基于优势分析法的暴雨综合致灾指数及阈值研究--以张家界市保险业为例

Study on torrential rain comprehensive disaster-causing index and its threshold based on Dominance Analysis Method:A case of insurance industry in Zhangjiajie City

首先,利用张家界2016—2020年降水资料和暴雨灾情资料,以中国人民财产保险股份有限公司湖南省分公司财货险(以下简称“财货险”)理赔案件为研究样本,运用优势分析法确定致灾因子的影响权重,构建暴雨综合致灾指数(I)模型。然后,采用指数函数拟合财货险出险乡镇(街道)数与I的关系,获得不同出险等级对应的I阈值。最后,选用2021年张家界汛期暴雨致灾理赔案例对I及其阈值进行验证。结果表明:张家界市暴雨造成的财货险出险次数总体呈东西少、中间多的特点。暴雨致灾因子中96 h累积降雨量对出险范围影响权重最大,达28.6%。模拟财货险出险范围、理赔金额、出险次数的等级与实际灾情等级均通过0.01显著性水平检验。阈值检验显示,用其对暴雨致灾案例的预测等级与实际等级的一致率(即准确率)为71.4%,其中,轻度、中度、重度灾情等级的预测准确率分别为70%、70%、100%,因而建立的I阈值可用于张家界市财货险专业气象服务。

First,based on the precipitation data and torrential rain disaster data in Zhangjiajie City of Hunan Province from 2016 to 2020,taking the claim cases of the property and cargo insurance(hereinafter referred to as"property and cargo insurance")from Hunan Branch of the People's Insurance Company of China as the research sample,the Dominance Analysis Method was used to determine the influence weights of disaster-causing factors to establish a comprehensive disaster-causing index(I)model of torrential rain.Second,an exponential function is used to fit the relationship between the number of town or street which filed claims of property and cargo insurance and I,then to determine the threshold of I corresponding to different accident levels.Finally,the claim case caused by torrential rain disaster in Zhangjia⁃jie in the flood season of 2021 were selected to verify the I and its threshold.The results show that the number of property and cargo insur⁃ance accidents caused by torrential rain in Zhangjiajie is generally low in east and west but high in middle areas.Among the disaster-caus⁃ing factors,the weight of the 96-hour accumulated precipitation on the scope of accident is the largest,reaching 28.6%.The simulated grades of the scope of accident,the amount of claim and the number of accidents of property and cargo insurance have a high correlation with the grades of actual disasters,and all passed the test at the 0.01 significance level.The threshold test results show that the consistency rate or accuracy between the predicted level and the actual level of torrential rain disaster-causing cases is 71.4%,in which the predicted values of accuracy for the mild,moderate and severe disaster levels are 70%,70%and 100%,respectively.Therefore,the threshold of I established in this study can be used for the industrial meteorological services related to the property and cargo insurance in Zhangjiajie.