摘要
为开展湖北省巴东县滑坡灾害防治,针对该区域进行滑坡易发性评价。在分析滑坡灾害与地形地貌、基础地质、水文环境、人类工程活动等相关因子统计关系的基础上,采用支持向量机(SVM)、高斯朴素贝叶斯(GNB)和随机森林(RF)3种机器学习模型对巴东县滑坡灾害进行易发性评价,并通过ROC曲线精度分析方法对比分析3种模型的评价结果。结果表明:①RF模型滑坡易发性评价结果的精度更高,更符合巴东县滑坡灾害实际情况;②巴东县滑坡极高易发区和高易发区面积约占巴东县总面积的26.3%,中等易发性面积约占巴东县总面积的24.5%,低易发区和极低易发区面积约占巴东县总面积的49.2%;③地层岩性、水库缓冲区、道路缓冲区、坡度、水系缓冲区、土地利用类型和断层缓冲区是研究区中较为重要的7个因素,其中地层岩性是控制因素,水库缓冲区和道路缓冲区是主要影响因素,因而滑坡极高、高易发区主要分布在长江、清江两岸及其支流地带、道路两侧和切坡建房附近。研究成果可为巴东县防灾减灾、合理规划土地资源以及同类研究提供参考。
In order to better prevent and control the landslide in Badong County of Hubei Province,we conducted a susceptibility assessment for landslides in this area.Based on the analysis of the statistical relationship between landslide disasters and related factors such as topography,basic geology,hydrological environment,and engineering activities,three machine learning models of Support Vector Machine(SVM),Gaussian Naive Bayes(GNB),and Random Forest(RF),were employed to assess the susceptibility of landslide disasters in Badong County.The evaluation results of the three models were compared and analyzed by using the ROC curve accuracy analysis method.The results showed that:①The RF model has a higher accuracy in evaluating landslide susceptibility,which was more in line with the actual situation of landslide disasters in Badong County.②The area of extremely high and high landslide susceptibility in Badong County accounted for about 26.3%of the county′s total area,the moderate susceptibility area accounted for about 24.5%,and the low and extremely low susceptibility areas accounted for about 49.2%of the county′s total area.③Lithology,reservoir buffer zones,road buffer zones,slope,river buffer zones,land use types and fault buffer zones were the seven most important factors in the study area,among which,lithology was the controlling factor,and reservoir buffer zones and road buffer zones were the main influencing factors.Therefore,the areas of extremely high and high landslide susceptibility were mainly distributed along the banks of Yangtze River and Qingjiang River and their tributaries,alongside roads,and near areas where slopes were cut for housing construction.The research findings can provide references for disaster prevention and reduction,reasonable planning of land resources in Badong County and similar studies.
作者
蒋根
肖诗荣
杨璇喆
马思哲
JIANG Gen;XIAO Shirong;YANG Xuanzhe;MA Sizhe(College of Civil Engineering&Architecture,China Three Gorges University,Yichang 443002,China;Changjiang Three Gorges Survey Institute Co.,Ltd.,Wuhan 430070,China)
出处
《水利水电快报》
2024年第11期48-55,共8页
Express Water Resources & Hydropower Information
基金
国家自然科学基金面上项目(41272310)。
关键词
滑坡易发性
机器学习
支持向量机
高斯朴素贝叶斯
随机森林
巴东县
三峡库区
landslide susceptibility
machine learning
Support Vector Machine
Gaussian Naive Bayes
Random Forest
Badong County
Three Gorges Reservoir area
