Estimation of daily reference evapotranspiration with limited climatic data using machine learning approaches across different climate zones in New Mexico

Reference evapotranspiration (ET 0 ) is a major factor for water resource management. Although the FAO Penman–Monteith model is the highly recommended for estimating ET 0 , its requirement of a complete climatic variables has made the application of this model complicated. The objective of this study was to investigate the potential of four machine learning (ML) models, namely extreme learning machine (ELM), genetic programming (GP), random forest (RF), and support vector regression (SVR), for estimating daily ET 0 with limited climatic data using a tenfold cross-validation method across different climate zones in New Mexico. Four input scenarios, namely S1 ( T max (maximum air temperature), T min (minimum air temperature), RH ave (average relative humidity), U 2 (wind speed at 2 m height), R S (total solar radiation)), S2 ( T max , T min , U 2 , R S ), S3 ( T max , T min , R S ), and S4 ( T ave , R S ), were considered using climatic data during the 2009–2019 period from six selected weather stations across different climate zones. The results showed that the estimated daily ET 0 differed significantly following ML model types and input scenarios across different climate zones. The ML models under S1 scenario showed the best estimation accuracy during the testing stage in climate zones 1 and 5 (RMSE and MAE < 0.5 mm day −1 ). The ML models under S3 and S4 scenarios were found to be more preferred at climate zones 1, 5, and 8 (RMSE and MAE < 1 mm day −1 ). The estimation accuracy of ML models was decreased with lack of RH ave and U 2 data in input scenarios although the ML models based on S4 scenario (only T ave and R S ) showed acceptable ET 0 estimations particularly in the climate zone 5 (0.5 mm day −1 < RMSE < 0.6 mm day −1 ). The SVR and ELM were the best ML models for all input scenarios in the studied climate zones where these models showed the best stabilities in the testing stages.