基于秩统计量的枯水期径流时序变点的非参数识别  

Nonparametric change point recognition for dry season runoff time series based on rank statistic method

作　　者：涂新军[1] 陈晓宏[1] 

机构地区：[1]中山大学水资源与环境研究中心,广东广州510275

出　　处：《水利学报》2009年第5期603-607,共5页Journal of Hydraulic Engineering

基　　金：国家自然科学基金重点项目(50839005);中山大学青年教师科研启动基金资助

摘　　要：本文提出了枯期径流时序变点的识别原理和基于秩统计量的非参数识别方法,该方法能很好地解决分布函数难以确定时的时序变点识别问题。以东江流域为例,对其干流的枯期径流量年系列(1957—2000年)进行了变点识别,结果显示1974年为东江流域枯期径流的显著性时序变点。通过对比分析变点前后的统计参数,1974年后的东江流域枯期径流量年均值明显增加,年际变化的离散程度增大。The rank statistic method for recognizing the change points of time series is suggested. By using this method the recognition problem occurred to hydrological elements analysis due to the uncertainty of distribution function can be solved. The method is applied to analyze and locate the change point of the annual dry season runoff time series in Dongjiang River according to the hydrological data obtained from three gauge stations. It is found that the change points of runoff time series of these stations occurred to 1970＇s obviously. The annual dry season runoff increases significantly and fluctuates remarkably after 1974.

关 键 词：变点 秩统计量 非参数识别 枯期径流量 

分 类 号：P333[天文地球—水文科学]