Evaluation of Machine Learning versus Empirical Models for Monthly Reference Evapotranspiration Estimation in Uttar Pradesh and Uttarakhand States, India

Reference evapotranspiration (ETo) plays an important role in agriculture applications such as irrigation scheduling, crop simulation, water budgeting, and reservoir operations. Therefore, the accurate estimation of ETo is essential for optimal utilization available resources on regional global scales. The present study was conducted to estimate monthly at Nagina (Uttar Pradesh State) Pantnagar (Uttarakhand stations by employing three ML (machine learning) techniques including SVM (support vector machine), M5P (M5P model tree), RF (random forest) against empirical models (i.e., Valiantzas-1: V-1, Valiantzas-2: V-2, Valiantzas-3: V-3). Three different input combinations C-1, C-2, C-3) were formulated using 8-year (2009–2016) climatic data wind speed (u), solar radiation (Rs), relative humidity (RH), mean air temperature (T) recorded both stations. predictive efficacy evaluated based five statistical indicators i.e., CC (correlation coefficient), WI (Willmott index), EC (efficiency RMSE (root square error), MAE (mean absolute error) presented through a heatmap along with graphical interpretation (Taylor diagram, time-series, scatter plots). results showed that SVM-1 corresponding C-1 combination outperformed other Moreover, had lowest (0.076, 0.047 mm/month) (0.110, 0.063 mm/month), highest (0.995, 0.999), (0.998, (0.999, 1.000) values during validation period stations, respectively, closely followed model. Consequently, SVM) found be more robust, reliable can used promising alternative locations.