using 15 N tracer methods, will be necessary to clarify the mechanisms responsible for the soil N 2 O emissions.
Supporting information S1 File. Data set underlying the findings. (XLSX)