Application of mass-balance and flow simulation calculations to interpretation of mixing at Äspö, Sweden

The hydrogeology of a vertical fracture zone at 70 m depth at the access tunnel to the Äspö Hard Rock Laboratory was monitored over 3 a for hydrochemical changes that could be effected by construction of a deep repository for high-level nuclear waste. Tunnel construction dramatically disturbed the hydrogeological system, but this provided an opportunity to integrate hydrogeochemical and hydrological evaluation of the zone. The objective of this study was to evaluate hydrogeochemical evolution, groundwater flow and surface water intrusion during the experiment using an integrated approach of geochemical mass-balance calculations and numerical flow simulations. The dilution of major ions was the dominant hydrochemical trend. However, HCO3 and SO4 showed significant enrichment. Increasing activity of 14C suggested that oxidation of organic C was the likely source of HCO3. Any mineral source dissolving during the experiment seemed insufficient to account for changes in SO4 and current intrusion of sea water was excluded according to the data. Cation exchange as well as minor calcite reactions in fractures were assumed probable in such temporary chemical conditions. Conservative two end-member mixing models with shallow groundwater in the zone and initial groundwater at tunnel level also assumed remarkable mass transfer (several mmol/l). Therefore a third SO4-rich end-member, a regional shallow groundwater type which may mix by lateral flow in the system, was tested. This was also expected from hydraulic measurements and preliminary flow simulations assuming homogeneity. Three end-member mixing calculations using Cl and SO4 as conservative tracers give a constant proportion of lateral water in all boreholes after 300 days, which is consistent with the steady state character of the flow field in the late part of the experiment. To predict reactions on plausible levels needs significant adjustments of initial and final waters, indicating uncertainties in the hydrochemical information of the fracture zone. In the flow simulations the transmissivities were selected so that the chemical mixing proportions would match simulated portions of flow as closely as possible. The simulated total recoveries (drawdowns) differ from the measurements mainly due to overly simple parametrisation of the transmissivity in the fracture zone. However, integrating hydrochemistry in flow modelling is considered encouraging in producing additional information of the heterogeneity of a flow structure.