Model for Prediction of Evapotranspiration Using MLP Neural Network

Evapotranspiration is one of the main components of the hydrologic cycle. This complex process is dependent on climatic factors. There are different methods to predict Reference crop evapotranspiration. Artificial neural networks in recent decades, and studies for modeling complex systems and nonlinear features have shown ability very high. In this study multi-layer perceptron networks (MLP) were used for estimating Reference crop evapotranspiration. By using meteorological data between 2000 and 2010 stations of Eghlid plain in Iranwas calculated the average values of evapotranspiration using the Penman-Monteith (PM). Then using these values as the output target, different networks with different structures were defined and taught. Finally, the network was evaluated to estimate evapotranspiration (Using part of the data are not used in the design of the network). By comparing the results of the ten networks was determined that MODEL 10 in the estimation of reference crop evapotranspiration is relatively more accurate.