岷江流域降水极值概率分布研究

以岷江流域15个气象站1962-2012年的年极大降水序列（AM）和超门限峰值降水序列（POT）为基础，选取广义极值分布（GEV）和广义帕雷托分布（GPD），采用L-矩参数估计法，研究降水极值的统计特征，根据K-S检验结果确定的最优分布计算不同重现期水平下的降水量。结果表明：GEV分布和GPD分布分别更适用于AM和POT序列。在雨量较少的地区，两种分布推算的不同重现期降水量相差不超过6mm，对于多雨地区，当重现期为50、100年时，这种差值要大于20mm，通过分析，认为GPD分布推算的结果更为可靠。

Research on Probability Distribution of Precipitation Extremes in the Minjiang River Basin

Abstract:Based on the annual maximum(AM) and peak over threshold(POT) extreme precipitation series from 15 stations in the Minjiang basin during 1962-2012,the generalized extreme value distribution and generalized Pareto distribution are used to investigated statistic characteristics by the L-moments method,and precipitation extremes with different return period were calculated by the robust distribution which determined with K-S test.Research results show that GEV fits better AM series and GPD fits better POT series.The difference of precipitation extremes with different return period is not more than 6mm in regions with little rainfall,and in rainy regions,the difference is more than 20mm when the return period is 50 or 100 years.It is believed that GPD is more reliable in this basin.