INFLUENCING FACTORS OF SPACE-TIME DISTRIBUTION OF PRECIPITATION AND COMPARISION OF INTERPOLATION METHODS FOR MOUNTAIN AREAS IN SOUTHWEST CHINA

摘要

Accurate estimation of precipitation in mountain areas is limited by the scarcity of rainfall stations in these areas. Meanwhile, adapting to the evident spatial differentiation of precipitation in mountain areas is difficult because the traditional method of spatial interpolation of precipitation only considers the influence of a single topographic factor on precipitation. In this study, the precipitation data of 30 ground stations in Ailao Mountain Area during 10 years (2005-2014) are measured. According to the temporal distribution characteristics of rainfall in the study area, the study period is divided into the annual, dry, and wet seasons. Spatial auto-correlation analysis of the affecting factors of precipitation is conducted. The precipitation interpolation region is divided using the K-means clustering analysis method, and the main affecting factors of precipitation are determined. Interpolation precision is evaluated by the leave-one-out method. The optimal method of spatial interpolation of precipitation for different seasons and districts is determined through analysis and validation. Results show that (1) the main affecting factors of mountain precipitation are elevation, latitude, slope, slope direction, and prevailing wind-direction effect index (PWEI). (2) The main affecting factors of annual and seasonal precipitation changes are elevation, latitude, slope, and aspect; those affecting wet season precipitation are elevation, latitude, slope, and PWEI; and those affecting dry season rainfall are elevation, slope, latitude, and aspect. (3) In the interpolation process, the affecting factors of precipitation are introduced and combined with the characteristics of the mountain environment. The accuracy of the interpolation results is significantly improved using the clustering analysis and partitioning methods. Therefore, according to the temporal and spatial distribution characteristics of precipitation, the improved interpolation method can increase estimation accuracy. The said method can also provide significant supplementary data and reference value for mountain areas which lack precipitation monitoring.