Development of landslide-risk assessment model for mountainous regions in eastern Korea

Abstract

Recently, damage from landslides caused by severe rain storms and typhoons is reported often. Now that due to domestic geomorphological characteristics, many cities are developed and infrastructure such as roads and railways are constructed nearby mountain area the damages of landslides are very serious in Korea. It is important to figure out areas with high landslide-risk and arrange the proper measures in order to prevent such damages effectively. Mountainous region in eastern Korea, which experiences landslides every year, especially requires regional preparation model that reflects its climate characteristic. In this study, landslide-related information such as geomorphological characteristics, soil properties, rainfall, vegetation and forest fire history of 423 sites was investigated through field survey, disaster reports and existing GIS database. Then landslide-risk assessment model was suggested by performing logistic regression analysis on them. The results of comparing calculated probabilities of landslide with the occurrence of actual landslides showed a classification accuracy of 91%.