Passive seismic monitoring uses seismic waves generated by natural sources (wind, rivers, ocean storms, etc.) to reconstruct the temporal and spatial evolution of the subsoil. The principle of noise monitoring consists in correlating the ground motion recordings between two points to bring out the Green function over successive time windows. If the medium does not change over time, the Green's function obtained by noise correlation is the same. Contrarily, if the environment changes, the speed of the waves will also change: Green's function is modified. This method has successfully been applied to monitor groundwater levels of shallow aquifers.
The objective of this research project is to monitor the groundwater in the Fontaine-de-Vaucluse catchment area (>1100 km2). Water level depths average about 700 meters, with fluctuations that can locally exceed 100 metres after intense rainfall events. The geological complexity including karst and the hydrological behaviour in the Fontaine de Vaucluse aquifer constitute a major challenge for seismic noise methods. We present here the preliminary results of an ongoing seismic survey of 20 stations distributed over the catchment area. We show the ability of our network to track the changes of groundwater content due to short term episodes of recharge.
The ambition of this project is to transform seismological velocity variations into additional constraints for hydrogeological models. Several approaches are being considered, including i/ the integration of spot measurements of groundwater table variations by intra-correlation of the different components of the same station; and ii/ the integration of spatially integrated measurements between two different seismological stations. This effort requires poroelastic modelling of speed variations and a correct knowledge of geology. This knowledge then makes it possible to specify the depth of the measured speed variations, and therefore the position of the groundwater table at our different measuring points. We then have virtual piezometers, complementary to the few real piezometers, whose water level measurements can be compared with each other and with the results of the models of flow. Thus, the contribution of new dynamic measurements from the virtual piezometers will help to specify (i) the large-scale permeability and porosity field, (ii) the hydrological impact of certain discontinuities (faults and low-permeability lithological levels, probably flow barriers) and the Banon and Sault ditches.
The authors thank TOTAL for funding this R&D project and giving permission to publish this paper.