Estimation of Reference Crop Evapotranspiration Using ANN and ANFIS in Semi-Arid and Dry Climates

Estimation of evapotranspiration, which is one of the components of the hydrological cycle and also an effective parameter in irrigation planning, is very important, but often the limitation or lack of meteorological data prevents to use of the proposed method of FAO Penman-Monteith. In the present study, Artificial Neural Network (ANN) and Adaptive-Network-Based Fuzzy Inference System (ANFIS) are discussed in Improvement of estimates the ETo data in the absence of data. Climatic data in this study include radiation hours, wind speed, minimum, maximum, and average air temperature, which were used as model inputs based on the FAO56-PM equation in three stances. The data were obtained from 12 synoptic stations in arid and semi-arid climates. The results of evaluating the models showed that artificial intelligence methods provide better results than experimental methods. Also, both ANN and ANFIS models provide the best results with average temperature and wind speed input data, followed by minimum and maximum temperature inputs, medium temperature inputs and irradiation hours, respectively. Between ANN and ANFIS models, the ANFIS model Showed better results.