Evapotranspiration (ET), as a major component of the hydrological cycle, is an important issue in water resources management. In most indirect methods to calculate ET, it is necessary to estimate at first the reference evapotranspiration (ETO) at first. In this study, three different meteorological data management scenarios (temporallocal, temporal-local-external and temporal-external) were examined to estimate the ETO by employing adaptive neurofuzzy inference system (ANFIS), FAO-Peman-Montieth (FPM) and three empirical models (Hargreaves, Priestley-Taylor and Makkink) under semi-arid conditions. In this investigation, 18 synoptic weather stations data (16 stations around the Shiraz city in south west, one station in center and another station in the north east of Islamic Republic of Iran) were used. Shiraz station was assumed as the main station. The results showed that Hargreaves model with temporal-local scenario produced an acceptable ETO estimation (NRMSE= 21.5%). With the temporal-local-external scenario, stations located at the distance less than 150km from the main station, all models resulted in an acceptable ETO estimations (10%
