High-Resolution LiDAR-Derived DEM for Flood Inundation and Risk Assessment in a Floodplain Area of Bangladesh

This dataset contains high-resolution UAV LiDAR, digital elevation models and aerial orthophotography covering approximately 250 km² of floodplain in and around Delduar Upazila, Tangail District, Bangladesh. The dataset was acquired between July and December 2024 using a DJI Matrice 300 RTK equipped with a DJI Zenmuse L2 LiDAR sensor and Zenmuse P1 45-megapixel camera. The dataset includes classified LiDAR point clouds (LAS), 1 m Digital Terrain Models (DTM), 1 m Digital Surface Models (DSM), 5-cm-resolution orthomosaics, LiDAR intensity, slope, aspect, hillshade, topographic wetness index, canopy-cover layers, first- and last-return surfaces, 0.25 m and 0.5 m elevation contours, survey boundaries, flight-grid boundaries, ground-control information and independent checkpoint observations. The survey was conducted over 130 flight grids at approximately 475 ft (≈145 m) above ground level, using a minimum front overlap of 75%, side overlap of 60% and approximately 10% LiDAR strip overlap. Survey control was established using static GNSS observations tied to Survey of Bangladesh reference marks, with orthometric heights referenced to EGM2008. Independent accuracy assessment using eight GNSS checkpoints produced a vertical RMSE of 0.118 m and horizontal RMSE of 0.080 m. The dataset therefore provides a high-resolution terrain reference suitable for floodplain and agricultural applications. The data are intended for flood inundation and drainage modelling, flood-depth analysis, micro-topographic assessment, agricultural land-type delineation, crop zoning, land suitability assessment, terrain classification, and validation of national and global digital elevation models. The dataset covers approximately 19 km × 21 km and uses WGS 84 / UTM Zone 46N for horizontal coordinates and EGM2008-referenced orthometric heights for vertical coordinates. Among Bangladesh's many natural hazards, flooding remains the most recurrent and damaging, and low-lying floodplains such as Delduar Upazila, Tangail District experience disproportionately greater impacts. Because effective disaster risk reduction, land-use planning, and climate-resilient management all depend on knowing where and how deeply water will collect, this work builds a high-resolution LiDAR-derived Digital Elevation Model (DEM) into a two-dimensional (2D) hydrodynamic framework and tests what that combination adds. Historical water-level records were fitted to the Log-Pearson Type III distribution to obtain design flood levels for the 5-, 10-, 50-, and 100-year return periods, and calibrated against observed data, was then used to reproduce flood depth, inundation extent, and duration. Because the LiDAR DEM resolves micro-topographic detail-flood-control embankments, drainage networks, raised homestead land that coarser, conventional DEMs simply miss, the resulting terrain surface tracked the Sentinel-1 SAR record of the 2019 flood closely. The gap between data sources was large: The National DEM put the inundated area at 16,600 ha, roughly 75% more than the 9,460 ha produced by the LiDAR-based model. The LiDAR-driven simulations also generated depth and duration patterns that were more physically consistent, which in turn made the derived hazard and risk outputs more dependable. Taken together, the actual results indicate that pairing high-resolution LiDAR terrain data with 2D hydrodynamic modelling meaningfully sharpens flood simulation and offers a firmer basis for disaster preparedness, infrastructure planning, and sustainable management of Bangladesh's floodplains.