水文时间序列突变点分析的启发式分割方法  

Heuristic segmentation method for change-point analysis of hydrological time series

作　　者：汪丽娜[1] 陈晓宏[1,2] 李粤安[3] 林凯荣[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 [2]广东省近岸海洋工程重点实验室,广东广州510275 [3]广东省水利厅,广东广州510635

出　　处：《人民长江》2009年第9期15-17,共3页Yangtze River

基　　金：国家自然科学基金重点项目(50839005);国家自然科学基金资助项目(50809078);国家科技支撑计划课题(2006BAB14B07)

摘　　要：水文时间序列突变点的检测是水文统计分析中的一项重要内容,但由于水文时间序列的非线性、非平稳性,导致对检测水文时间序列突变点的技术要求较高。建立了时间序列变点分析的启发式分割数学模型,以此来分析水文时间序列均值的突变。以理想的水文时间序列为仿真基础,结合合水水库的月径流序列,采用启发式分割算法得出去除周期成分后的月径流序列的突变点,以此证明了启发式分割算法检测水文时间序列突变的可行性和优越性。结果表明:4个突变点中,3个突变点集中于水库建成后不久,说明人类活动的剧烈影响为主要驱动力,加剧了对入库月径流量的影响,造成月径流的频繁波动和数次突变点的产生。The abrupt change point testing of hydrology time series is very important.But it requests higher technology supporting for the reason of nonlinear and non-stationary in hydrology time series.This paper presents an approach to study abrupt change of the mean value of hydrological time series by establishing a math model of the heuristic segmentation method.It takes the ideal hydrology time series as the simulation foundation,the monthly runoff series of Heshui Reservoir and the abrupt change points of elimination cycle ingredient monthly runoff series counted by Heuristic segmentation algorithm to indicate the method is feasible and superior for studying the abrupt change of hydrological time series.The results show that,the 3 abrupt change points are closer to the time of reservoir establishment for the all 4 points,which means the impact of human activities is the main reason that causes the frequent fluctuation of monthly discharge and the forming of abrupt change point

关 键 词：突变 启发式分割 月径流序列 水文时间序列 

分 类 号：TV121[水利工程—水文学及水资源]