汛期洪水的二变量联合分布与应用

Application of Bivariate analysis on flooding in flood season

为在气候变化下更好地描述流域汛期洪水的概率变化特性,达到防洪和洪水资源化的统筹安排,考虑水文事件的相依性和多元化,采用Copula函数理论,结合第Ⅰ型Gumbel分布,研究淮河蚌埠站以上流域1915年到2008年的汛期洪水。研究结果表明:汛期洪量和洪峰的理论联合概率与实测资料联合频率拟合很好;同时探索了洪水联合分析的实际应用:一方面根据联合分布提出"等效水文情景"的概念,以更好地体现水文事件的不确定性和水文变量组合的概率等效性;另一方面论证联合重现期可作为单变量重现期的区间估计,为二元分析的洪水量级划分和洪水的联合风险分析以及工程设计提供一条新思路。

In order to better describe the characteristics of runoff under the context of climate change,the Copula function was adopted to investigate the dependency and pluralism between hydrology events in this study.The flooding data in flood season during the period of 1915 to 2008 in Huaihe basin were collected and analyzed.Both the margin distribution of the flooding volume and the maximum flow in flood season was simulated using Type I Gumbel distribution,and the Gumbel-Hougard Copula was used to show the relationship.The study shows that the bivariate distribution simulations honor the empirical probability very well.Also in this paper,the concept of hydrologic scene was presented to demonstrate the uncertainty and consociation of hydrology events,and provide a more overall mean for decision making.The joint return period was considered to be the interval estimation of traditional return period.From the distribution model and the isolines of return period,it can distinguish the flood magnitude,to some degree,be used as a new way to analyze flooding risk and engineering design.