Exploiting Intelligent Models for Predicting Reference Evapotranspiration in a Semi-Arid Region

In the soil- water- plant- atmosphere system, water enters to atmosphere from soil surface or plant directly. The transfer of water from the soil surface to the air is called evaporation(E), and the outflow of water from the plant is called transpiration(T), and the sum of E and T is called evapotranspiration (ETO). Estimating (ETO) plays a significant role in water resources management and irrigation scheduling. Moreover, estimating ETO in irrigation and design systems is vital to determine crop water needs. The present study is aimed at examining ETO base on Hargreaves- Samani approach in Shiraz synoptic station during the period in (۱۹۵۱- ۲۰۱۹). Additionally, the efficiency of artificial neural network (ANN), genetic algorithm-based ANN (GANN) models are investigated to determine the nonlinear relationship between the inputs and outputs variables. To find the efficiency of all models, mean square error (MSE) are calculated and compared. The results indicate that the ANN model could provide more reliable prediction over the anther model.

Keywords:

Artificial neural network , Evapotranspiration , Genetic algorithm-based ANN , Hargreaves - Samani Equation.