Assessment of three satellite precipitation products for hydrological studies in a data-scarce context: Ouarzazate basin, southern Morocco
Khalid En-Nagre, Mourad Aqnouy, Abdelmounim Bouadila, Chaimaa Et-Takaouym, Morad Chahid, Ismail Bouizrou, Ismail Hilal, Jamal Eddine Stitou El Messari, Aqil Tariq
Abstract
Arid and semi-arid regions, such as southern Morocco, face challenges due to limited precipitation data availability, posing significant obstacles for climate and hydrological studies. To bridge this data gap, this study directly evaluates the reliability of Satellite Precipitation Products (SPPs) from the latest version (V06) of the 3-hourly Integrated Multi-satellite Retrievals (3IMERG-V06) for Global Precipitation Measurement (GPM), The Tropical Rainfall Measuring Mission (TRMM-3B42-V7), and Climate Hazards Group Infrared Rainfall Station (CHIRPSV2.0). The assessment involves comparing ground-based measurements in the Ouarzazate basin to estimate precipitation characteristics. Additionally, SPPs are indirectly evaluated for their ability to simulate runoff using two hydrological models (GR4J and HEC-HMS). The analysis employs the point-to-pixel method, and various performance criteria, including R 2 , NSE, PBIAS, RMSE, and Distance between Simulation and Observation Indices (DISO), are utilized for result comparison. Findings indicate that all three products struggle to accurately replicate daily precipitation amounts, with R 2 values < 0.5 and positive PBIAS values suggesting precipitation overestimation. However, when evaluating hydrological performance, the HEC-HMS model correctly simulates the observed flow when using ground-based data, with NSE (R 2 ) values of 0.76 (0.63) during the validation phase. In contrast, the GR4J model performs poorly, with NSE (R 2 ) values of 0.30 (0.45). When the three SPPs forced both models, HEC-HMS acceptably simulates flows using TRMM data, with NSE (R 2 ) values of 0.57 (0.56) during the calibration phase and 0.63 (0.67) during the validation phase, along with positive PBIAS values below 15.30%. Additionally, flow irregularities and outliers affect simulation results in continuous mode, which can lead to biased estimates. However, the hydrological evaluation method is considered a complementary approach to assess the performance of SPPs, particularly in arid regions where data is scarce.