EFFECTIVENESS OF LANDSLIDE SUSCEPTIBILITY MAPPING USING THE MAXIMUM ENTROPY MODEL AND WEIGHTS OF EVIDENCE MODELLING IN THE KUNINGAN REGENCY, WEST JAVA, INDONESIA

Abstract

Kuningan is one of the regencies in the West Java region, which has had a problem with landslides every year for the last decade. In this area, there were 124 landslides recorded from 2011 to 2022. It is necessary to have extensive knowledge of the variables impacting the indicators used to geographically classify landslide susceptibility. This research attempts to create maps of landslide susceptibility based on the relationship between the parameters and inventory data of landslides. In this case, we present landslide susceptibility mapping in the Kuningan area using two methods, namely maximum entropy (MaxEnt) and weights of evidence (WoE). The results showed that for a variety of landslide susceptibility models, the two approaches generated comprehensive susceptibility distributions. Even though the two models’ AUC parameters are nearly identical, the MaxEnt approach produces maps with larger low-susceptibility zones than the WoE method, according to a comparison of the maps created using the two approaches. This research offers preliminary recommendations for zonation prone to landslides, which is helpful for spatial design. In order to create landslide susceptibility maps that are more exact, accurate, and dependable in forecasting landslide events, additional studies need to be done.

Keywords

• susceptibility
• landslides
• WOE method
• MaxEnt
• AUC