应用地理探测器改进地面沉降危险性评估模型的研究

地面沉降危险性评估对受地面沉降灾害影响的城市有着重要的参考意义。目前应用较广泛的因子叠加模型和模糊层次分析模型存在着评估指标权重的确定人工干预较强,影响评估结果的问题。本文采用地理探测器分别对因子叠加模型和模糊层次分析模型进行改进,根据各评估因子的分布与沉降分布的关联性定量地计算各项评估指标的权重。本文利用改进后的三个模型:地理探测器--最大值因子叠加模型、地理探测器--加权因子叠加模型以及地理探测器--模糊层次分析模型分别进行地面沉降危险性评估,并与最大值因子叠加模型、加权因子叠加模型、模糊层次分析模型进行对比,结果表明:改进后的模型总体评估准确度分别为75.23%、89.56%、76.38%,均高于未改进模型的60.51%、53.73%、65.05%。

Hazard assessment model of land subsidence based on geographical detector

Risk assessment of land subsidence is a significant indicator of the potential risk for urban safety caused by land subsidence. The widely used factor overlapping model and fuzzy analytic hierarchy process model have a problem that is the determination of index weights of the models is affected significantly by human experience, which resulting in relatively subjective assessment result. We propose amethodby combining factor overlapping model and fuzzy analytic hierarchy process model with geographical detector separately. Index weights are determined by the relationship of geographical distribution between factors and land subsidence. Three models combined with geographical detector( Geographical Detector-Maximum Value Factor Overlapping, Geographical Detector-Weighted Factor Overlapping, Geographical Detector-Fuzzy Analytic Hierarchy Process) were applied to assess the hazard separately and validated by comparing the three models with three originalmodels. The result shows that the overall assessment accuracy of improved models are 75. 23%,89. 56% and 76. 38%,the overall assessment accuracy of originalmodels are 60. 51%,53. 73%and 65. 05%. Models combined 00 with geographical detector have a higher overall assessment accuracy.