水文频率分析中抽样误差的统计试验研究  

Statistical Test Study on Sampling Error in Hydrological Frequency Analysis

作　　者：唐林[1] 任智慧[2] 

机构地区：[1]黄河勘测规划设计有限公司,河南郑州450003 [2]黄河水利科学研究院,河南郑州450003

出　　处：《水文》2013年第6期6-10,共5页Journal of China Hydrology

摘　　要：基于统计试验方法,探讨水文频率分析中抽样误差的分布问题。试验结果表明:小样本容量时,设计值抽样误差存在"系统性"偏小,误差均值小于零,随样本容量的增加,偏差逐渐缩小,趋于零均值正态分布;当Cv、Cs值较大时,抽样误差明显增大,小样本情况下还须额外考虑抽样误差均值的修正。Based on statistical test method, sampling error distribution in hydrological frequency analysis has been studied. The test results show that： a） Design value is estimated ＂systemic＂ smaller than actual value according to smaller sample size. The sam- pling relative error（SRE） is below zero. With the sample size increasing, deviation gradually reduce, and SRE is characteristics of zero mean normal distribution. b） SRE would significantly increase with large value of Cv and Cs. So additional revision of SER should be considered in small sample size condition.

关 键 词：水文频率分析 抽样误差 统计试验 

分 类 号：P333[天文地球—水文科学]