关键词: 降水量, 空间插值, Kriging插值法, 长江流域, 黄河流域, 珠江流域, 中国

Abstract: The spatial distribution information of precipitation is of great value in the fields of water resource management, drought and flood disaster prediction and sustainable development. Based on precipitation data (1981~2010) and DEM data of 1915 meteorological stations in China, we produce a nice spatial distribution of precipitation of China by using eight interpolation methods, including Inverse Distance Weight method, Radial Basis Function method, Global Polynomial Interpolation method, Local Polynomial Interpolation method, Ordinary Kriging method, Simple Kriging method, Universal Kriging method and Empirical Bayesian Kriging method. By using cross-validation, we find that the Empirical Bayesian Kriging method is optimal for the whole China, and the optimal methods for Yellow River basin, Yangtze River basin and Pearl River basin are Universal Kriging method, Ordinary Kriging method, Radial Basis Function method, respectively. Through verification for cities, it′s proved that the optimal spatial interpolation methods for basins are valid and instructive.

Key words: precipitation, spatial interpolation, Simple Kriging method, Yellow River basin, Yangtze River basin, Pearl River basin；China