Characteristics of Particulate-Phase Polycyclic Aromatic Hydrocarbons ( PAHs ) in the Atmosphere over the Central Himalayas

PAH concentrations were measured in total suspended particle (TSP) samples collected from six sites along two southnorth transects across the central Himalayas from April 2013 to March 2014. The annual average TSP and PAH (especially 5and 6-ring compounds) concentrations were found to decrease noticeably northwards along both transects. At rural and urban sites, the TSP and PAH concentrations showed clear seasonal variations, with the lower concentrations around the mid-monsoon season and the higher values in the winter season. Meanwhile, at the remote sites (e.g., Nyalam and Zhongba), these pollutants generally remained constant throughout the year but with relatively higher levels during the pre-monsoon season. Both IndP/(IndP + BghiP) and Fla/(Fla + Pyr) ratios suggested that atmospheric PAHs from urban and rural sites were mainly associated with emissions from biomass burning, coal burning and petroleum combustion. However, the contribution of biomass burning increased at remote sites. Similar compositions of PAHs were found at three remote sites located on both sides of the Himalayas (Jomsom, Zhongba, and Nyalam), suggesting that the northern side of the Himalayas may be affected by anthropogenic emissions from the Indo-Gangetic Plain (IGP) via long-range atmospheric transport. This work provides a database of PAHs in central Himalayas for further assessing environmental risk of air pollution in the remote regions.


INTRODUCTION
Atmospheric aerosols have an important influence on global climate (Bonasoni et al., 2010;IPCC, 2013).They are the main culprits in visibility degradation (Srivastava et al., 2008) and are effective carriers of toxic chemicals such as polycyclic aromatic hydrocarbons (PAHs) (Kaur et al., 2013;Sarkar and Khillare, 2013).PAHs, a large group of organic compounds with two or more aromatic rings, are generated by incomplete combustion or pyrolysis of materials such as coal, oil, gas, refuse, and biomass (Nielsen et al., 1996;Liu et al., 2007;Rajput et al., 2014).PAHs have received considerable attention owing to their persistence and toxicity, especially their carcinogenic and/or mutagenic properties (Bhargava et al., 2004).The World Health Organization (WHO) recommends guidelines in terms of a carcinogenic slope factor, and the European Union indicative limit value is set at 1 ng m -3 of benzo [a]pyrene (European Union, 2005;WHO, 2006).The United Kingdom has set an air quality standard of 0.25 ng m -3 benzo[a]pyrene (EPAQS, 1999).Since some PAHs (e.g., 4-6 ring PAHs) are persistent in the atmosphere, they can be transported over long distances to remote areas such as the Arctic or the Himalayas, thousands of kilometers away from their sources (Crimmins et al., 2004;Ding et al., 2007;Rajput et al., 2013).
The Himalayas is a remote and mountainous region where emissions from anthropogenic activities are not significant when compared to those in upwind adjacent regions in South Asia such as the highly populated, rapidly developing, and polluted Indo-Gangetic Plain (IGP).The IGP is a major regional/global emitter of organic pollutants into the atmosphere (NEERI, 2006;Singh et al., 2013;Chen et al., 2015b).It has been estimated that South Asia produced about 90 Gg y -1 of PAHs in 2004 (Zhang and Tao, 2009), and the transport of PAHs from the IGP is considered to be the most important pathway controlling the levels of PAHs over the Himalayas (Rajput et al., 2013;Gong et al., 2014).Increasing pollutant emissions associated with the fastgrowing economies of South Asian countries have led to the progressive increase of aerosol concentrations above the natural baseline, with a clearly measurable positive trend in the last 30 years (Gautam et al., 2009).The Indian monsoon circulation and westerly winds are responsible for the distribution and transport of air pollutants to the Himalayan region (Bonasoni et al., 2010).Annually, this region has a long dry season (Nov-May) when air pollutants accumulate, notably on the southern side of the Himalayan Mountains.The so-called "Atmospheric Brown Cloud (ABC)", a 3 kmthick brownish layer of pollutants, has been detected extending from the Indian Ocean to the Himalayas (Tripathi et al., 2005;Ramanathan et al., 2007).The aerosols that gather in the foothills can be lifted to high altitudes (Decesari et al., 2010;Qiu, 2013) and can even travel across the high Himalayas and reach the Tibetan Plateau (TP) (Bonasoni, 2008;Xia et al., 2011;Chen et al., 2015a;Cong et al., 2015;Lüthi et al., 2015), thus affecting the atmospheric quality and causing environmental risks in the high and remote regions.
Therefore, it is critical to understand their emission, transport, transformation, removal and impacts of various air pollutants, including PAHs in the Himalayas.Some studies of elements and carbonaceous particle concentrations have already been carried out in this region (Tripathi et al., 2005;Bonasoni et al., 2010;Kim et al., 2015;Li et al., 2016).
However, current knowledge of PAH compositions for this region remains deficient (Kishida et al., 2009;Rajput et al., 2013;Chen et al., 2015bChen et al., , 2016)), especially in terms of spatial and seasonal variations.Here, we present results of a year-long time series of TSP and particulate-phase PAHs observations at six sites (Lumbini, Pokhara, Jomsom, Zhongba, Dhunche and Nyalam) across two Himalayan transects to investigate characteristics of PAHs in this critical region (Fig. 1).This paper aims to provide a better understanding of the temporal and spatial characteristics of ambient TSP and PAH concentrations over the Himalayas.

Sampling Site Description
The TSP samples were collected along two south-north transects, extending from Lumbini to Zhongba and Dhunche to Nyalam across both sides of the central Himalayas (Fig. 1).Lumbini (27°29′N, 83°17′E, 100 m a.s.l.), lies on the border of India and Nepal, is in a rural area of the IGP.The major pollution sources are agricultural activities, domestic heating and cooking (entailing the burning of large amounts of crop straw and wood each year, especially in winter and the premonsoon season), eleven cement factories and more than fifty other industries along the nearby Lumbini-Bhairahawa industrial corridor (Chen et al., 2016).Pokhara (28°11′N, 83°59′E, 813 m a.s.l.), located about 100 km north of Lumbini, is Nepal's third largest city, with a metropolitan population exceeding 200,000.In recent years it has experienced rapid urbanization, accompanied by an increased number of vehicles.Jomsom (28°46′N, 83°43′E, 3048 m a.s.l.), a small town in the Mustang district of Nepal, is located in the Kali Gandaki Valley, a river valley that cuts across the Himalayas at the transition between South Asia and Central Asia.Here, major human activities are tourism and limited agriculture.The measurement station is about 100 m above Jomsom, and away from the only road that runs along the valley.Since the location is remote, with minimal local emissions, we consider the site to present, as far as is possible, a regional background in the Himalayas.Dhunche (28°7′N, 85°18′E, 2051 m a.s.l.) is a semi-urban small town situated in the Langtang National Park and is the headquarters of the Rusuwa district, about 50 km north of Kathmandu and 14 km south of the Chinese border.The major sources for pollution are biomass burning, vehicle emissions, tourism, and agricultural activities surrounding the town.Zhongba (29°42′N, 83°59′E, 4704 m a.s.l.) is located on the northern side of the Himalayas, and is characterized by yak husbandry.The sampling site in Zhongba is around 20 km away from the town with a minimum influence of local emissions into the atmosphere.Nyalam (28°10′ N, 85°59′ E, 4166 m a.s.l.) is a town in the Tibet-Himalayas with some local emissions, including biomass combustion for cooking and heating and limited vehicular traffic.

Meteorology and Backwards Air-Mass Trajectories
The southern and northern sides of the Himalayas have differing climates.At sites in Nepal, the dominant surface wind direction is southerly during the monsoon season.In other seasons, westerly and southwesterly winds prevail due to the influence of the westerly winds (Bonasoni et al., 2010).At sites on the TP, southwesterly winds near the surface dominate during the monsoon season, while northeasterly winds prevail in other seasons (Ma et al., 2011).All sites experience four distinct seasons: winter (December-February), pre-monsoon (March-May), monsoon (June-September), and post-monsoon (October-November) (Bonasoni et al., 2010).
To identify trajectories of air masses arriving at sampling sites on the north side of the Himalayas, five-day air mass backward trajectories were calculated using the Hybrid Single-particle Lagrangian Integrated Trajectory (HYSPLIT) model and Global Data Assimilation System (GDAS) data provided by the Air Resources Laboratory of the National Oceanic and Atmospheric Administration (http://www.arl.noaa.gov/ready/hysplit.html).The cluster trajectories were calculated at the height of 500 m starting at 00:00 Coordinated Universal Time on the sampling day.

Sample Collection
A total of 190 aerosol samples were collected at these six sites during the period of April 2013 to March 2014.They were collected on pre-combusted (550°C, 6 h) quartz fiber filters (90 mm in diameter, Whatman plc., Maidstone, UK.) using six samplers fitted with TSP cyclone which can collect all the suspended particles (flow rate: 100 L min -1 , KC-120H: Qingdao Laoshan Applied Technology Institute, Qingdao City, China).Samples were collected every three to seven days for 24-h periods at Lumbini, Pokhara, Dhunche and Jomsom, and for 48-h periods at Zhongba and Nyalam.The samplers were installed on rooftops at varying heights of 2-10 m above the ground (Table 1).Field blank filters were placed in the sampler for 24-h with no air flow.In some periods, particularly the monsoon season, some samples could not be collected due to equipment breakdown associated with frequent precipitation or electricity supply failures.The pre-and post-sampling weights of all quartz filters were measured with an electronic microbalance with 10 µg sensitivity (AUW220D, Shimadzu), after equilibration at constant temperature and humidity (20°C, 39%) for at least 24-h.All samples were weighed at least three times to get mean concentrations in both pre-and post-weighing.The samples were stored at -20°C prior to extraction and chemical analysis.The volume of air passing through each filter was recorded by the sampler itself which converted to standard conditions using atmospheric pressure and ambient temperature monitored at each site.

Statistical Analysis
All of the statistical analysis was conducted using SPSS software (Statistical Package for the Social Sciences, Ver.16.0) and included a regression analysis.Comparisons and correlations were considered statistically significant when p < 0.05.The data are summarized as the mean ± one standard deviation.

Spatial Distribution and Seasonal Trend of TSP and PAH Concentrations
The TSP and measured PAH levels were remarkably different among the six sampling sites, with a clear decreasing trend from the southern to the northern side of the Himalayas (Table 2).The ANOVA test showed that the concentrations at the six sites were significantly different at the 0.05 level.A previous study has reported that Lumbini had the highest TSP and PAH concentrations (TSP: 209.1 ± 113.4 µg m -3 ; PAHs: 91.6 ± 54.6 ng m -3 ) (Chen et al., 2016); these values are similar to those reported in many other South Asian cities, such as Delhi (Sarkar and Khillare, 2013) and Agra (Rajput and Lakhani, 2010) but lower than those observed in Kathmandu, the largest city in Nepal (Chen et al., 2015b).Going north from Lumbini, the annual average TSP and PAH concentrations at Pokhara decreased to 123.5 ± 98.1 µg m -3 and 20.7 ± 14.7 ng m -3 , respectively.Pokhara has experienced increased urbanization and increased numbers of vehicles in recent decades due to increases in the numbers of tourists and local residents.It is also a region with significant rural biomass burning, and may be partially decoupled from pollutants in the IGP surface air mass by the blocking effect of mountains situated between Lumbini and Pokhara.Both Dhunche and Jomsom are located in rural areas of the central Himalayas.However, the TSP and PAH concentrations of Dhunche (TSP: 132.3 ± 73.7 µg m -3 ; PAHs: 18.6 ± 5.7 ng m -3 ) were similar to those of Pokhara due to relatively intense local anthropogenic activities.TSP and PAH concentrations of Jomsom (TSP: 96.2 ± 40.8 µg m -3 ; PAHs: 11.1 ± 2.9 ng m -3 ) were found to be comparable with those of Barapani (PAHs: 14.1 ng m -3 ), located in the foothills of the Himalayas, which are clearly impacted upon by the transport of pollutants from the IGP in winter (Rajput et al., 2013).Considering that local emissions in Jomsom are very low, the measured pollutants probably accumulate by the up-valley winds that transport pollutants from the IGP.The lowest TSP and PAH concentrations were observed in Zhongba and Nyalam, located on the Tibet-Himalayas, with average concentrations comparable with those in earlier reports from other sites on the TP (Wang et al., 2014) and from background regions of Central Europe (Dvorska et al., 2012).
The time series of TSP and measured PAH concentrations from Lumbini and Pokhara in Nepal showed clear seasonal variations, with high concentrations in the pre-monsoon season gradually decreasing to minimum concentrations around mid-monsoon season, and gradually increasing again through post-monsoon to the maximum concentrations in the winter season (Fig. 2 and Table SI-1).At other sites, the seasonal variations of TSP and measured PAH concentrations were slightly higher during the non-monsoon season than those in the monsoon season (Table SI-1), although the ANOVA test showed that these seasonal differences were not statistically significantly at the 0.05 level.Intriguingly, the measured PAH concentrations in Nyalam and Dhunche had peak values in the pre-monsoon season of 2013, which might be associated with a pollution event (e.g., ABCs).
There is a variety of sources of TSP and PAHs in the IGP.For example, biomass (such as agricultural waste, animal dung and wood) is burned throughout almost the whole year mainly for cooking.Burning of large amounts of agro-residue also occurs in the IGP, especially in the non-monsoon seasons (Ram and Sarin, 2010), emitting a large amount of PAHs and other air pollutants (Sinha et al., 2014).Forest fires also increase significantly in the pre-monsoon season in this region (Vadrevu et al., 2012).Moreover, the combination of unique weather conditions in this region (long dry season extending from November to May) and geophysical features promotes the accumulation of atmospheric pollution during the non-monsoon season.This leads to the formation of regional scale plumes of air pollutants, known as ABCs, resulting in high concentrations of pollutants on the southern side of the Himalayas (Rengarajan et al., 2007;Ram et al., 2012).Thus, the increase in TSP and PAH concentrations in the non-monsoon season is likely to be due to the combined effects of source strength and accumulation of atmospheric pollutants (Datta et al., 2010).In contrast, high rainfall during the monsoon season washes out most of the PAHs and other pollutants from the atmosphere.Furthermore, the ambient temperature considerably affects the gas-particle partitioning of PAHs (Sitaras et al., 2004;Tham et al., 2008).Usually, high temperatures enhance the evaporation of particle phase PAHs to gas phases, whereas low temperatures strengthen the condensation of gas phase PAHs onto atmospheric particles, leading to higher concentrations of particle-phase PAHs in the winter and lower concentrations in the summer.Despite decreasing markedly from south to north along these two transects, the TSP and PAH concentrations in Zhongba and Nyalam exhibited similar seasonal variations to those in Nepal, suggesting that the northern side of the Himalayas might have similar sources for atmospheric PAHs.

Altitudinal Dependence of TSP and PAH Concentrations
In addition to the distance to the IGP, altitude is another important factor influencing the distribution of TSP and PAHs in the study region (Chen et al., 2008).Although several studies have investigated TSP and PAH concentrations in the IGP (Ram et al., 2010;Rajput et al., 2013), these investigations mainly focused on single sitesespecially those in urban/rural regions -and paid little attention to the PAHs, which are good markers for longrange transport pollutants.
The TSP and PAH concentrations showed altitudinal gradients in the study region (Fig. 3).For example, the annual average TSP and PAH concentrations decreased by factors of approximately 16 and 3 from the lowest to the highest altitude sites.Representation of the PAH and TSP concentrations vs. altitude shows a decreasing trend with statistically significant level (p < 0.05), especially in the non-monsoon seasons (Fig. 3).Previous studies have reported a clear decreasing pattern with increasing elevation of persistent organic pollutants in remote mountain areas, either in air, soil, or plants (Gong et al., 2014).For example, Gong et al. (2014) observed that the concentrations of persistent organic pollutants (POPs) such as hexachlorobenzene (HCB) and polychlorinated biphenyls (PCBs) in the atmosphere

Composition of PAHs
PAH profiles are most likely influenced by the emission source (Liu et al., 2007;Xu et al., 2012) and the amount of photo degradation due to their different atmospheric reactivity (Butler and Crossley, 1981;Ding et al., 2007).The half-Fig.3. Concentrations of TSP (µg m -3 ) and PAHs (ng m -3 ) along the altitudinal gradient in the monsoon and non-monsoon seasons.
lives of PAHs vary significantly, with Phe, Fla, Chr, and Pyr having longer half-lives than BghiP, IndP, Ant, and BaP (Butler and Crossley, 1981;Behymer and Hites, 1985;Kamens et al., 1988).Therefore, the composition profiles of PAH should be influenced by the transport distance to some extent.
Fig. 4 shows the percentage contributions of PAHs with different ring numbers to total concentrations at each site.At Lumbini, the major contributors to the measured PAHs were 5-and 6-ring PAHs (Fig. 4), with contributions to be 38.6% and 39.1%, respectively (Table SI-2).High molecular weight compounds with low vapor pressure significantly dominated the PAH composition profile, indicating that the particulate PAHs in Lumbini were derived directly from the surrounding emission sources with only short-range transport (or little photo-degradation) (Li et al., 2006;Liu et al., 2007).This profile pattern is consistent with that of aerosol samples collected from post-harvest biomass burning emissions in the IGP (Rajput et al., 2011), indicating that biomass combustion is the main source of particulate PAHs in Lumbini.However, this is in contrast to Kathmandu where there is a relatively large contribution of 4-ring PAHs, reflecting the dominant influence of fossil fuel emissions, for example, vehicle engine exhausts and coal combustion (Chen et al., 2015b).
There was a different pattern in Pokhara with that in Lumbini, showed higher contributions of 3-ring (17.3%) and 4-ring (29.3%)PAH.The proportion of high molecular weight PAHs to measured PAHs decreased significantly in samples from the foothills of the Himalayas, with the exception of Dhunche, indicating the important contribution of local emissions at Dhunche (Fig. 4).The three background sites of Jomsom, Zhongba and Nyalam showed relatively similar contributions to the 4 groups of PAHs (Fig. 4 and Table SI-2).Generally, the majority of high molecular weight PAHs is present in the particulate phase due to their low vapor pressures.As a result, high molecular weight PAHs were rarely returned to the atmosphere after scavenging by the intensive dry/wet deposition processes which occur frequently during long-range atmospheric transport (Simonich and Hites, 1995;Wania and Mackay, 1996).In addition, partitioning of PAHs between gas and particle is detemined by TSP concentrations (Liu et al., 2007).Similar TSP concentrations at remotes sites bring about approximate gas-particle partition coefficient.Thus, the simliar TSP concentrations and remoteness from emission sources at these three sites should be the reason for their similar contribution profiles.Wang et al. (2014) reported the compositions of PAHs in soil and air samples across the inland TP, finding that the atmospheric PAHs were dominated by 3-ring PAHs (80%).This is different from our study, where the corresponding contributions were 28.6% and 25.2% for Zhongba and Nyalam, respectively.The burning of yak dung for domestic activities is common in populated areas across the TP (Li et al., 2012), yet the PAH profiles in Zhongba and Nyalam were very different to those of yak dung combustion aerosols over the TP (which had high concentrations of 3-and 4ring PAHs such as Phe, Fla, Pyr, and Chr) (Li et al., 2012).In contrast, the PAH profiles in Zhongba and Nyalam were similar to that in Jomsom (Fig. 4), which might also reflecting the long-range transport of pollutants from south to north across the Himalayas.

Source Identification
Parent PAH ratios are frequently used to identify the origin of PAHs (Yunker et al., 2002;Rajput et al., 2014).Generally, various ratios are used simultaneously to crosscheck the results and to reduce uncertainties.In this study, the concentration ratios of IndP/(IndP + BghiP) and Fla/(Fla + Pyr) in the particulate-phase were selected as indicators to investigate emission sources (Table 2).In general, a ratio of IndP/(IndP + BghiP) lower than 0.2 indicates non-combusted petroleum as the primary input, a value between 0.2 and 0.5 indicates petroleum combustion, and values higher than 0.5 indicate that biomass and coal burning are the primary inputs.For Fla/(Fla + Pyr), a ratio lower than 0.4 indicates unburned petroleum, 0.4-0.5 indicates petroleum combustion, and higher than 0.5 indicates grass, wood or coal combustion (Yunker et al., 2002).The average IndP/(IndP + BghiP) ratio of Lumbini was 0.55 ± 0.037 during the sampling period (Chen et al., 2016), which is consistent with the mixed combustion of biomass and coal.The mean Fla/(Fla + Pyr) ratio was 0.49 ± 0.026, similar to the corresponding value for wheat burning, implying that biomass burning, especially agro-residue burning, is an important contributor to atmospheric PAHs at Lumbini.The IndP/(IndP + BghiP) (0.68 ± 0.096) and Fla/(Fla + Pyr) (0.42 ± 0.113) ratios from Pokhara differed from those of Lumbini (Table 3) and are consistent with values from biomass, coal burning and petroleum combustion, respectively.There may be a local contribution from Pokhara city, owing to the increased number of vehicles in recent decades as a result of increasing of tourists and local residents (the population is now 200,000).In addition, a significant level of rural biomass burning occurs around the city.Thus, the PAHs in Pokhara were derived from the combined impacts of local contributions from Pokhara city (both biomass and fossil burning) and polluted air mass transport from upwind IGP which often arrived during the winter and pre-monsoon seasons.
The ratios of IndP/(IndP + BghiP) (0.52 ± 0.023) and Fla/(Fla + Pyr) (0.48 ± 0.029) from Dhunche were found to be similar to those at Lumbini, indicating that Dhunche was also seriously affected by biomass, coal, and petroleum fuel combustion.However, the ratios of IndP/(IndP + BghiP) in Jomsom, Zhongba, and Nyalam were similar (0.44 ± 0.021, 0.44 ± 0.016, and 0.45 ± 0.038, respectively) and indicative of wheat burning (0.43; Rajput et al., 2011).The same pattern was also found in the ratio of Fla/(Fla + Pyr) (0.47 ± 0.013, 0.48 ± 0.029, and 0.48 ± 0.012, respectively), indicating that biomass burning might be the major contributor in these background regions.However, these two ratios at the three remote sites were also in the range of petroleum combustion, indicating that liquid fuel combustion might be another source of PAHs.Since these locations are remote with minimal local emissions, we consider these sites to be largely influenced by pollutants undergoing long-range transport.
It should be noted that while ratios can be somewhat helpful in distinguishing petrogenic from combustion-derived sources, the diversity of fuels and combustion conditions is likely to produce variations in ratios from a single source, hindering the identification of biomass versus fossil fuel combustion inputs.Additionally, PAHs can be transformed by atmospheric processes so that diagnostic ratios measured in atmospheric samples can differ greatly from those reported for the original sources.As a result, source diagnostic ratios should be used with care and in the context of the studied area.

Transport of PAHs across the Himalayas
The PAH concentrations over high mountain regions are generally influenced by long-range transport from nearby continental sources.China and India were identified as important sources of PAHs, contributing 30% (114 Gg y -1 ) and 23.6% (90 Gg y -1 ) of the total global emissions, respectively (Zhang and Tao, 2009).In addition, Pakistan was another major emitter of PAHs (12 Gg y -1 ).The study region is surrounded by these three countries.However, it is difficult to identify the potential source regions or to compare the relative strengths of different sources affecting the sensitive mountain regions.Therefore, the HYSPLIT back trajectories were used to identify possible source regions and to assess the influence of long-range transport of aerosols to the northern side of the Himalayas.Taking Zhongba as an example, the calculated trajectories were consistent with other descriptions of atmospheric circulation patterns corresponding to the Indian monsoon regime (Fig. 5) (Cong et al., 2009).In the monsoon season, most of the air masses come from the northern (45%) and eastern India (27%), and therefore carry pollution aerosols and moisture.In the non-monsoon seasons, the transport pathways of air masses arriving at Zhongba were dominated by the strong westerlies that also pass over the IGP (> 70%).The backward trajectories analysis indicated that the IGP is the major source region, but there was considerable uncertainty due to the coarse resolution of the meteorological fields (Koracin et al., 2011).In addition, the HYSPLIT trajectories do not take into account thermally driven flows through Himalayan valleys.Therefore, topographic effects on air pollutant transport should also be considered.A diurnal valley wind system often occurs in mountainous terrain, with an upvalley wind during the day reversing to a down-valley wind during the night.In this situation, a significant temperature difference exists between the mountain tops and lowlands.A previous study has reported that the wind regime at the Nepal Climate Observatory-Pyramid, located on the southern side of the Himalayas, was characterized by an evident diurnal cycle of mountain/valley breezes (Bonasoni et al., 2010).During the daytime, the southerly up-valley winds reach a maximum speed in the afternoon and can deliver air pollutants from the foothills to higher altitudes.However, a different mountain-valley breeze circulation was observed on the northern side of the Himalayas, with a down-valley wind dominating in the daytime, especially in the afternoon.Therefore, the local mountain/valley breeze circulation acts as the connection for air pollutants crossing the Himalayas (Cong et al., 2015).In addition, the large valleys (e.g., Mustang valley between Jomsom and Zhongba) provide potential channels for the transport of PAHs crossing the Himalayas (Marinoni et al., 2010;Xia et al., 2011;Lüthi et al., 2015).

CONCLUSIONS
This study investigated the PAH concentrations of atmospheric aerosols along two south-north Himalayan transects.The TSP and PAH concentrations decreased significantly from the southern to the northern side of the Himalayas, with the highest concentrations found in Lumbini (TSP: 210 ± 113 µg m -3 ; PAHs: 91.6 ± 54.6 ng m -3 ) and the lowest concentrations in Nyalam (TSP: 59.1 ± 62.0 µg m -3 ; PAHs: 5.57 ± 3.36 ng m -3 ).In addition to latitude, the PAH and TSP concentrations exhibited a decreasing pattern with increasing elevation, especially in non-monsoon seasons.Moreover, the PAH and TSP concentrations showed a clear seasonal cycle, with higher concentrations during the non-monsoon season than those in the monsoon season, reflecting the different transport and dispersion of emissions.The high percentages of 4-6 ring PAHs in Lumbini, Pokhara and Dhunche indicated that local emissions play an important role in the accumulation of pollutants, while the increase of 3-ring PAHs at the other three remote sites might reflect the higher deposition efficiency of high molecular weight PAHs during long-range transport.The evaluation of diagnostic molecular ratios indicated that atmospheric PAHs in Nepal originate mainly from the use of coal, biomass fuels and vehicular emissions.Similar compositions and seasonal variations were found at Jomsom and Zhongba, suggesting that the northern side of the Himalayas may be affected by anthropogenic emissions from the IGP.Based on the backward air-mass trajectories, we also found that the IGP is the main potential source region for pollutants sampled on the northern side of the Himalayas.

Fig. 1 .
Fig. 1.Location of the study area and sampling sites.

Table 1 .
Location and description of six sampling sites.

Fig. 4 .
Fig. 4. Percentage composition of 3, 4, 5, and 6 ring PAHs in aerosols of six sampling sites and Bode (27.69°N, 85.40°E, located at nearly the center of the Kathmandu Valley).Data from Bode site were reported by Chen et al. (2015b).

Table 2 .
Diagnostic ratios of PAHs in aerosols of six sampling sites and their source profiles.