The study was conducted in a semi-arid grassland in Duolun County, Inner Mongolia Autonomous Region, China. The sampling site is situated in a meteorological observation station of the Chinese Academy of Sciences (42° 22′ N, 116° 49′ E; 1324 m a.s.l.). This site features a typical semi-arid continental monsoon climate. The average annual precipitation is 380 mm and the maximum precipitation occurs in July. The growing season generally starts in May and continues until the end of September, with a length of about 150 days. Precipitation amount in the growing season accounts for 88% of the total annual precipitation. The annual average temperature is 2.2° C and the monthly average temperature ranges from -16.9° C in January to 19.4° C in July. The dominant soil type is chestnut soil. Livestock grazing and crop cultivation are the two major land-use practices in this area.

2.2.1 Dry N deposition

Reactive nitrogen (Nr) components, including NH₃, NO₂, HNO₃, particulate NH₄⁺ (pNH₄⁺) and particulate NO₃^- (pNO₃^-), were sampled monthly from July 2013 to December 2015. NH₃, HNO₃, pNH₄⁺ and pNO₃^- were measured with a DELTA (DEnuder for Long Term Atmospheric sampling) system. The DELTA system consisted of three parts: a circular sampling chain with diffusers and a particle collection device, a small low-volume pump (0.2-0.4 L/min) and a high-precision gas flow meter. Details about the DELTA system have described in detail previously (Flechard et al., 2011; Xu et al., 2015). Briefly, when ambient air passes through the DELTA system, HNO₃ and other acidic gases are captured by two long diffusion tubes coated on the inner wall with a mixed solution of 10 g/L K₂CO₃ and 10 g/L glycerol in methanol; and NH₃ is captured by another two short diffusion tubes coated with citric acid (5% (m/v) in methanol). Subsequently, the acidic particulate aerosols are collected by a filter membrane with basic adsorbent in the DELTA train membrane system, and the alkaline particulate aerosols are collected by a membrane with acidic adsorbent.

NO₂ was monitored with Gradko passive tubes (Gradko International Limited, Winchester, UK), which are composed of an acrylic acid tube (71.0 mm in length and 11.0 mm in internal diameter), two polyethylene caps (gray and white, located at both ends of the diffuser) and two layers of stainless steel wire. The NO₂ was absorbed into 20% triethanolamine solution coated onto two stainless steel wire meshes in the gray cap. The sampling heights of the DELTA system and NO₂ diffusion tubes were kept at 2 m above ground. Three NO₂ sampler replicates were deployed for monthly sampling. Exposed DELTA sample trains and NO₂ diffusion tubes were replaced with new ones at the end of a month and were preserved in a refrigerator at 4° C, and then analyzed immediately after collection. The gaseous HNO₃ collected in glass denuders of the DELTA sampling train system and the pNO₃^- on the filter membrane were extracted using 10 mL 0.05% H₂O₂ solution, and the NH₃ and pNH₄⁺ denuders and membranes were extracted using 6 and 10 mL high purity water, respectively. An AA3 continuous flow analyzer (Bran+Luebbe GmbH, Norderstedt, Germany) was utilized to measure NH₄⁺ and NO₃^- concentrations in the extracted solutions. NO₂ in the Gradko passive sampler was extracted with a mixed solution of sulfonamide, H₃PO₄, and N-1-naphthylethylene-diamine, and NO₂ concentration was determined by absorbance at a wavelength of 542 nm. For quality assurance purposes, we set up monthly blank samples both in the laboratory and in the field and analyzed them using the same procedures outlined above.

2.2.2 Wet/bulk N deposition

Wet/bulk N deposition was collected by a precipitation gauge (SDM6, Tianjin Weather Equipment Inc., China) at the sampling site from May 2013 to November 2015. The precipitation gauge was a passive sampler without power input, consisting mainly of a stainless steel funnel and a glass container. Precipitation was automatically collected by the precipitation gauge and the amount was measured with a graduated cylinder of the gauge. Before being taken to the laboratory of China Agricultural University for chemical analysis, the collected precipitation was recorded, mixed, and put into a clean polyethylene bottle (50 mL) inside a refrigerator (-18° C). It should be noted that we rinsed the precipitation gauge with high purity water after each collection to avoid sample contamination. The precipitation samples were filtered by 0.45 μ m filter membranes, then 15 mL filtered fluid was taken out, frozen and kept in a polyethylene bottle. NH₄⁺-N and NO₃^--N were determined by a continuous flow analyzer (Bran+Luebbe GmbH, Norderstedt, Germany), as mentioned above, within a month.

We calculated the wet/bulk N deposition flux based on NH₄⁺-N and NO₃^--N concentrations and precipitation amount in each precipitation event. The volume-weighted mean concentration of Nr components (C_w; mg N/L) was obtained with Equation 1:

${{C}_{w}}=\frac{\sum\nolimits_{i=1}^{n}{\left( {{C}_{i}}\times {{P}_{i}} \right)}}{\sum\nolimits_{i=1}^{n}{{{P}_{i}}}}, $ (1)

where C_i is the inorganic Nr concentration in the i^th precipitation event (mg N/L); P_i is the precipitation amount in the i^th precipitation event (mm); n is the number of precipitation events. The wet/bulk deposition flux of Nr components (D_w; kg N/hm²) was calculated using Equation 2:

${{D}_{w}}=0.01\times {{C}_{w}}\times \sum\nolimits_{i=1}^{n}{{{P}_{i}}}$. (2)

Dry N deposition is affected by various chemical and environmental factors like meteorological conditions, physicochemical properties of Nr components, roughness of underlying surface, and the ability of the underlying surface to capture and absorb Nr components (Flechard et al., 2011). Extra issues arise concerning characteristics of the bidirectional flow of N, making the quantification of dry deposition a challenge. Nr emission is the predominant process when the concentration in the surface is higher than that in the atmosphere, while deposition is expected when the gradient reverses. Inferential models have previously been extensively used as operational tools to determine dry N deposition in many monitoring networks all over the world, such as NADP (National Atmospheric Deposition Program) and EMEP (European Monitoring and Evaluation Program). Using the inferential method, we obtained the monthly dry deposition flux of Nr components by multiplying the concentrations of Nr components measured in our field samples with deposition velocities obtained from simulation results of a global atmospheric chemical transport model. The dry deposition flux of Nr components was expressed by Equation 3:

F_d=C_z× V_d, (3)

where F_d is the dry deposition flux of an Nr component (kg N/hm²); C_z is the measured atmospheric concentration of the Nr component (μ g N/m³); and V_d is the dry deposition velocity (cm/s). It should be noted that V_d was obtained from a nested-grid version of the Goddard Earth Observing System (GEOS)-Chem chemical transport model (Chem CTM) for Asia. The model has a spatial horizontal resolution of 0.50° latitude× 0.67° longitude and a temporal resolution of 6 h. The V_d estimation of dry N deposition follows a standard big-leaf resistance-in-series model, which is determined by meteorological conditions and land-use types (grassland in the present study). We averaged the hourly V_d data to obtain monthly values of dry deposition velocities of all Nr components. Details about the model description and the simulation of dry N deposition fluxes can be found in Zhang et al. (2012). It should be mentioned that a part of the wet/bulk and dry N deposition results (e.g., data from July 2013 to September 2014) in present work was published previously by Xu et al. (2015).

Monthly concentrations of NH₃, NO₂, HNO₃, pNH₄⁺ and pNO₃^- in atmosphere were in the ranges of 0.34-6.79, 0.24-6.92, 0.01-0.53, 0.84-5.63 and 0.33-1.48 μ g N/m³, respectively, from July 2013 to December 2015 (Fig. 1). The seasonal variation of NH₃ was remarkable with an annual mean concentration of 2.33 μ g N/m³. The highest concentration of NH₃ occurred in summer (June-August) while the lowest in winter (December-February). The annual mean concentration of NO₂ was 1.90 μ g N/m³ with the highest concentration in April and the lowest in January. The annual mean concentration of HNO₃was 0.18 μ g N/m³. Compared with NH₃ and NO₂, the monthly concentrations of HNO₃ were lower and also less variable, except that the concentration in September was notably higher than those in other months. The lowest concentration of HNO₃ occurred in May. The annual mean concentration of pNH₄⁺ was 1.42 μ g N/m³ with the maximum monthly concentration in July and the minimum in January. Similar to pNH₄⁺, the monthly concentrations of pNO₃^- also showed the highest value in July. The annual mean concentration of pNH₄⁺was 0.62 μ g N/m³.

Fig. 1

	Figure Option ViewDownloadNew Window
	Fig. 1 Monthly concentrations of Nr (reactive nitrogen) components (including NH3, NO2, HNO3, pNH4+ and pNO3-) in the atmosphere from July 2013 to December 2015

During the monitoring period (i.e., from May 2013 to December 2015), the monthly volume-weighted concentrations of NH₄⁺ and NO₃^-in precipitation varied from 0.27 to 12.19 and 0.10 to 11.68 mg N/L, respectively (Fig. 2). The average annual concentrations of NH₄⁺-N and NO₃^--N in precipitation were 2.71 and 1.99 mg N/L, respectively. Compared with NO₃^--N, the NH₄⁺-N concentration in precipitation was generally higher except for values in September and October, 2013. Higher concentrations of NO₃^--N and NH₄⁺-N were observed in the months (February-May) when the precipitation was lower, while lower concentrations were always accompanied by higher monthly precipitation (June-September). The precipitation amount, to a large extent, affected the concentrations of NO₃^--N and NH₄⁺-N.

Fig. 2

	Figure Option ViewDownloadNew Window
	Fig. 2 Monthly volume-weighted concentrations of Nr components (including NO3--N and NH4+-N) in precipitation from May 2013 to December 2015

Monthly dry deposition velocities of each Nr component are listed in Table 1. Based on these values, we calculated the annual dry deposition fluxes of NH₃, NO₂, HNO₃, pNH₄⁺ and pNO₃^- (Fig. 3). Specifically, annual dry deposition fluxes were 3.17, 1.13, 0.63, 0.91 and 0.36 kg N/(hm²• a) for NH₃, NO₂, HNO₃, pNH₄⁺ and pNO₃^-, respectively, during the monitoring period. It should be noted that the monitoring data in 2013 were not included in the calculation of annual dry deposition flux since the data were not complete for 2013. During the monitoring period (excluding the data of 2013), the summed annual dry deposition flux of the five atmospheric Nr components (i.e., NH₃, NO₂, HNO₃, pNH₄⁺ and pNO₃^-) averaged 6.21 kg N/(hm²• a). Due to the high concentration of NH₃, the dry deposition flux of NH₃ was also the largest one among the five atmospheric Nr components, followed by NO₂, HNO₃, pNH₄⁺ and pNO₃^-. Generally, NH₃ contributed 51% to the total dry N deposition, being higher than the combined contribution of the other four Nr components (49%).

Table 1

Table 1

Table 1 Monthly dry deposition velocities of different atmospheric Nr components Month NH3 NO2 HNO3 pNH4+ pNO3- (cm/s) January 0.26 0.01 0.15 0.13 0.13 February 0.30 0.01 0.17 0.15 0.15 March 0.30 0.03 0.37 0.18 0.18 April 0.30 0.07 0.87 0.22 0.22 May 0.40 0.24 2.06 0.26 0.26 June 0.55 0.38 2.44 0.27 0.27 July 0.55 0.38 2.36 0.23 0.23 August 0.53 0.34 2.26 0.24 0.24 September 0.46 0.26 2.19 0.22 0.22 October 0.34 0.13 1.59 0.16 0.16 November 0.27 0.03 0.67 0.13 0.13 December 0.27 0.01 0.14 0.13 0.13

Table 1 Monthly dry deposition velocities of different atmospheric Nr components

From 2013 to 2015, the annual wet/bulk deposition fluxes fluctuated in the ranges of 3.22-7.90 kg N/(hm²• a) for NH₄⁺-N and 2.72-3.38 kg N/(hm²• a) for NO₃^--N. The average annual wet/bulk deposition flux of NH₄⁺-N (5.37 kg N/(hm²• a)) was 1.7 times that of NO₃^--N (3.16 kg N/(hm²• a)). The annual wet/bulk deposition flux was averagely 8.53 kg N/(hm²• a) and the annual N deposition (dry deposition plus wet/bulk deposition) flux was averagely 14.73 kg N/(hm²• a).

Fig. 3

	Figure Option ViewDownloadNew Window
	Fig. 3 Seasonal variations of dry and wet/bulk deposition fluxes of Nr components (NH3, NO2, HNO3, pNH4+, pNO3-, NH4+-N and NO3--N)

In China, fossil fuel is intensively consumed to support the growth of population and economy, leading to large amounts of Nr emissions to the atmosphere and subsequent N deposition to the ecological ecosystems. At the same time, lack of effective measures to control atmospheric Nr emissions aggravates the increasing trend of N deposition to the ecological ecosystems (Liu et al., 2013). Consequently, China has become one of the world’ s hotspots in N deposition (Vet et al., 2014) and suffered from the threat of continued and serious Nr pollution.

The current study found that monthly average concentrations of Nr components in atmosphere and in precipitation in grassland of Inner Mongolia were considerably lower than those measured in two well-developed regions: North China Plain (Luo et al., 2013; Xu et al., 2016) and Southwest China (Kuang et al., 2016). Due to the burning of fossil fuels and the intensified agricultural activities and livestock production, the North China Plain has been a hotspot of NH₃ in China (Liu et al., 2016; Pan et al., 2016a) with the largest annual mean NH₃ concentration of 16.9± 5.9 μ g N/m³ in Quzhou (Xu et al., 2016). This value is far higher than the annual mean NH₃ concentration observed from this study (2.33 μ g N/m³). The results of atmospheric NH₃ concentrations observed in our study were higher than those observed in some European countries (Flechard et al., 2011). Although the annual NH₃ concentration at the sampling site was within the range of regional background NH₃ concentration in China (Meng et al., 2010), it exceeded or was very close to the critical level of 3 (± 1) μ g N/m³ that was set for protecting the most fragile regions (Cape et al., 2009). Therefore, the observed atmospheric NH₃ concentrations posed a potential environmental threat to the grassland in Inner Mongolia. During the monitoring period, the monthly NH₃ concentrations exhibited the highest value in summer and the lowest value in winter. The seasonal variations were caused by season-related human activities and meteorological factors. In summer, high temperature was conducive to enhancing microbial activities and strengthening microbial ammonification. Moreover, high temperature favored ammonia volatilization and also favored NH₃ emission from fertilizer application during potato cultivation near the grassland. During the growing season, herdsmen applied N fertilizer to the grasses to promote the growth of pastures, further elevating the NH₃ concentration in summer. In winter, however, the low temperature severely restricted biological activities and the fertilizer application was minimal. Consequently, less ammonia volatilized. High wind speeds in winter also favored the dispersion of pollutants, further lowering the NH₃ concentration.

It is commonly accepted that NO_x is mainly produced in various combustion processes, including those associated with transportation, industry, coal use and power plants (Holland et al., 2005; Paulot et al., 2013). Due to a low level development of industrialization in the sampling site, NO₂ concentrations were remarkably lower than those reported in most urban sites in China (Xu et al., 2016). However, the annual mean NO₂concentration during the monitoring period was more than two times higher than previous value measured by Shen et al. (2009). The recent increase in the numbers of vehicles can partially explain the increased NO₂ concentrations. The number of private cars in Inner Mongolia grew from 1.88× 10⁶ in 2011 to 3.35× 10⁶ in 2015 (National Bureau of Statistics of China, 2011-2015). The higher NO₂ concentrations were observed in spring and autumn, being inconsistent with typical seasonal variations (Xu et al., 2015). Higher NO₂ concentrations in spring might be attributable to coal combustion for domestic heating. HNO₃ exhibited high concentrations in autumn and low concentrations in summer, being consistent with the seasonal variations of NO₂.

The annual mean pNH₄⁺ concentration measured in the present study was much lower than many other observations in China (Xu et al., 2015) and in west Europe (Holland et al., 2005), but it was comparable to that measured at rural sites in USA (Li et al., 2016) and was higher than that reported in UK (Marner and Harrison, 2004). The concentration of pNH₄⁺ in summer was higher than those measured in spring, autumn or winter. The high NH₃ concentration and high relative humidity in summer were likely important causes of this phenomenon. High concentration of NH₃ and high relative humidity both favored the transition from gaseous NH₃ to particulate NH₄⁺, although increased temperatures did not favor particulate ammonium nitrate formation. Because of the low HNO₃ concentrations, the pNO₃^- concentrations were lower than most results measured in China (Pan et al., 2012; Xu et al., 2015). The seasonal pattern of pNO₃^- concentrations was the same as that of pNH₄⁺ concentrations with the higher values in summer and lower values in other seasons.

N in precipitation mainly derived from water-soluble atmospheric Nr components. These pollutants were scavenged from the air by rainfall or snowfall. The annual concentrations of NH₄⁺-N and NO₃^--N in wet/bulk deposition in the sampling site were also lower than those reported in many previous studies conducted in urban and rural sites in China (Pan et al., 2012; Huang et al., 2013), but they were higher than those observed in a semi-arid region in Urumqi (Li et al., 2012). In our study, the highest monthly average concentrations of NH₄⁺-N and NO₃^--N occurred in summer and the lowest in winter, showing a significant negative correlation between N concentration and precipitation. This pattern suggests a dilution effect of precipitation on N concentrations (Jia and Chen, 2010; Kuang et al., 2016). Compared with previously published data (Xu et al., 2015), annual concentration of NH₄⁺-N in precipitation increased by 19% while annual concentration of NO₃^--N reduced by 6% in grassland of Duolun County.

A different seasonal pattern of N deposition was observed at the sampling site (Fig. 4). The highest deposition fluxes of Nr components occurred in summer and the lowest values always appeared in winter with an exception of HNO₃ that had the highest deposition flux in autumn. In summer, NH₃ deposition fluxes were markedly higher than those in spring, autumn and winter, accounting for 57% of the annual NH₃ deposition flux. This is consistent with the seasonal variation of NH₃ concentration. Deposition fluxes of NO₂ and pNH₄⁺in summer dominated the annual deposition. Deposition of pNH₄⁺ in spring, autumn and winter contributed equally to the annual deposition, and the seasonal variations of pNO₃^-deposition were similar to those of pNH₄⁺ deposition. Compared with other Nr components, the flux of HNO₃ dry deposition showed different seasonal variations with the highest value in autumn, contributing 56% of estimated annual HNO₃ deposition. This pattern can be partially explained by the higher concentrations and higher deposition velocities of HNO₃ in autumn. Due to different emission sources and different meteorological conditions, the percentage contribution of individual Nr component to the total dry N deposition varied with geographical locations. In the current study, NH₃ contributed the most to the total dry N deposition (51%). This value was significantly higher than the contributions of NO₂ (18%), HNO₃ (10%), pNH₄⁺ (15%), and pNO₃^- (6%) to the total dry N deposition. The gaseous Nr components accounted for 79% of the total dry N deposition in our study, being highly consistent with the finding in North China Plain (Xu et al., 2016).

Similar to dry N deposition, wet/bulk deposition fluxes of NH₄⁺-N and NO₃^--N and total N deposition were higher in summer than in other seasons, being in a good agreement with the seasonal distributions of precipitation at the sampling site. In summer, precipitation accounted for 62% of the annual total, while in winter, precipitation was relatively small. Positive relationships between precipitation amounts and wet N deposition fluxes have been reported in many previous observations (e.g., Pan et al., 2012; Huang et al., 2013). Deposition fluxes of NH₄⁺-N and NO₃^--N in summer (average values of 2.64 and 1.41 kg N/(hm²• a), respectively) contributed 49% and 44% to the annual wet/bulk deposition, respectively. Both higher precipitation amounts and high atmospheric concentrations of Nr components contributed to the high N deposition fluxes in summer. The deposition flux of NH₄⁺-N was 1.87 times that of NO₃^--N in summer, 1.85 times in spring, 1.07 times in autumn and 1.07 times in winter. The ratio of NH₄⁺-N concentration to NO₃^--N concentration is widely perceived as an indicator of the sources of atmospheric Nr components (Huang et al., 2013). Specifically, when the ratio is larger than 1, it is commonly accepted that Nr mainly originated from agricultural activities; whereas it is commonly accepted that Nr mainly originated from industrial activities when the ratio is lower than 1 (Xu et al., 2015). In this study, the ratios of NH₄⁺-N concentration to NO₃^--N concentration in summer and spring indicated that N in wet/bulk deposition mainly originated from agricultural activities. In autumn and winter, the contributions of agricultural and industrial activities to wet/bulk N deposition were equal.

During the monitoring period, the annual mean fluxes of oxidized (e.g., HNO₃ and pNO₃^-) and reduced (e.g., NH₃ and pNH₄⁺) dry N deposition were 2.12 and 4.09 kg N/(hm²• a), respectively. Dry N deposition estimated in the current study was obviously lower than those observed in northern China (Shen et al., 2009; Pan et al., 2012; Luo et al., 2013) but higher than those monitored at rural sites in Canada (Zhang et al., 2009). The reduced N accounted for 66% of dry N deposition, being consistent with previous studies conducted in China (Xu et al., 2015). NH₃deposition accounted for 78% of dry reduced N deposition. It’ s universally acknowledged that the major sources of atmospheric NH₃ are livestock wastes and agricultural N fertilizers (Duan et al., 2016). According to Kuang et al. (2016), the amounts of NH₃ emitted from chemical fertilizer use and animal husbandry in China increased from 4.9 Tg N in 1980 to 8.0 Tg N in 2012. The amount of agricultural fertilizer application increased by 13% from 2011 to 2015 in Inner Mongolia (National Bureau of Statistics of China, 2011-2015). NO₂ was the largest contributor to oxidized N in dry deposition, followed by HNO₃ and pNO₃^-. The mean deposition fluxes of NO₂, HNO₃ and pNO₃^- were 1.13, 0.63 and 0.36 kg N/(hm²• a), respectively, corresponding to 53%, 30% and 17% of oxidized N in dry deposition. NH₃ and NO₂ were the two major contributors of dry N deposition at the sampling site, together contributing to 69% of dry N deposition. NO₂ is a precursor to HNO₃. NH₃ can react with acid gases such as H₂SO₄ and HNO₃ in the air, promoting the formation of PM_2.5. PM_2.5 reduces atmospheric visibility and long-term exposure to elevated PM_2.5 concentrations can result in respiratory and cardiovascular diseases (Zheng et al., 2015). Due to strict standards of implementation and technical improvements to limit atmospheric pollution, NO_x emissions almost dropped by 41% from 1990 to 2010 in USA (Li et al., 2016). Butler et al. (2005) quantified the effect of reduced emission of NO_x on N deposition and the results showed that when NO_x emissions were reduced by 50%, the total NO₃^- deposition would be decreased by 37% and the total N deposition would be dropped by 25%. This work well manifested the importance of controlling NO_x emissions, especially in regions upwind of sensitive ecosystems. China also has made efforts to control NO_x emissions. For example, in the 12^th Five-Year-Plan (2011-2015), NO_x emission was proposed to reduce by 10%. And, in late 2012, a stricter NO_x emission standard was released by Ministry of Environmental Protection of the People’ s Republic of China (http://www.zhb.gov.cn/). However, we still have a long way to go to controlling N deposition.

Deposition fluxes of NH₄⁺-N and NO₃^--N in precipitation at the sampling site were lower than those in regions with serious N pollution in northern China (Pan et al., 2012). However, compared with the values estimated in Europe, the wet/bulk deposition fluxes in the current study were higher (Cape et al., 2012). Wet/bulk N deposition fluxes in the current study were also higher than those in Urumqi of Xinjiang (Li et al., 2012). Zhu et al. (2015) found that wet N deposition, precipitation, N fertilizer application, and energy consumption were closely related and, therefore, wet N deposition could be reduced to acceptable levels by cutting down energy consumption and also by lessening N fertilizer application. Reduced N dominated wet/bulk deposition with a contribution of 63%, being similar to the results reported in USA (Li et al., 2016). Li et al. (2016) pointed out that wet N deposition in USA changed in recent decades from being nitrate-dominated to ammonium-dominated and that the changes were due to decreases in nitrate wet deposition fluxes in most regions (consistent with NO_x emission reductions) and increases in ammonium wet deposition. In China, the reduced N also dominates the total N deposition, but the percentage of oxidized N in total deposition is increasing rapidly (Liu et al., 2016), demonstrating the importance of implementing continuous NO_x emission control measures.

Total inorganic N deposition (dry and wet/bulk) was 14.7 kg N/(hm²• a) in the present study, being close to the upper limit of the N critical load (10-15 kg N/(hm²• a)) (Bobbink et al., 2010). In this study, about 58% of the total N deposition was from wet/bulk deposition and 42% from dry deposition. The annual mean N deposition measured from this study was slightly higher than that estimated at Ireland grasslands (Henry and Aherne, 2014), but was lower than the values in most studies carried out in different sites in China (Pan et al., 2012; Xu et al., 2015). The fractional contribution of dry N deposition in the current study (42%) was lower than those in most studies in China (Pan et al., 2012; Wang et al., 2013; Xu et al., 2015; Kuang et al., 2016).

Reduced N constituted 64% of the total N deposition in the current study, 1.5 times higher than the contribution from oxidized N. Experimental studies have shown that reduced N is more effective than oxidized N in driving the decline of biodiversity in acidic and mesotrophic grasslands (Stevens et al., 2011). The currently recognized N saturation critical load in terrestrial ecosystems is 10-15 kg N/(hm²• a), a value derived from field experiments (Bai et al., 2010). Total N deposition obtained from the current study exceeded the lowest critical load value and may potentially pose threats to ecosystems. Field N addition experiments conducted on temperate grassland in Inner Mongolia indicated that increased N deposition led to a decrease of plant species richness (Zeng et al., 2016) and an alteration of species composition of plant communities (Duprè et al., 2010), posing a threat to ecosystem stability. In addition, when N deposition increases, a sizable species richness loss is expected in regions with low N deposition (Stevens et al., 2011).