Li et al. (2026) Influence of open-source topographic data on basin-scale flash flood modelling in High Mountain Asia

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2026
• Date: 2026-03-30
• Authors: Zhu Li, Yan-Fang Sang, Vijay P. Singh, Yueling WANG, Xilin Xia, Yichi Zhang
• DOI: 10.1016/j.ejrh.2026.103396

Research Groups

• Key Laboratory of Water Cycle and Related Land Surface Processes, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing, China
• Yarlung Zangbo Grand Canyon Water Cycle Monitoring and Research Station, Linzhi, China
• Key Laboratory of Compound and Chained Natural Hazards, Ministry of Emergency Management of China, Beijing, China
• University of Chinese Academy of Sciences, Beijing, China
• School of Engineering, University of Birmingham, Birmingham, United Kingdom

Short Summary

This study investigates the influence of four widely used open-source topographic datasets on basin-scale flash flood modeling in High Mountain Asia, revealing significant biases in simulated hydrological and hydrodynamic characteristics, as well as delayed early warning times, underscoring the critical need for high-precision topographic data.

Objective

• To investigate and quantitatively compare the influence of different open-source topographic data sources on the simulation of hydrological (flood extent, depth) and hydrodynamic (velocity, force) characteristics, as well as temporal characteristics (arrival time, peak time), of flash floods in mountainous areas.

Study Configuration

• Spatial Scale: Woka River basin in southern High Mountain Asia (HMA), covering an area of 1600 square kilometers. Flash flood modeling focused on a 5.92-kilometer river reach downstream from a landslide dam.
• Temporal Scale: Flash flood events simulated under three outburst flood scenarios corresponding to design rainfall with 50-, 100-, and 300-year return periods. Analysis of flood propagation at 5, 10, and 15 minutes, and temporal characteristics like flood arrival time and occurrence time of maximum indicators.

Methodology and Data

• Models used: A fully 2D hydrodynamic model based on an explicit Godunov-type finite volume scheme for discretizing shallow water equations on a Cartesian uniform grid. A GPU-accelerated Python package for multi-hazard simulations was employed.
• Data sources:
  • Topographic data:
    • 5-meter (m) measured topographic data (Unmanned Aerial Vehicle - UAV) as benchmark.
    • Four open-source Digital Elevation Models (DEMs): 30-m Advanced Spaceborne Thermal Emission and Reflection Radiometer Global Digital Elevation Model (ASTER GDEM), 30-m Shuttle Radar Topography Mission DEM (SRTM DEM), 30-m Copernicus DEM (COP DEM), and 12.5-m Advanced Land Observing Satellite Phase Array type L-band Synthetic Aperture Radar DEM (ALOS PALSAR DEM).
  • Land use data: 30-m Landsat TM image data.
  • Rainfall data: Design rainfall depths for 50-, 100-, and 300-year return periods, estimated from the "Map of Statistical Parameters of Heavy Rainfall on the Tibet Plateau".
  • Model parameters: Manning’s coefficient set at 0.032 (calibrated within 0.01 to 0.1 for sensitivity analysis).
  • Hydrodynamic force calculation: Formula P = KAγV²/2g, where P is hydrodynamic force (kilonewtons, kN), K is pier shape coefficient, A is water-blocking area (square meters, m²), γ is water volume weight (10 kilonewtons per cubic meter, kN/m³), V is flow velocity (meters per second, m/s), and g is acceleration due to gravity (meters per second squared, m/s²).

Main Results

• The 30-m ASTER GDEM data produced unreasonable simulation results and is not recommended for flash flood modeling in mountainous areas.
• At the basin scale, 30-m SRTM, 12.5-m ALOS PALSAR, and 30-m COP DEM data underestimated the maximum flood depth by -16.04% to -5.18%.
• The 30-m COP DEM data underestimated the maximum flood extent by -3.51% to -4.63%.
• Simulated maximum flood velocity and hydrodynamic force exhibited biases ranging from -16.38% to 11.92% and -99.46% to -5.87%, respectively. The maximum hydrodynamic force was consistently underestimated with the largest biases.
• At three specific infrastructure sections, the biases in simulated maximum flood depth and velocity varied considerably due to differing precision of the open-source topographic data.
• All simulated temporal characteristics (flood arrival time, occurrence time of maximum flood depth, velocity, and hydrodynamic force) showed consistent delays ranging from 1.23% to 180.22% compared to the benchmark. Temporal biases for hydrodynamic characteristics were larger than for hydrologic characteristics.
• Calibration of Manning’s coefficient offered only slight improvement in simulation accuracy and could not effectively mitigate the biases caused by the low quality of open-source topographic data.

Contributions

• Provided the first comprehensive investigation into the influence of open-source topographic data on the simulation of hydrodynamic characteristics of flash floods in mountainous areas, extending beyond traditional hydrological characteristics.
• Quantified significant biases in flood extent, depth, velocity, and critically, hydrodynamic force and temporal characteristics caused by different open-source DEMs in a data-scarce High Mountain Asia basin.
• Demonstrated that adjusting Manning's coefficient cannot effectively compensate for biases introduced by low-quality topographic data.
• Offered a theoretical foundation for high-precision flash flood simulation and early warning in data-scarce mountainous regions, providing scientific reference for disaster prevention and mitigation policy-making.
• Highlighted that these biases lead to overestimated rainfall thresholds, underestimated risk levels, and delayed early warnings, emphasizing the critical importance of reliable topographic data.

Funding

• Science and Technology Projects of Xizang Autonomous Region, China (XZ202401JD0001, XZ202501ZY0004)
• National Natural Science Foundation of China (42471029)

Citation

@article{Li2026Influence,
  author = {Li, Zhu and Sang, Yan-Fang and Singh, Vijay P. and WANG, Yueling and Xia, Xilin and Zhang, Yichi},
  title = {Influence of open-source topographic data on basin-scale flash flood modelling in High Mountain Asia},
  journal = {Journal of Hydrology Regional Studies},
  year = {2026},
  doi = {10.1016/j.ejrh.2026.103396},
  url = {https://doi.org/10.1016/j.ejrh.2026.103396}
}