Mapping Landslide Vulnerability using Spatial Analysis of Weight of Evidences (WoE) in Central Mamuju, West Sulawesi Province, Indonesia Ilham Alimuddin (ab*), Muhammad Ichsan(b), Cahyadi Ramadhani(c)
(a)Disaster Study Center of Hasanuddin University
Kampus Unhas, Jalan Perintis Kemerdekaan Km 10, Tamalarea
*ialimuddin[at]gmail.com
(b)Department of Geological Engineering
(c)Disaster Study Center of Hasanuddin University
Abstract
Mamuju Tengah Regency, West Sulawesi Province, is an area with a relatively high level of geological disaster risk, particularly landslides. Over the past five years, there have been several significant landslide incidents that have not only caused infrastructure damage but have also disrupted social and economic activities of the community. Spatial analysis to identify locations and map areas with high risk/hazard levels can be one of the efforts to assist in anticipating such disasters by serving as a basis for decision-making. The multivariate statistical method known as Weight of Evidences can be used to evaluate the influence of each factor on landslide hazards by overlaying landslide distribution data and then comparing it separately with various thematic data layers from eleven factors, including slope direction, elevation, distance from roads, distance from the river, slope gradient, geology, peak ground acceleration, distance from geological structures, soil type, land type, and land cover. The validation results shown in the Central Mamuju area are presented in the form of attribute tables and prediction level graphs, namely AUC (Area under Curve), which is one type of accuracy statistic for prediction models (probability) in assessing/analyzing the level of landslide vulnerability in the Central Mamuju area, which is 0.9303 (Excellent Model). It is known that the three most influential parameters in landslide events in this area are slope gradient, slope direction, and land cover.
Keywords: Landslide Vulnerability, Weight of Evidence (WoE), Central Mamuju
