Estimation of Evapotranspiration Using Empirical Models, modeling it with Artificial Neural Network and Their Comparison with Lysimeter Data (Case Study: Salman Farsi Agro-Industry CO)

Evapotranspiration determination is a key factor for irrigation scheduling, water balance, irrigation system design and management and crop yields simulation. Reference Evapotranspiration is a complex and multivariate phenomenon that depends on climatic factors and most accurate way to estimate it is lysimeter but using Lysimeter requires a lot of time and money, hence the Evapotranspiration estimation is done by meteorological parameters and applying empirical models. These models have the coefficients that each coefficient is representative of regional conditions that equation is calibrated in that area. Nowadays Artificial Neural Networks (ANN) are being applied in several problems of water engineering where there is no clear relationship between effective parameters on the estimation of phenomenon. The purpose of this study was to evaluate Artificial Neural Networks and Experimental models in the estimation of evapotranspiration for Salman Farsi Agro-Industry. based on daily meteorological data and 3-years (March 2016 to March 2019) data from Lysimeter of the station, Evapotranspiration was calculated to above methods. The results of calculations showed that the Artificial Neural Network has better performance than all the empirical models, it has a RMSE, MAE and R2 respectively is equal to 1.25, 0.24 and 0.97 mm/day Also among the empirical models, the Penman-FAO-Monteith model with RMSE, MAE and R2 equal to 2.07 , 3.09 and 0.91 mm/day is a priority. Also, 10 scenarios were defined to evaluate the accuracy of the neural network model by reducing the climatic parameters.