地震工程学报2025，Vol.47Issue(5)：1090-1101,12.DOI:10.20000/j.1000-0844.20230811003
基于斜坡单元和信息量-逻辑回归的滑坡易发性评价
Landslide susceptibility assessment based on slope units and information value-logistic regression:a case study of the Xiangxi and Zhaxi River basins in the Three Gorges Reservoir area
摘要
Abstract
Frequent landslides in the Xiangxi and Zhaxi River basins of the Three Gorges Reser-voir area pose significant threats to life,property,and transportation safety.To investigate the causes and development of landslides in this area,this study selected 11 influencing factors,namely elevation,elevation difference,slope gradient,slope aspect,engineering geological rock groups,slope structure,water contact degree,distance to water systems,distance to roads,rainfall,and land use.The results showed the following:(1)Rainfall and engineering geological rock groups exhibited the strongest positive correlation with landslide occurrence.(2)Suscepti-bility patterns differed significantly between the two basins because of geological and lithological variations,i.e.,the right and southern left banks of the Xiangxi River showed higher susceptibili-ty than the northern left bank,whereas both banks of the Zhaxi River exhibited high to very high susceptibility.(3)Susceptibility levels decreased significantly with the increase in distance to riv-ers;meanwhile,areas with intensive human activities generally exhibited higher susceptibility.The model achieved an accuracy of 0.840 under the receiver operating characteristic curve,confir-ming its reliability in landslide prediction.Slope units exhibited superior alignment with actual landslide distributions compared with grid-based approaches.关键词
斜坡单元/信息量/逻辑回归/滑坡/易发性Key words
slope unit/information value/logistic regression/landslide/susceptibility分类
天文与地球科学引用本文复制引用
蔡环宇,易庆林,邓茂林,宋雨航,费霈妍..基于斜坡单元和信息量-逻辑回归的滑坡易发性评价[J].地震工程学报,2025,47(5):1090-1101,12.基金项目
国家自然科学基金面上项目(42172303) （42172303）
