Estimation of the pan evaporation coefficient in cold and dry climate conditions via the M5 regression tree model

Abstract

In this study, class A pan coefficient (K_p) values were simulated via the M5 tree model, by using daily meteorological data of four stations in the East Azerbaijan province, which has arid and cold climate in the northwest of Iran. Firstly, the FAO-24 and FAO-56 methods, which are commonly used to calculate K_p values, were taken into consideration in the study. The K_p values calculated in the second stage were assumed to be observed values and were taken as the outputs of the M5 model. Four different training datasets consisting of 66, 70, 75 and 80% of the original data were tested. The best results were obtained when 70% of the data was used for training and 30% for testing. Results indicated that a K_p value was easily simulated with simple linear equations with high accuracy rate (R² = 0.99) in all the stations. Furthermore, the K_p value was easily simulated using only two meteorological variables (relative humidity and wind speed), without the need for complex tables and equations. The most important finding of this study was the easy estimation of the K_p with a number of linear functions obtained from the M5 model; as a result, the simulated K_p can help us to calculate evapotranspiration accurately for more effective irrigation planning. The proposed method offers advantages as it is simpler and easier than the existing approaches in the literature.