This study develops and compares a lumped and a distributed hydrological model for the Loukkos river basin in northern Morocco, using the HEC-HMS platform with the Soil Moisture Accounting (SMA) loss method. The Loukkos basin (3740 km2) is prone to recurrent flooding and requires reliable rainfall-runoff simulation tools for early warning purposes. Both models were calibrated and validated on independent periods containing documented flood events. A sensitivity analysis identified Soil Storage, Tension Storage, and Maximum Infiltration as the most influential SMA parameters. The lumped model achieved a Nash-Sutcliffe Efficiency (NSE) of 0.537 during calibration and 0.624 during validation, while the distributed model reached 0.749. Despite higher computational costs, the distributed model captures spatial variability in precipitation and land cover more effectively. The sub-basin-level comparison indicates that current performance differences are primarily attributable to calibration constraints rather than to fundamental modelling limitations. Results suggest that further calibration refinement, supported by improved spatial data, can enhance flood forecasting capabilities in the basin.
Sebari et al. (Thu,) studied this question.