Assessing the suitability of hybridizing the Cuckoo optimization algorithm with ANN and ANFIS techniques to predict daily evaporation

Estimation of evaporation is of indispensable significance for management and development of water resources. This study aims to identify the suitability of hybridizing the Cuckoo optimization algorithm (COA) with two well-known approaches of artificial neural networks (ANN) and adaptive neuro-fuzzy inference system (ANFIS) for prediction of daily pan evaporation. For this aim, two hybrid models of ANN–COA and ANFIS–COA are developed and their performances are compared with single ANN and ANFIS. As case study, the daily climate parameters including the average air temperature (Tavg), sunshine hours (S), relative humidity (Rh), wind speed (W) and pan evaporation (E) measured and collected for three Iranian stations of Zabol, Iranshahr and Shiraz have been utilized. The used data sets are divided into three parts so that 60, 20 and 20 % of the data are applied for training, testing and prediction phases, respectively. The achieved results prove that the models’ performances are variable among cities. It is found that combining the COA with ANN and ANFIS techniques does not enhance the precision of the developed ANN and ANFIS models noticeably in all considered stations. In fact, the results demonstrate that hybridizing the COA with ANN and ANFIS cannot be a viable option for estimation of daily evaporation. Overall, the study results indicate that further accuracy can generally be achieved by the ANN model; consequently, the ANN model can be sufficiently used in the prediction of daily evaporation.