中文摘要：

目的分析陕西月、季、年降水量的正态分布特征，为气候分析与预测中正确选用统计方法和统计模型提供依据，进而提高气候分析与预测能力。方法应用偏度系数、峰度系数检验和X2检验，对陕西82个测站1958--2007年月、季、年降水量的正态分布性进行检验分析。结果表明陕西各测站年降水量基本遵从正态分布；大部分气象观测站夏季降水量遵从正态分布，部分观测站春季和秋季降水量遵从正态分布，有一半的测站冬季降水量遵从正态分布；4月份和7月份月降水量遵从正态分布的测站数较多，冬季月降水则多不遵从正态分布。渭河流域月降水量遵从正态分布的月份最多，陕北长城沿线降水遵从正态分布的月份最少。结论陕西降水量遵从正态分布的测站具有明显季节变化和地域性，降水量偏度系数未能达到阈值标准是非正态分布的根本原因。

英文摘要：

Aim To analyze the distribution of precipitation in a month, a year and the reason, which can offer a right way to select a right statistics method and model and then improve the ability in climate diagnosis and forecast. Methods The normal precipitation of 82 meteorological stations in Shaanxi province was analyzed by calculating the coefficient of skewness, coefficient of kurtosis and Chi-square test. The result shows： the yearly precipitation of each station in Shaanxi is subject to normal. Results But they have obviously seasonal change, and regional, most stations in summer and partial stations in spring and autumn are subject to normal, and only half stations in winter are poorly subject to normal. The distributions of monthly precipitation have apparently variation. Spring, early summer and early autumn show the best subjection to normal distribution, but most stations in winter are different. The monthly precipitation subject to normal is easy to see in the stations of HanJiang watershed, and the precipitation is small in the stations along the Great Wall in north Shaanxi. Conclusion The coefficient of skewnesses doesnt reach the threshold,which is the basic reason for the samples which don＇t pass the test. This study suggests we should choose the proper statistics models based on the distribution of the meteorological elements in the climate analysis and forecast.