Egypt's water future: AI predicts evapotranspiration shifts across climate zones

Egypt is a country located in northeastern Africa. The research evaluated the random forest (RF) and extreme gradient boosting (XGB) as single models and the models' hybrid to predict the ETo for the baseline and future (2015–2099) period from Shared Socioeconomic Pathways (SSP1–26, SSP2–45 and SSP5–85) based on 18 GCMs models. The hybrid model has performed better than single models; compared RF and XGB to RF-XGB, the RMSE values were decreased in all zones esepically in zone 3 by 16.2 %, these results indicate that the highest performances of all models are observed in the middle and south Egypt, which exhibit the strongest correlation between temperature and ETo. For the SSP5–8.5 scenario, the ETo increased over the years for all zones; the ETo will increase by 4.38 %,3.71 %, 4.27 %, 2.16 %, 3.26 %, 1.35 %, 5.22 % at the year 2099 compared to the year 2015 for zone 1, 2, 3, 4, 5, 6 and 7 respectively. The T min and T max are the most critical factors that affect the ETo in all zones in the baseline and future scenarios. This study provides important insights into applying machine learning models to estimate ETo and its implications for future water management strategies. Such models hold promise for significantly enhancing regional agricultural water-resource planning and management. • The hybrid models (RF-XGB) have performed better than single models for predicting the reference evapotranspiration. • The T min and T max are the most critical factors that affect the ETo in all zones in the baseline and future scenarios. • For the SSP5–8.5 scenario, the ETo increased over the years for all zones in comparison with the baseline.