Evaluating Three Supervised Machine Learning Algorithms (LM, BR, and SCG) for Daily Pan Evaporation Estimation in a Semi-Arid Region

Evaporation is one of the main components hydrological cycle, and its estimation crucial important for water resources management issues. Access to a reliable estimator tool evaporation simulation in arid semi-arid areas such as Iran, which lose more than 70% their received precipitation by evaporation. Current research employs Bayesian Regularization (BR) Scaled Conjugate Gradient (SCG) algorithms training Multilayer Perceptron (MLP) model (as MLP-BR MLP-SCG) comparing performance with Levenberg–Marquardt (LM) algorithm MLP-LM). For this purpose, 16 meteorological variables were used on daily scale; including temperature (5 variables), air pressure (4 relative humidity (6 variables) input data sets, pan target variable MLP model. The surveys conducted during period 2006–2021 Fars Province region has many natural lakes. Various combinations input-target pairs tested several learning algorithms, resulting seven scenarios: (1) temperature-based (T), (2) pressure-based (F), (3) humidity-based (RH), (4) temperature–pressure-based (T-F), (5) temperature–humidity-based (T-RH), (6) pressure–humidity-based (F-RH) (7) temperature–pressure–humidity-based (T-F-RH). results indicated superiority three-component scenario T-F-RH, considerable weakness single-component RH compared others. best root mean square error (RMSE) equal 1.629 1.742 mm per day Wilmott Index (WI) 0.957 0.949 (respectively validation test periods) belonged Additionally, amount R2 (greater 84%), Nash-Sutcliff efficiency 0.8) normalized RMSE (less 0.1) all indicate reliability estimates provided In comparison between studied two BR SCG, most cases, showed better powerful common LM algorithm. obtained suggest that future researchers field consider SCG supervised numerical