Developing and performance assessment of a regression equation for estimating reference crop evapotranspiration based on FAO Penman-Monteith in central Iran

So far, several methods have been presented in different parts of the world to estimate reference crop evapotranspiration. The most popular and prestigious of them can be combined methods of Penman family, modified Blany-Criddle, Hargreaves - Samani and thornthwaite. Due to the lack of Lysimeter data in the most regions of Iran, providing regional equations based on local conditions with the appropriate accuracy is very important. The main purpose of this research was developing a new regional model to estimate ET0. Firstly, the main variables affecting ET0 were identified using three approaches of artificial neural network (ANN), factor analysis (FA) and multiple-regression in four main synoptic stations (Yazd, Kerman, Isfahan and Shiraz) in central Iran. The results showed that the average monthly temperature was the most influencing factor on the amount of ET0. The regression equation between monthly average temperature (Tmean) and ET0 was extracted (ET0 = -0.0118 + 0.263 Tmean) for the entire study area. The result of obtained equation was compared with FAO Penman - Monteith, Blaney - Criddle, Hargreaves - Samani and Thornthwaite methods. The extracted equation was tested and evaluated by the data of four main stations and also Tabas, Fasa, Naein and Bam stations in Central Iran. The RMSE, MAE and R2 statistics were applied to compare the methods. The obtained equation was only depended on the mean monthly temperature and had very good accuracy to estimate the ET0 in comparison with Hargreaves-Samani, Thornthwaite, and even somewhere Blaney-Cridle methods in Central Iran.