Résumé
In order to cope with future climatic and anthropogenic pressures on hydro systems and water resources, it is essential to propose engineering tools for recognising and managing karstic reservoirs. Karst reservoirs supply 25% of the world's population with drinking water. They, therefore, represent an important underground water resource, but one that could, in many cases, be exploited more fully, particularly on the shores of the Mediterranean. Over the last 50 years, our understanding of the processes that control their hydrodynamics has progressed enormously. This progress has made it possible to improve their modelling, but major obstacles remain. These relate to characterising their internal structure and the organisation of flows. Although karst reservoirs are particularly anisotropic and heterogeneous, an organised and hierarchical network of conduits concentrates and controls the flow towards one or more outlets. At present, however, surveillance of these conduits is limited to the use of artificial tracers that give a speed as the crow flies between a water inlet and an outlet (spring) or to exploration by diving, which can only cover a tiny proportion of the conduits. Faced with these difficulties, geophysical imaging and stochastic simulators are among the current approaches, but these methods are too poorly constrained to provide a robust topography of the conduits. The lack of more precise information on the geometry of the conduits, the flow conditions through them, or changes in water quality along the drains is a real obstacle to developing management models for karstic aquifers.