Transport and retention from single to multiple fractures in crystalline rock at AEspoe (Sweden): 2. Fracture network simulations and generic retention model

摘要

Hydrogeologic characterization of crystalline rock formations on the field scale is important for many applications but still presents a multitude of challenges. In this work we use comprehensive hydrostructural information and present a detailed simulation study of flow and advective transport in a discrete fracture network (DFN) that replicates the Tracer Retention Understanding Experiments (TRUE) Block Scale rock volume at the AEspoe Hard Rock Laboratory (Sweden). Simulated water residence time r and hydrodynamic retention parameter β are used as independent constraints for estimating material retention properties as presented in paper 1 of this series, whereas simulated mean water residence times are compared with observed values. We find that the DFN simulations reproduce water residence times reasonably well, indicating that the characterization data are sufficient and that the DFN model does capture dominant features of the flow paths analyzed. The empirical quadratic law that relates aperture and transmissivity seems to better reproduce calibrated mean water residence times than the theoretical cubic law for the five flow paths. The active specific surface area (β/τ) [1/L] as inferred from simulations is used for defining a generic retention model for the dominant rock type (AEspoe diorite) that matches fairly well the entire range of calibrated retention parameters of the TRUE tests. The combination of paper 1 and this work provides a general, comprehensive methodology for evaluating tracer test results in crystalline rock where a comparable amount of information is available; critical to this methodology is that tracer tests are carried out using tracers with sufficiently different sorption affinities (of factor 10-100).