Neuro-fuzzy estimation of reference crop evapotranspiration by neuro fuzzy logic based on weather conditions

• Published in: Computers and electronics in agriculture 2020-06, Vol.173, p.105358, Article 105358
• Main Authors: Petković, Biljana, Petković, Dalibor, Kuzman, Boris, Milovančević, Milos, Wakil, Karzan, Ho, Lanh Si, Jermsittiparsert, Kittisak
• Format: Article
• Language: English
• Subjects: Agricultural management; Artificial neural networks; Climate models; Evapotranspiration; Fuzzy logic; Hydrology; Neuro fuzzy; Parameters; Reference crop; Regression; Regression models; Water resources management; Weather
• ISSN: 0168-1699; 1872-7107
• DOI: 10.1016/j.compag.2020.105358

•Society is becoming drastically sensitive to weather conditions and climate change.•To analyze relationships between different factors of the weather and climate change.•Reference evapotranspiration (ET0) is an important parameter for climatological.•ET0 estimation is very difficult to achieve due to too many input parameters.•Global radiation has the strongest influence on the ET0. Reference evapotranspiration (ET0) is considered and one of the most valuable parameter for hydrological, climatological investigation and water resources management as well. In this article the evapotranspiration was determined with the simplified equation of Makkink. There is need for precise approximation of the reference crop evapotranspiration in order to determine the water requirement in irrigated agriculture. However ET0 estimation is very difficult to achieve due to too many input parameters. Therefore the primary objective of the research was to establish regression models of the ET0 in regard to several input weather parameters. The regression models will be created by input/output data pairs. The main aim is to achieve predictive capable models for the ET0. Also according to the regression models precision one can determine the input parameters influence on the ET0. Hence one king of ranking process will be performed in order to select which factors have the most influence on the ET0. The repression models will be created by neuro fuzzy logic procedure since the procedure could handle high nonlinearity between input and output data pairs. According to the results Global radiation has the strongest influence on the ET0. Combination of Daily average temperature and Global radiation is the optimal combination for the ET0 estimation.