Added value of integrated satellite precipitation products for event-based flood modeling in semi-arid regions

Accurate flood simulation and forecasting are often hindered by the limited availability of high-resolution rainfall observations, particularly at sub-daily scales. This challenge is especially pronounced in developing countries such as Morocco, where dense rain gauge networks are lacking. Satellite-based precipitation products offer a valuable alternative for analyzing extreme hydrological events in poorly gauged regions. This study evaluates the applicability of three low-latency satellite precipitation products and, for the first time in Morocco, investigates the integration of two satellite products to improve flood modeling performance. The event-based MISDc hydrological model was applied using hourly streamflow observations from September 2000 to August 2015 across 20 sub-basins in southern Morocco. Flood events were identified using a 95th percentile discharge threshold, resulting in 558 events for model calibration and validation. Results demonstrate that integrated satellite products consistently outperform individual products, particularly during the validation phase. Positive hourly Kling–Gupta Efficiency (KGE) values were obtained at 14 out of 20 stations (70%), with a median KGE of 0.2. Model performance was influenced by basin physiographic characteristics, with larger basins showing improved simulations, while steeper slopes and higher elevations were associated with reduced accuracy. Overall, this study highlights the added value of integrated satellite products for sub-daily flood modeling in semi-arid and arid basins lacking hourly rainfall observations.