Integrating High-Frequency Monitoring and Earth Observation for Characterizing Groundwater Dynamics in Northwestern Tunisia

Groundwater in semi-arid agricultural regions is increasingly threatened by the combined effects of climate variability and intensified anthropogenic water use. This study investigates groundwater abstraction dynamics and aquifer response in the Kalaa Khasba Plain (Northwestern Tunisia) using high-frequency groundwater level observations, complemented by climate indicators and Earth observation (EO) datasets. The study period (2019–2024) captured an ongoing prolonged drought and persistent groundwater depletion in the basin. A novel event-based segmentation of high-frequency groundwater-level data was applied to identify pumping and recovery cycles from pumping induced observation. The pumping segments were used to analyze abstraction behavior across diurnal, seasonal and inter-annual scales.

The results reveal that pumping is strongly seasonal, with peak activity in July-August, and exhibits a pronounced diurnal cycle characterized by shutdowns during evening electricity peak tariff hours. Groundwater levels show a clear long-term decline, and a strong negative relationship with pumping hours, confirming that abstraction is the dominant driver of groundwater depletion in this semi-arid setting. Aquifer transmissivity and storativity were estimated by fitting multi-cycle Theis solutions to the observed drawdown-recovery sequences. This demonstrates that high-frequency groundwater monitoring can capture operational pumping significantly well and can function as a “passive” pumping test while still yielding realistic aquifer parameters, even though some non-uniqueness remains. Integration with EO data further clarifies the links between hydrological conditions and pumping behavior. ERA5-Land soil moisture exhibits robust seasonal cycles and a moderate negative correlation with monthly abstraction, while Sentinel-2 NDVI/NDWI reveal shifts in cropping and irrigation practices and lagged vegetation responses to pumping.

Overall, the study shows that high-frequency groundwater monitoring, when combined with EO, climate indicators and model results, provides a powerful and cost-effective diagnostic framework for understanding groundwater-agriculture interactions in data-scarce, semi-arid regions. The findings highlight the need for improved monitoring, better integration of ground- and satellite-based data with modeling outputs, and targeted management strategies to mitigate long-term groundwater depletion under increasing climatic and anthropogenic pressures.

Acknowledgment: This work was supported by the OurMED PRIMA Program project funded by the European Union’s Horizon 2020 research and innovation under grant agreement No. 2222, and by the project SMART Medjerda: Capacity building in monitoring for intelligent management of the Medjerda water resources, funded through the program of Wallonia Brussels International and Tunisia under grant No. 1.1.2.