基于Copula方法的干旱历时和烈度的联合概率分析  

Joint Probability Analysis of Drought Duration and Severity Based on Copula Approach

作　　者：许月萍[1] 张庆庆[1] 楼章华[1] 刘德地[1] 

机构地区：[1]浙江大学建筑工程学院水文与水资源研究所,杭州310058

出　　处：《天津大学学报》2010年第10期928-932,共5页Journal of Tianjin University(Science and Technology)

基　　金：国家自然科学基金资助项目(50809058);浙江省自然科学重点基金资助项目(Z5080048);教育部博士点基金资助项目(200803351029)

摘　　要：采用自回归马尔可夫模型来延长干旱数据,以解决干旱数据短缺的问题,在此基础上获取长序列干旱数据;应用Copula方法模拟干旱历时和干旱烈度之间的相依关系,并用自助抽样法检验Copula函数的拟合效果;最后得出边际分布分别为皮尔逊Ⅲ型和伽马函数的两元联合分布,并计算干旱历时和干旱烈度的联合分布概率.模拟结果表明,Clayton Copula能较好地模拟两变量之间的相依关系.根据Copula联结函数来模拟水文干旱极限事件,可考虑水文干旱极限事件不同变量之间的相依性,方法简单合理,可成为水文干旱极限分析的一个有效工具.To solve the problem of the shortage of drought data, autoregressive Markov model was used to extrapolate original time series and long time series of drought data are obtained. Then Copula method was employed to simulate the dependence between drought duration and severity, and bootstrap method was used to check the validity of Copula. In the end, a joint distribution of drought duration and severity was obtained with P-Ⅲ and Gamma distributions as margins respectively. The simulation results show that Clayton Copula can model the dependence of the two variables pretty well. Using the Copula approach to model various events of hydrological drought is a simple but reasonable method, which puts different variables of extreme drought events into consideration and this approach can serve as a very useful tool for drought extreme analysis.

关 键 词：自助抽样法 自回归马尔可夫模型 COPULA方法 干旱极限分析 

分 类 号：TV122.5[水利工程—水文学及水资源]