Spatial Modeling in Landslide Susceptibility

Abstract

Landslide Susceptibility Mapping is one of the mitigation effort to detect vulnerable areas for minimizing the risk of landslide disasters. This paper describes spatial model development for assessing landslide susceptibility by considering human and biophysical factors. The main objective of this research is to develop a spatial modeling of landslide susceptibility, particularly in several regencies of West Java Province. The data analysis include data pre-processing, regression analysis, correlation analysis, score development, and weight determination using Principal Component Analysis (PCA).  The study found that  the most important factors that contributed to landslide susceptibility within the research area is the Landuse/Landcover, then  followed by Slope, Distance to River, Soil Type, Annual Rainfall, Population Density, Geology Age, Climate Type, and Geomorphology. The first three highest factors occupied almost 53% of the total weight. The model successfully estimated the spatial distribution and degree of landslide susceptibility into 3 classes having overall accuracy of about 68%.