Landslide susceptibility analysis using an artificial neural network (Ann) model: A case study in yushan national park, Taiwan

Landslide phenomenon continues to be one of the worst natural disasters around the world. The clear need for accurate landslide susceptibility mapping has led to multiple approaches. This study aims to perform landslide susceptibility analysis for the Yushan National Park (YNP) in central Taiwan based on an Artificial Neural Network (ANN) Model using Remote Sensing data and Geographical Information System (GIS). In recent years, Machine Learning (Artificial Intelligence) and Data Mining techniques have been introduced as efficient tools in hazard and susceptibility analysis. ANN is one of the commonly used not only because it can deal with complex and non-linear relationships between slope stability and conditioning factors, but also minimize subjectivity. To perform ANN analysis, besides the static (predisposing) factors of landslide occurrence including topographic slope, aspect, curvature, elevation, topographic index, distance to geological lineament, some dynamic (triggering) ones have been selected in this study such as vegetation index (NDVI) and precipitation (rainfall). All factors are analyzed with back - propagation training method to generate the landslide susceptibility map for the YNP. A landslide inventory map available for this study is used to validate the model. The results show where landslide is more likely to occur and highlight important factors that can explain the slope stability in YNP. Finally, this work can be used as a reference to assist slope failure, slope management and tourism planning considering landslide susceptibility in YNP.