Hydrological and geophysical characterization and prediction of flow and transport processes in fractured rock pose extraordinary challenges. The discrete nature of fractured rock implies that physical properties and the geometry of individual fractures cannot be imaged using the same geophysical or hydrological data and approaches that are typically used in sedimentary environments. In a previous project, we have shown that reflections from high-frequency electromagnetic signals that are emitted and received in boreholes makes it possible to obtain important constraints about fracture geometry away from the boreholes. In a case study, we acquired such single-borehole and cross-borehole Ground Penetrating Radar (GPR) data at a well-characterized fractured rock aquifer located in Brittany, France. We then acquired five successful saline tracer injection tests in known transmissive fractures while monitoring single-borehole GPR in another borehole to track the resultant tracer movement stimulated by pumping and injection, as well as the natural flow regime. After developing a suitable processing scheme, we found for the different experiments that we could image the tracer movement over tenths of meters through a network of connected fractures.
The present one-year prolongation of the previous project focuses on developing a method to identify the geometry and properties of transmissive fractures in low to moderately fractured rock by jointly combining (inverting) different types of geophysical and hydrological data. The approach will be stochastic (it relies on Bayesian Theory), which implies that we seek many models that can explain the available data, but possibly with very different geometries. The method will be applied to the field site discussed above. The outcomes of the proposed inversion will be probability distribution functions of all model parameters (up to 100) describing possible geometries and hydrological properties of the fractures in which tracer movement occurs given available hydrological and geophysical data and a priori constraints. The application of the inversion algorithm to the existing field data will not only provide quantitative insights in the relative value of different data types in hydrological characterization of fractured rock, but also provide fundamental insights in the possible transport pathways and dispersion mechanisms that can explain hydrological data acquired in boreholes. By providing the full range of permissible parameter values, the proposed method is also suitable for risk and uncertainty assessments for a wide range of application areas related to contaminant transport, petroleum engineering, and geothermics.