基于CF多元回归模型的地震滑坡危险性研究——以香港屯门为例

Risk Assessment of Earthquake-induced Landslide Based on CFMulti-variable Regression Model——A Case Study of Tuen Mun,Hong Kong

本文根据前人的研究成果以及对自1300年以来地震滑坡震例研究的基础上,选取了岩性、断裂距离、地震烈度、地形坡度、高程、水系等6个因子作为地震诱发滑坡的影响因子,并确定出各个因子在地震滑坡事件中的影响权重。通过GIS技术将滑坡确定性系数CF与回归模型相融合,建立CF值多元回归模型,以解决滑坡评价过程中影响因子的选择及量化的问题。最后,将模型应用于香港屯门地区,进行了该区的地震滑坡空间分布及稳定性初步分析。

According to the previous investigations and study on the more than one hundred earthquake-induced landslides happened from 1303, the factors contributing to landslide occurrences include lithology, seismic intensity, faults, slope angle, slope aspect, elevation , river etc. In this paper we select these factors and then the weights of different factor contributions to landslide are calculated. By integrating landslide certainty factor （CF） and multi-variable regression model, some problems can be solved including the selection and quantification of landslide factors. Earthquake-induced landslide spatial distribution and stability analysis are carried out for Tuen Mun in Hong Kong using the CF multi-variable regression mode 1.