Le et al. (2025) Impacts of elevation bias and topographic uncertainty on flood modeling: model robustness and floodplain sensitivity mapping in a lowland River Basin

Identification

• Journal: Journal of Hydrology
• Year: 2025
• Date: 2025-12-23
• Authors: Xuan-Hien Le, Naoki Koyama, Tadashi Yamada
• DOI: 10.1016/j.jhydrol.2025.134832

Research Groups

• Research and Development Initiative, Chuo University, Bunkyo City, Tokyo, Japan

Short Summary

This study evaluates the impact of Digital Elevation Model (DEM) quality on flood simulation and sensitivity mapping in a low-relief basin, identifying the most robust global DEMs as alternatives to high-resolution LiDAR data.

Objective

• To evaluate the impacts of DEM quality (specifically elevation bias and topographic uncertainty) on flood simulation and sensitivity mapping in the low-relief Huong River Basin, Vietnam.

Study Configuration

• Spatial Scale: Huong River Basin, Vietnam
• Temporal Scale: Sequence of storm-induced flood events in October 2020

Methodology and Data

• Models used: RRI (Rainfall-Runoff-Inundation) model
• Data sources:
  • High-resolution LiDAR DEM
  • Global DEMs: ASTER GDEM v3, AW3D30, MERIT, HydroSHEDS
  • Observed water levels for validation
  • 66 flood depth marks for validation

Main Results

• The LiDAR-based simulation achieved the highest accuracy (Root Mean Square Error (RMSE) < 0.31 m; Nash-Sutcliffe Efficiency (NSE) > 0.88).
• Among global DEMs, MERIT showed the best overall performance, closely matching LiDAR results in hydrograph and inundation depth.
• HydroSHEDS also performed well, providing stable water level simulations and coherent floodplain representation.
• ASTER and AW3D30 were affected by elevation artifacts, leading to fragmented flood extent and significant depth overestimation.
• Elevation error analysis showed ASTER with the highest RMSE (5.66 m) and MERIT with the lowest bias (1.13 m).
• DEM-induced variations propagated into flood sensitivity zoning: ASTER overestimated high-risk zones by up to 18 %, while MERIT remained within 5 % of the LiDAR baseline.
• Vertical elevation error is confirmed as a dominant source of uncertainty in flood modeling for low-relief basins.

Contributions

• Provides a comprehensive evaluation of multiple global DEMs against high-resolution LiDAR data for flood modeling and sensitivity mapping in a low-relief, data-scarce region.
• Quantifies the propagation of DEM-induced errors from flood depth to flood sensitivity zoning.
• Identifies MERIT as a viable alternative to LiDAR in data-scarce regions for flood modeling, provided its residual biases are recognized and adequately accounted for.

Funding

• [No funding information provided in the excerpt.]

Citation

@article{Le2025Impacts,
  author = {Le, Xuan-Hien and Koyama, Naoki and Yamada, Tadashi},
  title = {Impacts of elevation bias and topographic uncertainty on flood modeling: model robustness and floodplain sensitivity mapping in a lowland River Basin},
  journal = {Journal of Hydrology},
  year = {2025},
  doi = {10.1016/j.jhydrol.2025.134832},
  url = {https://doi.org/10.1016/j.jhydrol.2025.134832}
}