GIS-aided statistical landslide susceptibility modeling and mapping of antipolo rizal (Philippines)

Abstract

Slope instability associated with heavy rainfall or earthquake is a familiar geotechnical problem in the Philippines. The main objective of this study is to perform a detailed landslide susceptibility assessment of Antipolo City. The statistical method of assessment used was logistic regression. Landslide inventory was done through interpretation of aerial photographs and satellite images with corresponding field verification. In this study, morphologic and non-morphologic factors contributing to landslide occurrence and their corresponding spatial relationships were considered. The analysis of landslide susceptibility was implemented in a Geographic Information System (GIS). The 17320 randomly selected datasets were divided into training and test data sets. K- cross fold validation is done with k= 5. The subsamples are then fitted five times with k-1 training data set and the remaining fold as the validation data set. The AUROC of each model is validated using each corresponding data set. The AUROC of the five models are; 0.978, 0.977, 0.977, 0.974, and 0.979 respectively, implying that the models are effective in correctly predicting the occurrence and nonoccurrence of landslide activity. Field verification was also done. The landslide susceptibility map was then generated from the model. It is classified into four categories; low, moderate, high and very high susceptibility. The study also shows that almost 40% of Antipolo City has been assessed to be potentially dangerous areas in terms of landslide occurrence.