Comparison and validation of spatial reference evapotranspiration datasets over Africa

Reference evapotranspiration (ET0) is an important variable for water resources management and agricultural planning. Some regions, including Africa lack sufficient in-situ meteorological measurements to represent the climatic conditions. Open-access Global ET0 data sets present a viable alternative that could potentially fill the gap. This study compares eight spatial ET0 data sets against ET0 estimated from 165 weather stations across Africa. Performance was assessed using statistical metrics, including R2, Bias, RMSE, and RBias. Findings reveal that high-resolution data sets align better with in-situ data in temperate and tropical climates compared to low-resolution data sets. Results for arid regions appear to show low performance for all data sets, but results are less certain due to the availability of stations in this climate. This study also reveals that the input data contribute to 60–70% of the variability between data sets, with the remainder contributed by different model implementation, indicating the importance of good quality of input data.