In two locations of Ecuador with similar latitude but different altitude (Portoviejo, Coastal zone and Riobamba, Andean zone) the Holdrigde (Ho) and Hargreaves-Samani (H-S) models and three modifications of these (daily, monthly and annual with constants C HO and KE and KT, respectively) were fitted in order to estimate crop reference evapotranspiration (ETo) for a range of moving averages in time, for a three day period. Once C HO, KE and KT were calibrated those ETo estimates were compared with ETo estimated by the Pan Evaporation Method (ETo Pan). Among C HO (Ho) and KE (H-S) coefficients, the best performance was achieved relating C HO and KE to vapor pressure deficit, since it corrected the vertical thermal gradient and increased the precision in ETo prediction, while relating the KT (H-S) coefficient to air temperature allowed the correction of the solar radiation. When comparing the three modifications of both models with ETo Pan, the daily C HO model and the monthly H-S model generated slightly better ETo estimates in Portoviejo and Riobamba, respectively. Finally, in terms of quality of ETo estimates, any of the three modifications of both models were better than those with Penman-Monteith in both locations.
reference evapotranspiration; altitude; temperature; vapor pressure deficit; Hargreaves-Samani
