Daily reference evapotranspiration estimation under limited data in eastern Africa

The objective of this study was to evaluate the FAO-56 Penman-Monteith (FAO-PM) reference evapotranspiration (ETo) equation and two Valiantzas equations for estimating daily reference evapotranspiration under limited data and four other ETo equations across Tanzania and the southwestern Kenya. The results showed the applicability of the FAO-PM equation under missing solar radiation (Rs), relative humidity (RH), and wind speed (u2) data with regression slopes varying from 0.68 to 0.89, from 0.79 to 1.00, and from 0.79 to 0.96, respectively, and root mean squared error (RMSE) lower than 0.63, 0.53, and 0.44 mm/day under the respective conditions. Under lacking relative humidity data, the simplified method provided very good ETo estimates. There were large discrepancies in ETo estimates with the FAO-PM equation when two or three weather variables were missing. The Valiantzas 2 equation with full data provided the most accurate ETo estimates under the eastern Africa conditions with coefficient of determination R²>0.97, regression slope ranging from 0.96 to 1.05, RMSE<0.23 mm/day, MBE ranging from -0.03 to 0.17 mm/day, and very low relative error (RE<5.5%). The Irmak, Abtew, Hansen, and Hargreaves equations produced moderately accurate ETo estimates with RMSE as high as 0.91, 0.74, 0.74, and 0.66 mm/day for the respective equations, and relative error as high as 16.3%.