Evaluation of empirical methods for estimating reference evapotranspiration in Central High Lands and Arid Western Lowlands of Eritrea

FAO Penman-Monteith (FAO56-PM) method remains difficult to implement across Eritrea due to severe shortages of standardized meteorological data. This study evaluated the accuracy of five alternative empirical methods by comparing them with the FAO56-PM model using established performance metrics (R², RRMSE, NSE, %MBE, and MAPE). Cumulative Performance Index (CPI) was used to determine the overall performances of five alternative ETo methods. The study identified the modified Hargreaves-Samani (CPI=3.6), Romanenko (CPI=3), and Schendel (CPI=2.6) methods as the most viable simplified alternatives for the data-scarce Central Highlands. However, no method proved optimal for the Arid Western Lowlands. Hargreaves-Samani and Blaney-Criddle methods performed poorly, with combined CPI values of 1.7 and 1.4, respectively. The findings suggest that the modified Hargreaves-Samani and Romanenko methods can effectively replace the FAO56-PM model for estimating crop water requirements in both irrigated and rainfed agricultural systems across all crop types in the Central Highlands. However, the study underscores the critical need for rigorous local calibration and validation of the Hargreaves-Samani, Blaney-Criddle, and Schendel methods to enhance their accuracy. © 2025 Elsevier B.V., All rights reserved.