摘要
概述了枯水频率分析的研究进展,采用PE3分布、Gumbel分布、GL分布、GP分布和Weibull分布等5种线型拟合枯水流量系列,所有分布均采用线性矩法估计参数。统计试验方法用于比较上述分布拟合枯水流量的优劣,结果表明在变差系数较小时,二参数分布能获得较优的结果；当变差系数较大时,三参数分布的模拟精度高于二参数分布。基于实测枯水流量系列的比较研究结果证明了统计试验的结论,且在水文系列变差系数较小的流域能获得较高的模拟精度。
This paper briefly snmmarizes the advancements of low flow frequency analyzing, in which five distributions, including Pearson Type Ⅲ, Gumbel, Generalized Logistic, Generalized Pareto and Weibull, are selected to fit low flow series, and the Lmoment method is used to estimate the parameters of those distribution functions. Based on the Monte Carlo simulation and annual minimum low flow series of different averaging intervals, the descriptive and predictive ability of those distributions is tested and compared. The results indicate that two-parameter distributions have small bias and root mean square error in low flow quantile estimates if Cv is small. Otherwise, three-parameter distributions are better than two-parsmeter distributions. Application of the recorded low flow data series also show that a better curve fitting can be achieved in a watershed where the Cv of low flow series is small.
出处
《水文》
CSCD
北大核心
2006年第1期28-33,共6页
Journal of China Hydrology
基金
国家重大基础研究前期专项资助项目(2003CCA00200)
关键词
枯水流量系列
分布线型
统计试验
比较研究
low flow series
distribution functions
statistical tests
comparative study
