Performance of Low-Complexity Hydrodynamic Models for Dam-Break Flood Mapping: Trade-offs between Full Momentum, Diffusive-Wave, and Local-Inertial Models

Flood management under dam-break scenarios requires skillful and computationally efficient models for rapid decision-making. Full momentum solvers are the most conceptually accurate approaches for highly variable flows, but are computationally demanding. A trade-off emerges when comparing full momentum models with simplified shallow water equation solvers. This study presents an extensive performance comparison between different model configurations under dam-break scenarios. The Algodões Dam, located in Brazil, was selected as a representative case study, where the performance of the HEC-RAS 2D full momentum and diffusive models was compared with the local-inertial HydroPol2D model. Additionally, DEMs with high (5-meter, LiDAR) and low (30-meter, satellite radar) spatial resolutions were assessed to understand their impact on simulations. Performance indicators were assessed for maximum flood depths and flood extent maps over time, with results indicating a Critical Success Index (CSI) consistently above 0.9. Sensitivity analyses indicated that dam height is the key geometric parameter influencing peak discharge and time to peak, while Manning’s roughness had a limited impact on flood extent except in low-resistance areas. To further explore trade-offs, ten dams with diverse physiographic settings were simulated using HydroHP-1D and HydroPol2D. Results revealed that HydroHP-1D tends to produce higher flood depths and faster times to peak compared to the two-dimensional HydroPol2D model. These findings demonstrate that simplified hydrodynamic models can provide timely and sufficiently reliable flood predictions to support disaster risk reduction efforts, particularly in dam-break scenarios where rapid decision-making is critical, making them valuable tools for the development of flood emergency action plans.