Md. Mahfuzar Rahman

Cities face significant flooding risks due to dense populations and valuable assets, making comprehensive flood risk assessment crucial for effective management. This study evaluates floodi...

Flood susceptibility mapping (FSM) has become a key area of research in hydrology and disaster management, providing critical insights for identifying vulnerable zones and supporting flood ...

Flood is one of the predominant natural hazards, inducing catastrophic impacts on lives, structures, and agriculture in low-elevated and riverine areas like Feni, Bangladesh, highly susceptible to flooding (flash and riverine floods), amplified by intense and prolonged rainfall in the monsoon. This study aims to assess flood-causing factors, risk profile by susceptibility maps, contrasting multi-criteria decision-making (MCDM) models, and sensitivity analysis for the Feni district. The novel aspect includes a comparison of eight MCDM models (subjective, objective, and ranking-based) for flood susceptibility mapping using traditional and DeLong-bootstrap receiver operating characteristic (ROC) curves, sensitivity, and confusion matrix analyses. This study explores twelve fundamental hydro-meteorological factors (Elevation, Slope, Aspect, Curvature, Drainage density, Topographic wetness index, Rainfall, Normalized difference vegetation index, Distance from river, Topographic position index, Land use land cover, and Stream power index), where each factor illustrated statistical significance according to a multicollinearity test applying MCDM methods: Analytical Hierarchy Process (AHP), Fuzzy AHP (FAHP), Analytic Network Process (ANP), Decision Making Trial and Evaluation Laboratory (DEMATEL), Technique for Order Preference by Similarity to Ideal Solution (TOPSIS), VIse Kriterijumska Optimizacijai Kompromisno Resenje (VIKOR), Entropy, and Evaluation Based on Distance from Average Solution (EDAS). The results indicate that 28-62.5% of the area is moderately susceptible to flooding, 0.22-16.84% is very highly susceptible, as assessed by the accuracy using the ROC curve and the area under the curve (AUC) value. The traditional ROC analysis yielded AUC values of 0.93, 0.93, 0.93, 0.93, 0.82, 0.70, 0.92, and 0.85 for AHP, FAHP, ANP, DEMATEL, TOPSIS, VIKOR, Entropy, and EDAS, respectively. Among these methods, DEMATEL, a subjective MCDM approach, exhibited the best predictive performance, achieving the highest AUC (0.935) when further validated using the DeLong bootstrap test, sensitivity, and confusion matrix analysis with supporting data from the Feni District. This study facilitates a novel methodology for flood susceptibility assessment in data-deficient Feni and analogous regions, guiding policymakers and governments for sustainable flood management and increasing the community's resilience.