在介绍极值理论基础、重尾分布判别方法的基础上,选用3种常用极值指数估计量,以山西省太原站月降水序列为例研究极值指数估计方法在月降水频率分析中的应用。结果表明:太原站月降水序列的样本分布属于重尾分布;选取序列强降水部分进行拟合时,Moment估计量对该段经验点据的拟合效果相对最佳,且计算简便,可为当地月降水序列的频率分析提供参考。
This paper studied the extreme value index estimation methods and their application in monthly precipitation frequency analysis. Based on the introduction of extreme value theory and discrimination method of heavy-tailed distribu- tion, three kinds of EVI estimators were selected. Monthly precipitation series of Taiyuan station in Shanxi province was chosen as an example for application of extreme value index estimation methods in monthly precipitation frequency analy- sis. The results showed that the distribution of monthly precipitation series of Taiyuan station was heavy-tailed distribu- tion. Heavy precipitation section of series was best fitted by Moment estimator. Simple in calculation, high in accuracy, Moment estimator can provide reference for monthly precipitation frequency analysis in Taiyuan.