FLOW AND NUCLIDE TRANSPORT IN FRACTURED MEDIA - THE IMPORTANCE OF THE FLOW-WETTED SURFACE FOR RADIONUCLIDE MIGRATION

Radionuclides which migrate from a repository for nuclear waste in crystalline rock are transported by the water in a very complex fracture network. One of the main retardation mechanisms is that of uptake by diffusion and sorption in the rock matrix. The uptake is strongly influenced by the size of the contact surface between the mobile water and the rock. It is also strongly influenced by the residence time distribution of the water in contact with these surfaces. In recent years, several large-scale field experiments and observations have been made that show strong channeling effects. A considerable fraction of the water may flow in preferential pathways without good mixing with the rest of the water. This leads to the question if the traditional advection-dispersion-based equations are sufficient to describe the transport. In this paper the advection-dispersion model is compared with a pure channeling model where the different channels have different flow-rates. In the channeling model no mixing between the waters in the channels takes place. In the advection-dispersion model there is very frequent mixing. These two models would seem to form the two extremes of the properties of fracture network or channel network models. A recently developed channel network model where the individual channel members form a three-dimensional network is also used to simulate tracer transport. The channel network model has stochastically varying conductivities of the different channel members. The channel network can be thought of as connecting the channels in the channeling model and allowing the waters to mix along the flow paths. In all models the magnitude of the flow-wetted surface has a very strong impact on the arrival times for sorbing species. The time to attain a certain concentration at a point is proportional to the magnitude of the flow-wetted surface squared. Considering that the flow-wetted surface in rock is one of the entities which is poorly known and is difficult to assess, this suggests that there is a need for more field data and for methods to assess this entity. The three models need different types of data. All need the magnitude of the flow-wetted surface. The channeling and channeling network models need the conductivity distribution of the channels and channel members respectively. These data can be obtained by hydraulic tests, The advection-dispersion model needs some measure of the dispersivity. This information could be obtained by tracer tests using nonsorbing tracers. It is shown in the paper that if the hydraulic system in the rock is best described by a channel network then the use of the advection-dispersion model can give ambiguous results when used to predict transport of sorbing tracers.