Abstract:By using the basic 30 m resolution elevation data in Ya'an, Sichuan Province, the aspect and gradient parameters of the grids were extracted. The main 50 times of precipitation collected from the 307 automatic weather stations in Ya'an during the flood season of 2017 (June-September) were divided into 16 times of large-scale precipitation and 34 times of middle-and small-scale precipitation. At the same time, by using ECMWF 0.25°×0.25° resolution reanalysis wind field data, the composite average wind field is synthesized according to the height of different sites, and the dynamic uplifting is calculated by the aspect and gradient values of each site. The daily sunshine hours and astronomical solar radiation data were also used to calculate the thermal uplift function of the terrain. The multiple linear regression was applied to the precipitation distribution. According to the standard coefficient values of regression, the influence of each variable on the precipitation distribution was determined. The following conclusions are drawn. (1) In the middle-and small-scale precipitation, the thermal uplift has the most important effect on the precipitation distribution, followed by the altitude, and then by the dynamic uplift of the terrain. (2) On the contrary, in the large-scale precipitation, the dynamic uplift of the terrain has the greatest influence on the precipitation distribution, followed by the altitude, and then by the thermal uplift. (3) The maximum daily precipitation values mostly appeared at the stations with 1000 m high, which corresponds well to the lifting condensation level. (4) From the long-term statistics, the dynamic uplift of the terrain and the vegetation conditions on the surface have the most significant influence on the precipitation distribution. According to different precipitation scales, attention should be paid to different dynamic and thermodynamic functions to predict the location of areas with frequent precipitation.