ABSTRACT Flash flooding is amongst the most severe natural hazards, causing widespread socioeconomic impacts across both wet regions and drylands. In ungauged mountainous basins, effective risk warning based on hydrodynamic modelling is challenging due to sparse hydrological observations and complex terrain. Rainfall forecasts can enable timely alerts despite the computational demands of modelling, but their inherent uncertainties further complicate predictions. This study explores the potential of a Flash Flood Risk Index (FFRI), which integrates hydrodynamic simulation outputs with socio‐economic exposure and vulnerability indicators, to provide actionable early risk signal under data‐scarce conditions. The 2022 Datong flash flood in China is used as a case study. Grid‐based hydrodynamic simulations were conducted across varying key parameters and rainfall scenarios to discuss model uncertainty. Model performance was evaluated using UAV‐derived inundation extents, achieving high F1 scores (0.88–0.90), indicating reliable reproduction of flood extents. Simulated water depths and river discharges, however, exhibited substantial discrepancies, particularly in downstream convergence zones, which highlights the critical influence of parameter and rainfall uncertainty on hydrodynamic outputs. The FFRI proposed in the study mitigated these uncertainties, consistently identifying high‐risk areas, especially at the administrative (village) scale across all scenarios. These findings demonstrate that, in data‐limited basins, integrating hydrodynamic modelling with socio‐economic indicators is a practical way to provide actionable risk signal, supporting early‐warning and emergency response where traditional calibration and detailed observations are unavailable.
Guo et al. (Mon,) studied this question.