**2. Data**

### *2.1. Hydro-Estimator Satellite Rainfall Estimates*

The rainfall rate data from Hydro-Estimator (HE) Satellite Rainfall Estimates covering June, July, and August 2014–2020 were used in this study. The HE rainfall rate estimates were produced using the data from NOAA's Geostationary Operational Environmental Satellites and also using available geostationary data over Europe, Africa, and Asia. The datasets were downloaded from ftp://ftp.star.nesdis.noaa.gov/pub/smcd/ emb/f\_f/hydroest/world/world/archive/ (accessed on 1 January 2022). The horizontal resolution of HE dataset is ~4 km, and the temporal resolution is 1 h. The HE dataset has been widely used in various studies [21–23]. HE provides the most stable result for several kinds of algorithms that combine infrared and passive microwave data over mountainous regions [32–34].

### *2.2. Ground Observational Data*

The data from 50 ground meteorological observation sites belonging to the China Meteorological Administration were chosen, covering June, July, and August 2014–2020. Among them, 36 were located in the TP region and 14 in the SB region. The data included precipitation, air temperature at 2 m, humidity, and surface pressure, with a temporal resolution of 1 h, that were employed. Detailed station information is provided in Tables S1 and S2 in the Supporting Information. Additionally, the data from 12 radiosondes, including temperature, wind, pressure, and lifting condensation level (LCL) were chosen. The remaining observations were from TIPEX3 and a research program entitled 'The interaction between the earth and atmosphere of the TP and its influence on the weather and climate in the downstream' [4,7] (Figure 1). Cloud radar data, including CBH, cloud cover, and LWC (temporal resolution: 1 min), at the sites of Naqu (30.46◦N, 90.59◦E; 4730 m above mean sea level (MSL)), Yushu (33.01◦N, 96.56◦E; 3689 m MSL), and Linzhi (29.46◦N, 94.44◦E; 3326 m MSL), from July and August 2014–2020, were also used. The LWC was retrieved by using the cloud radar data with the equation LWC = 3z0.5, where LWC is a power relationship with reflectivity z. For details about the cloud radar instrument and the cloud radar LWC retrievals, see [36,37]. The locations of the observation sites are shown in Figure 1.

**Figure 1.** (**a**) Topography of the Tibetan Plateau and Sichuan Basin, in which the black dots indicate the locations of the ground meteorological observation sites of the China Meteorological Administration; '×' represents the radiosonde stations; and pentagrams represent the cloud radar observation sites. (**b**) Elevation goes along the line of 31.2◦N marked with the dashed blue line in (**a**).

### *2.3. ERA5 Reanalysis Dataset*

We obtained hourly estimates of u-wind, v-wind, temperature, precipitation, CBH, cloud cover, cloud IWC, cloud LWC, dewpoint spread, zero degree level, and CAPE from the ERA5 reanalysis dataset. ERA5 combines vast amounts of historical observations into global estimates using advanced modeling and data assimilation systems. ERA5 has been widely used in various studies [38]. The ERA5 precipitation rate is greater than that in the observations when the precipitation rate is less than 10 mm day−<sup>1</sup> over the TP [39]. The data were downloaded from https://www.ecmwf.int/en/forecasts/datasets/reanalysisdatasets/era5 (accessed on 1 January 2022) and spanned the period 2014–2020 for the summer months of June, July, and August. The horizontal resolution of ERA5 is ~31 km.

### **3. Methods**

The summer precipitation in China can be divided into two stages: the mei-yu period and the midsummer period. The 'mei-yu' rain, also called plum rain or the East Asian rainy season, is caused by precipitation along a persistent stationary front known as the

Meiyu front for nearly two months during the late spring and early summer in East Asia. These weather systems can produce heavy rainfall and flooding. The typical meiyu period is generally at the beginning of mid or late June and at the end of early or middle July [12–16]. Therefore, two time periods were chosen: (1) the mei-yu period, during 1–25 June in the monsoon phase and (2) the midsummer period, during 1 July–10 August. In terms of the study domain, the region of the eastern TP and its downstream area (28◦–34◦N, 90◦–110◦E) were chosen (Figure 1) [15], which were then further separated into two subregions with different elevations (framed areas in Figure 1): 90◦–100◦E and 100◦–110◦E, which represented the TP and SB. To describe the diurnal cycle during summer, the ERA5 and observational data were grouped into the mei-yu and midsummer seasons. The precipitation feature (PF) number is defined as the number of hours with precipitation in the diurnal cycle during the observational period when the precipitation observed was larger than 0.02 mm h−1. The numbers of the PF in the two subregions during the two different seasons are listed in Table 1.

**Table 1.** The number of precipitation features (Unit: hour) from the Hydro-Estimator Satellite Rainfall Estimates dataset in the two subregions within 28◦–34◦N shown in Figure 1 in each of the two chosen seasons.


The ground and cloud radar observations were temporally averaged to 1 h. Local time (LT) was defined as Coordinated Universal Time (UTC) + 7 h. To match the ground observations with the satellite and ERA5 data in spatial terms, four grid values near the ground observations including precipitation and cloud parameters at different levels from the ERA5 dataset and satellite were interpolated using the bilinear interpolation method to produce the value at the ground observation site [40,41]. We use Hovmöller diagrams [23] to show the diurnal cycle of CBH, precipitation, dewpoint spread, IWC, and LWC and their changes with latitude. Typically, longitude is plotted along the x-axis, and time is recorded on the ordinate; then, the contour values of a named physical field are presented through color or shading. In addition, the height of the LCL can be calculated as *Z*lcl = 123(*T* − *T*d), where *T* is the air temperature at 2 m, and *T*d is the dewpoint temperature, in which the LCL is determined by the dewpoint spread.
