关键词: 滑坡易发性预测;影响因子;信息量模型;Logistic回归模型;信息量-Logistic回归耦合模型;

Abstract: The prediction of regional landslide susceptibility can accurately reflect the spatial probability characteristics of landslide distribution in a specific research area.In this paper, the coupling model based on information value model and logistic regression model was used to predict the landslide susceptibility in Chongyi County, Jiangxi Province.First of all, 10 influence factors including elevation, slope gradient, slope structure, plane curvature, profile curvature, topographic relief, distance from water system, lithology, Normalized Difference Vegetation Index(NDVI) and Normalized Difference Built-up Index(NDBI)were selected.Then, the logistic regression model was constructed by using the information value of each factor.In addition, the information value model and the logistic regression model were used as comparison models to explore the evaluation results of the landslide susceptibility of the three models.The results showed that the coupling model had the best prediction performance(AUC = 80.4%),and the rest were logistic regression model(76.8%) and information model(72.8%);the distribution rules of landslide susceptibility predicted by each model had certain similarities, namely most of the landslides occurred in areas of low elevations, close to water systems, carbonate lithologic strata, low vegetation coverage and dense buildings.

Key words: landslide susceptibility prediction; influence factor; information value model; logistic regression model; information value-logistic regression coupling model;