Document Type : Original Article
Authors
1
Water Engineering Department, Agricultural Engineering College, Sari Agricultural Sciences and Natural Resources University, Sari, Iran.
2
Researcher, Soil Conservation and Watershed Management Research Department, Mazandaran Agricultural and Natural Resources Research and Training Center, Agricultural Research, Education and Extension Organization, Sari, Iran
10.22044/jhwe.2026.16883.1081
Abstract
Evaporation is among the main components of water cycle in nature, which plays an essential role in agricultural studies, hydrology, meteorology, reservoir operation, irrigation and drainage system design, irrigation scheduling and water resources management. This study investigated various methods, including empirical equations and neural network (ANN), for four meteorological stations around Shahid Rajaei Dam in Sari during a 10-year period. The results obtained from model statistical indicators, distribution diagram, estimated daily evaporation rate and observations showed the neural network method could estimate daily evaporation at the four studied stations with good accuracy. However, the best structure of neural network models was selected for the four stations of Soleiman Tangeh, Sari office, Farim Sahra and Telamadre with seven input variables, one hidden layer and 12, 8, 10 and 12 neurons, respectively, based on MSE and R2. Correlation coefficients of daily data at Soleiman Tangeh, Sari office, Farim Sahra and Telamadre stations were obtained as 0.88, 0.91, 0.92 and 0.89, respectively. Moreover, results of monthly evaporation simulation revealed ANN could estimate monthly evaporation at Soleiman Tangeh, Sari office, Farim Sahra and Telamadre stations with good accuracy by obtaining correlation coefficients of 0.98, 0.98, 0.99 and 0.99 at the confidence level of 95%, respectively. Results of evaluating the empirical equations for evaporation calculation showed Meyer’s method was the best for estimating daily evaporation at the studied stations, followed by Ivanov equation. MSE, R2 and EF values of monthly evaporation estimation were obtained as 436.9 (mm day⁻¹)², 0.97 and 0.3 at Soleiman Tangeh station, respectively, which was the closest station to Shahid Rajaei Dam using Meyer’s method. Among the empirical methods used in this study, USBR was determined as the poorest equation for estimating monthly evaporation at Soleiman Tangeh station. Finally, results of daily, monthly and annual evaporation simulation indicated ANN could accurately estimate evaporation at Soleiman Tangeh, Sari office, Farim Sahra and Telamadre stations with a high correlation coefficient compared to empirical equations.
Keywords
)