Comparison of some crop reference evapotranspiration models in a cold semiarid climate to optimize the use of radiation models

The estimation of reference evapotranspiration (ETo) is of great importance due to its applications in water resource management as well as irrigation scheduling. Difficulties associated with using lysimeters have encouraged researchers to use various ETo models, while the shortage of actual radiation data seems the main obstacle for users of radiation-based models. In this research the output of four radiation-based evapotranspiration models including: Penman-Montieth-FAO56 (PMF56), Penman-Montieth FAO-Irmak (PMFI), modified Jensen-Haise (JH1), and Jensen-Haise (JH2) are evaluated for a cold semi-arid climate. The daily ETo values were generated for 16 different scenarios and the results were compared against a two-year lysimeter data during the growing season (May to November). Deviations of model results were investigated using mean of R2, RMSE, MBE and t-test criteria. The results indicated that the JH2 model which uses radiation model of Daneshyar, can generate the most accurate ETo values (R2>0.85, P<O.05) in cold semi-arid climates. Although, the Angstrom radiation models are widely being used to generate radiation data in recent years, this research showed that other radiation models can provide more accurate radiation data for radiationbased ETo models. In places with lack or shortage of meteorological data, using an accurate radiation model might significantly reduce the errors generated by certain ETo models. The mentioned radiation models and ETo models should be examined for other climates.