An Assessment of the Effective Geofactors of Landslide Susceptibility: Case Study Cibeber, Cianjur, Indonesia

Abstract

Landslides in Indonesia cause severe disasters, significantly in populated areas and their vicinities. Landslide susceptibility assessment (LSA) is a vital tool for its mitigation. In this study, a bivariate statistical method called the Weights of Evidence (WoE) was applied to characterize landslide-prone areas on a landslide susceptibility map. The objective of this study was to estimate the spatial relationships between effective geofactors and landslide inventory data, and their relationship with landslide susceptibility. The three geofactors: slope gradient, lithology, and the landuse, together with a landslide inventory were found effective in LSA at Cibeber Sub-District, Cianjur Regency. After that a geographic information system (GIS) was applied to analyze and produce geofactor maps. The validation of the generated maps was carried out using areas under the curve (AUC). The results of modeling indicated a positive relationship between slope gradient, lithology, and land use with the occurrence of landslides, with the success rate of 81.69, 80.1 and 81.39%, respectively. Finally, the success and prediction accuracy of the WoE model for the landslide susceptibility map was 82 and 80.4%, respectively. Following this, the study area was classified into four levels of landslide susceptibility: very low, low, moderate, and high. According to the validation results, the landslide susceptibility map produced by the WoE model exhibited satisfactory properties. This landslide susceptibility map can be used for preliminary landslide hazard prevention and mitigation, as well as for proper planning of landuse and infrastructure development.