Landslide Risk Assessment Using a Combined Approach Based on InSAR and Random Forest

Landslide risk assessment is important for management and loss–damage reduction. Herein, we assessed landslide susceptibility, hazard, in the urban area of Yan’an City, which located on Loess Plateau China affected by many loess landslides. Based 1841 slope units mapped study area, a random forest machine learning classifier eight environmental factors influencing landslides were used susceptibility assessment. In addition, differential synthetic aperture radar interferometry (DInSAR) technology was hazard The accuracy 0.903 under receiver operating characteristics (ROC) curve 0.96. results show that 16% 22% classified as being at very high high-susceptibility levels landslides, respectively, whereas 24% high-hazard respectively. obtained based map only 26% high-risk these are mainly concentrated centers. Such zones should be taken seriously their dynamics must monitored. Our expected to provide information planners help them choose appropriate locations development schemes improve integrated geohazard mitigation City.