Optimization and Performance TestIng of Models with increasing complexity for Karst Aquifer Resources and Scarcity AssessmenT

• Applicants: Professor Dr. Gabriele Chiogna; Dr. Ibrahim Al Khoury
• Subject Area: Hydrogeology, Hydrology, Limnology, Urban Water Management, Water Chemistry, Integrated Water Resources Management
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 565767911

Karst aquifers supply freshwater to nearly 25% of the world's population. The depletion of karst water resources is a looming concern, driven by climate change and anthropogenic pressure. Mathematical models are indispensable tools to better understand the functioning of karst watersheds and support karst water resources management. Yet, the assessment of water resources remains a challenge in karst aquifers due to their heterogeneity, duality of the recharge and discharge mechanisms, anisotropy and flow nonlinearity. Existing watershed-scale models often fail to represent interbasin groundwater flow across surface topographic divides, typically observed in binary karst watersheds, and the interplay between flow processes of the karst and non-karst units of the groundwater recharge basin of binary karst systems. Semi-distributed models combining spatially distributed recharge with the karst aquifer dominant flow components offer a promising solution to assess flow spatial variability even with limited data on the karst structure. Concurrently, the growing number of parameters and model complexity associated with this approach could result in overparameterization, equifinality and large prediction uncertainty. This project aims to develop and optimize three mathematical models (KarstMod,LuKARS,ISPEEKH) that conceptualize with increasing complexity the hydrological processes of karst systems and their dominant controls using either lumped or spatially variable approaches. We aim to create an efficient computational environment for model calibration via dimensionality reduction, sensitivity analysis, parameter optimization and uncertainty quantification using surrogate models of KarstMod,LuKARS and ISPEEKH. We also aim to explore the effect of different level of functional complexity present in the three models on recharge and flow spatial variability. We will model three karst watersheds in Europe (Kerschbaum, Ouysse and Touvre) characterized by increasing heterogeneity of the landscape properties, recharge mechanisms, discharge signatures and anthropogenic impact. A multi-model performance comparison will show the impact of integrating spatial variability of groundwater recharge and discharge on the overall hydrological simulations. We use the calibrated models to assess future water resources in the three sites under scenarios of climate change and anthropogenic pressures, generating a multi-model water resource estimate. The systematic approach adopted in this project shall improve the simulation of the hydrodynamic functioning of karst systems by providing efficient karst hydrological models and exploring the tradeoff between hydrological model complexity (lumped versus semi distributed models) and hydrological model performance. The scenario-based multi-model assessment can assist in detecting the probability of water scarcity in karst watersheds and guide the development of evidence-based policies for the sustainable use of karst water resources.

• DFG Programme: Research Grants
• International Connection: France
• Co-Investigator: Professor Dr. Andreas Hartmann
• Cooperation Partner: Professor Dr. David Labat