基于加权信息量的地质灾害易发性评价——以四川省泸定县为例

Geologic Hazards Susceptibility Assessment Based on Weighted Information Value--A Case Study in Luding County,Sichuan Province

[目的]对四川省泸定县进行地质灾害易发性评价,为该区地质灾害预防预测提供依据。[方法]借助谷歌影像解译,获取崩塌、滑坡、泥石流等地质灾害隐患点279处。选取地形地貌、岩性构造、气象水文、土壤与土地利用(LULC)4个方面构建评价指标体系。运用确定性系数法确定因子权重,并结合信息量法构成加权信息量模型。通过ArcGIS空间分析平台,开展灾害易发性评价。[结果]研究区极高和高易发区分别占总研究区面积的13.54%,26.49%。地质灾害点共225处落在极高易发区和高易发区内,占总样本灾害点的80.65%。通过受试者工作特征曲线(ROC)的线下面积(AUC)进行检验,其值为0.793,评价模型精度良好。[结论]对四川省泸定县的地质灾害易发性进行了等级划分,采用的加权信息量方法的易发性评价结果可信。

[Objective]The susceptibility of geological disaster was evaluated,in order to provide a theoretical basis for geological disaster prevention in Luding County,Sichuan Province.[Methods]With the help of Google image interpretation,279 potential geological disaster points such as collapse,landslides and debris flows were obtained.The evaluation index system was constructed from four aspects including topography,lithologic structure,meteorological hydrology,soil and land use(LULC).The weight of the factors is determined by the deterministic coefficient method,and the weighted information model was constructed by the information quantity method.Disaster vulnerability was assessed by the ArcGIS spatial analysis.[Results]The extremely high and high risk areas accounted for 13.54%and 26.49%of the total study area,respectively.A total of 225 geological disaster sites were located in the extremely high and high risk areas,accounting for 80.65%of the total sample disaster points.The test was performed by the offline area(AUC)of the receiver operating characteristic curve(ROC),and its value was 0.793,and the accuracy of the evaluation model was excellent.[Conclusion]The geological disasters in Luding County was classified,and the risk assessment results of the weighted information method were reliable.