Tiltmeters as Tools for Characterizing Geometrical and Hydrodynamical Properties of Fractured Crystalline Aquifers and Fault Zones

Abstract:

In many geological reservoirs, open fractures or fault zones generally induce high spatial variability of hydrodynamical properties and shape the main deep-seated flow paths. It is of crucial interest to determine their structure and properties in order to achieve a sound and sustained exploitation of resources or to estimate the risk of failure of any underground storage. Tiltmeters have emerged as new tools to observe deformation generated by groundwater flow. As such instruments are highly sensitive to pressure gradients, they are perfectly suited for monitoring channelized flow in connected fractures and fault zones. Hence, they provide a unique insight of these reservoirs' geometry and dynamics over broad time scales. Here we demonstrate that continuous tilt data from surface long baseline tiltmeters (LBT) can be used alone to evaluate the general functioning of a fractured hardrock system and estimate the hydraulic properties of its main conductive features. The study is applied to the pumping site of Ploemeur observatory (Brittany, France) which is well documented and instrumented, and therefore forms a convenient setting for introducing LBT as tools for fractured media hydrology. On the short term, tilt signals are strongly correlated with pumping cycles and associated head level changes in well-connected boreholes. Besides, when pumps are stopped the maximal tilt direction is systematically perpendicular to a subvertical fault zone whose azimuth of strike has thereby been refined down to degree precision. By using a semi-analytical model of deformation, we establish the link between tilt and pressure change during pumping interruptions which then allows for hydraulic properties estimation from tilt measurements only. Finally, we validate our results with previous estimates obtained from other studies and discuss the orientation of future work that could enhance these estimates.