Characteristics of the summer atmospheric boundary layer height over the Tibetan Plateau and influential factors

Based on intensive sounding, surface sensible heat flux, solar radiation, and soil moisture observational datasets from the Third Tibetan Plateau Atmospheric Scientific Experiment and the routine meteorological operational sounding and total cloudiness datasets in the Tibetan Plateau (TP) for the period 2013-2015, we investigate the features of summer 10 atmospheric boundary layer (ABL) over the TP and its major influential factors. It is found that the convective boundary layer (CBL) and the neutral boundary layer (NBL) show remarkable diurnal variations over the TP, while the stable boundary layer (SBL) diurnal variation is weak. In the early morning, the ABL height distribution is narrow, with a small west-east difference. The SBL accounts for 85% of the TP ABL. At noon, there is a wide distribution in the ABL height up to 4000 m. The CBL accounts for 77% of the TP ABL, with more than 50% of the CBL height above 1900 m. The ABL 15 height exhibits a large west-east difference, with a mean height above 2000 m in the western TP and around 1500 m in the eastern TP. In the late afternoon, the CBL and SBL dominate the western and eastern TP, respectively, resulting in a larger west-east difference of 1054.2 m between the western and eastern TP. The high ABL height in a cold environment over the western TP (relative to the plain areas) is similar to that in some extreme hot and arid areas such as Dunhuang and Taklimakan Deserts. For the western (eastern) TP, there is low (high) total cloud coverage, with large (small) solar radiation 20 at the surface and dry (wet) soil. These features result in high (low) sensible heat flux and thus promotes (inhibits) the local ABL development.

2014; Pal et al., 2015). Therefore, the accurate specification of the ABL height is essential to develop weather, climate, and 30 air pollution prediction models.
The ABL can generally be divided into three types, that is, the convective boundary layer (CBL), the stable boundary layer (SBL), and the neutral boundary layer (NBL) (Stull, 1998). During the daytime, when sufficient solar radiation reaches the surface, the CBL usually dominates (Chen et al., 1997), with unstable thermal stratification and vigorous turbulence. At 35 night, infrared cooling at the surface results in the nocturnal SBL (Zhang et al., 2011a;Miao et al., 2015), with ground-based temperature inversion and weak turbulent motion. The NBL occurs mainly in high wind conditions, particularly combining with thick and extended cloud coverage, and also occurs during the sunrise and sunset transition periods, with turbulence of almost the same intensity in all directions (Blay-Carreras et al., 2014).

40
The ABL height can be calculated from temperature, humidity, and wind profiles (Seibert et al., 2000;Seidel et al., 2010;Davy, 2018). The CBL height is generally less than 2000-3000 m and the SBL thickness is not more than 400-500 m (Garratt, 1992). The ABL height shows an obvious spatial variation due to differences in topography, thermal properties of the underlying surface, and weather conditions. For example, the CBL can grow to as high as 4700 m in New Delhi before the outbreak of the South Asian monsoon, whereas it may only reach 900 m in Bangalore during the monsoon period 45 (Raman et al., 1990). Seidel et al. (2012) pointed out that a large east-west spatial gradient of the ABL height at sunset in the United States may be related to a difference in the solar altitude angle with respect to latitude. Guo et al. (2016) identified three large-scale spatial patterns in the ABL height in China, that is, a west-east gradient during sunrise, an east-west gradient during sunset, and a south-north gradient at noon. The reasons for the first two patterns are similar to those observed in the United States shown in Seidel et al. (2012), while the south-north gradient may be related to local surface and 50 hydrological processes Zhang et al., 2017).
The Tibetan Plateau (TP) with an average elevation exceeding 4500 m has complex land surface processes and boundary layer structures (Tao and Ding, 1981;Yanai and Li, 1994;Xu et al., 2002;Yang et al., 2004;Li et al., 2007;Sun et al., 2007;Zhao et al., 2018). The ABL height is generally higher compared to some plains areas (with the ABL height of 1000-1500 m) 55 (Zhao et al., 1992). Some studies have addressed that the ABL height in the TP can be as high as 2000-3000 m (Ye et al., 1979;Xu et al., 2002;Zhang et al., 2003). Song et al. (1984) examined the ABL height at Gaize station of the western TP is above 3000 m, while the ABL heights obtained by Li et al. (2000), Liu et al. (2001), and Lü et al. (2008) are lower in the central TP (between 400 m and 1800 m at Dangxiong station and 1750 m at Namucuo Lake). Moreover, there is a significant difference in the ABL height over the TP between dry and rainy seasons (Zuo et al., 2004). Li et al. (2011) found that the 60 ABL height at Naqu station is 2211-4430 m in the dry season and 1006-2212 m in the rainy season. Chen et al. (2013Chen et al. ( , 2016 observed the super-thick ABL at Gaize station during winter, with the ABL height above 5000 m. These results show that the ABL height over the TP varies greatly with position and season. https://doi.org/10.5194/acp-2020-787 Preprint. Discussion started: 24 September 2020 c Author(s) 2020. CC BY 4.0 License.
Although observations in the TP and studies on the local ABL features have made progress, routine meteorological 65 operational sounding observations are scarce over the western TP, owing to its high elevations, naturally harsh environmental conditions, and less-developed logistics. The previous studies on the ABL over the western TP are often based on a short-time experimental observational data at Gaize in one summer (Song et al., 1984;Chen et al., 2013). Thus the interpretation of their results has certain limitations. Moreover, there are significant differences in surface properties and climatic conditions between the eastern and western TP . However, few studies have examined this west-70 east difference in ABL due to the scarce data in the western TP. To obtain a longer observational data in the western TP, the Third Tibetan Plateau Atmospheric Scientific Experiment (TIPEX-III) beginning in 2013 has deployed routine sounding systems at Shiquanhe, Gaize, and Shenzha stations of the western TP ( Fig. 1), which fills in the data gaps in the operational sounding network over the western TP (Zhao et al., 2018). Compared to the previous field experiments over the TP, the TIPEX-III has a wider and longer coverage of sounding observations over the western TP, providing valuable observational 75 data for studying the ABL features.
This study utilizes the TIPEX-III sounding observational data to investigate the features of the ABL height over the TP, compares differences in the ABL height between the western and eastern parts of the TP, and analyzes the major factors affecting the ABL height. The remainder of this paper is organized as follows. Main features of data and methods are 80 described in Section 2. In Section 3, the characteristics of the ABL height over the eastern and western TP and their regional differences are analyzed in detail. Section 4 gives major factors affecting the ABL height over the TP and the west-east differences. Discussions and conclusions are given in Section 5.

Data and analysis methods
The TIPEX-III carried out the intensive routine meteorological sounding observations at Shiquanhe (SQH), Gaize (GZ), and 85 Shenza (SZ) stations of the western TP (marked by red dots in Fig. 1) since the 2013 summer (Zhao et al., 2018). These intensive observations have been applied in research on the vertical structure of the upper troposphere and lower stratosphere at Gaize station during the rainy season and effects of assimilating the intensive sounding data on downstream rainfall (Hong et al., 2016;Yu et al., 2018;Zhao et al., 2018;Zhao et al., 2019b). These intensive sounding data and the routine meteorological operational sounding data at 16 stations of the central-eastern TP from the China Meteorological 90 Administration (marked by black dots in Fig. 1) are utilized in this study. The sounding observations were carried out at 08:00 Beijing Time (BJT; 00:00 UTC), 14:00 BJT (06:00 UTC), and 20:00 BJT (12:00 UTC) each day for the above intensive and operational sounding stations, including vertical profiles of temperature, humidity, and wind direction and speed. After the quality of the sounding observational data, we finally select the periods from 15 June to 31 July 2013, from ( Fig. 1a) and 4757, 2049, and 4841 profiles separately at 08:00 BJT (Fig. 1b), 14:00 BJT (Fig. 1c), and 20:00 BJT (Fig. 1d) for 19 stations over the TP. Meanwhile, it is noted that there is a large difference in the sample number between the intensive and operational observation records at 08:00 BJT and 20:00 BJT ( Fig. 1b and d), which is called the original dataset for convenience. Thus we also select the operational observation records that correspond to the intensive observation records (called the test group dataset for convenience) at these two times and repeat the related analyses, obtaining the similar results 100 (shown in Section 3.2), which shows that the difference in the sample number between the intensive and operational observation records does not change our conclusions.
To analyze the factors affecting the ABL in the TP, we use the TIPEX-III 30-min mean surface sensible heat flux (SHF), downward solar radiation, and 5-cm soil volume moisture content at SQH (bare soil with few obstacles), Naqu (NQ; alpine 105 steppe), and Linzhi (LZ; alpine meadow with few shrubs and trees) stations in the 2014-2015 summers Zhao et al., 2018;Li et al., 2019Li et al., , 2020. In addition, the operational observation of total cloudiness (cloud cover) at 02:00 BJT, 08:00 BJT, 14:00 BJT, and 20:00 BJT from the China Meteorological Administration are also used.
We use the potential temperature gradient method proposed by Liu et al. (2010) to calculate the ABL height. Here σ is a critical value. Consistent with Liu et al. (2010) and Zhang et al. (2017), σ is set to 1.0 K. The ABL height is also calculated on the basis of their methods. For both the CBL and NBL, the ABL height is calculated as the height at which an air parcel rising adiabatically from the surface becomes neutrally buoyant (Stull 1988). The SBL height is defined as the lower of the heights of both the thermal stable layer from the surface and the maximum wind shear in the low-level jet stream. 120

A general characteristic of the ABL height
Diurnal variation is an important feature of the ABL, which consists of different periods of daytime, nighttime, and diurnal transitions (from day to night and from night to day) (Liu et al., 2010). In the central TP (near 90°E), 08:00 BJT, 14:00 BJT, and 20:00 BJT correspond to 06:00 (the early morning), 12:00 (noon), and 18:00 (the late afternoon) local standard time, respectively. To reveal a difference in the ABL height between the eastern TP (ETP) and the western TP (WTP), we divide all sounding stations over the TP into two groups. One is for the WTP (to the west of 92.5°E) with 8 stations, and the other is for the ETP (to the east of this longitude) with 11 stations. respectively. In the early morning (08:00 BJT) (Fig. 2a), the ABL height is generally low (<450 m) over the TP and shows a relatively homogeneous feature. At this moment, the distribution of the ABL height is narrow, with a frequency peak of 35% at the ABL height of 300 m (Fig. 2d) and 78.5% (99.6%) of the ABL height below 500 m (1000 m) (Fig. 2e). Figure 2f displays the zonal sections of the ABL height along 32°N. In this figure, the ABL height varies between 218.4 m and 433.9 m from east to west, not showing a remarkable west-east difference. 135 At noon (14:00 BJT) (Fig. 2b), the ABL height remarkably increases over the TP, with an average of 1887.7 m, and exhibits a large west-east difference. There is a wide distribution of the ABL height up to 4000 m, with a relatively flat peak between 900 m and 2900 m (Fig. 2d) and only 17.8% (more than 50%) of the ABL height below 1000 m (above 1900 m) (Fig. 2e).
The regional mean ABL height is 2124.2 m in the WTP and 1693.5 m in the ETP, with a mean difference of 430.7 m 140 between the WTP and the ETP. Along 32°N, the ABL height remarkably increases from 1379.4 m at GanZ station to 2504.2 m at SQH station, with the west-east difference exceeding 1200 m (Fig. 2f).
In the late afternoon (20:00 BJT) (Fig. 2c), the ABL height continues to increase in the WTP, with the regional mean height > 2000 m, while it begins to decrease in the ETP, with the regional mean height < 1000 m. This result indicates a larger west-145 east difference of 1054.2 m between the WTP and the ETP. Especially, the ABL height is 602 m at HY station and 2920.6 m at SQH, with a difference above 2000 m between these two stations ( Fig. 2f). At this moment, the frequency of the high ABL height decreases. The frequency peak is 12.8% at the ABL height of 300 m (Fig. 2d) and 50% of the ABL heights are less than 1000 m (Fig. 2e). It is evident that in daytime the ABL height exhibits an increasing trend from east to west in the TP. This west-east difference increases from noon to the late afternoon. The ABL height reaches the maximum in the late 150 afternoon.
This feature is likely related to a difference in the solar elevation angle with respect to longitude because night begins earlier 170 in the WTP than in the ETP, which supports an earlier transition from the daytime CBL to the nighttime SBL in the east (Seidel et al., 2012;Guo et al., 2016). However, the NBL shows a relatively weaker diurnal variation over the TP (Fig. 4d- with the mean occurrence frequency of 6.4%, 20.0%, and 25.5% at 08:00 BJT, 14:00 BJT, and 20:00 BJT, respectively. Figure 5 shows the ABL height-occurrence frequency relationship for the SBL, NBL, and CBL at 08:00 BJT, 14:00 BJT, 175 and 20:00 BJT. For the SBL, the frequency distribution of the ABL height shows the similar feature at three measurement times ( Fig. 5a-c) and is characterized by a narrow single mode, with the frequency peaks of 39.0%, 28.1%, and 36.6% at the boundary layer height of 200 m, 300 m, and 300 m at 08:00 BJT, 14:00 BJT, and 20:00 BJT, respectively. It is found that the SBL height above 80% is < 600 m and the cumulative frequency of the SBL height exceeding 1000 m is little (near zero) at these three times (Fig. 5d, e, and f). The ABL heights of the NBL and CBL show large variations. At 08:00 BJT (Fig. 5a), 180 the distributions of the NBL and CBL height are narrow, with the frequency peaks of 27.5% and 35.1% at the ABL height of 300 m for NBL and CBL, respectively, similar to that of the SBL, which is possibly due to the initial development of the CBL and NBL. At 14:00 BJT (Fig. 5b), there is a wide distribution of the ABL height up to 4000 m, with a relatively flat peak between 1000 m and 3000 m. This result is remarkably different from a single peak of the SBL. The frequency of the NBL height is generally less than 5% between 500 m and 3000 m (Fig. 5b), with a frequency peak of 6.1% at 1000 m, and 185 more than 50% NBL height exceeds 1700 m (Fig. 5e). The height of the CBL is higher, with a frequency peak near 4.5% between 1500 m and 2500 m (Fig. 5b). More than 50% CBL height is above 2000 m (Fig. 5e). This feature is also seen at 20:00 BJT ( Fig. 5c and 5f). The distributions of the NBL and CBL heights are still wide but the frequency of the high ABL height decreases, with the frequency peak below 500 m.
190 Because the CBL height shows a remarkable west-east difference in the TP at both 14:00 BJT and 20:00 BJT (shown in Fig.   3h-i), we analyze the cross section of the occurrence frequency of the CBL height along 32°N for these two times (Fig. 6a), in which the cross section includes SQH, GZ, SZ, NQ, CD, GanZ, and HY stations. In this figure, generally speaking, the high CBL tends to occur more frequently in the WTP than in the ETP. For the ETP, the high frequency of the CBL height mainly occurs below 1400 m, with the peak of 14.4% at the height of 350 m. The occurrence frequency of the CBL height 195 above 2000 m is low (<2%). For the WTP, there are two main peaks of the occurrence frequency. One strong peak (4%-10%) corresponds to the high CBL between 2500 m and 3500 m, especially at SQH station, and another weak peak appears between 200 m and 1000 m. The occurrence frequency is low for the CBL height between 1000 m and 2500 m. Figure 6b shows the cross section of the cumulative occurrence frequency of the CBL height along 32°N. As illustrated in Fig. 6b, the cumulative frequency contours gradually go upward from east to west. The eastern TP is dominated by a low CBL height,200 with the 50% CBL height below 1000 m, and there is only the 5% CBL height above 2500 m. For the western TP, there are larger CBL heights, with almost 50% CBL extending upward to more than 2500 m, almost 10% reaching 4000 m or higher, and only 15% CBL below 1000 m. The above features indicate a significant difference between the ETP and the WTP in the frequency distribution of the ABL height over the TP.

205
To investigate an effect of differences in the sample profiles shown in Fig. 1b and d, we use the test group dataset to repeat the above analyses. Figures 7a and b show the scatter plots of the occurrence frequency of the SBL, NBL, and CBL from the original and test group datasets at each of 19 stations at 08:00 BJT and 20:00 BJT, respectively. It is seen that the correlation coefficients between the two datasets are 0.92-0.99, with root-mean-square errors (RMSEs) between 1.1% and 2.7%. The similar results are also seen in the SBL, NBL, and CBL heights at 08:00 BJT (Fig. 7c) and 20:00 BJT (Fig. 7d). The 210 correlation coefficients in the ABL height are 0.90-0.99. The RMSE of the SBL height is 14 m and 25 m at 08:00 and 20:00 BJT, respectively. The RMSE of the CBL and NBL heights are 54-59 m at 08:00 BJT and 99-107m at 20:00 BJT. These high correlations and little errors show that the difference in the sample number does not change our conclusions.
From the foregoing analysis, the ABL height in the TP shows a remarkable diurnal variation and spatial difference, 215 especially for both NBL and CBL. From noon to the late afternoon, there are larger NBL and CBL heights in the WTP compared to the ETP, with the differences of 668.9 m and 955.0 m between the WTP and the ETP at 20:00 BJT, respectively.
Then, which factors contribute to these differences in ABL between the WTP and ETP? In the following section, we examine some factors that may be responsible for to the ABL height over the TP.

Factors responsible for the ABL height over the TP 220
A lot of studies have addressed effects of surface sensible heat flux (SHF), soil volume moisture content (VWC), downward solar radiation flux (DSR), and the total cloud cover (CLD) on ABL height (Liu, et al., 2004;Brooks and Rogers, 2006;Zhao et al., 2011;Sanchez-Mejia and Papuga, 2014;Rihani et al., 2015;Lin et al., 2016;Zhang et al., 2017; western TP, we utilize the TIPEX-III SHF, DSR, VWC and CLD observations at SQH, NQ, and LZ stations, analyzing the 225 relationships between the above variables and the ABL height. Figure 8a-c shows the scatter plots between the 6-hour mean SHF and the ABL height at SQH, NQ, and LZ stations. As shown in this figure, the correlation is 0.80, 081, and 0.71 (significant at the 99% confidence level) at the three stations, respectively. That is, when SHF is strong, the turbulent motion is strong and the ABL height develops, which supports the 230 result of Zhang et al. (2011b). Their result shows a significant correlation of 0.78 in the arid area of Northwest China between the ABL thickness and the cumulative SHF. Figure 9a and b further present the features of the ABL height and SHF at SQH, NQ, and LZ stations. The maximum value of SHF is 270 W/m 2 , 165 W/m 2 , and 100 W/m 2 at SQH, NQ and GZ stations, respectively, and has a large difference (105 W/m 2 ) between SQH and NQ stations. This result indicates a decreasing trend of SHF from west to east in the TP, consistent with a reduction of the ABL height from SQH via NQ to LZ 235 station (shown in Figs. 3 and 9a). In addition, Fig. 10 demonstrates the diurnal variations of SHF and the ABL height at SQH, NQ, and LZ stations. It is clear that the peak of the SHF diurnal variation occurs earlier compared to that of the ABL height at SQH station. The maximum ABL height occurs near 20:00 BJT (approximately 17:20 LST), corresponding to a stronger SHF. At LZ station, however, the ABL height at 20:00 BJT (18:20 LST) corresponds to a negative SHF. This difference in SHF between SQH and LZ stations is possibly associated with more cloud cover (reducing the solar radiance at the surface) 240 and one-hour time difference (an earlier transition to night at LZ station). Consequently, the difference in the ABL height between the WTP and ETP may be attributed to a west-east difference in SHF that is a direct thermal factor affecting the ABL development in the TP.
The solar radiation at the surface is an important component of the surface energy budget, affecting land surface temperature 245 and SHF. We show the scatter plots between the 6-hour mean DSR and the ABL height at SQH, NQ, and LZ stations ( Fig.   8d-f). The ABL height is highly correlated with the 6-hour average of DSR at these stations, with the correlation coefficients of 0.86, 081, and 0.73, respectively, which is equivalent to those of SHF. The mean DSR shows a decreasing trend from SQH (510 W/m 2 ) to LZ (200 W/m 2 ) station. Since the solar irradiance at the surface is negatively associated with the local cloud cover (Guo et al., 2011;Lin et al., 2016;Li et al., 2017;Zhang et al., 2017), the cloud cover is also correlated to the 250 ABL height. Figure 8g-i show that the 6-hour mean CLD has significant correlations of -0.56, -0.65, and -0.54 with the ABL height at SQH, NQ, and LZ stations, respectively. A decrease of the mean ABL height from SQH to LZ station (Fig. 9a) is corresponded to an increase of cloud cover (Fig. 9d) and a decrease of DSR (Fig. 9c). When cloud cover is between 0 and 20%, the mean ABL height for the NBL and CBL is 2019 m /2732 m in the ETP/WTP and when cloud cover is >80%, the ABL height decreases to 741 m /1626 m in the ETP/WTP (Fig. 11). Therefore, the increased cloud cover inhibits the 255 development of both the NBL and CBL. The difference in cloud cover between the WTP and ETP contributes to the westeast distribution of DSR and SHF, also finally contributing to the ABL development. Corresponding to more cloud cover in the ETP, the local ABL is more closely associated with atmospheric moisture processes. Soil moisture is also an important factor affecting SHF. Low soil moisture generally coincides with a high surface sensible 260 heat flux, which facilitates the ABL development (e.g., McCumber and Pielke, 1981;Sanchez-Mejia and Papuga, 2014;Rihani et al., 2015). Figure 8j-l shows that the relationship between the ABL height and the 6-hour mean VWC at SQH, NQ, and LZ stations. The ABL height at LZ station is negatively correlated to the local soil moisture, with a significant correlation coefficient of -0.45, which indicates that the ABL height is lower when surface soil is moister. However, the negative correlation is weaker at SQH station, with a correlation coefficient of -0.21. This difference between the WTP and 265 the ETP is associated with the climatic feature of the local soil moisture. The surface type transitions from alpine meadow with few shrubs and trees or alpine steppe in the ETP to bare soil with few obstacles in the WTP (Lin et al., 1981;. Accordingly, soil moisture decreases gradually from the ETP to the WTP (Fig. 9e), with a mean soil moisture below 0.10 m 3 /m 3 at SQH station and 0.38 m 3 /m 3 at LZ station. Little soil moisture in the WTP has a weak modulation to the local surface heat flux, which results in a weak correlation between the ABL height and soil moisture in the WTP. 270

Summary and discussion
Using the summer TIPEX-III intensive and meteorological operational observational datasets, we examine the ABL features and the relationships of the ABL height with surface sensible heat flux, solar radiation, cloud cover, and soil moisture in the TP region. The main conclusions are summarized as follows.

275
The ABL height exhibits diurnal variations and regional differences in the TP, which are weak for SBL and remarkable for both NBL and CBL. In the early morning, the ABL height is generally low over the TP, not showing a large west-east difference, and the distribution of the ABL height is narrow, with 78.5% of the ABL height < 500 m. At noon, the CBL and NBL heights remarkably increase and have a wide distribution in the ABL height up to 4000 m, with more than 50% of the ABL height exceeding 1900 m. The height exhibits a large west-east difference. At this moment, the distribution of the SBL 280 height is also narrow, with the cumulative frequency of 80% at the height of 500 m, and the west-east difference is not obvious. In the late afternoon, there are a narrow distribution of the SBL height and wide ones of the NBL and CBL heights.
The ABL height continues to increase in the WTP, while it begins to decrease in the ETP. This feature results in a larger west-east difference in the ABL height. In spite of a cold environment in the TP (relative to plain areas), the WTP still has the ABL height above 2000 m, which is similar to some extreme hot and arid areas such as Dunhuang and Taklimakan 285 Deserts. In the ETP, the ABL is similar to that in North China (1500-1900 m) and is generally larger compared to the East Asian summer monsoon region (＜1500 m) such as the Yangtze River Delta and the Pearl River Delta (Zhang et al., 2011;Guo et al., 2016;Zhang et al., 2017;Qiao et al., 2019). For the SBL and CBL, the occurrence frequency shows remarkable diurnal variations. Most (few) of the SBL (CBL) in the 290 TP occur in the early morning and its occurrence frequency rapidly decreases (increases) at noon, accounting for 3.6% (76.9%) of the ABL. Moreover, possibly due to a difference in the solar elevation angle with respect to longitude in the late afternoon, the SBL and CBL dominates the ETP and WTP, respectively. However, the NBL shows a relatively weak diurnal variation over the TP, with the mean occurrence frequency of 6.4% in the early morning and around 20% at noon and in the late afternoon. 295 The ABL height is significantly correlated to SHF, DSR, and cloud cover in the TP and to soil moisture in the ETP. The west-east reduction in the ABL height is closely associated with the decreasing trends in both SHF and DSR and the increasing trends in both cloud cover and soil moisture from west to east. The above factors affecting the WET and ETP ABL heights are summarized in Fig. 12. That is, in the WTP (ETP), with low (high) cloud cover, there is larger (smaller) 300 downward solar radiation at the surface. Meanwhile, corresponding to bare soil (alpine meadow or steppe) in the WTP (ETP), there is a dry (wet) soil condition. These features cause high (low) sensible heat flux, thus promoting (inhibiting) the local ABL development.
It is noted that the peak ABL height can drop from 3000-4000 m for a deep CBL in the WTP to 1000-2000 m for a shallow 305 CBL in the ETP. Such a steep inhomogeneous distribution of the ABL height may trigger local mesoscale circulation and precipitation (Segal et al., 1992;Goutorbe et al., 1997;Huang et al., 2009;Zhang et al., 2019;Qiao et al., 2019). Therefore, the influences of west-east differences in the ABL height over the TP on the local weather and climate should be further studied in the future.

310
Code and data availability. All data used are available from the TIPEX-III for providing the data available on its homepages (http://data.cma.cn/tipex).