Optical and Physical Characteristics of the Lowest Aerosol Layers over the Yellow River Basin

Abstract

Studying the presence of aerosols in different atmospheric layers helps researchers understand their impacts on climate change, air quality, and human health. Therefore, in the present study, the optical and physical properties of aerosol layers over the Yellow River Basin (YERB) were investigated using the CALIPSO Level 2 aerosol layer products from January 2007 to December 2014. The Yellow River Basin was divided into three sub-regions i.e., YERB₁ (the plain region downstream of the YERB), YERB₂ (the Loess Plateau region in the middle reaches of the YERB), and YERB₃ (the mountainous terrain in the upper reaches of the YERB). The results showed that the amount (number) of aerosol layers (N) was relatively large (>2 layers) in the lower part of the YERB (YERB₁), which was mainly caused by atmospheric convection. The height of the highest aerosol layer top (HT_H) and the height of the lowest aerosol layers base (HB₁) varied significantly with respect to the topography of the YERB. High and low values of aerosol optical depth (AOD) were observed over the YERB₁ (plain area) and YERB₃ (elevated area) regions, respectively. Population, economy, and agricultural activities might be the possible reasons for spatial variations in AOD. AOD values for the lowest aerosol layer were high—between 0.7 and 1.0 throughout the year—indicating that aerosols were mainly concentrated at the bottom layer of the atmosphere. In addition, the integrated volume depolarization ratio (0.15–0.2) and the integrated attenuated total color ratio (~0.1) were large during spring for the lowest aerosol layer due to the presence of dust aerosols. The thicknesses of the lowest aerosol layers (TL₁) did not vary with respect to the topographic features of the YERB. Over the sub-regions of the YERB, a significant positive correlation between the AOD of the lowest aerosol layer (AOD₁) and the thickness of the lowest aerosol layer (TL₁) was found, which indicates that TL₁ increases with the increase of AOD₁. In the whole YERB, a positive linear correlation between the N and HT_H was observed, whereas a negative correlation between N and the portion of AOD for the lowest aerosol layer (PAOD₁) was found, which revealed that the large value of N leads to the small value of PAOD₁. The results from the present study will be helpful to further investigate the aerosol behavior and their impacts on climate change, air quality, and human health over the YERB.

1. Introduction

Aerosols are colloids of fine solid particles or droplets in the air [1,2,3] and can come from natural sources, such as fog, dust, and sea salt, or anthropogenic sources, such as haze, black carbon, and smoke [4,5,6,7,8,9,10]. The aerodynamic diameters of these solid or liquid particles mostly range between 0.001 and 100 μm, while larger particles can precipitate, resulting in suspension, and mixing in the atmosphere [11]. The World Health Organization pointed out that particulate matters with aerodynamic diameters less than 10 μm (PM₁₀) can penetrate into the lungs to the deepest part of the bronchus (such as the bronchioles and alveoli), causing disease [12]. The previous study has pointed out a certain relationship between the aerosol extinction coefficient and lung cancer mortality [13]. Within the field of environmental science, the Intergovernmental Panel on Climate Change reported that atmospheric aerosols are responsible for one of the largest uncertainties in the estimation of climate forcing, and they can scatter and absorb solar shortwave and longwave radiation, directly affecting the Earth’s radiation balance [14]. They can also act as cloud condensation nuclei, changing the cloud’s physical characteristics, life cycle, and indirectly affecting the Earth’s climate system [15,16,17]. Therefore, the long-term, accurate, and regular monitoring of aerosol optical properties must be carried out continuously. In China, an increase in industrial activities and urbanizations was observed in recent years with the rapid economic development. Increasing numbers of aerosol particulates have been discharged into the atmosphere, leading to frequent air pollution incidents [18]. In recent years, severe regional haze pollution events have occurred in China’s Pearl River Delta, the Yangtze River Delta, the Beijing–Tianjin–Hebei city cluster (BTH), and the northwest region of China, and corresponding aerosol observations have been conducted [19,20,21,22,23,24]. Recently, China has experienced many severe air pollution events [9]; for example, the Yellow River Basin (YERB) has experienced serious local pollution emissions due to coal mining, urbanization, excessive coal consumption, and the development of heavy industry (such as iron, steel, and cement). However, scientific studies related to these events over the YERB are limited.

Ground-based sun photometers such as the Aerosol Robotic Network (AERONET) and Chinese Aerosol Remote Sensing Network (CARSNET) provide continuous observation of the physical and optical properties of aerosols for a long period [25,26,27,28]. However, the vertical distribution of aerosols cannot be obtained using site measurements (sunphotometer, nephelometer, aethalometer, and aerosol spectrometer, etc.). These instruments provide point-based aerosol information; however, information about aerosol vertical distribution is very important for understanding the quantitative characteristics of aerosols. In contrast, aerosol characteristics cannot be obtained at a large scale using in situ measurements, as these measurements are limited to the point locations. For these reasons, Light Detection and Ranging (LiDAR) has gradually become the preferred remote sensing technology for atmospheric aerosol observations [29,30]. LiDAR is an active remote sensing detector, which emits laser beams and uses the echo signal by the scattering effect between the laser beam and atmospheric aerosol particles to perform quantitative retrieval [31,32,33]. Therefore, LiDAR operates continuously during the night and day and provides the vertical distributions of aerosol particles. In addition, LiDAR has a high spatial resolution (covering the range from the Earth’s surface to the stratosphere with meter resolution) and temporal resolution (repeating frequency can reach up to kilohertz). In addition, it has a high detection sensitivity, and the signal received range can reach more than 10 orders of magnitude [34,35]. NASA launched the Cloud-Aerosol LiDAR and Infrared Pathfinder Satellite Observation (CALIPSO), the first active remote sensing sensor in 2006. It acquires not only aerosol information over a large regional scale but also the vertical distribution of different layers of atmospheric aerosols, which provide a new outlook on aerosol observations [36,37]. Therefore, this study intends to use the CALIPSO satellite to investigate the optical and physical characteristics of the lowest aerosol layers over the YERB. As the lowest aerosol layers are close to the ground surface and strongly related to human activities, it is, therefore, necessary to understand the physical and optical characteristics of the lowest layers of aerosols. However, only a few studies have been conducted over the region using optical remote sensing data [38,39], and no one has focused on the vertical distributions of the physical and optical properties of aerosol layers over the YERB. In addition, these have not characterized aerosols into different layers. To the best of our knowledge, this is the first study that has investigated different aerosol optical properties and their vertical distributions. Therefore, the objective of this study is to provide an understanding of the characteristics of the lowest aerosol layers over the YERB using the CALIPSO level 2 aerosol layer product from 2007 to 2014. The paper is organized as: the study area is discussed in Section 2, the materials and methods are described in Section 3, the results are discussed in Section 4, and the conclusions are provided in Section 5.

2. Study Area

Figure 1 shows a topographic map of the YERB region, which is located at 30° N–45° N, 95° E–120° E, including the plains in the downstream reaches of the YERB (YERB₁: 31° N–42° N, 112° E–119° E), the Loess plateau region in the middle reaches of the YERB (YERB₂: 32° N–43° N, 105° E–112° E), and the mountainous terrain in the upstream reaches of the YERB (YERB₃: 30° N–39° N, 94° E–105° E). The YERB is the second-longest river in China, which originates from the northern foot of the Bayan Har Mountains on the Tibet Plateau, and flows through nine provinces including Qinghai, Sichuan, Gansu, Ningxia, Inner Mongolia, Shanxi, Shaanxi, Henan, and Shandong. The basin area is 7.95 million km² and is about 1900 km long and 1100 km wide. The YERB is a vast territory and contains many mountains, different landforms, and a complicated geographical environment; there is also a great disparity between the east and west [40]. In addition, the YERB is located at middle latitudes, and the effects of atmospheric circulation and monsoon circulation on it are complicated. Therefore, the climate of different regions in the YERB differs significantly, and the annual and seasonal changes of climate elements are large [41]. From west to east, the YERB includes the eastern Tibet Plateau (upstream), Loess Plateau (middle), and North China Plain (downstream), and the east-west altitude difference is 4480 m. Therefore, the study of the optical and physical characteristics of aerosols in different topographic environments of the YERB can help to understand the specific characteristics of aerosols and can provide evidence for the study of aerosol radiation forcing, which is of great scientific significance.

3. Materials and Methods

CALIPSO is a small earth-probing satellite project jointly carried out by NASA’s Langley Research Center (LaRC) and the National Space Research Center of France in 2002 [42]. CALIPSO was launched on 28 April 2006 and can provide not only aerosol and cloud characteristics that vary with latitude and longitude, but also information about the vertical distribution of aerosols and clouds. CALIPSO can identify aerosol types and cloud water/ice phase states and can obtain vital data such as radiation flux and atmospheric state. This data can be used to assess the impact of aerosols and clouds on radiation, and thus improve the ability to predict climate change [43,44]. The orbital altitude of the CALIPSO satellite is 705 km and the dip angle is 98.2°, which has the advantage of a wide field of view. CALIOP is the most important piece of detection equipment on the CALIPSO satellite, including the laser transmitting and receiving system, which is located on a T-type optical platform, thus ensuring the stability of the system. The CALIOP telescope receives backscattered signals divided into three channels. One channel measures the backscattering intensity of 1064 nm, while the other two channels measure the orthogonal polarization backscattering signal of 532 nm. CALIOP can obtain the backscattering coefficients of two wavelengths (532 and 1064 nm) during the day and night, and can also provide vertical profiles of the 532-nm volume depolarization ratio, which is the ratio of the 532-nm vertical backscattering intensity to the 532-nm parallel backscattering intensity. It reflects the irregularity in the measured particles. CALIOP can also provide the color ratio, which is the ratio of the 1064-nm backscattering intensity to the 532-nm total backscattering intensity. The size of the particles can be identified by this color ratio [45,46]. The CALIPSO dataset is processed, stored in a hierarchical data format, and released by the Atmospheric Sciences Data Center of Langley.

CALIPSO datasets are mainly composed of two levels, i.e., Level 1 and Level 2 data products [44], whereas, in this study, Level 2 aerosol layer products at 532 nm were used. The CALIPSO aerosol Level 2 products include aerosol layers, aerosol classifications, and aerosol profile products. The horizontal resolution of the aerosol layer products is 5 km, and the revisit time is 16 days. The aerosol layer products are generated based on the raw CALIPSO profile data using the selective iterated boundary locator (SIBYL) algorithm. Using SIBYL, the feature layers are detected in each raw profile data, and then the amount (number) of aerosol feature layers (N), and the height of the feature layer’s base (HB_N) and top (HT_N) is obtained. The data contain at most eight vertical layers (N ≤ 8) in the whole atmosphere in every raw profile data. During the next time, the Scene Classification Algorithm (SCA) is applied on each vertical layer to classify into six different species, including clean marine aerosol, dust aerosol, polluted continent aerosol, clean continent aerosol, polluted dust aerosol, and smoke aerosol. Different species in different layers are endowed with different LiDAR ratios and retrieved to extinction coefficient $\mathsf{\delta }\left(\mathrm{z}\right)$ and backscatter coefficient $\mathsf{\beta }\left(\mathrm{z}\right)$ profiles based on Hybrid Extinction Retrieval Algorithms (HERAs). Each raw profile dataset has different aerosol feature layers, and each feature layer has an extinction coefficient profile. Hence, the aerosol optical depth (AOD) of the Nth aerosol layer (AOD_N) is calculated as the integral of the extinction coefficient in different layers (Equation (1)). Then, the integrated volume depolarization ratio of the Nth aerosol layer (DR_N) and the integrated attenuated total color ratio of the Nth aerosol layer (CR_N) can also be calculated using Equations (2) and (3), respectively. The uncertainty analysis of AOD_N, DR_N, and CR_N are also performed based on a relative error, and the results are stored in the data products with respect to different quality flags. In this study, only the highest quality (uncertainty ≤ 3) datasets were considered and resampled at 1° × 1° spatial resolution.

This study aimed to explore the vertical physical and optical characteristics of the aerosol layers, and the variables used include the amount (number) of aerosol feature layers (N), the height of the highest aerosol layer top (HT_H), the height of the highest aerosol layer base (HB_H), the AOD of the Nth aerosol layer (AOD_N), the sum of the AOD from all the aerosol layers (AODs, Equation (4)), the AOD proportion of the nth aerosol layer (PAOD_N, Equation (5)), the integrated volume depolarization ratio of the Nth aerosol layer (DR_N), and the integrated attenuated total color ratio of the Nth aerosol layer (CR_N). The details of every variable can be found in Abbreviations.

$${\mathrm{AOD}}_{\mathrm{N}}=\underset{{\mathrm{HT}}_{\mathrm{N}}}{\overset{{\mathrm{HB}}_{\mathrm{N}}}{{\displaystyle \int }}}\mathsf{\delta }\left(\mathrm{z}\right)\mathrm{dz};\mathrm{N}=1,2,\dots ,7,8$$

(1)

$${\mathrm{DR}}_{\mathrm{N}}=\frac{{{\displaystyle \sum }}_{{\mathrm{HT}}_{\mathrm{N}}}^{{\mathrm{HB}}_{\mathrm{N}}}{\mathsf{\beta }}_{532,\perp }\left(\mathrm{z}\right)}{{{\displaystyle \sum }}_{{\mathrm{HT}}_{\mathrm{N}}}^{{\mathrm{HB}}_{\mathrm{N}}}{\mathsf{\beta }}_{532,\Vert }\left(\mathrm{z}\right)}; \mathrm{N}=1,2,\dots ,7,8$$

(2)

$${\mathrm{CR}}_{\mathrm{N}}=\frac{{{\displaystyle \sum }}_{{\mathrm{HT}}_{\mathrm{N}}}^{{\mathrm{HB}}_{\mathrm{N}}}{\mathsf{\beta }}_{1064}\left(\mathrm{z}\right)}{{{\displaystyle \sum }}_{{\mathrm{HT}}_{\mathrm{N}}}^{{\mathrm{HB}}_{\mathrm{N}}}{\mathsf{\beta }}_{532}\left(\mathrm{z}\right)}; \mathrm{N}=1,2,\dots ,7,8$$

(3)

$${\mathrm{AOD}}_{\mathrm{s}}={{\displaystyle \sum }}_1^{\mathrm{N}}{\mathrm{AOD}}_{\mathrm{N}}; \mathrm{N}=1,2,\dots ,7,8$$

(4)

$${\mathrm{PAOD}}_{\mathrm{N}}=\frac{{\mathrm{AOD}}_{\mathrm{N}}}{{\mathrm{AOD}}_{\mathrm{s}}}; \mathrm{N}=1,2,\dots ,7,8$$

(5)

In this study, the AOD of the lowest aerosol layer (AOD₁) and the AOD proportion of the lowest aerosol layer (PAOD₁) were mainly analyzed in the next section. The mean annual and seasonal values of each variable over the different sub-regions of the YERB were calculated to investigate the interannual and seasonal variations and characteristics of the aerosol layers.

4. Results and Discussions

4.1. Interannual Variations and Characteristics of Aerosol Layers over the YERB

Figure 2a shows that the annual mean values of AOD_S were high over the YERB₁ (0.4–0.6) compared to the YERB₂ (0.2–0.3) and YERB₃ (0.1–0.2) regions. This might be due to a large urban population, relatively developed economy, and large aerosol emissions, as YERB₁ is located in the North China Plain (NCP) [47] and close to urban agglomerations of the BTH region, which is one of the most polluted regions with high aerosol contents in China [48,49]. These results are consistent with the previous study, which has reported a mean AOD mean value of 0.5 over the Beijing and XiangHe AERONET sites [50]. YERB₃ is located in the upstream region of the YERB, including the eastern Tibet Plateau, which is sparsely populated; therefore, low aerosol contents were expected [51,52,53]. The previous study of aerosol layers over the Tibet Plateau has reported mean AOD values less than 0.2 due to low aerosol loadings over the region [54]. Annual mean AOD values were between 0.2 and 0.3 over the YERB₂, which were close to those observed at the Semi-Arid Climate and Environment Observatory of Lanzhou University (SACOL) station. At SACOL, AOD observations from a sky radiometer were between the 0.2 and 0.4 range at SACOL station during spring [55], and annual mean AOD was observed between 0.3 and 0.4 at SACOL [56]. AOD values at SACOL represented the single point location, whereas the AOD over YERB₂ was the average of a specific region, which might be the reason for differences between annual mean AOD values at SCAOL and YERB₂. Furthermore, the geographical environment of SACOL is a semi-arid climate with high-frequency sandstorms. Thus, the AOD values are slightly higher in SACOL compared to the values observed over the YERB₂ [57].

The values of N exhibit strong regional characteristics (Figure 2b). The highest values of N (2.1 to 2.4) were observed over the YERB₁ compared to the YERB₂ (1.7–2.0) and YERB₃ (1.4–1.6). This might have happened as YERB₁ is located in the monsoon climate region; the air temperature is relatively high and the atmosphere aerosol loadings are relatively large, which might lead to aerosol vertical stratifications in the atmosphere [58,59]. However, YERB₃ is located at the eastern edge of the Tibet Plateau, and the underlying surfaces are mainly composed of mountains and plateaus, where aerosol loadings are relatively low, and the air temperature also remains low [60]. Thus, this scenario might have very little chance for aerosol vertical stratifications in the atmosphere. Therefore, low values of N were observed over the YERB₃.

The HT_H vary significantly with respect to the topography of the three sub-regions; i.e., YERB₃ has relatively high altitudes because of the plateaus and mountains (Figure 2c). Therefore, the annual mean values of HT_H were between 6.0 and 7.5 km over the YERB₃ followed by YERB₂ (4.5–5.5 km) and YERB₁ (3.5–5 km).

Similar to the integrated AOD, the highest values of AOD₁ (0.3–0.4) were observed over YERB₁ compared to the YERB₂ (0.1–0.2) and YERB₃ (~0.1) (Figure 2d). Figure 2e shows that the PAOD₁ values were very large (0.7 to 0.85) over all the sub-regions, which indicates that atmospheric aerosols were mainly concentrated in the bottom atmospheric layer. The PAOD₁ values over YERB₃ (0.8–0.85) were large compared to YERB₂ (0.72–0.77) and YERB₁ (0.72–0.75). This phenomenon might be due to the high altitudes and low temperatures in the YERB₃ region, or due to the weak effects of aerosol vertical convection and aerosol stratification [53,54,60]. Figure 2f shows that the values of HB₁ significantly change with respect to the topography, as the highest values of HB₁ (4–5 km) were observed over the YERB₃ region compared to the YERB₂ (2–3 km) and YERB₁ (1–2 km) regions. As shown in Figure 2g, the values of TL₁ were slightly higher over YERB₃ (1.2–1.4 km) than those observed over the YERB₂ (1.15–1.2 km) and YERB₁ (1.14–1.2 km) regions. This might be due to the vertical convection effects and stratifications of the aerosol layer, which were stronger over YERB₁ and YERB₂ than YERB₃ [56,57,58,59].

The aerosol layer-integrated volume depolarization ratio can reflect the non-spherical characteristics of aerosol particles. Higher values of DR₁ (0.095–0.12) were observed over YERB₃ and YERB₂, indicating the non-spherical characteristics of the lowest aerosol layers. This might be because of dust aerosols, which occur more frequently in the eastern edge of the Tibet Plateau (YERB₃) and the Loess plateau (YERB₂), leading to more priority being given to those aerosol types regarding the dust over those regions [53,54,55,56,57]. In addition, the shapes of the dust particles are non-spherical; therefore, DR₁ values were large over these regions. These results were consistent with the previous study [61]. Further, small values of DR₁ (0.07–0.09) were small over the YERB₁, indicating spherical characteristics of the lowest aerosol layers. The aerosol layer-integrated attenuated total color ratio can reflect the size of the aerosol layer particles. Figure 2i shows the interannual variations in the integrated attenuated total color ratio of the lowest aerosol layers (CR₁) over the YERB. The CR₁ values were lower (0.45–0.6) over YERB₃ than those observed over YERB₂ (0.55–0.7) and YERB₁ (0.6–0.7). These results suggested that the lowest aerosol layer contains aerosol particles with a small size over YERB₃ compared to YERB₂ and YERB₁. This might be because the aerosol loadings of the eastern main body of the Tibet Plateau (YERB₃) are low all year round [53,54].

4.2. Seasonal Variations and Characteristics of Aerosol Layers over the YERB

Seasonal variations were analyzed in AOD_S over the three sub-regions of the YERB (Figure 3a and Figure 4a,b). During each season, the seasonal mean AOD_S values were high over the YERB₁ (0.4–0.6) region compared to the YERB₃ (0.12–0.2) region. This might be due to the regional effects, as the YERB₁ at the North China Plain (NCP) is close to the BTH, which is a region of high aerosol loadings [24,47,48,49]. Therefore, the seasonal mean values of AOD_S were high over YERB₁. However, YERB₃ is located at the eastern Tibet Plateau, which is a region of low aerosol emissions that led to low mean AOD_S values [52,54,60]. Overall, the values of AOD_S were higher in summer (~0.6) (Figure 4b) followed by autumn (~0.5) (Figure 4c), winter (~0.4) (Figure 4d), and spring (~0.4) (Figure 4a) over the YERB₁ region. This might be due to the large relative humidity in summer [9,39,62]. Figure 3b and Figure 4e–h show the seasonal variations in N over the three sub-regions of the YERB, and the seasonal mean values of N were greater than or equal to 2 over YERB₁ (2.0–3.0) compared to YERB₂ (1.7–2.0) and YERB₃ (1.4–1.7). YERB₁ is located in a region with a temperate monsoon climate, and the atmosphere possesses a stronger vertical movement than that over YERB₃ and YERB₂, which leads to a clear vertical stratification of the atmosphere and increases the amount of aerosol stratification [50,58,59,63,64]. However, YERB₃ is located at the eastern edge of the Tibet Plateau, which is mainly composed of mountains and plateaus with low air temperature and weak atmospheric vertical convection [51,65]. In addition, the values of N were larger in spring (Figure 4e) followed by summer (Figure 4f), autumn (Figure 4g), and winter (Figure 4h). Especially over the YERB₁ region, N was significantly greater (~2.5) in spring compared to the other regions. This might be due to high temperature and strong vertical convection, which led to a significant stratification of aerosols, resulting in greater values of N over YERB₁ [24,48,49,58]. Similarly, the greater values of N were observed over YERB₂ compared to YERB₃, which might be due to the frequently occurred sandstorms over the Loess Plateaus and a significant effect of atmospheric convection [55,57].

The seasonal mean values of HT_H (6–7.5 km) over the YERB₃ during each season were greater than those values observed over the YERB₂ (4–5 km) and YERB₁ (3–5 km) due to the high altitudes of the plateaus (Figure 3c and Figure 4i–l) [47,53,54,57]. At the seasonal scale, large values of HT_H were observed in spring and summer, whereas small values were observed in autumn and winter. This might be due to the frequently occurred straw burning phenomena, which led to large aerosol emissions and increased the HT_H [9,63,66].

Similar to the integrated AOD_S, high seasonal mean AOD₁ values were found over the YERB₁ compared to the other sub-regions (Figure 3d and Figure 4m–p). At the seasonal scale, seasonal mean values of AOD₁ were higher in summer (~0.4) followed by autumn (~0.38), winter (~0.35), and spring (~0.28) over the YERB₁. This might be due to the grain harvest season and straw-burning activities, which frequently occur in summer and autumn [24,47]. The seasonal mean values of PAOD₁ were large over the YERB₁ region compared to the other sub-regions (Figure 3e and Figure 5a–d). These results indicate that atmospheric aerosols were mainly concentrated in the lowest aerosol layer over the sub-regions. It was noticed that the PAOD₁ slightly decreased over the YERB₂ and YERB₁ regions in spring and summer, which might be due to the large values of N, as discussed above. However, large values of PAOD₁ (0.79–0.83) were observed in autumn and winter over all the sub-regions, which might be due to low temperature and weak vertical convection and aerosol stratification [51,57,58].

High seasonal mean values of HB₁ were observed over the YERB₃ (4–4.6 km) region compared to the YERB₂ (2–2.4 km) and YERB₁ (1–1.5 km) regions (Figure 3f and Figure 5e–h). At the seasonal scale, high values of HB₁ were observed in spring and summer, whereas low values were observed in autumn and winter. In spring and summer, air temperature is relatively high compared to autumn and winter with strong vertical convectional effects, which increased the height of the lowest aerosol layer base [58,66]. No significant variations in TL₁ were observed with respect to topographical variations (Figure 3j and Figure 5i–l). Seasonal variations showed large values of TL₁ in spring (1.35 to 1.5 km) followed by summer (1.2–1.4 km), and autumn and winter (0.9–1.2 km).

Seasonal variations showed high values of DR₁ in spring (0.12–0.13) over the YERB₂ and YERB₃ regions, which indicate the non-spherical characteristics of the lowest aerosol layers. However, lower values were observed in summer (0.06–0.08), autumn (0.07–0.1), and winter (0.09–0.12), which indicates spherical shapes of the aerosols in the lowest aerosol layers. Seasonal variations of CR₁ showed high values in spring (0.7) over the YERB₁ and YERB₂, which indicate the large size of aerosol particles in the lowest layers. However, relative small values were observed for other seasons, especially over the YERB₃ (0.5–0.6).

4.3. The Correlations of the Lowest Aerosol Properties over the YERB

In order to better understand the seasonal spatial characteristics of aerosol layers over the YERB, correlation analyses of aerosol layer parameters are worth studying at seasonal and regional scales, which can help to better understand the causes of spatiotemporal characteristics of aerosol layers over the YERB. Correlation analyses were performed between AOD₁ and TL₁ over each sub-region (Figure 6), which showed the direct relationship, i.e., the thicker the TL₁, the higher the AOD₁. This was consistent with theoretical understanding, as AOD is the integral of extinction coefficients with respect to altitude, so the thicker the TL₁, the greater the integral distances, and the higher the values of AOD₁. According to previously discussed results, AOD₁ was high over the YERB₁ region; therefore, a higher correlation was expected over the YERB₁. Figure 6 showed a higher correlation (R²) of 0.59 in spring, 0.58 in summer, 0.78 in autumn, and 0.76 in winter over the YERB₁ sub-region compared to the YERB₂ and YERB₃ regions. These results were consistent and supported by the previous study over the Tibet Plateau [54].

Figure 7 showed a significant positive linear correlation (R² ≥ 0.90) between N and HT_H over all the sub-regions of the YERB. These results were expected because of significantly vertically distributed aerosol layers in the whole atmosphere, which led to the higher values of HT_H.

Similarly, a significant negative correlation was observed between N and PAOD₁ over the YERB region (Figure 8). This was expected because, for a large amount (number) of aerosol feature layers (N), small values of PAOD₁ were observed, which led to a significant negative correlation.

Figure 9 shows the probability distribution scatter diagram between DR₁ and CR₁ over the YERB, and the color bar represents the probability values. It can be noticed that the data points were mainly concentrated in the lower-left corner over each region during each season. The range of DR₁ and CR₁ was between 0 and 0.2, and 0 and 1, respectively. These results indicate smaller particle sizes with non-spherical shapes in the lowest aerosol layers over the whole YERB. However, the CR₁ values over all the sub-regions were generally higher in spring (1–1.5) compared to the other seasons. This might be due to the presence of dust aerosols with large particle sizes over the YERB. It is well-known that dust layers generally exist at high altitudes with large particle sizes with non-spherical characteristics.

5. Conclusions

In this study, the optical and physical properties of the aerosol layers over the YERB were investigated using the CALIPSO Level 2 layer products from January 2007 to December 2014. This paper can be a reference to further investigation of the physical and optical properties of aerosols over the YERB or over the other regions. The main outcomes are:

(a)

AOD_S values and relative large values of aerosol amount (N > 2) were observed over the YERB1 region, as YERB₁ is the most polluted region, and the atmosphere possesses strong vertical movement, which can lead to a clear vertical stratification.

(b)

High values of the HT_H were observed over the YERB. In addition, the HT_H and the HB₁ significantly vary with respect to the topography and seasons as high values observed in spring and summer compared to autumn and winter. However, the TL₁ did not vary with topographical features, and low values of the TL₁ values (approximately 1–1.5 km) were observed over the YERB.

(c)

PAOD₁ was quite large (0.7–1.0) over the YERB, which suggested that atmospheric aerosols were primarily concentrated in the bottom layer.

(d)

The values of DR₁ (0.15–0.2) and CR₁ (~1) were large over the YERB in spring because of dust aerosols.

(e)

A positive correlation between AOD₁ and TL₁ and A and HT_H was observed over the whole YERB, whereas N and PAOD₁ showed a significant log-negative correlation.

Overall, this study provides a better understanding of air pollution in the YERB region. In the future study, aerosol types and their microphysical properties over the region will be considered. Also, the vertical feature mask (VFM) products of CALIPSO will be considered to investigate the vertical structure and layer stratification characteristics of different aerosol species.