基于GIS与逻辑回归的清平滑坡易发性评价

Landslide Susceptibility Assessment Based on GIS and Logistic Regression Model

四川省绵竹市清平地区在"5·12"汶川大地震作用下形成了大量的滑坡等地质灾害。基于Arc GIS平台,运用遥感及地理信息技术获取滑坡信息,选择高程、坡度、坡向、水系、断层、地层岩性作为滑坡易发性评价因子,采用逻辑回归模型对该区域滑坡易发性进行了分析,并将研究区划分为高易发区、中等易发区和低易发区,并对这评价结果进行分析。结果表明:(1)滑坡高易发区主要位于研究区的东部、西北部和南部区域;(2)高中易发区结果的精度为0.936,表明这逻辑回归的预测能力较好。

A large number of landslides and other geological disasters were triggered by the Wenchuan earthquake in the Qingping area,Mianzhu City,Sichuan Province.Based on the ArcGIS software,this paper used remote sensing and geographic information technology to obtain landslide information and selected elevation,slope angle,slope aspect,river system,fault and lithology as the input parameters.The logistic regression model was used to analyze the landslide susceptibility in the Qingping area.The susceptibility was divided into three classes:high,moderate and low.The result of the susceptibility maps showed that the high susceptibility area was located at eastern,northwestern and southern part of the study area.The accuracy of the results in the high and middle prone area was 0.936,which indicated that the predictive ability of logistic regression was much better.