Atmospheric Nitrogen Deposition to a Southeast Tibetan Forest Ecosystem

: With atmospheric reactive nitrogen (Nr) emissions increasing globally, research into Nr deposition has attracted increasing attention, especially in remote environments. These ecosystems are very sensitive to global change, especially enhanced Nr deposition. Forest environments, in particular, are highlighted because of their important ecological function. We quantiﬁed atmospheric Nr concentrations and deposition over four years of continuous monitoring in a southeast Tibetan boreal forest ecosystem, an ecosystem in which forest biomass and carbon density are high around the world. Average annual bulk Nr deposition was 3.00 kg N ha − 1 y − 1 , with those of reduced and oxidized species estimated at 1.60 and 1.40 kg N ha − 1 y − 1 , respectively. Bulk deposition of both NH 4 + and NO 3 − were controlled by precipitation amount: both Nr deposition and precipitation were highest in summer and lowest in winter. Dry deposition of NH 3 and NO 2 were 1.18 and 0.05 kg N ha − 1 y − 1 , respectively. Atmospheric NH 3 concentrations were in the range 1.15–3.53 mg N L − 1 , highest in summer and lowest in winter. In contrast, no clear trend in seasonal NO 2 concentrations was observed. Monthly NO 2 concentrations were 0.79–1.13 mg N L − 1 . Total Nr deposition (bulk plus dry) was 4.23 (3.00 + 1.23) kg N ha − 1 y − 1 in the forest. Reduced nitrogen was the dominant species. In conclusion, Nr deposition was in the range at which forest net productivity and carbon sequestration are sensitive to any variation in nitrogen input, so quantiﬁcation of Nr deposition should continue and with greater detail.


Introduction
With rapid economic growth (e.g., agricultural, industrial, urbanization), atmospheric reactive nitrogen (Nr) deposition has increased significantly and has caused a series of ecological and environmental problems [1,2], especially in China [3] and many developing countries in recent years [4]. In order to evaluate the potential effects caused by additional nitrogen inputs via atmospheric deposition, Nr deposition has been measured in different environments around the world [5][6][7], including forest ecosystems [8][9][10].
Nr deposition to forests can result in both harmful and beneficial effects. Generally, additional Nr inputs to already N-rich ecosystems result in soil acidification, nutrient imbalance and a decrease in plant diversity [11,12]. However, positive effects, such as additional carbon sequestration, can be caused by enhanced N deposition in N-limited environments [13]. In China, tropical and subtropical forests are N-rich ecosystems, but temperate and boreal forests are N-limited [14]. Due to the high population density and intensive human activity in tropical and subtropical regions, Nr deposition is relatively high in these areas [15]. Therefore, the impacts of increasing Nr deposition and its impacts have been quantified in south China, where these forests are found [16,17]. In contrast, quantification of fluxes and impacts of Nr are rare in boreal forests because Nr deposition and human disturbance have been traditionally considered as low. However, Nr pollution has no bounds [18]. It has been reported that tree growth rates have accelerated in boreal forests, even where human activity is limited [19], and carbon sequestration was significantly enhanced in forests when Nr deposition was between 3 and 11 kg N ha −1 yr −1 [20]. In conclusion, although no substantial deleterious effects of Nr pollution have been observed in low N deposition regions, quantification of Nr deposition fluxes in boreal forest ecosystems is crucial because both net productivity and carbon sequestration are sensitive to additional nitrogen inputs.
The Qinghai-Tibet Plateau is considered to be an N-limited ecosystem, sensitive to global change, such as increased Nr deposition [21,22]. To date, quantification of Nr deposition using long-term in situ monitoring is rare in this region because of limited access and technical resources. Sporadic research has revealed that Nr deposition is relatively low but has increased in Nyingchi City, the main city in the region, during recent years [23]. In fact, large spatial variation in Nr deposition exists in the plateau, as high as 18 kg N ha −1 y −1 in urban areas and as low as 0.44 kg N ha −1 y −1 in remote areas [24,25], but none of the previous research has focused on forest ecosystems in the plateau. Understanding Nr deposition is essential, especially in the southeast Tibetan boreal forest ecosystem, which has one of the highest levels of carbon storage and carbon density in forest ecosystems around the world [26]. We, therefore, established an in situ monitoring site and continuously measured Nr deposition for four years in a boreal forest ecosystem at Sejila Mountain in the Qinghai-Tibet Plateau, with the aim of: (1) measuring current atmospheric Nr concentrations and Nr deposition to the forest; (2) quantifying the dominant Nr deposition species. The results will help us to predict the positive or negative effects that would occur in a southeast Tibetan boreal forest ecosystem from enhanced Nr deposition.

Sampling Location
The Nr deposition monitoring site was established at the National Field Scientific Observation Station of Alpine Forest Ecosystem (29 • 38 N, 94 • 42 E, 3950 m a.s.l.) ( Figure 1). This is part of the Chinese Forest Ecosystem Research Network (CFERN) and National Ecosystem Research Network of China (CEORN). The annual average precipitation and average temperature were approximately 1000 mm and −0.73 • C, respectively. The station is surrounded by local primary fir forest, and there is no Nr pollution source nearby, except for one state road (#318). The dominant vegetation species is Abies georgei var. Smithii. Most of these trees are more than one hundred years old. The undergrowth is Sorbus (Sorbus rehderiana var. rehderiana), Rosa (Rosa omeiensis, Rosa sikangensis), and Lonicera (Lonicera saccata; Lonicera angustifolia) as well as other herbaceous plants.

Precipitation and Gaseous Nr Measurements
Precipitation was collected with simple precipitation gauges from January 2016 to December 2019. This method has been widely used in previous studies [27]. Briefly, a gauge (SDM6, Tianjin Weather Equipment Inc., Tianjin, China) was placed on the roof at the observation station where there were no surrounding obstacles. The gauge was continuously open, and precipitation (rainwater and snow) collected by hand once a week. (The samples will therefore contain wet and some dry deposition). Collected samples were transferred to 50 mL pre-cleaned plastic bottles and stored at −18 • C until analyzed. After each sampling, deionized water was used to wash the steel funnel and glass bottles of the rain gauge to eliminate contamination.

Precipitation and Gaseous Nr Measurements.
Precipitation was collected with simple precipitation gauges from January 2016 to December 2019. This method has been widely used in previous studies [27]. Briefly, a gauge (SDM6, Tianjin Weather Equipment Inc., China) was placed on the roof at the observation station where there were no surrounding obstacles. The gauge was continuously open, and precipitation (rainwater and snow) collected by hand once a week. (The samples will therefore contain wet and some dry deposition). Collected samples were transferred to 50 mL pre-cleaned plastic bottles and stored at −18 °C until analyzed. After each sampling, deionized water was used to wash the steel funnel and glass bottles of the rain gauge to eliminate contamination.
Atmospheric NH3 and NO2 only were measured using passive samplers. This is commonly accepted as providing an approximate measure of atmospheric Nr concentrations [28]. NH3 was captured using adapted low-cost, high absorption (ALPHA; designed by the Center for Ecology and Hydrology, Edinburgh, UK) passive samplers; NO2 was collected using Gradko diffusion tubes (Gradko International Limited, UK). ALPHA passive samplers and Gradko diffusion tubes were fixed beside the rain gauge at the height of 2 m above the roof and exchanged with new samplers or diffusion tubes on the first day of each month. There were three duplicate ALPHA samplers and diffusion tubes. Both NH3 and NO2 samples were measured within 48 h of sampling.

Analytical Procedures
Precipitation samples were analyzed as in a previous study [28]. Briefly, samples were thawed and filtered through a syringe filter (0.45 µm, Tengda Inc., Tianjin, China). NH4 + and NO3 -in the filtered solutions were measured by AA3 continuous-flow analysis (Bran + Luebbe GmbH, Norderstedt, Germany). The detection limits were 0.01 mg N L −1 .
Gaseous NH3 captured by ALPHA samplers was extracted with 10 mL-high-purity water and measured by AA3 continuous-flow analyzer as above. Gaseous NO2 captured by Gradko diffusion tubes was extracted with a solution containing sulfanilamide, H3PO4 and N-1-Naphthylethylen-diamine, and the concentration of NO2 − determined by colorimetry at a wavelength of 542 nm. A blank sample was measured at the same time to eliminate errors. Atmospheric NH 3 and NO 2 only were measured using passive samplers. This is commonly accepted as providing an approximate measure of atmospheric Nr concentrations [28]. NH 3 was captured using adapted low-cost, high absorption (ALPHA; designed by the Center for Ecology and Hydrology, Edinburgh, UK) passive samplers; NO 2 was collected using Gradko diffusion tubes (Gradko International Limited, Winchester, UK). ALPHA passive samplers and Gradko diffusion tubes were fixed beside the rain gauge at the height of 2 m above the roof and exchanged with new samplers or diffusion tubes on the first day of each month. There were three duplicate ALPHA samplers and diffusion tubes. Both NH 3 and NO 2 samples were measured within 48 h of sampling.

Analytical Procedures
Precipitation samples were analyzed as in a previous study [28]. Briefly, samples were thawed and filtered through a syringe filter (0.45 µm, Tengda Inc., Tianjin, China). NH 4 + and NO 3 − in the filtered solutions were measured by AA3 continuous-flow analysis (Bran + Luebbe GmbH, Norderstedt, Germany). The detection limits were 0.01 mg N L −1 . Gaseous NH 3 captured by ALPHA samplers was extracted with 10 mL-high-purity water and measured by AA3 continuous-flow analyzer as above. Gaseous NO 2 captured by Gradko diffusion tubes was extracted with a solution containing sulfanilamide, H 3 PO 4 and N-1-Naphthylethylen-diamine, and the concentration of NO 2 − determined by colorimetry at a wavelength of 542 nm. A blank sample was measured at the same time to eliminate errors.

Calculations and Statistical Analysis
Bulk Nr deposition (N b ) was calculated as the Nr concentration (N c ) multiplied by precipitation amount (P) in each precipitation event using: Monthly bulk Nr deposition (N s ) was calculated by summing all bulk deposition fluxes collected in a month, as follows: where i represents the number of samples in the month.
The monthly Nr concentration (N c ) in precipitation was calculated by dividing N s by precipitation amount (P s ) in the month: where i represents the number of precipitation events in the month. Dry deposition was calculated as the atmospheric Nr concentration multiplied by the deposition velocity. Monthly atmospheric Nr concentrations were obtained from the measured NH 3 and NO 2 concentrations and deposition velocities in forests from Flechard et al. [10]; these have been widely used to estimate Nr deposition in forest environments [29,30]. Deposition velocities of NH 3 and NO 2 were 1.6 cm s −1 and 0.16 cm s −1 , respectively.
We defined March to May as spring, June to August as summer, September to November as autumn, and December, January and February as winter.
Variance analysis was performed by SPSS 19.0. The Duncan test was used to test the significance between monthly precipitation amounts, NH 4 + deposition, NO 3 − deposition, and atmospheric NH 3 and NO 2 concentrations. Figures were drawn with originPro8. and NH4 + (c) deposition. Different letters above the error bar means that the difference reaches a significant level. Correspondingly, the same letters exists above the error bar that means the difference is insignificant.

Variations in the Monthly Volume Weighted Mean NO3 − and NH4 + Concentrations and Their Relationship
Monthly volume-weighted mean concentrations of both NO3 -and NH4 + showed similar patterns of high in winter and low in summer, with little variation from March to September. Monthly concentrations of NO3 − and NH4 + were in the range of 0.12-0.39 and 0.16-0.38 mg L −1 , respectively (Figure 3a,b). In order to understand the predominant Nr deposition forms in different months, monthly concentrations of both NO3 − and NH4 + were used in a linear fitting procedure. The result revealed that monthly NO3 − and NH4 + concentrations in precipitation are positively significantly correlated (R 2 = 0.951) (Figure 3c), and the dominant Nr form shifted from NH4 + to NO3 − as the bulk Nr concentration increased. letters above the error bar means that the difference reaches a significant level. Correspondingly, the same letters exists above the error bar that means the difference is insignificant.

Variation in the Monthly Atmospheric NO 2 and NH 3 Concentrations
Monthly mean concentrations of atmospheric NO 2 and NH 3 were in the ranges of 0.79-1.13 and 1.15-3.53 mg N L −1 , respectively (Figure 4). There was no obvious trend of monthly atmospheric NO 2 concentrations throughout the year. In contrast, atmospheric NH 3 concentrations increased from January to May and then decreased from August to December. The highest NH 3 concentrations occurred from March to August, corresponding to the plant growing season.

Variation in the Monthly Atmospheric NO2 and NH3 Concentrations
Monthly mean concentrations of atmospheric NO2 and NH3 were in the ranges of 0.79-1.13 and 1.15-3.53 mg N L −1 , respectively (Figure 4). There was no obvious trend of monthly atmospheric NO2 concentrations throughout the year. In contrast, atmospheric NH3 concentrations increased from January to May and then decreased from August to December. The highest NH3 concentrations occurred from March to August, corresponding to the plant growing season.  . Variation in monthly atmospheric NO2 (a) and NH3 (b) concentrations. Different letters above the error bar means that the difference reaches a significant level. Correspondingly, the same letters exists above the error bar that means the difference is insignificant.   . Variation in monthly atmospheric NO2 (a) and NH3 (b) concentrations. Different letters above the error bar means that the difference reaches a significant level. Correspondingly, the same letters exists above the error bar that means the difference is insignificant.

Bulk Nr Deposition and Monthly Precipitation
Bulk Nr deposition was determined from precipitation amount and Nr concentrations in precipitation [28]. We found that Nr deposition was mainly controlled by precipitation, increasing with precipitation. It has been reported that Nr deposition is logarithmically related to precipitation amount because atmospheric NH 4 + and NO 3 − are scavenged by rainfall, and scavenging decreases as rainfall intensity, duration and amount increases [31]. At our site, the relationship between precipitation amount and Nr deposition fitted both linear and logarithmic equations. However, with a linear relationship for both NH 4 + and NO 3 − giving the better fit (Figure 6a,b). This could be because atmospheric NH 4 + and NO 3 − concentrations were both so low as to be almost unaffected by rainfall intensity, duration and amount. Although we could not measure gaseous/particulate NH 4 + and NO 3 − concentrations because of limited technical resources, the low NH 3 and NO 2 concentrations support this idea to some extent (Figure 4a,b). Overall, precipitation amount was the dominant factor determining Nr deposition, both Nr deposition and precipitation showing similar patterns of monthly variation.
1.00, 1.29 and 1.33 kg N ha −1 y −1 , respectively, and of NO2 was 0.05, 0.04, 0.05 and 0.05 kg N ha −1 y −1 , respectively. The average total annual Nr deposition (bulk plus dry) was 4.23 (3.00 + 1.23) kg N ha −1 y −1 . Reduced nitrogen was dominant, accounting for 65.7 percent of the total deposition. The percentage of reduced nitrogen was very high in the dry deposition, being 96.1 percent of the annual dry Nr deposition.

Bulk Nr Deposition and Monthly Precipitation
Bulk Nr deposition was determined from precipitation amount and Nr concentrations in precipitation [28]. We found that Nr deposition was mainly controlled by precipitation, increasing with precipitation. It has been reported that Nr deposition is logarithmically related to precipitation amount because atmospheric NH4 + and NO3 − are scavenged by rainfall, and scavenging decreases as rainfall intensity, duration and amount increases [31]. At our site, the relationship between precipitation amount and Nr deposition fitted both linear and logarithmic equations. However, with a linear relationship for both NH4 + and NO3 − giving the better fit (Figure 6a,b). This could be because atmospheric NH4 + and NO3 − concentrations were both so low as to be almost unaffected by rainfall intensity, duration and amount. Although we could not measure gaseous/particulate NH4 + and NO3 − concentrations because of limited technical resources, the low NH3 and NO2 concentrations support this idea to some extent (Figure 4a,b). Overall, precipitation amount was the dominant factor determining Nr deposition, both Nr deposition and precipitation showing similar patterns of monthly variation.

Variation of Monthly Bulk Nr Concentrations and Atmospheric Nr Concentrations
Bulk Nr concentrations are usually inversely correlated with precipitation amount [28]. We found this, with high concentrations of both NH4 + and NO3 − in November to February when precipitation amounts were relatively low (Figures 2a and 3a,b). This trend could have been

Variation of Monthly Bulk Nr Concentrations and Atmospheric Nr Concentrations
Bulk Nr concentrations are usually inversely correlated with precipitation amount [28]. We found this, with high concentrations of both NH 4 + and NO 3 − in November to February when precipitation amounts were relatively low (Figure 2a and (Figure 3a,b). This trend could have been amplified by the collection of some dry deposition (e.g., dust) in the continuously open precipitation gauges, which would have had the biggest effect when precipitation was low. Monthly variation of NH 4 + concentrations in precipitation was comparatively less than those of both precipitation amount and atmospheric NH 3 concentrations from January to September. Both atmospheric NH 3 concentrations and precipitation amounts were high in Spring and Summer (Figure 2a and (Figure 3b). However, bulk NH 4 + concentrations were not low, i.e., not diluted by large amounts of precipitation, because of high atmospheric NH 3 concentrations. However, the dominant Nr species in precipitation changed from Atmosphere 2020, 11, 1331 9 of 14 NH 4 + to NO 3 − as bulk Nr concentrations increased (Figure 3c). Low monthly bulk Nr concentrations mainly occurred in spring and summer. NH 3 emissions are usually the largest in summer due to intensive agricultural activities and higher temperatures [32]. Atmospheric NH 3 concentrations were highest in summer at our site ( Figure 4b) and in a range of different ecosystems, as shown by in situ monitoring [23,30]. Ambient NO 2 is mainly attributed to industrial sources (e.g., fossil fuel burning; automobile exhausts; power plants), and so is highest in the most populated areas [33,34]. Although some research has shown that NO 2 exhibits seasonal variation [29,35] because of fossil fuel burning in winter, there were no NO 2 pollution sources near our site other than the state road (#318), and so NO 2 concentrations showed little seasonal variation.  [10]. We measured NH 3 and NO 2 concentrations of 2.34 and 0.92 µg N m −3 , respectively. Atmospheric NH 3 concentrations in southeast Tibet were, therefore, higher than the average in Europe, but those of NO 2 lower. This is not surprising because the Tibetan forest is in a remote area, surrounded by dense primordial forest. Low NO 2 concentrations can be attributed to the fact that atmospheric Nr sources are mostly natural.

Reduced and Oxidized Nitrogen Deposition in Bulk and Dry Deposition
High atmospheric NH 3 concentrations can be explained by fog events that frequently occur in the research area, resulting in high NH 4 + concentrations [37]. Average atmospheric NH 3 and NO 2 concentrations across China are 6.1 and 6.8 µg N m −3 [28], almost 2.6 and 7.4 times than those in our research area, respectively. Clearly, the atmosphere in the southeast Tibetan forest is relatively "clean", especially for NO 2 . However, ammonia is likely to be more ecologically and biologically toxic than oxidized species [38,39], and the annual deposition of reduced nitrogen was higher than that of oxidized nitrogen. Therefore, this southeast boreal forest ecosystem could be threatened by Nr deposition dominated by reduced species.

Comparison of Nr Deposition and Potential Ecological Effects with Other Regions
It is interesting to compare the current Nr deposition to this southeast Tibetan forest ecosystem with bulk deposition to other forest ecosystems around the world published in the last decade (Table 1). Clearly, bulk Nr deposition to our site is very low. It has been reported that external Nr inputs can increase long-term carbon sinks in boreal, temperate and tropical forests [20,40]. Thomas reported that forest carbon sequestration was significantly enhanced when Nr deposition was between 3 and 11 kg N ha −1 y −1 [20]. It had been reported that the critical nitrogen load of deciduous conifers in cold-temperate and temperate mountains was in the range 5-10 (or 10-15) kg N ha −1 y −1 [11,41]. Considering the low annual Nr deposition to this Tibetan forest ecosystem, which is lower than the Nr deposition critical load, the changes of forest net productivity should be given more attention.

Implications and Uncertainty Analysis
Although bulk Nr deposition was more precisely measured than dry Nr deposition, there are errors in the former. For example, both NO 3 − and NH 4 + in precipitation are absorbed by the forest canopy, especially in lower Nr deposition regions [9]. Additionally, total bulk Nr deposition is underestimated when dissolved organic nitrogen (DON) is not measured, as was the case in this work. Izquieta-Rojano et al. reported that the contribution of DON to total bulk Nr deposition ranged from 34% to 56% in forest ecosystems [55]. For dry deposition, deposition velocities are influenced by several factors and vary across ecosystems [9,10]. The deposition velocity of NH 3 is often overestimated in forest ecosystems because NH 3 exchange to foliar surfaces is bidirectional, so NH 3 deposition can be overestimated. Due to poor technical resources at our research site, we could not estimate the actual Nr deposition velocities of different Nr species. We used deposition velocities of NH 3 and NO 2 to "background environments" across China to calculate Nr deposition (NH 3 :0.47 cm s −1 ; NO 2 :0.17 cm s −1 ) [28], making the deposition of NH 3 and NO 2 0.33 and 0.05 kg N ha −1 y −1 , respectively. Reduced nitrogen deposition still dominates.
Flechard et al. reported that the percentages of NH 3 in total dry reduced nitrogen deposition and NO 2 in total dry oxidized nitrogen deposition were 77.4% and 21.1% [10], respectively. Clearly, NH 3 was the dominant species in dry reduced nitrogen deposition, but NO 2 was not so important in dry oxidized nitrogen deposition. Dry deposition of NH 3 and NO 2 were 1.18 and 0.05 kg N ha −1 y −1 in our research area. We used these percentages (77.4% and 21.1%) to convert NO 2 deposition to total oxidized nitrogen deposition and NH 3 deposition to total reduced nitrogen deposition giving values of total reduced nitrogen and oxidized nitrogen of 1.52 and 0.24 kg N ha −1 y −1 , respectively. Reduced nitrogen was the dominant Nr deposition species in this southeast Tibet forest ecosystem.

Conclusions
Based on four years of continuous monitoring, we found that Nr deposition was small in a southeast Tibet boreal forest. Average concentrations of atmospheric NH 3 and NO 2 were 2.34 and 0.90 µg N m −3 , respectively. Bulk deposition of NH 4 + and NO 3 − was 1.60 and 1.40 kg N ha −1 y −1 , respectively.
Nitrogen deposition showed a clear seasonal variation-high in summer and low in winter-dependent on the monthly variation in precipitation. Additionally, atmospheric NH 4 + concentrations were high in summer and low in winter. Annual Nr deposition was less than the critical load (5-10 kg N ha −1 y −1 ), being 4.23 kg N ha −1 y −1 and dominated by reduced nitrogen. Deposition was mainly concentrated in the months of plant growth. As a result, carbon sequestration and net productivity in this southeast Tibet boreal forest ecosystem deserve more research because these regions may benefit from slightly increased Nr deposition. More comprehensive methods, as used in Nr deposition monitoring, should be employed in these regions sensitive to global change.