Toward Reliable Precipitation Estimation in Data-Scarce Semi-Arid Regions: A Multi-Source Evaluation and Bias Correction Study over Morocco

Reliable precipitation monitoring is crucial for hydrological and water resource management, particularly in semi-arid regions like Morocco where ground-based observation networks remain sparse. In this context, the present study evaluates and validates three distinct precipitation products, specifically the soil moisture-derived SM2RAIN-ASCAT, the reanalysis-based ERA5, and the satellite-based CHIRPS, against observed data from 36 synoptic stations distributed across Morocco's diverse climatic zones over the period 2007–2022. Different results for the different timescales have emerged, with performance differing markedly by both source and temporal scale. Despite strong detection capabilities, ERA5 showed the strongest overall performance, achieving the highest correlation and probability of detection (POD) throughout the study domain. In contrast, SM2RAIN-ASCAT exhibited a systematic overestimation, while CHIRPS showed a widespread tendency toward underestimation. Based on daily assessments, all products showed poor accuracy and elevated false alarm ratios, in particular during the dry summer months (JJA), where convective and sparse rainfall makes precise satellite retrieval especially challenging. A Quantile Mapping (QM) bias correction methodology was applied to address these disparities and improve the quantitative reliability of each dataset. The correction revealed that, particularly at the monthly and seasonal scales, the explained variance (R²) for ERA5 and CHIRPS increased significantly (R² > 0.6), indicating a tighter alignment with ground observations. While SM2RAIN-ASCAT showed improved consistency following correction, it remained less reliable than ERA5 and CHIRPS in representing temporal dynamics across the study area. Finally, these results highlight that bias-corrected ERA5 and CHIRPS are the most dependable sources for hydrological applications in Morocco, while bias-corrected SM2RAIN-ASCAT stands as a particularly valuable alternative for monthly assessments in data-scarce environments where soil moisture-based retrieval offers a distinct and independent estimation pathway.