Development of Boosted Machine Learning Models for Estimating Daily Reference Evapotranspiration and Comparison with Empirical Approaches

Proper irrigation scheduling and agricultural water management require a precise estimation of crop requirement. In practice, reference evapotranspiration (ETo) is firstly estimated, used further to calculate the each crop. this study, two new coupled models were developed for estimating daily ETo. Two optimization algorithms, shuffled frog-leaping algorithm (SFLA) invasive weed (IWO), on an adaptive neuro-fuzzy inference system (ANFIS) develop implement novel hybrid (ANFIS-SFLA ANFIS-IWO). Additionally, four empirical with varying complexities, including Hargreaves–Samani, Romanenko, Priestley–Taylor, Valiantzas, compared models. The performance all investigated was evaluated using ETo estimates FAO-56 recommended method as benchmark, well multiple statistical indicators root-mean-square error (RMSE), relative RMSE (RRMSE), mean absolute (MAE), coefficient determination (R2), Nash–Sutcliffe efficiency (NSE). All tested in Tabriz Shiraz, Iran studied sites. Evaluation results showed that yielded better than classic ANFIS, ANFIS-SFLA outperforming ANFIS-IWO. Among models, generally Valiantzas model its original calibrated versions presented best performance. terms complexity (the number predictors), obviously enhanced by increasing predictors. most accurate study sites achieved via full predictors, within 0.15 mm day−1, RRMSE 4%, MAE 0.11 both high R2 NSE 0.99 test phase at