Integrating Rainfall Return Periods in MCDA-Based Flood Risk Mapping: A Fuzzy-AHP Case Study in an Ungauged Watershed

Flooding is one of the most devastating hydrological disasters, severely impacting human lives and the environment. Effective flood risk analysis is crucial for mitigation, as it identifies areas at higher risk of flooding. One common approach is to combine Multi-Criteria Decision Analysis (MCDA) with Geographic Information Systems (GIS), which allows decision-makers to map vulnerable areas even when observational data are limited. However, previous studies often neglected the probabilistic nature of extreme events. This study aims to fill the gap by incorporating rainfall return periods into the Fuzzy Analytic Hierarchical Process (Fuzzy AHP), a popular MCDA method, to evaluate its impact on flood risk mapping. The framework considers rainfall scenarios together with key factors that affect flooding. These factors include elevation, slope, river density, distance to rivers, Topographic Wetness Index (TWI), soil type, land use/land cover, population density, female ratio, poverty ratio, and road density. Six rainfall return periods (2, 5, 10, 25, 50, and 100 years) with three distinct intensity-duration patterns are included in the analysis. In total, eighteen rainfall scenarios were generated by combining short-duration–high intensity, moderate-duration–moderate intensity, and long-duration–low intensity events. Including rainfall return periods gave a more balanced view of flood risk factors, with rainfall, elevation, and slope showing the strongest correlations (± 0.7). Validation with Sentinel-1 SAR data showed that by incorporating rainfall return periods into Fuzzy AHP, produced a more robust result. Over 90% of flooded pixels in the Sentinel-1 SAR imagery were correctly classified as the three highest risk classes: Moderate to High, High, and Very High. In contrast, models that did not embed the rainfall return periods misclassified more than 70% of flooded pixels into lower-risk classes. Our findings highlight the importance of considering rainfall return periods for accurate regional flood risk assessment.