Transport and retention from single to multiple fractures in crystalline rock at AEspoe (Sweden): 1. Evaluation of tracer test results and sensitivity analysis

摘要

We evaluate.the breakthrough curves obtained within a comprehensive experimental program for investigating the retention properties of crystalline rock, referred to as Tracer Retention Understanding Experiments (TRUE). The tracer tests were conducted at the Aspo Hard Rock Laboratory (Sweden) in two phases jointly referred to as TRUE Block Scale (TBS); the TBS tests comprise a total of 17 breakthrough curves with nonsorbing and a range of sorbing tracers. The Euclidian length scales are between 10 and 30 m, compared to 5 m for the earlier tests TRUE-1. The unlimited diffusion model is consistent with measured breakthrough curves and is adopted here for evaluation. The model has four independent parameters, two of which are related to advection and dispersion, one which is related to diffusion-sorption, and one which is related to surface sorption; the individual retention parameters or properties cannot be inferred from breakthrough curves alone and require additional constraints. The mean water residence times for the TBS tests are in the range 15-250 h, whereas the coefficient of variation of the water residence times is in the range 0.4-0.6. A consistent trend is found in the calibrated retention parameters with the sorption affinities of the tracers involved. Using Bode sensitivity functions, it is shown that sensitivity increases for the retention parameter with increasing sorption affinity; for nonsorbing tracers, diffusion and hydrodynamic dispersion are shown to "compete," exhibiting similar effects; hence, their estimates are uncertain. The analysis presented here exposes a few fundamental limitations and sensitivities when evaluating diffusion-controlled retention in the subsurface; it is general and applicable to any site with comparable tracer test data. In part 2, it will be shown how discrete fracture network simulations based on the hydrostructural information available can be used for further constraining individual retention parameters, in particular, the active specific surface area (s_f) and the rock matrix porosity (θ).