Optimal modelling approaches to quantify large-scale and long-term groundwater recharge and water resources in karst aquifers under Mediterranean climate.

The Mediterranean region is one of the "hotspots" of predicted shift in climate and will beaffected by increasing water scarcity in the near future. Karstified aquifers are vulnerable tochanges in the hydrological cycle due to their high hydraulic conductivity and low storagecapacity. We aim to design an optimal management strategy of carbonate aquifers underMediterranean climate and therefore we have to answer the following questions: what is theavailable quantity of groundwater, how can it be assessed appropriately, and how will thewater resource develop in the long-term? For this goal two numerical approaches for the Western Mountain Aquifer, atransboundary aquifer between Israel and the Palestinian Territories, with a size of 9,000 km2are compared: i) a process-based deterministic and ii) a stochastic numerical model. Theaquifer is developed since 1950 and constitutes today the main freshwater resource for bothcountries. HydroGeoSphere is used to simulate the hydrological-hydrogeological system witha deterministic multi-continuum approach. This concept represents characteristics of therock-soil landscape, local recharge along karst features, transmission losses of ephemeralstreams (wadis), and erratic precipitation pattern most accurately. However, due to the largenumber of hydraulic parameters simulation results are subjected to high parameteruncertainty. A stochastic single-continuum modelling concept using MODFLOW is developed toconsider parameter uncertainty, data scarcity, and the aquifer genesis. The single-continuummodel is parameterised with a stochastic prediction of the karst networks distribution.Parameterization and location of the karst features are defined with probability densityfunctions. Within the stochastic model the unsaturated zone is assigned as boundarycondition based on a spectral analysis of long-term spring discharge measurements. Thespectral analysis assesses the relation between precipitation and spring discharge to derive atransfer function that represents the entire karst system behaviour and its hydrologicalcharacteristics using a set of lumped parameters. Both numerical flow models are calibrated using an inversion of time series ofpiezometric pressure heads and spring discharge measured between 1951 – 2006. Climate projections show that during winter mean temperatures will rise in the rechargearea by up to 2˚ C increasing evaporation rates distinctively. However, total precipitation willdecrease by 20%. Climate modelling indicates that a larger proportion of rainfall will occurduring extreme events with an increase of the total amount of 10% affectingly mostly theNorthern recharge area. We therefore expect, that the impact of regional shifts in climate onrecharge rates and groundwater resources will be impressive. Therefore, based on thecalibrated groundwater models and newly developed TF the response of the groundwatersystem and recharge to changes in climate is analyzed. For this purpose, climate data from ahigh-resolution (∼8 km grid) regional climate model between 2041-2070 are currentlyinvestigated. The methods and findings of the case study are then generalized for the global transfer toother hard-rock aquifers under Mediterranean climate conditions, e.g. in California,employing an empirical hydro-pedotransfer functions (HPTFs). The newly developed HPTFswill give daily percolation rates, requiring only easy accessible input data like precipitationand potential evaporation. [ABSTRACT FROM AUTHOR]