مقايسه سيستم‌هاي هوش مصنوعي ANN و ANFIS در تخمين ميزان تبخير-تعرق گياه مرجع در مناطق بسيار خشك ايران

Comparison of Artificial Intelligence Systems (ANN & ANFIS) for Reference Evapotranspiration Estimation in the Extreme Arid Regions of Iran

Evapotranspiration is one of the basic components of the hydrological cycle and is essential for estimating irrigation water requirement. In recent years, the use of intelligent systems for estimation of hydrological related parameters has increased dramatically. The objective of this study was to evaluate the possibility of estimating daily reference evapotranspiration by using artificial intelligence systems, and to compare these systems together. The potential of using the adaptive neuro-fuzzy inference system (ANFIS) and the artificial neural network (ANN) techniques was investigated for estimating daily reference evapotranspiration (ET0). Various daily climatic data including daylight hours, air temperature, relative humidity, and wind speed from three synoptic weather stations (Esfahan, Kerman and, Yazd) located in the extreme arid climatic regions of Iran were used as inputs to the ANFIS and ANN techniques to estimate ET0 as calculated by the FA056 Penman-Monteith (F-P-M) equation. A comparison was made between the estimates provided by the ANN and ANFIS and those of the following empirical models: The Makkink, Priestley-Taylor, Hargreaves-Samani, FAO Blaney-Criddle, and Ritchie. Root mean squared errors, mean bias errors, and determination coefficient statistical indices were used for the evaluation of the performance of the models. Results revealed that the ANN and ANFIS techniques could be employed successfully for modeling daily ET0 process. The ANFIS85 model with three input parameters including solar radiation, maximum air temperature, and wind speed estimated ET0 with a higher degree of a accuracy than empirical models. The FAO Blaney-Criddle model was found to perform better than other empirical models included in this study. Keywords: Adaptive neuro-fuzzy inference system, Empirical models, Evapotranspiration, Extra arid climate, Iran