Anthropogenic Aerosol Pollution over the Eastern Slope of the Tibetan Plateau

Because of rapid urbanization and industrialization, an increasing amount of attention is being paid to the air pollution associated with aerosols (Kahn et al., 2005; Ramanathan and Carmichael, 2008; Xue et al., 2008; Huang et al., 2010). Aerosols significantly impact the radiation balance by scattering and absorbing radiation energy (Huang et al., 2009; Zhang et al., 2009; Liu et al., 2011; Che et al., 2018) and by altering cloud microphysical properties (Huang et al., 2006; Rosenfeld et al., 2008; Sakaeda et al., 2011; Zhang et al., 2015). Recently, many researchers have focused on Asian aerosols because of their high concentrations, compositional complexity and significant impacts on the weather/climate system (e.g., Huang et al., 2014; Liu et al., 2014; Deng and Xu, 2015; Zhu et al., 2018). The optical properties, long-range transport and climatic effects of Asian aerosols have been investigated through in situ measurements (VanCuren, 2003; Liu et al., 2011), aircraft- and ship-based measurements (Dickerson et al., 2007; Stith et al., 2009), remote sensing data (Liu et al., 2008; Jia et al., 2015; Li et al., 2015) and model simulations (Chin et al., 2007; Chen et al., 2011, Chen et al., 2013). To put it simply, Asian aerosols represent a significant problem for the weather/climate system and atmospheric environment.

Sometimes known as "the roof of the world", the Tibetan Plateau (TP) is a sensitive indicator and regulator of climate change and plays a significant role in driving the climate change in the Northern Hemisphere and even the globe through thermal and mechanical forcing (Lau et al., 2006; Wu et al., 2007). The TP contains fragile ecosystems and the headwaters of Asia's primary rivers (Yao et al., 2012; Ma et al., 2017), which sustain 40% of the world's population and centuries-old civilizations. Because of fragmentation, intersection and high heterogeneity of vegetation patches, the meteorological parameters over the TP show a wide range (Ma et al., 2011). Based on satellite observations and model simulations (e.g., Huang et al., 2007; Liu et al., 2008, Liu et al., 2015), aerosol accumulation has even been found over the TP. Besides the heating effect of absorbing aerosols on the atmosphere over the TP (Lau et al., 2010), they have also been found to induce a reduction in the snow albedo by 2.1% over the southern TP when the aerosol optical depth (AOD) increased by 0.1 (Lee et al., 2013, 2016). This so-called snow darkening effect induced by aerosols can drive the retreat and thinning of glaciers, which drastically alters the ecosystem, induces additional carbon release, creates a positive feedback and accelerates climate warming (Yang et al., 2010). Moreover, the addition of aerosols could enhance the cloud core updraft and precipitation by intensifying convection (Zhou et al., 2017; Liu et al., 2019). Also, due to the large scale of the topography, the TP acts as a channel for aerosol transport to the upper-tropospheric atmosphere and long-range transport around the Northern Hemisphere (Xu et al., 2018). Therefore, in view of the important role of the TP in the climate patterns, monsoon process and atmospheric circulation in Asia and the Northern Hemisphere (Ma et al., 2017), the addition of abundant aerosols in the air over the TP poses new climatic and environmental risks.

Most researches on the aerosols over the TP have focused on the aerosol properties over the northern and southern slopes. A large number of studies have investigated the aerosol transport from the Taklimakan Desert and Indian Peninsula to the TP by using satellite observations, reanalysis data and model simulations (Liu et al., 2008; Lau et al., 2006, Lau et al., 2010; Dumka et al., 2010; Chen et al., 2013; Xu et al., 2014, Xu et al., 2015; Jia et al., 2015; Liu et al., 2015). However, few studies have focused on aerosols over the eastern slope of the TP (ESTP, mainly the Sichuan Basin and the Hengduan Mountains). Those that have, found that the AOD values over the ESTP are even higher than those over the Taklimakan Desert, the North China Plain and Yangtze River Delta regions (Luo et al., 2014; Liu et al., 2015). Though there have been some studies on the optical and chemical properties of aerosols by field aerosol sampling in the city of Chengdu (Tao et al., 2013, Tao et al., 2014; Zhang et al., 2018), the sources of carbonaceous aerosols (Chen et al., 2014; Hu et al., 2016), and the distribution of organic aerosols over the Sichuan Basin (Li et al., 2013), most of them focused mainly on the precise chemical components and emission sources based on case studies through field sampling or source apportionment techniques. Moreover, research on aerosols over the Hengduan Mountains is even more lacking. Thus, studying the aerosol transport processes and physical mechanisms of aerosol accumulation on a large scale over the ESTP is needed.

In this study, using multiple satellite observations and reanalysis datasets, the spatial and temporal evolution of the AODs, emissions and column mass fluxes over the ESTP are investigated by category. The aerosol emissions are discussed with respect to human activities and burning processes over different land covers. The sources and transport processes of aerosols over the ESTP are investigated by analyzing the column aerosol mass fluxes and emissions along streamlines. Based on the analysis of a combination of aerosol emissions, transport processes and meteorological fields, the physical mechanism of aerosol accumulation over the ESTP is discussed in detail.

2.1.1. CALIPSO

As an important member of the A-Train constellation of satellites, the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO), launched in April 2006, fills the gap in observing the global three-dimensional behavior of aerosols and clouds. The CALIPSO level 2 dataset includes the "layer", "profile" and "vertical feature mask" (VFM) products. These products are archived and distributed by the Atmospheric Science Data Center. The level 2 VFM product provides distribution information on clouds and aerosols in the vertical and horizontal directions. On 8 November 2016, the CALIPSO project released its version 4.10 products with an improved "cloud and aerosol distinction" algorithm and definition of elevated layers and fringes (Omar et al., 2018), which allows for better classification of aerosol and cloud, especially at high latitudes and altitudes.

2.1.2. MISR

Observing at nine distinct zenith angles (70.5^°, 60.0^°, 45.6^°, 26.1^° and nadir for both backward and forward) and four narrow spectral bands (446, 558, 672, and 866 nm), the Multiangle Imaging SpectroRadiometer (MISR) provides ongoing global coverage with 36 spectral angular channels (9 cameras × 4 spectral bands). This unique design enables this instrument to have a high spatial resolution (17.6 km × 17.6 km), a wide range of along-track view angles, and a high-accuracy calibration, especially in reducing the dependence of explicit radiometric surface properties in aerosol retrieval algorithms (Martonchik et al., 2004). The AOD data of the MISR Level 3 product is retrieved from multiple orbits on geographic grids of 0.5^°× 0.5^° at the monthly scale.

2.1.3. MODIS

The Moderate Resolution Imaging Spectroradiometer (MODIS) aboard NASA's Terra and Aqua satellites makes near-global daily observations over the Earth ranging from 0.41 to 15 μm. The MODIS instrument views the entire surface of the Earth every one to two days with a viewing swath width of 2330 km. It achieves comprehensive observations of solar radiation, the atmosphere, oceans, cryosphere and land from a single series of polar-orbiting space platforms. The Level-3 MODIS Atmosphere Monthly Global Product MOD08_M3 contains monthly 1^°× 1^° grid values of atmospheric parameters related to aerosol particle properties, optical and physical cloud properties, atmospheric water vapor, atmospheric profile and stability indices, and the total ozone burden.

2.2.1. MERRA-2

The second Modern-Era Retrospective analysis for Research and Applications (MERRA-2) was the first long-term atmospheric reanalysis dataset, beginning in 1980. MERRA-2 was introduced to replace the original MERRA dataset because of advances in the Goddard Earth Observing System model and assimilation system (Gelaro et al., 2017). MERRA-2 was also the first satellite-era global reanalysis to assimilate space-based observations of aerosols and represent their interactions with other physical processes in the climate system, including aerosol analysis (Randles et al., 2017), meteorological fields (Reichle et al., 2017a, b), radiation budget (Collow and Miller, 2016), ozone (Wargan and Coy, 2016) and so on. It is provided by the Modeling and Assimilation Data and Information Services Center (https://disc.sci.gsfc.nasa.gov/datasets?page=1&keywords=MERRA-2) managed by the NASA Goddard Earth Sciences Data and Information Services Center. The MERRA-2 aerosol reanalysis has considerable skill in showing numerous observable aerosol properties (Gelaro et al., 2017; Randles et al., 2017), including dust, sulfate, organic carbon (OC), black carbon (BC), and sea salt aerosols (Chin et al., 2002; Colarco et al., 2010). The monthly aerosol diagnostics product of tavgm_2d_adg_Nx from MERRA-2 provides the emissions distribution of these five aerosol species and tavgm_2d_aer_Nx provides the AOD and aerosol column mass flux by category. The monthly surface flux diagnostics product of tavgM_2d_flx_Nx, which provides the distribution of total precipitation, is also used in this study. All data collections from MERRA-2 are on a longitude-by-latitude grid of approximately 0.625^°× 0.5^°. The variables of vertical velocity, air temperature, and the zonal and meridional wind components of horizontal wind from the instM_3d_asm_Np product are interpolated to 42 standard pressure levels in the vertical direction.

2.2.2. ECMWF

The European Centre for Medium-Range Weather Forecasts (ECMWF) is an independent intergovernmental organization established in 1975 to produce and disseminate numerical weather predictions. ERA-Interim, produced by ECMWF, is a global atmospheric reanalysis dataset from 1979 and is continuously updated in real time. It contains estimates of atmospheric parameters (e.g., air temperature, pressure and wind at different altitudes) and surface parameters (e.g., rainfall, soil moisture content, and sea surface temperature). Reanalysis datasets are widely used as surrogates for large-scale observations over the TP because of the sparse meteorological network in this region. The vertical velocity, air temperature, and zonal and meridional components of horizontal wind from ERA-Interim used in this study have a 0.5^°× 0.5^° (latitude × longitude) spatial resolution and 37 pressure levels in the vertical direction. The Copernicus Atmosphere Monitoring Service (CAMS) dataset provides information on atmospheric composition (such as the AOD of dust, sulfate, BC and so on) by combining atmospheric model simulations and Earth observation data. The AOD data from CAMS used in this study also have a spatial resolution of 0.5^°× 0.5^° (latitude × longitude).

Figure 1 shows the topographical distribution and overview of the TP and its vicinity, where the thin white lines indicate rivers. As shown in Fig. 1, the TP is located in the central and eastern part of the Eurasian continent, with the Indian Ocean to the south and the Pacific Ocean to the east. Unlike the southern and northern slopes of the TP, the ESTP (black dashed rectangle in Fig. 1) features more complex terrain, with steep high mountains (the Hengduan Mountains and Qinling Mountains) next to lowlands (the Sichuan Basin and Guanzhong Plain) and densely interspersed with rivers. Correspondingly, the ESTP features complex climate conditions and a fragile ecological system. Moreover, the addition of more human activities makes the situation more complicated. The black solid boxes in Fig. 1 represent the low-elevation areas of the ESTP (i.e., the Sichuan Basin) and the northern slope of the TP (i.e., the Taklimakan Desert), along with the important economic center in China known as the Beijing-Tianjin-Hebei (BTH) region.

Figure 1.  Topographical distribution and overview of the study area. The black solid rectangles indicate the typical regions (Taklimakan Desert, Sichuan Basin and BTH). The dashed rectangle denotes the eastern slope of the TP. The dashed black line indicates the trajectory of the CALIPSO satellite over the ESTP on 29 April 2016. The dotted line denotes the latitude of the altitude-longitude cross section of meteorological fields in Fig. 12. The thin white lines indicate rivers.

Reanalysis datasets and satellite observations are widely used as surrogates for large-scale observations over the TP because of the sparse meteorological network in this region. In this study, the AODs from MODIS, MISR and ECMWF CAMS are used to evaluate the MERRA2 datasets, and then the spatial and temporal evolutions of aerosol loading and emissions over the ESTP are analyzed by category. The inward and outward transports of aerosols over the ESTP are verified by categories via aerosol column mass flux and the aerosol sources can be determined by backstepping along the streamlines of aerosol column mass flux until reaching the large values of emissions. The sources, transport processes and accumulation of aerosols over the ESTP are discussed through analyzing the meteorological contours of the zonal and meridional components of wind speed, vertical velocity and temperature from MERRA-2 with comprehensive consideration of aerosol emissions and geographical conditions. An intercomparison and interverification of meteorological variables between ECMWF ERA-Interim and MERRA-2 data is performed to improve the reliability of the results.

Figure 2 shows the distribution of annual mean AOD obtained from MERRA-2 reanalysis data and satellite observations. Compared with MISR and MODIS observations, MERRA-2 data can describe the special distribution of aerosols well. Both reanalysis data and satellite observations indicate high AOD values over the ESTP, which are even higher than those over the Taklimakan Desert and the most developed industrial and populous regions. Additionally, the magnitude of the AOD from the MERRA-2 data (Fig. 2c) is smaller than that from the MODIS observations (Fig. 3a) but in good agreement with that from MISR observations (Fig. 3b). According to some previous studies (e.g., Abdou et al., 2005; Myhre et al., 2005), the AOD values produced by MODIS are systematically larger than those from MISR and Aerosol Robotic Network (AERONET) data over the land region. The AOD retrieval based on MODIS observation is highly dependent on the underlying land surface, and the AOD values are on average lower in forest and grassland areas but higher over urban and desert-like areas (Xie et al., 2011). By contrast, the MISR-retrieved AOD is more accurate and the MISR AOD data fall within the predicted uncertainties when validated against sunphotometer-observed AOD data and AERONET observations (Christopher and Wang, 2004; Kahn et al., 2005). Thus, on the whole, the AOD data from MERRA-2 can be used to describe the aerosol optical properties reliably.

Figure 2.  Annual mean AOD derived from (a) MISR, (b) MODIS and (c) MERRA-2 reanalysis data during the period 2000-18. The black solid rectangles indicate the typical regions (Taklimakan Desert, Sichuan Basin and BTH). The dashed rectangle denotes the eastern slope of the TP.

Figure 3.  Annual mean AOD contributed by sulfate, dust and carbonaceous aerosols derived from MERRA-2 and CAMS reanalysis data during the period 2000-18. The black solid rectangles indicate the typical regions (Taklimakan Desert, Sichuan Basin and BTH). The dashed rectangle denotes the eastern slope of the TP.

To accurately analyze the distribution of different aerosols, an intercomparison and interverification of the AODs derived from MERRA-2 and CAMS reanalysis data is performed by category (Fig. 3). Due to similar sources, OC and BC will be combined as carbonaceous aerosols in this study. Figure 3 shows that both the pattern and values of AOD from MERRA-2 and CAMS data are extremely consistent, except for some slight deviation of dust aerosol in northwestern China. The distribution of annual mean AODs shows that the dust aerosols are mainly distributed in the Taklimakan and Gobi deserts in northwestern China, while the sulfate and carbonaceous aerosols are mainly distributed over the foothills of the Himalaya (i.e., the southern slope of the TP), the Sichuan Basin and North China, where human activities are frequent. Overall, the aerosols over the ESTP are mainly sulfate, with a maximum over the Sichuan Basin, followed by carbonaceous and dust aerosols.

Figure 4 shows the spatial distribution of the annual mean horizontal wind vector and vertical wind speed at 850 hPa and 700 hPa derived from MERRA-2 and ERA-Interim data. The meteorological fields from these two datasets are highly consistent. There is a prevailing deep updraft over the Sichuan Basin. Due to the blocking of the Hengduan Mountains, it is difficult for the low-level airflow to reach the ESTP and the winds blow steadily from South China to the Sichuan Basin. The high-level wind field is the result of the dynamic effect of the TP.

Figure 4.  Spatial distribution of the annual mean horizontal wind vector (arrows) and vertical wind speed (colors; units: m s^-1) at 850 hPa and 700 hPa derived from MERRA2 and ERA-Interim from 1980 to 2018. The black solid rectangles indicate the typical regions (Taklimakan Desert, Sichuan Basin and BTH). The dashed rectangle denotes the eastern slope of the TP.

Figure 5 shows the vertical distribution of atmospheric classifications along the scanning orbit path of the CALIPSO satellite over the ESTP based on a typical case. As indicated in Fig. 5, the aerosols over the ESTP include smoke, polluted dust and polluted continental or smoke aerosols. Especially, over the ESTP, i.e., the Sichuan Basin, abundant smoke aerosols are lifted to altitudes greater than 5 km, whereas polluted dust and polluted continental aerosols remain in the lower layer. As indicated in Fig. 5, the aerosols over the ESTP are high in load and complex in type. These aerosols can be further lifted to high altitude and mixed with clouds. Similar phenomena are common in the CALIPSO observations from 2006 to 2018 (figures omitted) and therefore warrant further investigation. Although satellite observations have revealed abundant aerosol accumulation over the ESTP, few studies have investigated this phenomenon systematically. To investigate the mechanism of aerosol accumulation over the ESTP, multiple satellite observations and reanalysis data are combined to analyze the distribution characteristics, emissions and transport processes of different kinds of aerosol.

Figure 5.  Atmosphere classifications along the scanning orbit path of the CALIPSO satellite over the ESTP (as per the black dashed line shown in Fig. 1) on 29 April 2016.

To determine the contributions of different kinds of aerosols to the AOD over the ESTP and the spatial and temporal distributions better, further analyses are performed. Figure 6 shows the seasonal variations of sulfate, dust and carbonaceous aerosols. The results indicate that sulfate aerosols over the ESTP are the dominant type year round, with peak values in autumn and winter. Besides, there are large quantities of carbonaceous aerosols over the vicinity of the Hengduan Mountains and Sichuan Basin in spring. Additionally, dust aerosols are mainly present in the Taklimakan Desert and generate higher AOD values in spring and summer and lower AOD values in autumn and winter. Only in the spring, a small amount of dust aerosols appear in the Sichuan Basin. Generally, the emissions of aerosols are critical process in producing severe air pollution. To determine the reason for high aerosol loading over the ESTP, the seasonal emissions of sulfate, dust and carbonaceous aerosols are analyzed (Fig. 7). Overall, the patterns of aerosol emissions are similar to the corresponding AOD shown in Fig. 6, which indicate that the local emissions contribute directly to the aerosol loadings over the study area. The regions with large sulfate aerosol emissions correspond well to the areas with considerable human activity (Cheng et al., 2016). In the foothill of the ESTP, i.e., the Sichuan Basin, the pattern and intensity of sulfate emissions are almost the same year round. Notably, the emission of sulfate aerosols over the Sichuan Basin is significantly high but lower than that from the BTH region and some coastal areas. There are significant large emissions of carbonaceous aerosol in Southeast Asia and the Hengduan Mountains in spring, which corresponds to the wildfires in the forests, grasslands and savannas of Southeast Asia and South China (Zhu et al., 2017). In addition, the emissions of carbonaceous aerosols are mainly concentrated in the Sichuan Basin, the BTH region and the coastal areas, which exhibit the highest values in summer. Burning of agricultural residues and domestic biofuel usages are important sources of carbonaceous aerosols (BC and OC) and greenhouse gases (Streets et al., 2003; Duan et al., 2004). Considering the type of vegetation and urbanization, the local emissions of carbonaceous aerosol in the Sichuan Basin may be mainly from human activities, such as motor vehicles, cooking and the combustion of agricultural wastes (Gorin et al., 2006). Surrounding the TP, dust aerosols are emitted mainly from the northern slope of the TP, especially from the Taklimakan and Gobi deserts. The dust emissions over the northern slope of the TP peak in spring and summer and the ESTP produces almost no dust emissions. Sulfate aerosols are mainly from human industrial activities (Buchard et al., 2014), while the carbonaceous aerosols are mainly produced by agricultural activities and natural forest fires (Duan et al., 2004; Zhu et al., 2017). In this paper, sulfate and carbonaceous aerosols are roughly considered to be the anthropogenic aerosols and dust aerosols are considered to originate from natural emissions.

Figure 6.  Spatial distribution of multiyear averaged optical depths of sulfate (left column), dust (middle column) and carbonaceous (right column) aerosols derived from MERRA-2 for the period 1980-2018. The AOD of carbonaceous aerosols is multiplied by a factor of two. The black solid rectangles indicate the typical regions (Taklimakan Desert, Sichuan Basin and BTH). The dashed rectangle denotes the eastern slope of the TP.

Figure 7.  As in Fig. 6 but for the aerosol emissions. The emissions of sulfate and carbonaceous aerosols are multiplied by factors of 5 and 10, respectively. Units: g m^-2 s^-1. The black solid rectangles indicate the typical regions (Taklimakan Desert, Sichuan Basin and BTH). The dashed rectangle denotes the eastern slope of the TP.

According to above analyses, the aerosols over the Hengduan Mountains are mainly carbonaceous aerosols emitted from fires in the forests, grasslands, cropland and savannas, and the AOD is larger in spring. Additionally, the aerosols over the ESTP mainly accumulate in the Sichuan Basin. We compared the regional averaged AOD values over the Sichuan Basin, Taklimakan Desert and BTH region (black solid rectangles in Fig. 1) for the period 1980-2018, as shown in Fig. 8. The results show that the AODs over the Sichuan Basin and BTH region present similar trends from 1980 to 2018 and the AODs over the Sichuan Basin are slightly larger than those over the BTH region. Additionally, the AOD values over the Sichuan Basin and BTH region are both higher than those over the Taklimakan Desert, especially since 1995. During the period from 1980 to 2010, the AODs over the Sichuan Basin and BTH region increased and decreased after 2010, respectively. Thus, the Sichuan Basin may be an important source of anthropogenic aerosols over the high altitudes of the ESTP.

Figure 8.  Regional average AODs and aerosol emissions derived from MERRA-2 over the Taklimakan Desert, BTH region and Sichuan Basin from 1980 to 2018. The dust emissions over the Taklimakan Desert are multiplied by a factor of 0.3.

Furthermore, we compared the regional average AODs and emissions of three kinds of aerosols. As shown in Fig. 9, the AODs of sulfate, dust and carbonaceous over the Sichuan Basin are 0.31, 0.05 and 0.106, respectively. The AOD of dust over the Sichuan Basin is the lowest among the three regions. Unexpectedly, the AODs of sulfate and carbonaceous aerosol over the Sichuan Basin are larger than those over the BTH region, which is a booming industrialized region. However, the sulfate emissions (1.516× 10^-7 g m^-2 s^-1) in the Sichuan Basin are significantly less than those in the BTH region (4.094× 10^-7 g m^-2 s^-1), whereas the carbonaceous aerosol emissions in these two regions are similar. In short, the AODs over the ESTP are large relative to the corresponding emissions. As the emissions of sulfate and carbonaceous aerosols over the ESTP both show weak seasonal variations, the seasonal changes of AODs may be affected by transport processes and atmospheric circulation under specific meteorological and geographical conditions.

Figure 9.  Time series of regional average AOD derived from MERRA-2 data over the Taklimakan Desert, BTH region and Sichuan Basin from 1980 to 2018.

As illustrated above, the sulfate and carbonaceous AODs over the ESTP are comparable to those over the BTH region and deserts; however, the corresponding emissions are small. To provide a clear perspective of aerosols accumulating over the ESTP, the aerosol transport process and the meteorological factors are discussed below.

Figure 10 shows the seasonal column mass flux of sulfate, dust, and carbonaceous aerosols for the period 1980-2018. Over the ESTP, the sulfate column mass flux is low, which indicates that the transport is weak and the local sulfate emissions are dominant. The dust column mass flux in spring (Fig. 10b) is significantly stronger than that in other seasons, which means that strong transport of dust occurs in spring. In addition to local emissions, carbonaceous aerosols are mainly transported from Southeast Asia and southern China in spring. Furthermore, to compare the contribution of long-range aerosol transport in the three typical regions, the net flux is calculated by subtracting the outflow flux from the inflow flux. Table 1 shows the averaged aerosol column mass flux at the borders of the Taklimakan Desert, BTH region and Sichuan Basin. Both the negative inflows at the upstream border and positive outflows at the downstream border indicate that aerosols are transported from the study area to other surrounding regions. The net flux is calculated by subtracting the downstream outflow flux from the upstream inflow flux. Therefore, a positive value indicates that aerosols are transported from surrounding regions to the study area, while a negative value indicates that aerosols are transported from the study area to surrounding regions. For the main aerosols in these three regions, the outward transport of dust aerosols over the Taklimakan Desert is the greatest, followed by sulfate aerosols over the BTH region, whereas the transport of sulfate aerosols from the Sichuan Basin is weak. As listed in Table 1, the net mass fluxes of dust over the Taklimakan Desert are negative and large both in the longitudinal (-0.36038 g m^-1 s^-1) and latitudinal (-0.22651 g m^-1 s^-1) directions, which means that the Taklimakan Desert is a dust contributor to the surrounding areas. A similar situation is found for sulfate aerosols over the BTH region, except the net mass fluxes are smaller (-0.00110 g m^-1 s^-1 in the latitudinal direction and -0.04341 g m^-1 s^-1 in the longitudinal direction). In addition, Table 1 also shows that a large number of dust aerosols are transported to the Sichuan Basin in the latitudinal direction (with mass fluxes of 0.23678 g m^-1 s^-1). To some degree, Table 1 explains why the Sichuan Basin exhibits little (sulfate and carbonaceous) or no (dust) aerosol emissions but has large AOD values, as shown in Fig. 6.

Figure 10.  Seasonal column mass flux derived from MERRA-2 for sulfate, dust, and carbonaceous aerosols from 1980 to 2018. Units: g m^-1 s^-1. The column mass fluxes of sulfate and carbonaceous aerosols are multiplied by factors of 5 and 10, respectively. The black solid rectangles indicate the typical regions (Taklimakan Desert, Sichuan Basin and BTH). The dashed rectangle denotes the eastern slope of the TP.

Figure 11 shows the seasonal distributions of the horizontal wind vectors (arrows) and vertical wind speeds (colors) from 1980 to 2018 from MERRA-2. A perennial strong updraft is present over the ESTP at both 850 and 700 hPa, which is different from the conditions over the southern and northern slopes of the TP. Simultaneously, a perennial easterly flow is present in the eastern Sichuan Basin at 850 hPa, which may bring air masses with entrained pollution particles from the middle and lower reaches of the Yangtze River and even from East China (in summer and autumn) and South China (in winter and summer). As shown in Fig. 7, sulfate aerosol emissions in the Sichuan Basin are relatively greater than those in the surrounding areas year round; thus, the inward transport has relatively little effect on the sulfate aerosol concentration over the Sichuan Basin. The prevailing southerly wind in spring and winter has a greater impact on carbonaceous aerosols than sulfate aerosols. Due to the forest, grassland and savanna fires in the winter and spring, abundant carbonaceous aerosols are emitted in the vicinity of the Hengduan Mountains and Southeast Asia. Also, these carbonaceous aerosols can be transported to the ESTP by prevailing southerly wind. Similar to sulfate aerosols, the transport of carbonaceous aerosols is weak in the other seasons. Combined with the distribution of dust emissions (Fig. 7b) and column mass flux (Fig. 10b), the dust aerosols over the ESTP originate from the Gobi and Taklimakan deserts in Northwest China, as confirmed by the wind field at 700 hPa. In spring, due to the intensive dust emissions over the Taklimakan and Gobi deserts (Fig. 7a) and the strong northwesterly winds (Fig. 11e), some dust aerosols pass through the Qinling Mountains to the ESTP. Due to abundant anthropogenic aerosols over the ESTP, the transported dust aerosols are often polluted, which is consistent with the CALIPSO observations. However, in other seasons, no appropriate wind field exists to transport dust aerosols, resulting in almost no dust aerosols over the ESTP. Although the dust emissions in the Taklimakan Desert and the Gobi Desert are very large (Fig. 7e), the prevailing summer monsoon (Fig. 11 b, f) limits the transport of these aerosols to the ESTP.

Figure 11.  Spatial distribution of the horizontal wind vector (arrows) and vertical wind speed (colors; units: m s^-1) at 850 hPa (left column) and 700 hPa (right column) derived from MERRA2 data for (a, e) spring, (b, f) summer, (c, g) autumn and (d, h) winter from 1980 to 2018. The black solid rectangles indicate the typical regions (Taklimakan Desert, Sichuan Basin and BTH). The dashed rectangle denotes the eastern slope of the TP.

Because the perennial ascending current could influence the aerosol loading over the Sichuan Basin, further analysis is performed. Figure 12 shows the vertical section of the seasonal wind vector and the temperature lapse rate (contours). Considering the differences in the magnitude and unit between horizontal and vertical wind, the vertical wind velocity is multiplied by a factor of 50 in order to present the local circulation clearly. When straight westerlies flow across the TP, a local circulation is produced over the ESTP. Affected by the drag of upper-level air currents and the blocking by terrain, the easterly airflow at the lower level turns upward after encountering the TP, which triggers the formation of a local circulation. This terrain-driven circulation traps aerosols over the Sichuan Basin and makes diffusion difficult, resulting in prolonged pollution. Thus, the high values of AOD over the Sichuan Basin can be attributed to a combination of natural and anthropogenic factors.

Figure 12.  Altitude-longitude cross sections of wind vector (arrows) and temperature lapse rate (colors; units: ^°C km^-1) from MERRA2 along the dashed line in Fig. 1 for (a) spring, (b) summer, (c) autumn and (d) winter from 1980 to 2018. The gray shading indicates the topography.

Additionally, precipitation scavenging is a very important process to remove atmospheric aerosols (Hou et al., 2018). The relationship between the precipitation and aerosols in the Sichuan Basin is discussed further. Figure 13 shows the spatial distribution of total precipitation and the time series of monthly mean precipitation and AOD anomalies over the Sichuan Basin. In summer, the prevailing wind blows from the Bay of Bengal and the South China Sea to the interior (Figs. 11b and f), which provides sufficient water vapor. Strong updrafts (Fig. 12b) containing sufficient water vapor may be conducive to the formation of precipitation in the Sichuan Basin in summer (as shown in Fig. 13b). This suggests a significant negative correlation between the precipitation and AOD over the Sichuan Basin, with a correlation coefficient of -0.65 (Fig. 13e). More precipitation is beneficial for the wet deposition of the aerosols over the Sichuan Basin. Moreover, the largest temperature gradient and strongest updrafts are present in summer (Fig. 12b), which favors the dispersion of aerosols. Therefore, the sulfate and carbonaceous aerosols can disperse well over the ESTP in summer, and thus the AODs are the smallest (Figs. 7d and f). In contrast, the maximum AOD value in winter is partially due to the less precipitation and poor dispersion conditions in the ESTP.

Figure 13.  Spatial distribution of total precipitation derived from MERRA-2 data in the (a) spring, (b) summer, (c) autumn and (d) winter from 1980 to 2018. Time series of monthly mean precipitation and AOD anomalies over the Sichuan Basin during the above period are given in (e). The units of total precipitation are mm. The black solid rectangles indicate the typical regions (Taklimakan Desert, Sichuan Basin and BTH). The dashed rectangle denotes the eastern slope of the TP.

Aerosols in the atmosphere over the ESTP are substantial and even more than those over the most densely populated and industrialized regions and deserts. Compositionally, aerosols over the ESTP are mainly sulfate, followed carbonaceous and dust aerosols. Sulfate and carbonaceous aerosol emissions are mainly produced by human activities and therefore exhibit weak seasonal variation. The carbonaceous aerosols are mainly from emissions by local fires in the vicinity of the Hengduan Mountains in spring, which can be further transported by prevailing southwesterly winds to the ESTP. In spring, a small amount of dust aerosols are transported from the Taklimakan and Gobi deserts to the ESTP.

Besides local emissions and transport, another reason for high AODs over the ESTP is the limitation of dispersion. The terrain-driven local circulation can constrain aerosols within the Sichuan Basin, which makes it difficult for pollutant dispersion. In the summer monsoon season, when abundant moisture meets the ascending air movement enhanced by the plateau heat pump effect, there is abundant rainfall. The aerosol particles can be removed by wet deposition. Additionally, the larger vertical temperature gradients and the strong updrafts in the summer are also conducive to aerosol dispersion, resulting in minimum AOD values. On the contrary, the maximum AODs in winter are mainly due to less precipitation and poor dispersion conditions over the ESTP.

Thus, the causes of aerosol pollution over the ESTP can be summarized as local emissions, transport from outside the region, and accumulation of pollutants under specific geographical conditions. Inward transport and terrain-driven circulation contribute greatly to the accumulation of aerosols over the ESTP, which poses greater challenges to environmental governance than that from local emissions. In particular, these aerosols are uplifted into the upper atmosphere by the perennial updrafts.