Characterizations and source analysis of atmospheric inorganic ions 1 at a national background site in the northeastern Qinghai-Tibet 2 Plateau : insights into the influence of anthropogenic emissions on a 3 high-altitude area of China 4

12 Atmospheric particulate matter (PM) imposes highly uncertain impacts on both 13 radiative forcing and human health. While ambient PM has been comprehensively 14 characterized in China’s megacities; its composition, source, and characteristics in the 15 Qinghai-Tibet Plateau (QTP) are not yet fully understood. An autumn observational 16 campaign was conducted during the 1 15 October 2013 at a national background 17 monitoring station (3295 m a.s.l.) in the QTP. Real time concentrations of inorganic 18 water-soluble ions (WSIs) associated with PM2.5 were measured in addition to PM2.5 19 concentrations, gaseous pollutants, and meteorological parameters. SO4 was the 20 most abundant WSI (10.00 ±4.39 μg/m) followed by NH4 (2.02 ±0.93 μg/m), and 21 NO3 (1.65 ±0.71 μg/m). Observed WSI concentrations were lower as compared to 22 urban sites in eastern China; however, they were higher as compared to other QTP 23 monitoring sites. To better understand the potential sources of WSIs in the QTP, a 24 Positive Matrix Factorization receptor model was used. Results showed that mixed 25 factor including animal waste emission and biomass burning, crustal dust, salt lake 26 emissions, secondary sulfate and secondary nitrate were the major emission sources of 27 particulate inorganic ions at the study site. Correlation analysis between WSIs 28 revealed that NH4NO3, (NH4)2SO4, Na2SO4, and K2SO4 were the major atmospheric 29 aerosol components. High sulfate and nitrate oxidation ratios indicated strong 30 secondary formation of both SO4 and NO3. Both photochemical and heterogeneous 31 reactions contributed to the formation of particulate SO4, while the conversion of 32 NO2 to NO3 only occurred via photochemical reactions in the presence of high O3 33 concentrations and strong sunlight. 34


Introduction
Atmospheric aerosol has a significant impact on climate change and human health, the extent of which is determined by their physical and chemical properties.
High concentrations of aerosols are associated with rapid economic growth, urbanization, industrialization, and motorization, and have become a major environmental concern in China (Du et al., 2015).Extensive research has investigated the sources, chemical and physical properties, and evolution processes of aerosol particles at urban and rural sites in China during the last decade (Cao et al., 2007;Gong et al., 2012;He et al., 2011;Jiang et al., 2015;Sun et al., 2015;Sun et al., 2013;Wu et al., 2007).These studies indicated that fine particles are mainly composed of organics, sulfate, nitrate, ammonium, mineral dust, and black carbon.While these studies have greatly improved our understanding on the sources and physical/chemical properties of aerosol particles, they were predominantly conducted in developed areas of China, including Beijing-Tianjin-Hebei, the Pearl River Delta, and the Yangtze River Delta.In remote areas, such as the Qinghai-Tibet Plateau (QTP), studies on atmospheric aerosol properties are rare.
The QTP covers most of the Tibet Autonomous Region and Qinghai Province in western China, with an area of 5,000,000 km 2 and an average elevation over 4000 m.
The area is geomorphologically the largest and highest mountain region on earth (Yao et al., 2012).Described as the "water tower" of Asia, this area contains the headwaters of the Mekong, Yangtze, and Yellow Rivers.Therefore, climate variability and change in this region has fundamental impacts on a range of climate-related ecosystem services (McGregor, 2016).Due to its unique ecosystem, landforms, and monsoon circulation, the QTP has a profound role in regional and global atmospheric circulation, radiative budgets, and climate systems (Su et al., 2013;Kopacz et al., 2011;Yang et al., 2014;Jin et al., 2005).Limited anthropogenic activity, a sparse population, immense area, and high elevation mean that, alongside the Arctic and Antarctic, the QTP is considered one of the most pristine terrestrial regions in the world.Because of this, the region is an ideal location for characterizing background aerosol properties, regional and global radiative forcing, climate and ecological changes, and the transportation of global air pollutants.Thus, a comprehensive understanding of QTP aerosol chemistry is crucial for assessing anthropogenic influences and evaluating long-term changes in the global environment (Cong et al., 2015;Zhang et al., 2012).
Research relating to the chemical and physical characteristics of aerosols in the QTP is rare; hence, their sources, properties, and evolution processes are poorly understood.This lack of research is a result of the region's remoteness and challenging weather conditions.Most previous studies of aerosol chemistry in the QTP were conducted in the Himalaya (the southeastern or southern areas of the QTP) to assess the key roles of the Himalaya on regional climate and the environment, as well as the boundary transportation of air pollutants from South Asia (Cong et al., 2015;Zhao et al., 2013b;Wan et al., 2015;Shen et al., 2015).Conversely, the northeastern QTP, located in inland China, is likely to have very different atmospheric behaviors as compared to those of the Himalaya due to different climate patterns and aerosol sources between the two regions (Xu et al., 2015).
Two decades ago, natural emission and occasional perturbations from human activities were believed to be the major sources of particles in this area.(Wen et al., 2001;Gao and Anderson, 2001;Tang et al., 1999).With enlargement of human activities in recent years, several studies found that SO 4 2-, NO 3 -, NH 4 + , and Ca 2+ were major water-soluble ions (WSIs) (Xu et al., 2014;Li et al., 2013;Zhang et al., 2014), suggesting that both anthropogenic pollution and mineral dust contributed to the total mass of PM 2.5 .Recent real time observation also proved that sulfate, which is regarded as one of the marker of anthropogenic emission, was dominant in PM 1 mass in this area (Du et al., 2015).Despite some researches have been done in this area, there are still some scientific gaps on the aerosol chemical properties.Most of previous studies were based on the manual filter sampling, with the duration ranging from days to weeks (Du et al., 2015).These low temporal resolution data have limitations when being used for characterizing the rapid variations of chemical compositions and capturing secondary aerosol formations.Data from the high resolution instruments offer significant advantages over traditional filter-based measurements (Vedantham et al., 2014).To our knowledge, real-time measurements of aerosol particle composition at rural sites in the Tibetan Plateau are still rare (Du et al., 2015).
WSIs comprise a large portion of aerosol particles and may help understand chemical reactions in the atmosphere (Tripathee et al., 2017).They can provide important information for understanding chemical characterizations, sources, behaviors, and formation mechanisms; and hence, knowledge on the emission of gaseous precursors and the effect of regional and local pollution on ecosystem health (Wang et al., 2005;Tripathee et al., 2016).Furthermore, WSIs regulate the electrical properties of the atmospheric medium, participate in ion-catalyzed and ion-molecule reactions, and contribute to physicochemical interactions, including ion-induced new particle formation (Frege et al., 2017;Schulte and Arnold, 1990).Given the importance of WSIs on aerosol properties and the limited data on aerosol chemical compositions in the QTP, more WSIs data are needed to better characterize the chemical composition of aerosols in this area.
In our study, a real time monitor for WSIs associated with PM

Monitoring site
Figure 1 shows the location of the monitoring site at the peak of Daban Mountain, Menyuan Hui Autonomous County, Qinghai Province (37º36'30"N, 101º15'26"E; 3295 m a.s.l.).The site is owned by the Chinese national atmospheric background monitoring station system and is approximately 160 km north of Xining, the capital city of Qinghai Province.This monitoring site is also approximately 100 km northeast of Qinghai Lake, a saline and alkaline water body that is the largest lake in China.The area is characterized by a typical plateau continental climate with an annual temperature of 0.8 ºC and precipitation of 520 mm.Meteorological parameters during the observation period are summarized in Table 1.The site is surrounded by typical QTP vegetation, including potentilla fruticosa and kobresia.No strong anthropogenic emission sources exist in the adjacent area, with the exception of occasional biomass burning events and yak dung burning for residential cooking and heating.A national road (G227) and a provincial road (S302) to through the area of this site, however, the traffic volume around the site is small (Du et al., 2015).
Meteorological parameters (e.g.temperature, relative humidity, pressure, and wind speed and direction) were also recorded.

Oxidant ratio
Particulate sulfate and nitrate oxidation ratios (SOR and NOR, respectively), defined as the molar ratio of SO 4 2-and NO 3 -to total oxidized sulfur and nitrogen (Zhou et al., 2009), were used to evaluate secondary conversion from NO 2 and SO 2 to NO 3 -and SO 4 2-, respectively.High SOR and NOR indicate larger conversions of SO 2 and NO x to their respective particulate forms in PM 2.5 .In this study, NOR and SOR were calculated based on the following formulae: (1) (2) Given that there potentially exist the direct emissions of sulfate and nitrate (such as the sulfate released from the salt lake, which will be discussed in the source apportionment section), it is not suitable to apply their concentrations directly into the formulae ( 1) and ( 2).Therefore, we used the PMF model to calculate the concentrations of secondary sulfate and nitrate in each observation, and then take these modeled concentrations to estimate the SOR and NOR.The PMF model will be introduced in following section.

Ion balance
Ion balance was used to evaluate the acid-base balance of aerosol particles.We converted the WSIs mass concentration into an equivalent concentration, as follows: (3)

Source apportionment
Positive Matrix Factorization (PMF), developed by Paatero (Paatero and Tapper, 1994;Paatero, 1997), has been widely applied in source apportionment researches.In this model, a data matrix  !" , in which i is the sample and j is the measured chemical species, can be viewed as a speciated data set, and the concept of this model can be represented as: where  is the number of factors;  is the chemical profile of each source,  is the mass contribution of each factor to the sample;  !" is the source profile, and  !" is the residual for each species or sample.
PMF solves Eq (5) by minimizing the sum of the square of residuals weighted inversely with the error estimates of the data points, Q, defined as: where  !" is the uncertainty of chemical species j in sample i.
Furthermore, we used the results of the source apportionment to estimate the concentrations of secondary sulfate and nitrate and then applied them in the calculations of SOR and NOR.By running the PMF model, we identified the secondary sulfate and nitrate in factor  and , respectively.We also obtained the hourly contributions of each factor to the total ions, as well as the profile of each factor.Then the products of hourly contributions of factor  or  and the percentages of sulfate or nitrate in the profile of factor  or  are taken as the concentrations of secondary sulfate or nitrate.The calculation is defined as: Presume that factor s and t are the secondary sulfate and nitrate sources calculated by the PMF model (,  ≤ ), then: where  !,! is the contribution of factor  to sample  , while  !,! is the contribution of factor  to sample ; and  !,!"#$%&' is the ratio of  in the factor , while  !,!"#$%#& is the ratio of  in the factor .

Statistical analysis
Correlation analysis, analysis of variation (ANOVA), and linear regression were applied.All statistical calculations were preformed using R studio software packages (Version 0.99.903,RStudio, Inc.).

Descriptive analysis
Table 1 summarizes the concentrations of WSIs, PM 2.5 , and gaseous pollutants and data of meteorological parameters during the observation period.SO 4 2-accounted for 67.9% of the total WSIs mass, followed by NH 4 + (13.7%), and NO 3 -(11.2%).
SO 4 2-, NO 3 -and NH 4 + (SNA), accounting for 92.8% of the total WSIs mass, were the major components of secondary inorganic aerosols.To better understand the concentrations of WSIs, we compared our observations with other studies implemented in background sites or urban sites across China and high altitude areas around the world in Table 2. Our results are lower as compared to studies in Europe and the USA (VanCuren and Gustin, 2015;Moroni et al., 2015), and the high latitude Himalaya region (Carrico et al., 2003); however, observations are comparable with some urban area in Nepal (Carrico et al., 2003).Observed concentrations of SO 4 2-are also lower as compared to low altitude sites in China, for example urban sites in Beijing, Shanghai and Xi'an, and background sites in Shangdianzi (Beijing) and Lin'an (Zhejiang).Concentrations of NO 3 -were five to thirteen times lower as compared to those in low altitude areas (both urban and background sites), indicating that the influence of vehicle emissions in studying area is weak.NH 4 + levels were lower as compared to those in urban sites (three to six times lower), and also slightly lower as compared to background sites (less than three times lower).
SNA concentrations in this study were higher as compared to those at other sites in the QTP, including the southern edge (Cong et al., 2015), Qilian Shan Station (Xu et al., 2014), and Qinghai Lake in the northeastern QTP (Zhang et al., 2014;Zhao et al., 2015).Large differences in concentrations suggest that the monitoring site in this study appears to be more impacted by natural and human activities as compared to other sites in the QTP.To further examine the relationship between PM 2.5 and WSIs, we divided the PM 2.5 concentrations into four categories: a) C(PM 2.5 ) < 20µg/m 3 , (b) 20µg/m 3 ≤ C(PM 2.5 ) <30µg/m 3 , (c) 30µg/m 3 ≤ C(PM 2.5 ) <40µg/m 3 , and (d) C(PM 2.5 ) ≥ 40µg/m 3 and attributed each WSI measurement to its corresponding PM 2.5 category.Figure 3 shows the mean proportions of WSIs in PM 2.5 for the different categories.As the PM 2.5 concentration increases, the percentages of WSIs in PM 2.5 mass exhibited decreasing trends, suggesting that the contribution of WSIs to PM 2.5 increases was negligible.Therefore, more observational campaign should be implemented in the future to investigate the driver compositions on the increase of PM 2.5 mass concentrations.High NH 4 + and K + loading were observed in factor 1. Livestock feces, which is commonly found in the meadows around the sampling site is a possible source of NH 4 + .K + can be attributed to the combustion of biomass, which has both natural and anthropogenic origins.Li et al. (2015) found that occasional biomass burning events in the area contributed significantly to the formation of anthropogenic fine particles..
Consequently, factor 1 is attributed to mixed sources, animal emissions and biomass burning, and crustal materials.Factor 2 is identified as crustal materials with high loading of Mg 2+ and Ca 2+ (Ma et al., 2003).Factor 3 has high Na + loading and a moderate SO 4 2-loading, which is also shown in previous correlation analysis with the correlation coefficient between SO 4 2-and Na + is 0.76.Our monitoring site is approximately 100 km from Qinghai Lake, a saline and alkaline water body which is the largest lake in China.Zhang et al. (2014) collected total suspended particle (TSP) and PM 2.5 samples at Qinghai Lake, and found that the concentration of Na + was higher as compared to other mountainous areas.Furthermore, they found that SO 4 2was one the most abundant species in both TSP and PM 2.5 .Therefore, we attribute factor 1 to aerosols emitted from Qinghai Lake.Factor 4 and 5 are enriched with SO 4 2-and NO 3 -, respectively, which could be considered as the secondary sulfate and secondary nitrate.The precursor of SO 4 2-is SO 2 , which may originate from coal combustion, and NO 3 -is mainly converted from ambient NO x , emitted by both vehicle exhaust and fossil fuel combustion.

Diurnal variation analysis
Diurnal variations of WSIs in PM 2.5 , related gaseous pollutants (SO 2 , NO 2 , and O 3 ), and meteorological parameters (temperature, relative humidity, and wind speed) are shown in Figure 5. Figure5 Diurnal variations of WSIs in PM 2.5 , gaseous pollutants (SO 2 , NO 2 , O 3 ), as well as meteorological parameters (temperature, relative humidity and wind speed) during sampling period SO 4 2-concentrations begin to increase from midnight (Beijing time), reach peak levels at approximately 15:00, and then decrease gradually.SO 2 concentrations exhibit a bimodal trend, with peaks at 03:00 and 12:00; conversely, SO 4 2-exhibits an inverse trend.NO 3 -concentrations peak at midnight and in the early afternoon, with lowest levels occurring during the late afternoon.NO 2 displays high nighttime levels and low daytime levels.NH 4 + remains steady during the morning with a peak at 10:00, and then deceases until 16:00.O 3 , temperature, RH, and wind speed also display evident diurnal variations; O 3 , temperature, and wind speed are low (high) at night (day), while RH shows an inverse variation to this pattern.

Sulfate and nitrate oxidation ratio analysis
Average NOR and SOR during the whole measurement campaign were 0.16 and 0.55, respectively, suggesting potentially strong secondary formation of both SO 4 2and NO 3 -.Strong photochemical reactions and the existence of high O 3 concentrations would elevate the oxidant ratio from SO 2 and NO 2 to SO 4 2-and NO 3 -, despite the low intensity of local emissions.
Variation trends of SOR and NOR were compared to changes in PM 2.5 concentration and ambient RH, as shown in Figure 6.Previous research has shown that, in urban areas, both SOR and NOR increased with the PM 2.5 concentration, suggesting that heavy PM 2.5 pollution corresponds to high SOR and NOR (Xu et al., 2017).However, studies investigating SOR and NOR variations at background sites with low PM 2.5 concentration range are rare.Our results showed that increasing concentrations of PM 2.5 at low levels corresponded to decreasing SOR and NOR (Figure 6a), although the decreases were slight.This is consistent with our previous finding that concentrations of SO 4 2-and NO 3 -did not vary markedly with PM 2.5 increases at background sites, and provides further evidence to suggest that SO 4 2-and NO 3 -are not key drivers on the increase of PM 2.5 mass concentrations at low levels.
Crustal materials are either not responsible for it as shown in Figure 3.The simultaneous observations of Du et al. (2015), indicated that organics were thought to be the major driver.
In Figure 6b, SOR initially decreases and then increases as RH increases.Peak

Molecular composition of major ionic species
The molecular chemical forms of the major WSIs in PM 2.5 were identified using bivariate correlations based on individual WSI molar concentrations (Verma et al., 2010;Wang et al., 2005).In this study, we used equivalent concentrations for correlation analysis, and the coefficients are shown in Table 3. Figure 8

Ion acidity analysis
The ion balance, expressed by the sum of the equivalent concentration (µeq/m 3 ) ratio of cation to anion (C/A), is an indicator of the acidity of particulate matter (Wang et al., 2005).In this study, the ion balance ratio was 0.87, indicating that aerosols tended to be acid, in line with previous studies in the QTP (Xu et al., 2015;Zhao et al., 2015).However, the results across the QTP are quite different at different locations.In the studies of south edge of the QTP, the aerosols were found to be  et al., 2014;Xu et al., 2015).Another study at the Qinghai Lake also got slightly acidic result (PM 2.5 , C/A=0.8) (Zhao et al., 2015).In their study at the summer of 2012, Xu et al. ( 2015) also found that the equivalent balances of water-soluble species in different size modes indicate that the accumulation mode particles were somewhat acidic (with the linear regression slope of [NH4 + +Ca 2+ +Mg 2+ +K + ] vs. [SO 4 2-+NO 3 -] being 0.6) and that the coarse mode particles were almost neutral (the slope was 0.999), indicating that small size of particles show tendency of acid.As compared to the results at the south edge of the QTP that is mostly influenced by natural emission (such as mineral dust), the northeastern QTP suffers more anthropogenic emissions (e.g.SO 4 2-and NO 3 -), since it is more close to the areas with intensive human activities.
Figure 9 shows the scatter and linear regression plot of cations and anions (µeq/m 3 ).It is apparent that most points are below the 1:1 line, highlighting the acid tendency.The total equivalent anion concentration was regressed against the total equivalent concentrations of cations, and the slope of regression was 0.58.concentration was high.This provides further evidence to support our finding that SO 4 2-and NO 3 -did not contribute to PM 2.5 increases.

Conclusion
The QTP is an ideal location for characterizing aerosol properties.In this study, we investigated the characterizations of WSIs associated with autumn PM 2.5 at a background site (3295 m a.s.l.) in the QTP.In this study, we finished some analysis on WSIs by taking advantage of real-time data to: 1) analyze the diurnal variations of WSIs; 2) discuss the formation of secondary sulfate and nitrate at the QTP; and 3) investigate source apportionment on hourly data within short-term observation.All these above are difficult by using traditional manual PM2.5 sampling, given that it usually takes hours or even days for sample collection, and is unable to detect more variations on aerosol compositions and supply more data on finer temporal scale for further analysis.
During our observation, we collected real time concentrations of WSIs, and analyzed them together with PM 2.5 , gaseous pollutants, and meteorological parameters for investigating ion chemistry of aerosols in the QTP.SO 4 2-, NO 3 -, and NH 4 + (SNA) were the three most abundant WSI species, and crust-originated ions (Na + , Mg 2+ , K + , and Ca 2+ ) comprised a small fraction of total WSIs.As compared to similar studies in China, SNA concentrations in this study were lower as compared to low altitude urban areas, but higher relative to other sites in the QTP.NH 4 NO 3 , (NH 4 ) 2 SO 4 , Na 2 SO 4 , and K 2 SO 4 are found to be the major atmospheric aerosol components during our observation campaign.
Source apportionment using a PMF model identified five factors: mixed factor including animal waste emission and biomass burning, crustal dust, salt lake emissions, secondary sulfate and secondary nitrate.Based on the results of source apportionment, we found that the major sources of sulfate are salt lake emission and secondary transformation, while particulate nitrate is mostly from secondary conversion.After excluding the emission of sulfate from the salt lake, we investigated the possible formation pathway of SO 4 2-and NO 3 -, the concentrations of which showed evident diurnal variations.The results revealed that strong solar intensity and high O 3 concentrations combined with low daytime RH greatly enhanced the conversion of SO 2 and NO 2 to SO 4 2-and NO 3 -, respectively.Heterogeneous reactions were weak overnight, and contributed to SO 4 2-formation only.Our analysis suggests that photochemical reactions played a critical role in the secondary formation of SO 4 2and NO 3 -during our observation period.
To our knowledge, there is no such real-time measurement on WSIs associated with PM 2.5 at rural sites in the QTP yet.This study provides some preliminary results on aerosol ion compositions on the QTP, and proposes the potential formation mechanism of secondary sulfate and nitrate.These findings are supposed to be useful for further studies on aerosol chemistry in this area.

Figure 1
Figure 1 Location of sampling site

Figure 2
Figure 2 Correlation coefficients (r) between WSIs in PM 2.5 during sampling period

Figure 3
Figure 3 Mass portions of WSIs within different PM 2.5 level ranges 3.2 Source apportionment by PMF In this study, all WSIs and gaseous pollutants were introduced into the PMF model for source identification.Five factors were used in the PMF model.The distributions of the factor species and the percentage of total species are shown inFigure 4.
SOR occurs when RH reaches both its maximum and minimum levels, when RH is low (10-20%), O 3 is high(114.6µg/m 3 , approximately the 70 th percentile of O 3 concentrations), and vice versa (RH > 70% and O 3 93.8µg/m 3 , approximately the 30 th percentile of O 3 concentrations).The formation of particulate SO 4 2-can be achieved via aqueous-phase oxidation (heterogeneous reaction) or gas-phase oxidation (photochemical reaction).Normally, aqueous-phase oxidation from SO 2 to SO 4 2-is faster than gas-phase oxidation (Wang et al., 2016).When RH is low and O 3 is high, the photochemical formation of SO 4 2-via gas-phase oxidation should be considered the main oxidation pathway.Conversely, low O 3 and high RH are not sufficient to provide adequate oxidizing capacity; thus photochemical SO 4 2-formation becomes less important and aqueous-phase oxidation plays a more dominant role.NOR constantly decreases as RH increases.Particulate NO 3 -is predominantly formed by the gas-phase reaction of NO 2 and OH radicals during the day and by heterogeneous reactions of nitrate radicals (NO 3 ) at night (Seinfeld and Pandis, 2016).In this study, high (low) O 3 and low (high) RH lead to high (low) NOR, meaning that gas-phase reactions oxidized by high levels of O 3 are the major pathway for nitrate formation, while heterogeneous reactions play a less important role.(a) PM 2.5 (b) RH Figure 6 Variations of SOR and NOR as a function of PM 2.5 and RH.The vertical bars correspond to one standard error from the mean.

Figure 7
Figure 7 characterizes the diurnal variations of SO 4 2-, NO 3 -, SOR, NOR, O 3 , and RH.The variation of SOR is small, particularly as compared to the evident diurnal variation in SO 4 2-.Daytime gas-phase oxidation and nighttime aqueous-phase oxidation are thought to be equally important to the formation of SO 4 2-.NOR is high during the day and low at night; reflecting a strong positive correlation with O 3 (r=0.71,p<0.05) and a weak negative correlation with RH (r=-0.43,p<0.05).The high correlation between NOR and O 3 indicates that gas-phase oxidation via photochemical reactions is the main NO 3 -formation pathway.Trends of SOR and NOR with RH and O 3 suggest that both photochemical and heterogeneous reactions contribute to the secondary transformation of SO 2 , while only photochemical reaction drives the conversion of NO 2 to nitrate.It is apparent that photochemical reactions (dominated by O 3 oxidization) contribute markedly to the secondary conversion of both SO 2 and NO 2 , while heterogeneous reactions (promoted by the existence of aqueous phase) contributed to the formation of SO 4 2-and only had a weak effect on NO 3 -.Ma et al. (2003) found that fine nitrate particles (Dp < 2.0 µm) at Waliguan Observatory (150 km south of our monitoring site) were most likely produced via gaseous-phase reactions between nitric acid and ammonia, in line with our findings on nitrate formation.

Figure 9
Figure 9 Cation and Anion scatter plot and linear regression

Table 1
Descriptive statistics of WSIs species, gaseous pollutants and meteorological parameters

Table 2
Comparisons of WSIs concentrations with other high altitude and urban sites (mean, µg/m 3 )

Table 3
Correlation coefficients (r) between the equivalent concentrations of WSIs in PM 2.5