Approaches for delineating landslide hazard areas using receiver operating characteristic in an advanced calibrating precision soil erosion model

Abstract

Soil erosion is undesirable natural event that causes land degradation and desertification. Identify the erosion-prone areas is a major component of preventive measures. Recent landslide damages at different regions lead us to develop a model of the erosion susceptibility map using empirical method (RUSLE). A landslide-location map was established by interpreting satellite image. Field observation data was used to validate the intensity of soil erosion. Further, a correlation analysis was conducted to investigate the "Receiver Operating Characteristic" and frequency ratio. Results showed a satisfactory correlation between the prepared RUSLE-based soil erosion map and actual landslide distribution. The proposed model can effectively predict the landslide events in soil-erosion area. Such a reliable predictive model is an effective management facility for the regional landslide forecasting system.