Landslide hazard modeling using the artificial neural network approach in the Biang Loe River watershed

The Bantaeng Regency has an extreme topography and is located in a district prone to landslides.In addition, many watersheds have been damaged in the Bantaeng Regency area, one of which is the Biang Loe watershed.This study maps landslide vulnerability using an artificial neural network (ANN) model to provide a robust predictive framework for disaster mitigation.This study was conducted to identify the distribution of landslides in the Biang Loe watershed, analyze the factors that affect the occurrence of landslides and map the level of landslide vulnerability.Analysis of 103 landslide events (2018-2022) revealed that slope direction, lithology, slope steepness, curvature, and proximity to rivers are the primary drivers of instability.The model demonstrated high predictive performance with an area under the curve (AUC) of 0.811, categorizing it as "good" for regional hazard assessment.Results show that while 25.84% of the area falls into the very low vulnerability class, critical high-risk zones were identified in areas with slopes >45% and concave curvature.These findings provide a data-driven basis for regional risk management, enabling planners to prioritize slope stabilization and restrict infrastructure development in identified high-vulnerability corridors to minimize future economic and human losses.