Spatial prediction of flood hazard susceptibility level in Majene urban area, West Sulawesi, Indonesia
Abstract: This study models flood hazard susceptibility in the Majene urban area, West Sulawesi, Indonesia, for 2029–2049 using a GIS-based framework. It examines how land cover transformation associated with rapid urban growth and educational infrastructure expansion influences flood risk. A coupled Cellular Automata–Artificial Neural Network (CA–ANN) model was applied to simulate land cover change, while Multi-Criteria Decision Analysis (MCDA) was used to assess flood susceptibility. Land cover projections were derived from satellite imagery (2014, 2019, 2024). Validation against observed 2024 land cover produced an overall accuracy of 80.51% and a kappa coefficient of 0.62, indicating substantial predictive performance. Flood susceptibility was evaluated through weighted overlay of six biophysical parameters: land cover, slope, elevation, rainfall, distance to rivers, and soil type. Results indicate continued expansion of built-up and plantation areas, accompanied by reductions in grassland and open land. Susceptibility maps classify the area into low, medium, and high zones. Projections suggest a gradual decline in high-susceptibility areas and expansion of medium-susceptibility zones between 2029 and 2049. This shift is associated with increased vegetation cover from plantation growth, which may enhance infiltration and moderate runoff. The findings demonstrate the dynamic interaction between urban expansion and flood susceptibility, providing spatially explicit evidence to support risk-sensitive spatial planning, disaster risk reduction, and sustainable land management in rapidly developing coastal cities.
