Hydrogeological characterisation and modelling of deformation zones and fracture domains, Forsmark modelling stage 2.2

The work reported here collates the structural-hydraulic information gathered in 21 cored boreholes and 32 percussion-drilled boreholes belonging to Forsmark site description, modelling stage 2.2. The analyses carried out provide the hydrogeological input descriptions of the bedrock in Forsmark needed by the end users Repository Engineering, Safety Assessment and Environmental Impact Assessment; that is, hydraulic properties of deformation zones and fracture domains. The same information is also needed for constructing 3D groundwater flow models of the Forsmark site and surrounding area. The analyses carried out render the following conceptual model regarding the observed heterogeneity in deformation zone transmissivity: We find the geological division of the deterministically modelled deformation zones into eight categories (sets) useful from a hydrogeological point of view. Seven of the eight categories are steeply dipping, WNW, NW, NNW, NNE, NE, ENE and EW, and on is gently dipping, G. All deformation zones, regardless of orientation (strike and dip), are subjected to a substantial decrease in transmissivity with depth. The data gathered suggest a contrast of c. 20,000 times for the uppermost one kilometre of bedrock, i.e. more than four orders of magnitude. The hydraulic properties below this depth are not investigated. The lateral heterogeneity is also substantial but more irregular in its appearance. For instance, for a given elevation and deformation zone category (orientation), the spatial variability in transmissivity within a particular deformation zone appears to be as large as the variability between all deformation zones. This suggests that the lateral correlation length is shorter than the shortest distance between two adjacent observation points and shorter than the category spacing. The observation that the mean transmissivity of the gently-dipping deformation zones is c. one to two orders of magnitude greater than the mean transmissivities of all categories of steeply-dipping deformation zones may be due to the anisotropy in the stress field, where the maximum stress is horizontal and has an azimuth of c. 140 deg. The hypothesis is supported by the deformation zones that strike WNW and NW. These two categories of steeply-dipping deformation zones have, relatively speaking, higher mean transmissivities than steeply dipping deformation zones in other directions. Key hydrogeological aspects of the fracture domains modelled are: We find the geological division of the bedrock in between the deterministically deformation zones to fall into six fracture domains useful from a hydrogeological point of view. In fact, the suggested division is consistent with the hydrogeological modelling approach reported for modelling stage 1.2. The key aspect for Forsmark is that the corrected conductive fracture frequency for the potential fracture domain FFM01 shows very strong variations with depth, and so it is suggested that the Hydro-DFN be split into three layers: above the elevation -200, between the elevations -200 and -400, and below the elevation -400. FFM01 is also very anisotropic, being dominated by the HZ set, and only with a small contribution from the NE and NS sets. The top layer of fracture domain FFM01 is similar to the Hydro-DFN parameters for fracture domain FFM02. FFM03 has less variation with depth and is comparable to the middle section of FFM01, but is more isotropic. Data for fracture domain FFM06, which is also a part of the potential target bedrock, will be treated in modelling stage 2.3. Pending this information, it is envisaged that fracture domain FFM06 can be modelled in the same fashion as fracture domain FFM01. Fracture domains FFM04 and FFM05 lie in the periphery of the candidate area. Based on the statistical analysis, FFM05 seems to be similar to FFM03, while FFM04 is of slightly higher hydraulic conductivity, but the statistical significance of the data for these fracture domains is very limited, being based on about 120-150 m of borehole data. It is proposed that fracture domains FFM04 and FFM05 are assumed to have the same properties as FFM03. Finally, comments and recommendations are made in the report as a guidance for several aspects in forthcoming hydrogeological discrete fracture network and groundwater flow models. The comments and recommendations address the following matters: fracture set definitions reflecting observations made for all boreholes in modelling stage 2.2, i.e. the geological DFN results reported for modelling stage 2.2, semi-deterministic DFN modelling of so called possible deformation zones below the elevation -400 m RHB 70 in the potential target volume, and fracture domains outside the candidate area where there are no cored boreholes. For the conclusions drawn in the work reported here these three matters are of minor importance