Search

Availability of long-term and spatially high-resolution precipitation data is important for sustainable water resource management, drought monitoring and hydrological analysis. In Eritrea, in-situ precipitation measurements are too sparse to represent the highly variable rainfall on the ground. Thus, as alternative data sources, satellite-based (CHIRPS and GPM) and reanalysis (ERA5-Land) products were evaluated against 40 stations within the study period (1992--2018). The performance evaluation was made for monthly precipitation totals along the different elevations, agroecological zones and seasons. Overall, these products captured the seasonality of precipitation over the country's different zones. GPM performed well in all zones of the country (rs > 0.42, RMSE = 44.0 mm), except for the arid highlands (rs = 0.19, RMSE = 56.4 mm) during the summer season. The performance of GPM did not exhibit a distinct pattern with respect to elevation, amount, and nature of rainfall. CHIRPS demonstrated better estimates over areas where rainfall is of a convective nature during the summer period. These areas include the moist highlands (rs = 0.4, RMSE = 47.1), arid lowlands (rs = 0.48, RMSE = 39.0), arid highlands (rs = 0.31, RMSE = 48.1) and semi-desert (rs = 0.38, RMSE = 20.1) regions. However, there were no significant correlations over the sub-humid regions in the winter period (rs = 0.19, RMSE = 25.8), as rains originate from "warm" clouds. ERA5-Land highly underestimated the summer and spring rainfalls in all parts of the country. The summer percentage bias (PBIAS) of this product ranged between -22.6 and -80.1. Therefore, GPM could primarily be considered as an alternative rainfall data source for the summer and spring seasons over Eritrea. CHIRPS can also be used as a potential data source specifically for the areas with convective rainfall during the summer and spring seasons. However, ERA5-Land needs further improvement to represent precipitation in Eritrea. [ABSTRACT FROM AUTHOR]