Assessing cyclone-induced flood risk in the Sundarbans using Delft3D and multicriteria analysis

ABSTRACT Tropical cyclones and associated flooding are among the most destructive natural hazards globally. The Sundarbans, located along the southwestern coast of Bangladesh and directly facing the Bay of Bengal, is highly exposed to both hazards. Cyclone-induced flooding in Sundarbans poses severe risks to ecological integrity, coastal livelihoods, and regional resilience. Accurate assessment of cyclone-induced flood risk is challenging, as widely used global digital elevation models (DEMs) are often distorted by dense mangrove canopy, overestimating ground elevation and underestimation of flood hazard. This study addresses this limitation by developing two regional DEMs, a contour-based DEM and a survey-derived DEM from spot heights, to improve terrain representation and evaluate their influence on flood-risk mapping in low-lying coastal environments. These DEMs were integrated with a physics-based Delft3D storm surge model and an analytic hierarchy process (AHP)-based GIS framework to produce and compare two cyclone-induced flood-risk maps. Cyclone Bulbul, a recent severe cyclone that tracked across the Sundarbans, was used as a reference event. The Delft3D model achieved high accuracy (R2 0.85), with simulated surge heights closely matching observations. Flood-causative parameters included elevation, slope, normalized difference vegetation index (NDVI), precipitation, storm surge height, drainage density, and proximity to coastline. The contour-derived DEM indicates that ∼96% of the Sundarbans falls within moderate to high flood-risk zones, whereas the survey-derived DEM identifies ∼67% in these categories. Both assessments highlight the eastern Sundarbans as a multihazard hotspot. The study demonstrates that DEM accuracy and terrain representation substantially influence flood-risk delineation, with even subtle elevation differences redistributing risk classes in low-relief mangrove area. By coupling hydrodynamic surge modeling with ecologically relevant spatial indicators and improved terrain data, this study provides an integrated framework for cyclone-induced flood-risk assessment and emphasizes the importance of high-resolution elevation data for reliable mapping and climate-resilient management of mangrove-dominated coasts.