Dynamic coupling of hydrogeology and human decision-making for sustainable groundwater management in Konya Closed Basin, Türkiye

The Konya Closed Basin, a 50,000 km² semi-arid agricultural region in central Türkiye, with severe groundwater depletion due to intensive irrigation and uncontrolled well drilling. We developed and applied a coupled socio-hydrological model that integrates a process-based groundwater flow model (MODFLOW-UZF) with a system dynamics model (VENSIM) to simulate both hydrogeological processes and human decision-making. A Python-based coupling framework enables the bidirectional exchange of variables across different spatial and temporal resolutions. The coupled model, calibrated using historical groundwater level data from long-term observation wells, is used to test the impacts of various policy configurations, including crop pricing, extraction caps and surface water transfers. The coupled model captures both seasonal variability and long-term depletion trends more accurately than stand-alone models. We demonstrate that spatial and temporal disaggregation of policy measures can significantly enhance groundwater resource management. By incorporating feedback between aquifer conditions and farmer decisions, the model demonstrates how socio-economic and hydrological drivers jointly shape groundwater dynamics. Through this coupling, we emphasize the advantages of harnessing the respective strengths of each modeling approach to overcome the limitations of the other. These insights highlight the value of integrated modeling for sustainable groundwater governance in data-scarce, densely-irrigated regions. • A process-based hydrogeological model is coupled with a socio-economic system dynamics model. • Coupling bridges two models with different spatial and temporal resolutions. • The coupled model has enhanced capabilities for socio-hydrological policy analysis. • The model is applied to the Konya Closed Basin to assess policy interventions. • Spatiotemporal policies targeting vulnerable areas outperform aggregate approaches.