Assessment of shallow landslide susceptibility using an artificial neural network

Landslides are one of the most common and damaging natural hazards in mountainous areas. However, due to the complexity of the mechanisms that trigger landslides, it is very difficult to predict exactly when a landslide will occur. Therefore, research on landslide prevention and mitigation has mainly focused on predicting the locations of unstable slopes that are prone to landslides and are affected by multiple external forces. The prediction of the spatial distribution of unstable slopes, termed landslide susceptibility zonation, is important for government land-use planning and for reducing unnecessary loss of life and property. In this study, we investigated unstable slopes and established a GIS- and artificial neural network- (ANN-) based method to predict areas of potential landslides in the Silurian stratum in the Enshi region in China. Using failure mechanism analysis of typical existing landslides in the Silurian stratum, an evaluation index system was created that represents the most relevant factors influencing slope stability; an ANN was used to develop a susceptibility model to predict the spatial distribution of unstable slopes prone to landslides. The results were validated by remote sensing data and field investigations. This research proves that the proposed method for predicting the location of unstable slopes based on intelligence theory and GIS technology is accurate and reliable.