Abstract Modelling flash floods in arid regions is challenging due to the complexity of the rainfall–runoff process and the scarcity of hydrological data. This study presents a spatially distributed hydrological model to simulate hourly streamflow using a Horton-Bauer-based technique, with remote sensing data used to derive key model parameters. The final infiltration rate and the decay rate constant in the Horton equation are determined from soil characteristics. The model was applied to several storm events to estimate streamflow, and time to peak for 18 watersheds in Gansu, which ranged from 100 to 2854 km2 in area. The watersheds are divided into 1 km × 1 km cells, with cell-to-cell flow paths determined by the D8 algorithm and a digital elevation model (DEM) to route runoff to the basin outlet. The overland flow velocity is estimated using a simplified Manning's equation. The effective rooting depth and the overland velocity coefficient, the only two parameters, are calibrated using the Shuffled Complex Evolution (SCE) algorithm. The simulated values are compared with observed streamflow, and the model's performance is assessed using the Nash-Sutcliffe Coefficient of Efficiency (NSCE), Percent Bias (PBias), and Time to Peak Error (PeakT). Catchment-wise NSCE ranged from 0.74 to 0.92, whereas PBias varied from − 3.8% to 28.5%. Similarly, the Time to Peak Error ranged from − 4.8% to 14.6%. These results indicate that the model provides overall reasonable accuracy in predicting streamflow volumes and peak times, making it a viable tool for flood forecasting in data-scarce arid environments.
Anwer et al. (Tue,) studied this question.