Analysis of Landslide Susceptibility Levels Using Artificial Neural Network Method in the Suli Watershed, Luwu Regency

Abstract

The complex geographical characteristics of the Suli Watershed, including steep slopes, high rainfall, and a vulnerable geological structure, contribute to a high potential for landslides. Therefore, a specialized approach using an Artificial Neural Network (ANN) model is required. This study aims to identify landslide events during the 2020–2024 period, analyze their influencing factors, and map landslide vulnerability in the Suli Watershed using an ANN model. This quantitative study was conducted from June 2025 to January 2026 in the Suli Watershed and at the Watershed Management Laboratory of Hasanuddin University using spatial and statistical analysis. A total of 152 landslide points were identified for the 2020–2024 period through time-series image interpretation, with the highest incidence in 2024 
(42.76%) and the lowest in 2021 (7.89%). ANN modeling indicated that slope aspect is the most dominant parameter influencing landslides. Landslide susceptibility levels were divided into five classes, with the “High” class dominating 22.80% of the area, while the “Very Low” class covered the smallest area (17.38%). Slope aspect plays a significant role in landslide susceptibility within the Suli Watershed. The resulting hazard map is expected to serve as a basis for disaster risk mitigation and management planning in the region.