Black Carbon Aerosols at Mt . Muztagh Ata , a High-Altitude Location in the Western Tibetan Plateau

Measurements of equivalent black carbon (EBC) were conducted for 8 months from November 2009 to September 2010 at Mt. Muztagh Ata to determine its seasonal variation, transport, and potential contribution source areas. The daily EBC concentrations ranged from 33.6 to 330.2 ng m with an average of 133.1 ± 55.0 ng m during the period. Higher values were observed in summer and autumn (approximately 164.0 ng m) than in winter (approximately 96.5 ng m). The diurnal variation in all seasons was stable throughout the day but slightly increased during the nighttime. The results of the potential source contribution function analysis indicated four potential source areas for EBC, with the contributions of polluted trajectory clusters ranging from 4% to 50%. High EBC concentrations were found to be associated with regional circulations developed in high aerosol optical depth areas, resulting in recirculation and accumulation of EBC.


INTRODUCTION
The Tibetan Plateau, with an average elevation of more than 4,000 m above sea level (asl), plays a key role in Asian atmospheric circulation and climatology (Qin et al., 2006;Lu et al., 2008;Cao et al., 2009a).The atmosphere over the Tibetan Plateau is less affected by human activities than urban and rural regions in the middle latitudes of the Northern Hemisphere, providing a unique opportunity to study the effects of aerosols on global and regional climates.The chemistry and optical properties of aerosols over the Tibetan Plateau have thus been the subject of increasing interest (Carrico et al., 2003;Sagar et al., 2004;Pant et al., 2006;Ming et al., 2008;Cao et al., 2009aCao et al., , 2010;;Wang et al., 2012;Zhao et al., 2012;Cong et al., 2013).
Black carbon (BC) aerosols can efficiently absorb solar radiation and heat the atmosphere, thus crucially contributing to climate change (Hansen et al., 1997;Jacobson, 2001;Ramanathan et al., 2001;Jacobson, 2002;Menon et al., 2002;Koren et al., 2004;Wang, 2004;Ramanathan and Carmichael, 2008).The effects of BC are more pronounced in the Tibetan Plateau than in other areas because BC accumulates on snow and ice, which can reduce the surface albedo and consequently accelerate the melting of glaciers (Flanner et al., 2009;Xu et al., 2009;Menon et al., 2010;Xu et al., 2012).Previous studies have indicated that the main source regions of BC in the Tibetan Plateau are South and East Asia, but inconsistent results have been reported (Cao et al., 2010;Lu et al., 2012;Cong et al., 2013).Lu et al. (2012) reported that BC is mainly contributed by South Asia and that the intense BC emissions in eastern China (e.g., the Sichuan Basin) could influence the Tibetan Plateau in summer, which is inconsistent with the results of Cong et al. (2013) for the Nam Co region.Studies on BC in the western Tibetan Plateau are critical for determining the specific BC circulation and supplementing the scarce data (Cao et al., 2009a).
To investigate the variation of equivalent BC (EBC) in the western Tibetan Plateau, real-time EBC measurements were performed at Mt. Muztagh Ata.This study focused on the variation of EBC, its light absorption coefficient, and the potential source contribution areas, enabling a superior understanding of EBC levels and EBC regional transport.

Research Site
The study area was located at Mt. Muztagh Ata at 7,546 m a.s.l., which is the highest peak of the Kunlun Mountains in western China.The area is dominated by westerly winds and has a dry and cold climate.Mt.Muztagh Ata is a round-topped mountain, with year-round snow cover.The topography of the mountain is characterized by a gentle slope to the west and sharper gradients to the north and east.From the peak to the isohypse line at 5,200 m a.s.l., the mountain is covered by a perennial glacier.The area of snow and ice coverage extends for approximately 275 km 2 .
The sampling site used in our study (38.280°N, 75.023°E) was located at a field site of the Muztagh Ata Station for Westerly Environment, Observation, and Research, Chinese Academy of Sciences.It was situated close to the snowline on the western side of Mt.Muztagh Ata at approximately 4,500 m a.s.l.(Fig. 1).The wind direction (westerly) at Mt. Muztagh Ata depends on the general atmospheric circulation and seasonal climate.Temperature records from the sampling site indicate that the average monthly temperatures are -15.0,-13.0, -9.0, -5.2, -5.0, 3.5, 5.5, 5.3, 2.5, -4.0, -6.5, and -11.5°C from January to December, respectively.The four seasons are therefore divided: winter (November-February), spring (March-May), summer (June-August), and autumn (September-October) (Cao et al., 2009a).The annual precipitation averages < 200 mm and mainly comprises snow.This area is essentially free from local sources of residential pollution, but few farmers use the area for grazing during summer (Cao et al., 2009a).

EBC Measurement and Correction
BC that is measured optically is referred to as EBC according to a previous study (Petzold et al., 2013).EBC was measured continuously as 5-minute averages by using an aethalometer (Model AE-16, Magee Scientific Company, Berkley, CA, USA) in the present study (Hansen et al., 1984).The aethalometer uses a solid-state light source operating in near-infrared at a wavelength of 880 nm.The instrument measures the attenuation of a beam of light transmitted through a filter as particles are collected on the filter.The net attenuation is converted to an EBC concentration by using a mass absorption cross section (MAC) of 16.6 m 2 g -1 when λ = 880 nm, which is the manufacturer's default derived from comparison with thermal elemental carbon (EC) measurements (Watson et al., 2005;Park et al., 2006).The default was employed in this study to retain consistency with other comparison studies.In this study, the power for the instrument was supplied by a solar power system and had a flow rate of 4 L min -1 .The instrument was factory-calibrated and was also calibrated before use, with ± 2% accuracy (Hansen et al., 1984;Allen et al., 1999).The light absorption coefficient was backcalculated using the corresponding BC concentration with a MAC value of 16.6 m 2 g -1 .Measurements were performed from November 2009 to September 2010.No data were acquired from March to May 2010 because of the failure of the solar power system.The data were removed when the instrument did not function properly or unrealistic spikes were measured.
Aethalometers are one of the currently available devices employed for measuring BC, which have been used extensively (Watson et al., 2005;Park et al., 2006;Cao et al., 2010).Artifacts resulting from the filter-loading effect, sample-matrix effect, and light-scattering on the aethalometer filter have been reported in some previous studies (Weingartner et al., 2003;Schmidet al., 2006;Collaud Coen et al., 2010).The corrections of these artifacts were conducted using numerical methods (Weingartner et al., 2003;Arnott et al., 2005;Virkkula et al., 2007;Coen et al., 2010).Nevertheless, all correction methods have advantages and disadvantages under field conditions.For example, both aerosol scattering and absorption coefficients are needed for some correction methods, but most BC measurements using an aethalometer do not simultaneously acquire aerosol scattering coefficients.Virkkula et al. (2007) proposed a simple procedure by using a correction factor to correct the loading effects of BC (Virkkula et al., 2007).However, the assumption that BC concentration is stable in the ambient environment during the filter spot change is not always true.The correction factor is not a constant value; its value instead depends on the density of the particles deposited on the filter.The parameter in the correction method by Weingartner et al. (2003) substantially depends on the aerosol type (Weingartner et al., 2003).Conducting the correction was difficult because no scattering coefficient or aerosol type was available in the present study.Notably, a previous study reported that the greatest filter-loading corrections are found for the most polluted environments, indicating that the uncertainty in this study may be low because of the clean environment at the present site (Collaud Coen et al., 2010).The BC data were uncorrected.

Analysis of Air-Mass Back Trajectory and Potential Source Contribution Function
To determine the transport pathway of atmospheric EBC to Mt. Muztagh Ata, a backward trajectory analysis was conducted using the Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model.Five-day back trajectories at 100 m above ground level were obtained using the National Oceanic and Atmospheric Administration (NOAA) HYSPLIT model and used to calculate air parcel back trajectory clusters (http://www.arl.noaa.gov/ready/hysplit4.html)(Draxler and Rolph, 2003).The air mass back trajectories were classified into four clusters during the study period.The movement of an air parcel was described as a series of segment endpoints defined by their latitude and longitude during the period from November 2009 to September 2010.The potential source contribution function (PSCF) values were calculated on the basis of 5day back trajectories arriving at Mt. Muztagh Ata at 1000 LST (local standard time) of each day for three seasons (winter, summer, and autumn).
PSCF has been used to identify the potential locations of regional pollution sources (Polissar et al., 1999;Hsu et al., 2003;Liu et al., 2003;Zhou et al., 2004).In this study, the PSCF values were calculated using the pollutant concentration and backward trajectories produced by the HYSPLIT model (Draxler and Rolph, 2003).The PSCF value can be interpreted as the conditional probability that a critical concentration (i.e., greater than the criterion level) of a given analyte is related to the passage of air parcels through the ijth cell during transport to the receptor site.These cells indicate areas of "high potential" contributions.In this study, the number of endpoints found in the ijth cell was designated n ij .Moreover, the number of endpoints in the same cell, with arrival times at the sampling site corresponding to EBC concentrations higher than an arbitrarily set criterion, was designated as m ij .The selected criterion was the mean EBC value.The PSCF value for the ijth cell was then defined as The collective properties of a high number of endpoints are employed to estimate a conditional probability field that represents the likely contribution area of EBC.Assuming that the uncertainties in the endpoint locations are randomly distributed, the PSCF model should provide a reasonable estimate by using a sufficient number of endpoints that are distributed over the region.Thus, the contribution area of EBC can be obtained using the PSCF model.In the analysis, low n ij values can produce high PSCF values with large uncertainties, given the statistics of low numbers.In this study, to reduce the effect of low n ij values, the PSCF values were multiplied by an arbitrary weighting function W ij to more effectively reflect the uncertainty in the values for these cells (Polissar et al., 1999).The weighting function reduced the PSCF values when the total number of endpoints in a particular cell was less than three times the average value of the endpoints in each cell.In addition, this weighting function can reduce the contribution from the cells that represent unusual transport events.In this study, these weighting function limit values were obtained empirically by running the PSCF program numerous times and applying the trial-and-error method.

Aerosol Optical Depth According to Satellite Data
The aerosol optical depth (AOD) can be determined using the near-real time information contained within satellitederived aerosol data sets, such as those provided by NASA's moderate resolution imaging spectroradiometer (MODIS) instruments (King et al., 1999;Kaufman et al., 2002).The MODIS AOD can be used to indicate aerosol loading (Li et al., 2007;Zhao et al., 2013).The satellite aerosol product employed in this study was the level-2.0AOD from MODIS aboard the Terra satellite and Aura spacecraft (Torres et al., 2007).The MODIS AOD at 550 nm and the National Centers for Environmental Protection/National Center for Atmospheric Research (NCEP/NCAR) winds at 500 hpa were used for case studies.The NCEP/NCAR reanalysis data were provided by the National Oceanic and Atmospheric Administration (NOAA)/Oceanic and Atmospheric Research Laboratories (OAR)/Earth System Research Laboratory (ESRL) Physical Sciences Division, Boulder, Colorado, USA, (http://www.esrl.noaa.gov/psd/),and the corresponding MODIS AOD was obtained from http://disc.sci.gsfc.nasa.gov/giovanni/overview/index.html.

Seasonal Variation of EBC
Fig. 2 displays the daily concentration of EBC, measured by the aethalometer, from November 2009 to September 2010 at Mt. Muztagh Ata.A strong variation was observed in the daily EBC concentration, ranging from 33.6 to 330.15 ng m -3 during the study period.The monthly average EBC concentration ranged from 87.8 ng m -3 to 116.7 ng m -3 in November, December, January, and February.The levels of monthly EBC were slightly higher in summer and autumn (range: 138.5-176.6 ng m -3 ) than in winter (range: 87.8-116.7 ng m -3 ).The lowest values were observed in December, and the highest values were observed in August.In summer and autumn, the EBC concentrations were approximately 1.5 times as high as those in winter, which differs from the pattern observed in most cities or rural areas in which space heating is used in winter.The high EBC concentrations that occurred in summer and autumn were comparable with the previous EC results for the station (Cao et al., 2009a).In remote areas, local human influences are weak, and EBC and other pollutants are delivered by mid-and long-range transport (Liu et al., 2008).
The annual arithmetic means and standard deviations of EBC and the corresponding light absorption coefficient (back-calculated using the MAC value of 16.6 m 2 g -1 ) were 133.1 ± 55.0 ng m -3 and 1.02 ± 1.20 Mm -1 , respectively.The average seasonal values were 163.7 ± 37.4 ng m -3 , 164.2 ± 55.5 ng m -3 , and 96.5 ± 43.0 ng m -3 for EBC concentrations, and 2.67 ± 0.62 Mm -1 , 2.73 ± 0.91 Mm -1 , and 0.96 ± 0.97 Mm -1 for light absorption coefficients at 880 nm in summer, autumn, and winter, respectively (Table1).No data were obtained in spring because of the failure of the solar power system.The variation of EBC is most easily explained by  changes in the types of air masses sampled (Cao et al., 2009a).
The frequency distributions of EBC concentrations are displayed in Fig. 3 for the different seasons.The most frequently occurring EBC values were highest in summer (120.0-140.0ng m -3 ) and lowest in winter (40-60.0ng m -3 ).Wide seasonal variation with a unimodal distribution was observed in winter, and large seasonal variation with a unimodal distribution with one extremely gradual mode was observed in summer and autumn.The variation could be attributed to seasonal EBC transport and meteorological conditions.The frequency distribution in summer and autumn was more complicated.The most frequently occurring BC values in the summer months (June-August) were 80.0-100.0ng m -3 , 120.0-140.0ng m -3 , and 140.0-160.0ng m -3 , respectively.Those values in the autumn months (September-October) were 60.0-80.0ng m -3 , 120.0-140.0ng m -3 , and 200.0-220.0ng m -3 , respectively.Unimodal mode was observed in winter and in the total data, and the most frequently occurring EBC values were 40.0-60.0ng m -3 and 80.0-100.0ng m -3 , respectively.The frequency distributions were highly skewed with median values significantly lower than the mean values in winter and in the total data, but not lower than those in summer and autumn.The results for Mt.Muztagh Ata were typical of a remote site (Cao et al., 2009a), with a low EBC concentration and perturbations caused by the westerly and circular winds from the surrounding area.These atmospheric dynamics were chiefly responsible for the variation of EBC concentrations, especially in summer and autumn, which is discussed in Section 3.4.

Diurnal EBC Variation
The diurnal variation of EBC was investigated by calculating the average EBC concentration at the same time every day for each season (hourly averages calculated with the timestamp correspond to the beginning of the averaging period).(Fig. 4) The diurnal variation generated for the total data were fairly consistent throughout the day, but a slight increase in the concentration was observed during the nighttime.The minimum EBC values (107.7 ± 145.4 ng m -3 ) occurred at 1200 LST and in the afternoon, then increased to 172.0 ± 102.2 ng m -3 at 2200 LST during the study period.
The average diurnal variation was flat in all seasons, with a discernible predawn minimum in the morning and a broad peak from 2000 to 0200 LST.In summer, the average diurnal variation had a large low peak at 1200 LST and a clear afternoon commuting peak at 2200 LST, with a minimum and maximum concentration of 90.2 and 215.8 ng m -3 , respectively.The diurnal variation in autumn was comparable with that in summer, except for the high peak EBC concentration at 1200 LST (nearly 180.0 ng m -3 ), whereas the nighttime maximum exceeded 200.0 ng m -3 .In winter, the variation differed little throughout the day with a slight increase at midnight.In contrast to that of the urban area (Cao et al., 2009b), the less variable EBC concentration at the site was attributed to atmospheric transport rather than local emission sources in the Tibetan Plateau (Cao et al., 2009a).
The diurnal variation of EBC enables the role of mesoscale atmospheric processes and transport to be determined.During nighttime, when ambient temperatures and solar insolation are extremely low, atmospheric turbulence can occur.This turbulence can also occur when the stablenocturnal boundary layer is shallow after sunrise, as the land warms and thermals begin to develop.These processes can influence the transport and dispersion of EBC as well as diurnal variation.

Fig. 3.
Frequency distributions of the EBC mass concentration in winter (November, December, January, and February), summer (June, July, and August), and autumn (September and October).

Long-Range Transport of EBC and its Potential Source Regions
No local emission source is present at Mt. Muztagh Ata; EBC is transported over middle and long distances from the surrounding areas by the movement of air masses.Fig. 5 displays the PSCF plots for the four major source areas, in which PSCF values are displayed with regard to color scale.The potential source regions (PS1, PS2, PS3, and PS4) and the major polluted cluster pathways (C1, C2, C3, and C4), where the EBC was emitted, were likely located in the high PSCF value regions.According to the distributions of polluted cluster pathways and the corresponding PSCF values, the major contribution during the sampling periods was from PS3, which covered the central Asian area, including Kyrgyzstan, Tajikistan, Uzbekistan, Turkmenistan, and Kazakhstan.The high PSCF value in this area indicated anthropogenic influences from cities such as Alma-Ata, Bishkek, and Tashkent.The PS2 area was the region between Kyrgyzstan and China with high PSCF values in autumn and summer, and it included the cities of Kashgar, Aksu, Yili, Shihezi, and Urumchi.PS4 covered the northern area of Pakistan, India, and Afghanistan.Islamabad and Kabul are the major cities in this area.The contributions from PS4 were substantial in summer and winter.Long-range transport from Turkmenistan and Iran constituted the source area of PS1.The small area of PS1 and its distance from Mt. Muztagh Ata indicated a constant but slight EBC contribution.
PS3 was located in a boundary zone with high PSCF values, and most of the polluted air mass trajectories passed through this area, especially in winter.Clusters C1 and C3, which directly passed through the PS1 and PS3 regions (Fig. 5), contributed 55% of the polluted trajectories.PS3 had a greater influence on the EBC concentration at Mt. Muztagh Ata than the PS2 and PS4 source areas did because most polluted air parcels arriving at Mt. Muztagh Ata traveled over PS3 throughout the year, whereas air parcels traveled over PS2 only in autumn and summer.For the PS4 source area, the contribution was 16% for cluster C4 in winter and summer.Middle-range transport (C2, C3, and C4) was more significant than long-range transport (C1) regarding EBC contribution in this area.
Researchers have reported that Mt.Muztagh Ata differs from the main body of the Tibetan Plateau because the Asian monsoon has a much weaker influence in this region (Cao et al., 2009a).The results of this study revealed that EBC is mainly transported from the central and southern Asian countries, following the rapid economic growth of these developing countries.The population density and number of vehicles used in these regions are also increasing, and these areas include a high number of highly polluting industries such as power, steel, and cement (Tripathi et al., 2005).The increasing emission from PS3 and PS4 areas has increased the regional air pollution.This is also reflected in the PSCF results, which reveal high pollution values around the cities.More than 71% of the polluted clusters originated from central and southern Asian countries, whereas approximately 29% originated from the western area in Xinjiang Province.The results indicated that PS2 is also one of the major potential source regions in summer and autumn, a conclusion that is comparable with that of a previous study (Cao et al., 2009a).
To investigate the distributions of fires during the sampling period, seasonal fire hot spots were detected using Web Fire Mapper (http://maps.geog.umd.edu), which is a component of the MODIS Rapid Response System (Giglio, 2007).The results demonstrated that no significant active fire hot spots 0 : 0 0 2 : 0 0 4 : 0 0 6 : 0 0 8 : 0 0 1 0 : 0 0 1 2 : 0 0 1 4 : 0 0 1 6 : 0 0 1 8 : 0 0 2 0 : 0 0 2 2 : 0 0 occurred during the EBC episodes (discussed in Section 3.4) in this study (Fig. 6).The areas that contributed substantially to EBC levels, as identified by PSCF, included the fire hot spots in summer and autumn but not those in winter, indicating that fire event emissions may also be a seasonal but minor contributor to EBC at Mt. Muztagh Ata.The differentiation of natural and anthropogenic influences on EBC concentrations should be further analyzed.

AOD and Wind Patterns during EBC Episodes
The prevailing meteorological conditions during days with a high concentration of EBC were investigated in different seasons.During the sampling period, the days with the highest EBC concentration in each season were referred to as "episode days" (December 6, 2009, for winter;July 25, 2010, for summer;and September 6, 2010, for autumn).The relationships between high EBC concentration and prevailing synoptic scale/meteorological conditions were identified for these dates.The satellite aerosol product used in this study was the level-2.0AOD from MODIS aboard the Terra satellite and Aura spacecraft.A detailed discussion of the AOD measurements is provided in Section 2.4.
The westerly wind can be divided into northerly and southerly bypasses that circulate around the western Tibetan Plateau, which contribute to the stabilizing effect of the synoptic wind at the sampling site.According to Fig. 7(a), high wind speeds in excess of 20.0 m s -1 were observed for the northerly and southerly wind belts, whereas the observed wind speeds were < 6.0 m s -1 during the winter episode day.During the episode days in summer and autumn, the prevailing winds blew mainly from the west and south as well as formed an amphidromic region, with an extremely low wind speed at the site (< 4 m s -1 during summer episode days and 6 m s -1 during autumn episode days) (Figs. 7(b) and 7(c)).
Notably, the winds from different directions formed a "stagnant island," with a low wind speed at the sampling site on episode days.The winds were mostly from the west The results indicated that days with a high EBC concentration can be significantly influenced by synoptic circulation.During the seasonal episode days, the highest concentrations were found to be associated with stagnant conditions.Under these conditions, local and regional circulations developed in the high AOD area, resulting in the recirculation and accumulation of EBC.Notably, in this study, the AOD was not a direct manifestation of the EBC mass concentration, and the accuracy of the satellite derived AOD was considerably uncertain.

Comparison with Other Tibetan and Polar Areas
A comparison of the EBC mass concentration measured at Mt. Muztagh Ata with previous observations reported at other Tibetan and polar locations is provided in Table 2. Tibetan and polar locations are less affected by human activities than urban and rural areas; this is beneficial for investigating the regional differences in the magnitude of EBC concentrations over the Tibetan Plateau region and polar areas.The variation in the EBC mass concentrations may result from differences in meteorological conditions and periods of the year as well as disparities among sampling sites, particle sizes, and sampling methods.These previous studies (in Table 2) have not included original BC, EBC, or EC data, and information regarding the scattering coefficient measurements and aerosol types is unavailable; thus, conducting BC correction by using numerical methods is difficult.Previous studies have indicated that EBC determined by an aethalometer differed from filter EC depending on Fig. 6.Seasonal PSCF results, cluster analysis, and composite animations of MODIS hotspot and fire detection from the Terra and Aqua satellites (Web Fire Mapper) during the sampling period.
the location (Lavanchy et al., 1999;Hansenet al., 2000).Accounting for all known artifacts that occur during measurements is challenging.
The results of this study indicated a systematic increase in the concentration of EBC at lower altitudes in the Tibetan area.This was consistent with a previous report (Ming et al., 2013).The EBC concentrations at Mt. Muztagh Ata and the polar areas were lower than those measured in the southern area of the Tibetan Plateau.For example, the EBC concentrations at Langtang, Nepal, and Manora Peak, Himalayas exceeded 380.0 ng m -3 , but this did not occur at NCO-P, Nepal, which has a higher altitude.The EBC concentration at Mt. Muztagh Ata was approximately onethird of that reported for high alpine sites in the central Himalayas, presumably because of the stronger anthropogenic influences in those areas (Carrico et al., 2003).The sites in Nepal (Manora Peak and Langtang) were closer to sources of anthropogenic emissions and were possibly influenced by the Asian brown cloud (Carrico et al., 2003;Pant et al., 2006).Further comparisons of the data demonstrated that the concentration of EBC at Mt. Muztagh Ata was comparable with the concentration in northeastern Tibet (Wen et al., 2001) but lower than that in polar areas affected by local human activities (Polissar et al., 1998) and approximately 25 times higher than that in pristine polar areas (Wolff and Cachier, 1998;Pereira et al., 2006).A high EBC concentration was observed at Lulang, southeastern Tibet, suggesting that the valleys in the eastern Himalayan section are effective pathways by which EBC can enter the Tibetan Plateau (Zhao et al., 2013).Overall, the observations supported the few previous in situ measurements of EBC and also demonstrated that the specific circulation and topography of the region contributed to the measured level and transport patterns of EBC over the Tibetan Plateau (Cao et al., 2010;Zhao et al., 2013).Size fractions: TSP, total suspended particles; PM 10 and PM 2.5 , particles less than or equal to 10 and 2.5 µm, respectively.b Carbon analysis methods: TOR, thermal optical reflectance, TOT, thermal optical transmittance.

CONCLUSIONS
EBC concentrations were measured at Mt. Muztagh Ata.The daily EBC concentration ranged from 33.6 to 330.15 ng m -3 with an average value of 133.1±55.0ng m -3 during the sampling period.Higher EBC values were observed in summer and autumn (approximately 164.0 ng m -3 ) than in winter (approximately 96.5 ng m -3 ).The diurnal variation in all seasons was fairly stable but increased slightly during the nighttime.PSCF analysis confirmed that > 71% of the polluted clusters originated from central and southern Asian countries, whereas < 29% originated from the western area in Xinjiang Province.Four major potential EBC sources were identified, which included the major cities and fire hot spots in summer and autumn, but not in winter.The results demonstrated that medium-range transport (C2, C3,and C4) was more significant than long-range transport (C1) for EBC.High EBC concentrations were found to be associated with stagnant conditions and a regional circulation pattern that developed in the high AOD area, resulting in the recirculation and accumulation of EBC.Comparing this and previous studies revealed that EBC levels and transport vary widely throughout the Tibetan Plateau and should be investigated further in the future.

Fig. 1 .
Fig. 1.Location of the observational site at Mt. Muztagh Ata in the western Tibetan Plateau.

Fig. 2 .
Fig. 2. Daily average 1-hour EBC concentration (thin line) and 30-day running average (thick line) at Mt. Muztagh Ata from November 2009 to September 2010 (star indicates a lack of data for spring).

Fig. 4 .
Fig. 4. Diurnal variation of EBC in winter, summer, autumn, and for the total data.

)Fig. 5 .
Fig. 5. Potential source distribution of PSCF and the mean back trajectories of clusters.
in winter (Figs.7(a) and 7(d)) but were also southerly and easterly in summer and autumn.Southern sector winds blew across the area with a high AOD in summer (Figs.7(b) and 7(e)) and autumn (Figs.7(c) and 7(f)).According to NCEP/ NCAR wind trajectory analysis at 500 hpa and the satellite (MODIS) integrated AOD, a southerly circular wind covered the area with high AOD values, and the relatively low wind speed contributed to the accumulation of EBC at Mt. Muztagh Ata (Figs. 7(b), 7(c), 7(f), and 7(g)).

Fig. 7 .
Fig. 7. NCEP/NCAR wind trajectory analysis at 500 hpa and satellite (MODIS) integrated AOD for typical EBC episode days for different seasons during the sampling period: a, b, and c represent the NCEP/NCAR wind trajectory, and d, e, and f indicate the average AOD in winter, summer, and autumn, respectively.
a Ave: arithmetic mean value, b SD: standard deviation.c Max: maximum value, d Min: minimum value.

Table 2 .
BC concentration at Mt.Muztagh Ata and selected remote areas.