Flood Inundation Assessment on Agricultural Land: Integrating High Spatial Resolution Sentinel Data with LiDAR DEM

Abstract

Accurate flood inundation assessment is crucial for disaster management and agricultural resilience. This study integrates Sentinel-1 SAR, Sentinel-2 optical imagery, and high-resolution LiDAR DEM to develop a rapid flood mapping framework. Unlike traditional hydrodynamic models, we apply an image ratio method for flood extent detection and the Floodwater Depth Estimation Tool (FwDET) for flood depth mapping. The methodology was tested in Dien Chau district, Vietnam, where 6,917.55 ha of land was inundated, predominantly affecting paddy rice fields (6,369.05 ha submerged). The FwDET-derived flood depth map indicated a maximum depth of 3.14 m, with shallow flooding (0–0.5 m) being the most common. A Sentinel-2-based land use map achieved 93.05% accuracy (Kappa = 0.90), confirming its reliability for agricultural impact assessment. This study provides a cost-effective and scalable approach for flood mapping, supporting disaster response, land use planning, and agricultural resilience in flood-prone areas.