基于GIS和CF-Logistic回归模型地质灾害易发性评价:以青海湟中县为例  

Evaluation method of geological hazard susceptibility:A case study on GIS and CF-Logistic regression model in Huangzhong,Qinghai

作　　者：张晓博[1,3] 周萍 张焜 张兴 刘宝山[1,3] 邓辉 ZHANG Xiaobo;ZHOU Ping;ZHANG Kun;ZHANG Xing;LIU Baoshan;DENG Hui(Key Laboratory of the Northern Qinghai-Tibet Plateau Geological Processes and Mineral Resources,Qinghai Geological Survey Institute,Xining 810012,China;College of Earth Sciences,Chengdu University of Technology,Chengdu 610059,China;Qinghai Satellite Application Technology Center of Natural Resources,Xining 810012,China)

机构地区：[1]青海省地质调查院,青海省青藏高原北部地质过程与矿产资源重点实验室,青海西宁810012 [2]成都理工大学地球科学学院,四川成都610059 [3]自然资源青海卫星应用技术中心,青海西宁810012

出　　处：《沉积与特提斯地质》2023年第4期797-807,共11页Sedimentary Geology and Tethyan Geology

基　　金：青海省科技厅重点研发与转化计划(2019-SF-130);青海省科技厅应用基础研究(2018-ZJ-737)。

摘　　要：研究地质灾害易发性的评价方法,对地质灾害防治有着非常重要的现实意义。本文以青海省西宁市湟中县为研究区域,选取高程、坡度、坡向、地形起伏度、距河流距离、距断层距离和工程岩组7个评价因子,利用确定性系数与逻辑回归模型相结合的方法计算出每个单元格地质灾害发生的概率。同时利用ROC曲线和AUC值对模型分类精度进行验证,最终得到AUC值为0.863,说明该方法对湟中县地质灾害易发性评价具有较强的适用性和客观性。本文研究表明,高层、坡向、距河流距离和工程岩组4个因子对研究区地质灾害的影响最为显著。从地质灾害的空间分布来看,该方法计算结果表明极高、高易发区主要分布在湟水河及其干流两侧低山丘陵地区,低易发区主要分布在研究区北部和西南地区。从评价因子的角度分析,高易发区主要分布在离河流500m的松散冲洪积岩层和软弱层状碎屑岩岩层上。以上研究结果表明,基于CF-Logistic回归模型对研究区地质灾害易发性评价有较强的参考价值,能为研究区地质灾害的防治工作提供理论依据及方法。The evaluation method for the susceptibility of geological hazards has very important practical significance for the prevention and control of geological hazards.This paper takes Huangzhong County,Xining City,Qinghai Province as the research area to investigate the most suitable assessment method for the area’s geological hazard susceptibility.The results show that using seven evaluation factors(including elevation,slope,aspect,topographic undulation,distance from river,distance from fault,and engineering rock group)and the certainty factor and logistic regression model to evaluate the geological hazard susceptibility of Huangzhong County is suitable and reliable.Using the determinant coefficient and logistic regression model,we calculate the probability of geological hazard in each cell.Subsequently,we use the ROC curve and the AUC value to verify the classification accuracy of the model.The AUC value is 0.863,indicating that our method has good applicability for evaluating the susceptibility of geological disasters in the study area.The four factors of high-rise,aspect,distance from river,and engineering rock group have the most significant impacts on geological disasters in the study area.From the perspective of the spatial distribution of geological hazard susceptibility,extremely high and high susceptibility areas calculated using this method are mainly distributed in the low mountain and hilly areas on both sides of the Huangshui River and its main stream,while low susceptibility areas are mainly distributed in the northern and southern regions of the study area.From the perspective of geological hazard evaluation factors,high susceptibility areas are mainly distributed on loose alluvial strata and weak clastic rock strata 500 m away from the river.The research results indicate that the CF-Logistic regression model has strong reference value for evaluating the susceptibility of geological hazards in the study area and can provide theoretical and methodological basis for the prevention and control of geolo

关 键 词：地质灾害 GIS 易发性评价 确定性系数 LOGISTIC回归模型 湟中县 

分 类 号：P642.23[天文地球—工程地质学]