罗平县崩滑易发性评价模型对比研究

COMPARATIVE STUDY ON EVALUATION MODELS OF LANDSLIDE SUSCEPTIBILITY IN LUOPING COUNTY,YUNNAN PROVINCE

准确的地质灾害易发性分区评价结果,可为建立地质灾害监测预警系统及处理机制提供参考。依据崩滑地质灾害形成条件选取10个评价因子构建评价指标体系,基于共线性诊断和相关性分析检验评价因子以保证其相互独立。分别采用信息量模型(ICM)、归一化频率比模型(NFR)以及与逻辑回归(LR)耦合的信息量-逻辑回归(ICM-LR)耦合模型和归一化频率比-逻辑回归(NFR-LR)耦合模型对罗平县崩滑地质灾害进行易发性评价,并将评价模型结果划分为低、中、高和极高4个等级。采用ROC曲线对评价结果进行精度检验,其AUC值分别为0.820、0.796、0.882和0.840。得出ICM-LR模型的精度最高,且极高易发区主要分布在砂岩、碳酸盐岩组区域和水系延展区域。其低、中、高和极高的面积(分级比)分别为771.1 km^(2)(25.55%)、836.6 km^(2)(27.73%)、864.36 km^(2)(28.64%)和545.94 km^(2)(18.08%)。易发性分区结果与研究区崩滑地质灾害分布情况相符合,可为快速建立评价指标体系和区域崩滑易发性评价模型提供参考依据。

Accurate regional evaluation results of geological hazard susceptibility can provide reference for the establishment of geological hazard monitoring and early warning system and treatment mechanism.According to the formation conditions of geological disasters,ten evaluation factors were selected to construct an evaluation index system,and the evaluation factors were tested based on collinearity diagnosis and correlation analysis to ensure their independence.The susceptibility of landslide geological disasters in Luoping County was evaluated by information information model(ICM),normalized frequency ratio model(NFR),information information-logistic regression(ICM-LR)coupled with logistic regression(LR)and normalized frequency ratio logistic regression(NFR-LR)coupled model,the results of the evaluation model were divided into four grades:low,medium,high and extremely high.ROC curve was used to test the accuracy of the evaluation results,and the AUC values were 0.820,0.796,0.882 and 0.840,respectively.The results show that the ICM-LR model has the highest accuracy,and the highly susceptible areas are mainly distributed in the sandstone,carbonate formation area and drainage extension area.The low,middle,high and very high area(grading ratio)are 771.1km^(2),respectively^(2)(25.55%),836.6 km^(2)(27.73%),864.36 km^(2)(28.64%)and 545.94 km^(2)(18.08%).The results of susceptibility zoning are consistent with the distribution of landslide geological hazards in the study area,which can provide a reference for the rapid establishment of evaluation index system and regional landslide susceptibility evaluation model.