基于G-H Copula函数的两变量洪水重现期计算研究

Calculation of return periods of bivariate floods based on G-H Copula function

洪水事件的发生包含多种因素的影响,而传统单变量洪水设计独立控制洪水特征变量的防洪设计标准,忽视了其随时间变化的依存特征。本文选取新疆玛纳斯河流域为研究区,在考虑洪水峰量特征的基础上,利用G-H Copula函数进行联合设计研究。研究表明,GEV分布与GP分布能够很好的描述洪水的时间序列特征;G-H Copula函数考虑了洪水时间过程的联合特征,且二维联合设计方案是偏安全的,这有利于最大程度上防御汛限洪水。

The occurrence of flood events includes the influence of many factors such as flood peaks and flood volume.The traditional single-variable flood design independently controls the flood control design standard of flood characteristic variables,ignoring its dependence characteristics that change with time.In this paper,the Manas River Basin in Xinjiang is selected as the research area,and the G-H Copula function is used to conduct joint design research based on the characteristics of flood peaks and flood volume.Studies have shown that the generalized extreme value(GEV)distribution and the generalized Pareto(GP)distribution can well describe the time series characteristics of floods;the GH Copula function takes into account the joint characteristics of the flood time process,and the two-dimensional joint design scheme is On the safe side,this is conducive to preventing floods to the greatest extent.