基于Copula模型的暴雨综合指标分析——以万州为例  

Analysis of Comprehensive Index of Rainstorm Based on Copula Model——A Case Study in Wanzhou

作　　者：冯介玲 祝好 Feng Jieling;Zhu Hao(Chongqing Institute of Meteorological Sciences,Chongqing 401117;Chongqing Engineering Research Center of Agrometeorology and Satellite Remote Sensing,Chongqing 401117)

机构地区：[1]重庆市气象科学研究所,重庆401147 [2]重庆市农业气象与卫星遥感工程技术研究中心,重庆401147

出　　处：《气象科技进展》2021年第2期58-62,共5页Advances in Meteorological Science and Technology

基　　金：重庆市气象局青年基金项目(QNJJ-202101)。

摘　　要：利用重庆市万州区1967—2019年的降水资料,选取年暴雨量、年暴雨强度和年暴雨贡献率为暴雨要素,利用Archimedean Copula函数构建二维联合分布,并根据RMSE,AIC,Bias和OLS值进行优度检验,确立最优Copula函数,分析多要素联合下的暴雨概率以及重现期特征。研究表明,Clayton Copula函数适合构建年暴雨量和年暴雨强度、年暴雨强度和年暴雨贡献率的综合指标联合分布模型,Frank Copula函数适合构建年暴雨强度和年暴雨贡献率的综合指标联合分布模型;暴雨要素之间具有的正相关关系可由Copula函数直观清晰的反映,最优Copula函数适用于分析暴雨多要素的联合概率;Copula函数模型能够反映暴雨多方面的特征,基于暴雨多要素的联合重现期以及同现重现期可以呈现更多的暴雨要素信息,也可以反映暴雨灾害发生时的降水集中程度和暴雨的强度等级。因此,利用Copula函数能够有效反映暴雨的联合特征,提供更全面的暴雨信息。Based on precipitation data from 1967 to 2019 in Wanzhou District,this study selected annual rainstorm depth,annual rainstorm intensity and annual rainstorm contribution as rainstorm factors,built two-dimensional joint distribution by using the Archimedean Copula function,and used RMSE,AIC,Bias and OLS test to confirm the optimal Copula function,in order to analyze the probability of rainstorm and joint characteristic of return period with many combined factors.Results showed that,Clayton Copula function is suitable for building the comprehensive index joint distribution model of annual rainstorm depth and annual rainstorm intensity as well as annual rainstorm intensity and annual rainstorm contribution,Frank Copula function is suitable for building the comprehensive index joint distribution model of annual rainstorm intensity and annual rainstorm contribution.The positive correlation between rainstorm factors can be reflected directly and clearly by Copula function,and the optimal Copula functions is suitable for analyzing the joint probability of rainstorm factors.The Copula function model can reflect various characteristics of rainstorm,the internal information heavy elements can be presented by the joint return period and co-occurrence return period,and the degree of precipitation concentration and intensity level of rainstorm can be reflected as well.Therefore,the Copula function can be used to effectively reflect the joint characteristic and comprehensive information of rainstorm.

关 键 词：COPULA函数 暴雨 联合特征 重现期 万州区 

分 类 号：P426.62[天文地球—大气科学及气象学]