城市强降雨致涝风险评估与区划研究——以武汉市为例  

Research on Assessment and Zoning of Urban Heavy Rainfall Induced Flood Risk:A Case Study of Wuhan

作　　者：王德运[1,2] 吴祈[1] 张露丹 郭海湘 柯小玲[1,2] WANG Deyun;WU Qi;ZHANG Ludan;GUO Haixiang;KE Xiaoling(School of Economics and Management Science,China University of Geosciences,Wuhan 430074,China;Laboratory of Natural Disaster Risk Prevention and Emergency Management,China University of Geosciences,Wuhan 430074,China)

机构地区：[1]中国地质大学(武汉)经济管理学院,湖北武汉430074 [2]中国地质大学(武汉)自然灾害风险防控与应急管理实验室,湖北武汉430074

出　　处：《灾害学》2023年第4期107-113,共7页Journal of Catastrophology

基　　金：国家自然科学基金面上项目(72274186);湖北省自然科学基金联合基金项目(2022CFD128);中央高校基本科研业务费专项资金资助项目(CUG2642022006)。

摘　　要：以武汉市为研究区域,从遥感影像中提取相关淹没点数据并从致灾因子、孕灾环境及承灾体三方面出发选取12个洪涝灾害风险影响因素;基于随机森林、梯度提升决策树、XGBoost和逻辑斯蒂回归四种模型拟合淹没点与影响因素间的非线性关系,进而得到不同影响因素对风险水平的作用程度,并据此计算各指标权重,建立风险评估模型。基于所建立的模型,对武汉市洪涝灾害风险区进行识别。其结果为:①相比于随机森林、梯度提升决策树和逻辑斯蒂回归,XGBoost模型在拟合影响因素与淹没点之间的非线性关系方面具有更优的性能;②武汉市洪涝灾害风险水平主要受高程、人口密度和年平均降雨量的影响,高风险区域主要分布在中部主城区和东南部梁子湖附近;③承灾体暴露性急剧增加是武汉市洪涝灾害高风险地区形成的主要原因。This paper takes Wuhan as the research area,relevant inundation point data are extracted from remote sensing images,and 12 flood disaster risk indicators are selected from the following three aspects:disaster-inducing factors,disaster-breeding environment,and disaster-bearing body;Based on 4 models as the random forest,gradient boosting decision tree,XGBoost,and logistic regression,the non-linear relationship between the inundation point and the risk indicators are simulated,and the degree of influence of each indicator on the risk level is obtained,and consequently the weight of each indicator is calculated,and a risk assessment model is established.Based on the established model,the risk areas of flood disasters in Wuhan are identified.The conclusions are as follows:①Compared to random forest,gradient boosting decision tree,and logistic regression,XGBoost model has better performance in fitting the non-linear relationship between influencing factors and inundation points;②The flood risk level in Wuhan is mainly affected by elevation,population density,and annual average rainfall.High risk areas are mainly distributed in the central urban area and near Liangzi Lake in the southeast;③The sharp increase in exposure of disaster-bearing body is the main reason for the formation of high-risk areas of flood disasters in Wuhan.

关 键 词：洪涝灾害 风险评估与区划 影响因素 机器学习 武汉市 

分 类 号：X43[环境科学与工程—灾害防治] X915.5[天文地球—大气科学及气象学] P429