基于信息量模型的灵宝市地质灾害易发性分析与评估

Susceptibility Assessment of Loess Slope Collapses in Lingbao City Using the Information Value Model

摘要: 本研究以三门峡市灵宝地区为研究范围，综合选取坡度、坡向、坡型、地形起伏度、工程地质岩性组合、距水系距离、降水量、植被覆盖率及距道路距离等九项主要控制因子，作为区域地质灾害易发性分析与评估的指标体系。基于高分一号卫星遥感影像提取地质灾害相关的动态因子，构建信息量模型评价框架，并对各环境因子的单项信息量进行定量计算与综合分析，进一步探讨其空间分布特征及相对贡献度。在此基础上，划定不同危险等级的黄土崩塌易发性分区。结果表明，高易发性区域的崩塌频率为6.39，其次是中(0.93)、低(0.38)和极低(0.11)，与历史统计的崩塌点分布非常吻合，说明分析模型生成的崩塌易发区合理，具有较高的评价精度，有助于制定该地区防灾减灾策略。

Abstract: This study focuses on the susceptibility assessment of loess slope collapses in Lingbao City, Sanmenxia, a region characterized by extensive loess-covered terrain and highly erodible slopes. Nine conditioning factors were selected, including slope gradient, slope aspect, slope morphology, terrain relief, engineering geological lithological assemblages, distance to drainage networks, precipitation, vegetation coverage, and proximity to roads. Dynamic variables were derived from Gaofen-1 high-resolution remote sensing imagery. An Information Value (IV) model was employed to quantitatively evaluate the contribution of each factor, followed by a comprehensive analysis of their spatial distribution patterns and relative weights. A susceptibility zoning map was subsequently produced, classifying areas into different hazard levels. The results indicate that high-susceptibility zones exhibit a collapse occurrence frequency of 6.39, followed by moderate (0.93), low (0.38), and very low (0.11) susceptibility classes. The predicted distribution of loess slope collapses shows strong agreement with the historical collapse inventory, validating the robustness and high predictive capability of the applied model. These findings provide a scientific basis for region-specific geohazard prevention and mitigation strategies in loess-dominated landscapes.