Sensitivity of flood-prone areas to extreme rainfall using AHP and fuzzy AHP: A case study of Boussellam and K'sob watersheds, Algeria

ABSTRACT Mapping flood-prone areas’ sensitivity to extreme rainfall is crucial for disaster risk reduction. Using the analytical hierarchy process (AHP) and fuzzy AHP methods with Geographic Information Systems (GIS) data, including factors like slope, elevation, precipitation, land use, and vegetation indices, enhances disaster preparedness. This approach enables creating maps that support informed decision-making for community resilience. In this study, AHP and fuzzy AHP models were applied across different return periods (2, 5, 10, 100, and 1,000 years) and average annual rainfall, revealing that high-risk areas vary by model: the AHP model showed high susceptibility in 2.19–25.55% of the area, while the fuzzy AHP model identified 1.53–24.51% as very sensitive. These risks are concentrated in the low-slope zones of the Boussellam and K'sob watersheds, posing a threat to three out of four nearby cities for 100-year or longer events. The validation using ROC curves confirmed the reliability of the susceptibility maps, with AUC values of 0.86 for AHP and 0.87 for fuzzy AHP, demonstrating strong predictive performance.