Study of landslide susceptibility prediction based on information value model: a case study of Ningdu area

[摘要] The landslide susceptibility prediction(LSP) is of great importance to the prevention and control of regional landslide geology disasters. Taking Ningdu County of Jiangxi province as an case, this study obtains a total of 297 landslide locations in the study area and selects ten conditioning factors (elevation, slope aspect, slope, profile curvature, plan curvature, topographic relief, distance to rivers, lithology, NDVI, NDBI). The information value model(IVM) is used to predict the landslide susceptibility and the receiver operating characteristic curve (ROC) is adopted to evaluate the prediction accuracy of IVM. The result shows that the area under ROC (AUC) value of IVM is 0.838. It can be seen that IVM has a good prediction accuracy and also obtains a reasonable distribution characteristics of landslide susceptibility.

[学科分类] 材料科学（综合）