Groundwater chemistry around a repository for spent nuclear fuel over a glacial cycle. Evaluation for SR-Can

The chemical composition of groundwater in the rock volume surrounding a spent nuclear fuel repository is of importance to many factors that affect repository performance. The geochemical characteristics of present-day Swedish groundwater systems are governed by successive mixing events of several waters during the post-glacial evolution of the sites. The expected development of groundwaters at two Swedish sites - Forsmark and Laxemar - during a glacial cycle has been evaluated within the SR-Can project, and the results are presented in this report. For the temperate period following repository closure, an approach is proposed here to investigate the spatial and temporal evolution of groundwater geochemistry by coupling hydrogeological and geochemical models in a sequential way. The procedure combines hydrogeological results obtained with CONNECTFLOW within the SR-Can project with a mixing and reaction path simulation using PHREEQC. The hydrological results contain mixing proportions of four component waters (a deep brine, glacial meltwater, marine water, and meteoric infiltration) at each time step and at every node of the D regional model domain. In this work the mixing fractions are fed into PHREEQC using software developed to build formatted input files and to extract the information from output files for subsequent plotting and analysis. The geochemical calculations included both chemical mixing and equilibrium reactions with selected minerals: calcite, chalcedony and an Fe(III) oxy-hydroxide. Results for the Forsmark and Laxemar sites are graphically presented as histograms and box-and-whisker plots. Cross sections, where each node is colour-coded with respect to an important variable (pH, Eh or concentrations of main elements), are used to visualize the future evolution of the site. Sensitivity analyses are made to evaluate the effects of the different reactions and/or assumptions. The results reflect the progressive inflow of meteoric waters into the sites, and indicate that far-field groundwaters during the temperate period will not affect negatively the performance of the safety functions of the repository. For the permafrost and glacial periods, groundwater compositions are proposed based on the results from hydrological evaluations performed within the SR-Can project. It is concluded that for permafrost conditions groundwaters in the rock volume surrounding a repository will not affect negatively its performance. However, under glacial conditions meltwaters are expected to penetrate deep into the bedrock and, on the whole, groundwaters would then have such low salinities that it would affect negatively the stability of the bentonite buffer surrounding the canisters in the repository. An analysis of the possibility of penetration of O2-rich meltwaters down to repository depths during glacial periods is made based on studies performed within the SR-Can project and elsewhere. It is concluded the inflow of oxygen in single fractures is neutralised by the process of matrix diffusion and dissolution of Fe(II) minerals in the rock matrix. For fracture zones, with water advective times down to repository depths of only a few years, advancement of O2-rich waters to repository depth does not occur if variables are cautiously selected. For extreme situations, and given our present understanding of fracture zones, the occurrence of oxidizing conditions at repository depths within large fracture zones can not at present be completely ruled out, but at least they can be avoided in the repository design