Modeling monthly reference evapotranspiration process in Turkey: application of machine learning methods

In this study, the predictive power of three different machine learning (ML)-based approaches, namely, multi-gene genetic programming (MGGP), M5 model trees (M5Tree), and K-nearest neighbor algorithm (KNN), for long-term monthly reference evapotranspiration (ET₀) prediction were investigated. The input data consist of monthly solar radiation (R_s), maximum air temperature (T_max), and wind speed (W_s) derived from 163 meteorological stations in Turkey. Different input combinations were created and analyzed. The model's performance was evaluated using criteria such as Nash–Sutcliffe efficiency, Kling-Gupta efficiency, relative root mean squared error, mean absolute percentage error, and determination coefficient. Moreover, Taylor, radar, and boxplot diagrams were created. It was determined that the MGGP model outperformed both the M5Tree and the KNN models. The equation obtained from the MGGP model, for the best-performed combination of R_s-T_max-W_s, was presented. The best weather conditions were obtained as 0.029 to 31.814 MJ/m², ‑ 5.8 to 45.7 °C, and 0.140 to 5.086 m/s for R_s, T_max, and W_s, respectively. It was also found that the R_s was the most potent input variable for ET₀ estimation while W_s was the weakest.