长江中下游地区洪灾损失率函数构建及洪涝灾害损失评估  

Construction of flood loss rate function and flood loss assessment in middle and lower reaches of the Yangtze River

作　　者：王小杰 夏军强[1] 李启杰 周美蓉[1] 假冬冬[2] WANG Xiaojie;XIA Junqiang;LI Qijie;ZHOU Meirong;JIA Dongdong(State Key Laboratory of Water Resources Engineering and Management,Wuhan University,Wuhan 430072,China;The National Key Laboratory of Water Disaster Prevention,Nanjing Hydraulic Research Institute,Nanjing 210029,China)

机构地区：[1]武汉大学水资源工程与调度全国重点实验室,湖北武汉430072 [2]南京水利科学研究院水灾害防御全国重点实验室,江苏南京210029

出　　处：《水资源保护》2023年第5期88-98,共11页Water Resources Protection

基　　金：国家自然科学基金项目(51725902,41890823);中英牛顿高级学者基金项目(NSFC52061130219+NAF\R1\201156)。

摘　　要：采用Spearman相关系数和随机森林法分析了长江中下游各省市洪灾损失率与不同影响因素指标的相关性以及各指标的重要度,采用指数和乘幂2个多元回归函数构建了综合考虑降水、防洪减灾能力、社会经济发展与自然资源影响的洪灾损失率函数,通过对比2组函数的拟合结果选择最优函数作为预测模型,估算了2030-2100年5种共享社会经济路径(SSPs)下洪涝灾害损失。结果表明:长江中下游各省市洪灾损失率与降水总体呈正相关关系,与防洪减灾能力、社会经济发展和自然资源总体呈负相关关系;构建的多元回归函数曲线能较好地拟合洪灾损失率的变化规律,特别是上海、江苏、安徽和江西的经济损失率和死亡人口率函数曲线的拟合精度较高,决定系数大于0.85;在1991年和1998年降雨情景下,长江中下游地区未来洪涝经济损失较历史损失呈增加趋势,且最高分别增长603%和572%,受灾人口较历史损失呈减少趋势,且最低分别减少72%和52%;该地区未来在以适应挑战为主的SSP4路径下洪涝经济损失和受灾人口数量均最小且呈减少趋势,以经济发展为主的SSP5路径下洪涝经济损失最大且呈增加趋势,以区域竞争发展为主的SSP3路径下洪涝受灾人口最多且呈先减少后增加趋势。The correlation between flood loss rates,different indexes of influencing factors,and important degree of different indexes were quantitatively analyzed using the Spearman s correlation coefficient and random forest methods.Flood loss rate functions for different provinces in middle and lower reaches of the Yangtze River(MLRYR)were constructed using exponential and power multiple regression functions,comprehensively considering the impacts of rainfall,disaster prevention and mitigation ability,socio-economic development,and natural resources on the flood disaster.By comparing the fitting results of the two functions,the optimal function was selected as the prediction model,and flood loss of different provinces in MLRYR during the period 2030-2100 under five shared socioeconomic pathways(SSPs)was estimated.The results show that the flood loss rates of different provinces in MLRYR are positively influenced by rainfall,and negatively influenced by socio-economic development,natural resources,and disaster prevention and mitigation ability.The constructed multiple regression function curves can well fit the variation of flood loss rates,especially,the function curves of economic loss rate and death rate for Jiangsu,Anhui,and Jiangxi provinces and Shanghai municipality have high fitting precision,and the determination coefficients are greater than 0.85.Under the rainfall scenarios of 1991 and 1998,the flood economic loss in MLRYR in the future shows an increasing trend compared with the historical losses,with the maximum increasing rates of 603% and 572%,respectively,and the flood-affected population in the future shows a decreasing trend,with the minimum decreasing rates of 72% and 52%,respectively.The flood economic loss and flood-affected population in MLRYR are the least,with decreasing trends under SSP4 that is characterized by adaptation challenges,the flood economic loss is the largest with an increasing trend under SSP5 that is dominated by economic development,and the flood-affected population is the largest w

关 键 词：洪涝灾害 洪灾损失率函数 共享社会经济路径(SSPs) 长江中下游地区 

分 类 号：P429[天文地球—大气科学及气象学] TV122[水利工程—水文学及水资源]