Multi-criteria decision-making under comparison: benchmarking and optimal model selection for flood susceptibility...

Area of Science:

• Environmental Science
• Geographic Information Science
• Natural Hazard Assessment

Background:

• Flooding poses significant risks to low-lying areas like Feni, Bangladesh, impacting lives and infrastructure.
• Monsoon rainfall exacerbates flash and riverine flood susceptibility in the region.
• Effective flood risk assessment is crucial for sustainable development and community resilience.

Purpose of the Study:

• To assess flood-causing factors and create susceptibility maps for Feni district.
• To compare the performance of eight different multi-criteria decision-making (MCDM) models for flood susceptibility mapping.
• To identify the most accurate MCDM model for flood risk assessment in data-limited regions.

Main Methods:

• Utilized twelve hydro-meteorological factors including elevation, rainfall, and land use.
• Applied eight MCDM methods: AHP, FAHP, ANP, DEMATEL, TOPSIS, VIKOR, Entropy, and EDAS.
• Validated model performance using Receiver Operating Characteristic (ROC) curves, DeLong bootstrap tests, sensitivity, and confusion matrix analyses.

Main Results:

• Flood susceptibility mapping revealed that 28-62.5% of the area is moderately susceptible and 0.22-16.84% is very highly susceptible.
• The Decision Making Trial and Evaluation Laboratory (DEMATEL) model demonstrated the highest predictive accuracy with an Area Under the Curve (AUC) of 0.935.
• All tested MCDM models showed varying degrees of accuracy, with several achieving high AUC values.

Conclusions:

• The study presents a novel methodology for flood susceptibility assessment, particularly valuable for data-scarce regions like Feni.
• DEMATEL emerged as the superior MCDM model for flood susceptibility mapping in the study area.
• Findings provide crucial insights for policymakers to develop effective flood management strategies and enhance community resilience.