未来不同气候情景下的汉源县滑坡易发性评估  

Landslide susceptibility assessment in Hanyuan County under different future climate scenarios

作　　者：孙福路 朱红 崔成玲 孟健 吴馨悦 韩丹 SUN Fulu;ZHU Hong;CUI Chengling;MENG Jian;WU Xinyue;HAN Dan(College of Ecological Environment,Institute of Disaster Prevention,Sanhe 065201,China;Hebei Key Laboratory of Resource and Environment Disaster Mechanism and Risk Monitoring,Sanhe 065201,China;Beijing GEOWAY Info-tech Company,Beijing 100043,China;College of Earth Sciences,Institute of Disaster Prevention,Sanhe 065201,China;Ningxia Institute of Surveying and Mapping Geographic Information,Yinchuan 750004,China)

机构地区：[1]防灾科技学院生态环境学院,河北三河065201 [2]河北省资源环境灾变机理及风险监控重点实验室,河北三河065201 [3]北京吉威空间信息股份有限公司,北京100043 [4]防灾科技学院地球科学学院,河北三河065201 [5]宁夏回族自治区测绘地理信息院,宁夏银川750004

出　　处：《自然灾害学报》2025年第2期66-78,共13页Journal of Natural Disasters

基　　金：河北省高等学校科学研究项目(QN2025357);廊坊市科学技术研究与发展计划自筹经费项目(2024011025);河北省自然科学基金项目(D2020512001)。

摘　　要：汉源县滑坡地质灾害频发,对于未来不同气候情景下的滑坡地质灾害易发性情况的预测研究,可为未来气候变化下的防灾减灾工作提供参考依据。文中基于EC-Earth3模式下的3种未来不同气候情景(SSP1-2.6、SSP2-4.5、SSP5-8.5)数据,在Pearson相关性分析与多重共线性分析的基础上,最终选取坡度、坡向、剖面曲率、地形湿度、土地利用、距断层距离、降雨量和径流量共8个影响因子作为滑坡易发性评估指标,采用信息量-熵指数模型与支持向量机模型开展不同气候情景下滑坡易发性评估。结果表明:在SSP1-2.6、SSP2-4.5以及SSP5-8.5情景下,信息量-熵指数模型的曲线下面积(area under curve,AUC)值均为0.928,在SSP1-2.6、SSP2-4.5以及SSP5-8.5情景下,支持向量机模型的AUC值分别为0.957、0.967、0.969。支持向量机模型在未来不同气候情景下的滑坡易发性预测精度方面具有更强的鲁棒性,在SSP1-2.6、SSP2-4.5以及SSP5-8.5情景下发生滑坡灾害的区域面积分别为23.02%、21.09%、26.39%,表明在高排放、高发展的情景下,滑坡灾害发生的可能性将会更大。Landslides geological disasters frequently occur in Hanyuan County.Predictive research on the susceptibility of landslide disasters under different future climate scenarios can provide a reference for disaster prevention and mitigation efforts under future climate changes.This paper,based on data from three different future climate scenarios(SSP1-2.6,SSP2-4.5,SSP5-8.5)under the EC-Earth3 model,selects eight influencing factors for landslide susceptibility assessment,including slope gradient,slope aspect,profile curvature,topographic wetness,land use,distance to faults,rainfall,and runoff,based on Pearson correlation analysis and multicollinearity analysis.The evaluation of landslide susceptibility under different climate scenarios is conducted using both the information value-index of entropy(IV-IOE)model and the support vector machine(SVM)model.The experimental results show that the area under curve(AUC)values for the IV-IOE model are 0.928 across all scenarios.The AUC values for the SVM model are 0.957,0.967 and 0.969 under SSP1-2.6,SSP2-4.5 and SSP5-8.5 scenarios,respectively.The SVM model demonstrates stronger robustness in predicting landslide susceptibility under different future climate scenarios.The areas prone to landslide disasters under SSP1-2.6,SSP2-4.5 and SSP5-8.5 scenarios account for 23.02%,21.09%and 26.39%,respectively.The analysis concludes that the likelihood of landslide disasters will be greater under scenarios of high emissions and high development.

关 键 词：不同气候情景 信息量-熵指数模型 支持向量机模型 滑坡易发性评估 

分 类 号：P642.22[天文地球—工程地质学] X43[天文地球—地质矿产勘探]