基于多模型耦合的永嘉县滑坡易发性评价  

Landslide susceptibility assessment of Yongjia County based on multi-model coupling

作　　者：杨冰颖 缪海波[1] 马闯 崔玉龙[1] YANG Bing-ying;MIAO Hai-bo;MA Chuang;CUI Yu-long(School of Civil Engineering&Architecture,Anhui University of Science&Technology,Huainan 232001,China)

机构地区：[1]安徽理工大学土木建筑学院,安徽淮南232001

出　　处：《河南城建学院学报》2024年第4期91-99,127,共10页Journal of Henan University of Urban Construction

基　　金：国家自然科学基金(42277136);安徽省自然科学基金(2208085MD97);安徽高校自然科学研究项目(2022AH050806)。

摘　　要：区域性滑坡易发性评价是制定滑坡防灾减灾方案和开展国土资源规划利用的重要依据。以浙江省永嘉县为例,在分析滑坡孕灾条件的基础上,选取高程、坡度、曲率、地形位置指数(TPI)、地形起伏度、坡向、地层岩性、距断层距离、距水系距离、年均降雨量、年均风速、距道路距离、归一化植被指数(NDVI)等13个评价因子,采用逻辑回归-信息量(LR-I)耦合模型和逻辑回归-证据权(LR-WOE)耦合模型进行滑坡易发性评价。结果表明:坡向、距断层距离、距水系距离、曲率、坡度、距道路距离和地形位置指数(TPI)等7个因子对滑坡易发性评价的影响较大;利用LR-I模型得到的高易发区和极高易发区中,滑坡数量占比86.67%,滑坡面积占比36.96%,利用LR-WOE模型得到的高易发区和极高易发区中,滑坡数量占比86.19%,滑坡面积占比36.13%;两个耦合模型的准确率均在0.8以上,显示出较高的准确度和良好的预测性能。研究结果对该区域的滑坡防灾减灾具有一定的参考价值。Regional landslide susceptibility assessment is an important basis for formulating landslide disaster prevention and mitigation schemes and carrying out land and resources planning and utilization.Taking Yongjia County,Zhejiang Province as an example,on the basis of analyzing landslide disaster conditions,13 evaluation factors including elevation,slope,curvature,topographic position index(TPI),topographic undulation,slope aspect,stratum lithology,distance from fault,distance from water system,average annual rainfall,average annual wind speed,distance from road and normalized difference vegetation index(NDVI) were selected,and logistic regression-information content(LR-I) coupled model and logistic regression-weight of evidence(LR-WOE) coupled model were used to evaluate landslide susceptibility.The results show that seven factors,such as aspect,distance from fault,distance from water system,curvature,slope,distance from road and terrain position index(TPI),have great influence on landslide susceptibility assessment.Among the high-prone areas and extremely high-prone areas obtained by LR-I model,the number of landslides account for 86.67% and the landslide area account for 36.96%.Among the high-prone areas and extremely high-prone areas obtained by LR-WOE model,the number of landslides account for 86.19% and the landslide area account for 36.13%.The accuracy of both coupled models is above 0.8,showing high accuracy and good prediction performance.The research results have certain reference value for landslide disaster prevention and mitigation in this area.

关 键 词：滑坡 易发性评价 永嘉县 评价因子 多模型耦合 

分 类 号：P642.22[天文地球—工程地质学]