Modeling of Daily Reference Evapotranspiration in the Semi Arid Region using MLP and MLR Models

Study of reference evapotranspiration is important to find out the evaporative demand of the atmosphere, irrigation design, irrigation scheduling, cropping system model and water resource management. In this study the potential of Multilayer Perceptron (MLP) is investigated in modeling of daily reference evapotranspiration obtained using standard Penman-Monteith equation. Based on the intensive study of this paper, daily basis meteorological climatic data recorded from 2011 to 2015 were used to obtain the results. The study compares results obtained using Gradient Decent Momentum (GDM) with One Step Secant (OSS) learning algorithm,. Root mean square error, Coefficient of determination and Mean Absolute Percentage Error statics are used as criteria for evaluation of model performance. The developed GDM model for the ETo modeling shows good performance with RMSE index in the range of 0.301-0.311, MAPE between 8.982-9.749 and R 2 between 0.945-0.959 for training and testing period. Evaluated results show that MLP algorithm performs better than the MLR model.