University of Tabriz Journal of Geography and Planning 2008-8078 26 80 2022 07 23 Simulations reference evapotranspiration using artificial intelligence and comparison with experimental methods (Case Study: East Azerbaijan) Simulations reference evapotranspiration using artificial intelligence and comparison with experimental methods (Case Study: East Azerbaijan) 183 171 14632 10.22034/gp.2022.13830.1815 FA Monir Shirzad Master's student in Remote sensing, Tabriz University Hajar Feyzi Master's student in water resources engineering, Tabriz University Majid Rezaei Banafsheh Professor of Department of Climatology, University of Tabriz Journal Article 2016 05 24 Introduction Reference evaporation and transpiration is one of the important elements of the hydrological cycle, which plays an important role in agricultural studies, water resource management plans, irrigation and drainage network design and water structures (Nuri et al., 2013, Volume twenty, number five, page 12). Due to the small amount of precipitation and the limitation of water resources in Iran, the correct management of water resources is very important and it is necessary to be careful in using water. Data and Method In order to carry out this research, daily climatic data during the years 2014 to 2015 of East Azerbaijan (four stations of Maragheh, Midane, Jolfa and Ahar) were prepared from the regional meteorological organization. After normalization and determination of correlation, the data were used in MATLAB software with artificial neural network method with Lunberg-Marquardt training to 70-30 combination for training and simulation. The input data for the simulation of evaporation and transpiration (temperature, sunshine hours, humidity, wind speed) and the work evaluation criteria are RMSE, R2 and MAE, which we gave priority to the data with less error. Results and Discussion In this research, the method based on artificial intelligence (ANN) and three experimental models (Penman Monteith Fau (PMF56), Blaney Kridel (B-C) and Kimberly Penman (K-P) were used to model the non-linear transpiration evaporation system of the reference plant. The results showed that the artificial intelligence method has better accuracy and speed in estimating ET0 compared to experimental methods Conclusion The results showed that the artificial intelligence method has better accuracy and speed. Also, comparing the method of artificial neural networks with classical methods, the results indicate the appropriateness of the performance of artificial neural networks. Introduction Reference evaporation and transpiration is one of the important elements of the hydrological cycle, which plays an important role in agricultural studies, water resource management plans, irrigation and drainage network design and water structures (Nuri et al., 2013, Volume twenty, number five, page 12). Due to the small amount of precipitation and the limitation of water resources in Iran, the correct management of water resources is very important and it is necessary to be careful in using water. Data and Method In order to carry out this research, daily climatic data during the years 2014 to 2015 of East Azerbaijan (four stations of Maragheh, Midane, Jolfa and Ahar) were prepared from the regional meteorological organization. After normalization and determination of correlation, the data were used in MATLAB software with artificial neural network method with Lunberg-Marquardt training to 70-30 combination for training and simulation. The input data for the simulation of evaporation and transpiration (temperature, sunshine hours, humidity, wind speed) and the work evaluation criteria are RMSE, R2 and MAE, which we gave priority to the data with less error. Results and Discussion In this research, the method based on artificial intelligence (ANN) and three experimental models (Penman Monteith Fau (PMF56), Blaney Kridel (B-C) and Kimberly Penman (K-P) were used to model the non-linear transpiration evaporation system of the reference plant. The results showed that the artificial intelligence method has better accuracy and speed in estimating ET0 compared to experimental methods Conclusion The results showed that the artificial intelligence method has better accuracy and speed. Also, comparing the method of artificial neural networks with classical methods, the results indicate the appropriateness of the performance of artificial neural networks.

https://geoplanning.tabrizu.ac.ir/article_14632_87e37dc248e1c2f74a469b31433ac824.pdf