Calibration and validation of reference evapotranspiration models in semi-arid conditions

Abstract

Reference evapotranspiration (ETo) is an important parameter for climatological, hydrological and agricultural management. The FAO56 Penman-Monteith (FAO56-PM) model is one of the most accurate models. But it needs a detailed climate dataset from weather stations. Therefore, empirical reference evapotranspiration models (ETo) that need a reduced set of climate data can become an alternative approach. In this study, nine different evapotranspiration models were calibrated for the 1978-2000 period and validated based on the period between 2001-2017 with respect to standard FAO56-PM method based on the real climatic data obtained from Aseer metrological department, Saudi Arabia. The ranking of all the evaluated models based on the multi-criteria decision making was done in order to get the best alternative to the FAO56-PM Model. The result showed that Mahringer and Trabert models are the most appropriate with RMSE values of 2.13 mm/day and 2.47 mm/day, respectively and the value of percent error were 77.27% and 89.43%, respectively. Moreover, the values of mean bias error were found to be-2.03 mm/day and-2.35 mm/day, respectively. The calibration and validation of different ETo equations tend to increase their performance. Thus, the validated evapotranspiration model that used less climatic parameters could predict the ETo condition accurately for any region.