Assessment of six satellite precipitation products in a Moroccan arid area

Precipitation is the main component of the hydrological cycle; it is a crucial source of data in hydroclimate applications for water resources management. However, several regions, especially mountainous and arid regions, suffer from limited data from a ground measurement network. Remotely sensed data may provide a viable alternative for these regions. This study aims to evaluate six high spatio-temporal resolution satellite products (GPM-F, GPM-L, GPM-E, CHIRPS, PERSIANN-CCS-CDR and PDIR-Now) in the sub-Saharan regions of Morocco during the period September 2000-August 2020. The record data from 33 rain-gauge stations was used to evaluate these products on two spatial scales (pixel and basin scales) and three temporal scales (daily, monthly and annually), adopting a quantitative and qualitative evaluation. For all examined timescales, the results showed that the GPM-F product performed the best quantitatively, while at the detection capability tested for different threshold and at daily time scale, the GPM near real-time products (GPM-E and GPM-L) were better at detecting more intense rainfall events higher than 40 mm/day. At the daily time scale, GPM-E and GPM-L and, on monthly and annual scales, CHIRPS and PERSIANN-CCS-CDR, provided satisfactory precipitation estimates. Moreover, the evaluation based on the altitudes of rain gauges revealed a bias increasing from low to high altitudes. The findings also highlight that the continental and mountainous basins showed the lowest performance compared to the other locations closer to the Atlantic Ocean. The latitude-based analysis showed a decrease of bias and increase of correlation towards the most arid zones. These results provide valuable information for a scarcely gauged and arid regions, showing that GPM-F could be a valuable alternative to rain gauges.