中文摘要：

应用Archimedean Copula函数,探讨洪水多要素的联合概率分布和条件概率分布。以大渡河流域丹巴站49年的洪水资料为基础,根据图形分析法和AIC信息准则选择拟合效果好的Copula函数计算洪峰、洪量、历时的联合概率分布和条件概率分布,并绘制相应的条件重现期图。结果表明,Gumbel Copula可以较好的拟合洪峰与洪量,而Frank Copula对于洪峰与历时、洪量与历时及洪峰、洪量和历时的三维联合分布效果最好。通过情景设置分别计算不同条件下洪水特征变量的二维和三维条件概率分布和相应的条件重现期。Copula函数全面考虑了洪水特征量之间的相依性,可用于洪峰、洪量和历时联合概率分布的计算。

英文摘要：

Archimedean Copula function is used to analyze the joint probability distribution and the condition probability distribution of the three variables of flood including peak,volume and duration.Based on the 49years＇ flood data in Dadu River Basin,agood fitting Copula function is chosen to calculate the condition probability distribution of flood peak,volume and duration and drawn figs of corresponding return period.The results show that Gumbel Copula can fit the joint distribution of flood peak and volume while frank Copula is the best choice for the other joint distributions of 2-dimensional as well as the joint distribution of 3dimensional.Different scenarios are set to calculate two and three-dimensional conditional probability distributions and the corresponding return periods.Copula function that considers the dependency of the flood characteristics comprehensively can be used to calculate joint probability distribution among peaks,volumes and durations.