GIS支持下基于CF和CF-LR模型的恩施州滑坡灾害易发性评价

    LANDSLIDE SUSCEPTIBILITY ASSESSMENT FOR ENSHI, HUBEI PROVINCE: With GIS-based certainty factor and certainty factor-logistic regression coupling model

    • 摘要: 湖北省恩施州内地质条件复杂, 境内地质灾害数量众多, 尤以滑坡为甚. 以该州为研究范围, 择取了包括地表坡度、斜坡坡型、坡向、构造、道路、地表水系、地层岩性、植被覆盖率8个方面的影响因素, 基于ArcGIS平台统计分析空间数据的功能, 分别采用确定系数模型及确定系数和逻辑回归耦合模型的方法进行区域滑坡地质灾害易发性评价, 再通过对验证集灾害点在各个分区内的遍布情况和AUC值的比对进行两种模型的精度验证. 结果表明两种模型易发性分区结果大体上一致, 耦合模型的精度略高一筹. 基于该组合模型计算出的易发值, 将恩施州滑坡灾害易发性等级划为低易发区、中易发区、高易发区和极高易发区, 为该地区地质灾害防治提供支撑.

       

      Abstract: The geological conditions in Enshi Prefecture of Hubei Province are complex, with a high number of geological disasters, especially landslides. Taking the prefecture as the research area, eight influencing factors are selected, including surface slope, slope type, slope aspect, structure, road, water system, formation lithology and vegetation coverage. Based on the statistical analysis of spatial data on ArcGIS platform, the regional landslide susceptibility is evaluated by using the certainty factor (CF) and certainty factor-logistic regression (CF-LR) coupling model. Then the accuracy of both models is verified by comparing the distribution of disaster sites in the validation set in each zoning area and AUC values. The results show that the susceptibility zoning of the two models are generally consistent, although the accuracy of the coupling model is slightly higher. Based on the calculated values of the combined models, Enshi area is classified into low, medium, high and extra-high susceptible zones in terms of landslide susceptibility level, which can provide support for the prevention and control of geological hazards in the area.

       

    /

    返回文章
    返回