Predicting groundwater level based on remote sensing and machine learning: a case study in the Rabat-Kénitra region

Groundwater level prediction in the Rabat-Kénitra region of Morocco has been significantly advanced using machine learning, achieving unprecedented accuracy. A study integrating satellite data from GRACE and MODIS—such as land surface temperature, soil moisture, and vegetation indices—alongside precipitation and terrestrial water storage, demonstrated that Gradient Boosting Regression (GBR) models outperformed traditional methods with an R² value of 0.99, a Mean Absolute Error (MAE) of 1.94, and a Root Mean Squared Error (RMSE) of 2.98. This breakthrough highlights the potential of combining diverse environmental datasets with advanced machine learning techniques to enhance groundwater management, particularly in water-scarce regions vulnerable to climate change. The approach offers robust predictions, addressing the limitations of conventional methods that struggle with non-linear relationships and data noise.

Elmotawakkil, A., Sadiki, A., & Enneya, N. (2024). Predicting groundwater level based on remote sensing and machine learning: a case study in the Rabat-Kénitra region. Journal of Hydroinformatics, 26(10), 2639–2667. https://doi.org/10.2166/hydro.2024.494