Quantitative assessment of radionuclide retention in the near-surface system at Forsmark. Development of a reactive transport model using Forsmark 1.2 data

The main objective of this work is to assess the migration behaviour of selected long-lived radionuclides through the near-surface system of Forsmark, with special focus on the evaluation of the capacity of the Quaternary deposits and sediments for radionuclide retention. The work reported here is based on data and information from Forsmark Site Descriptive Model version 1.2. From the geological point of view, the near-surface systems in the Forsmark area consist of Quaternary deposits and sediments that overlay the granitic bedrock. Glacial till is the more abundant outcropping Quaternary deposit and the remainder is made of clayey deposits. These types of near-surface sediments show distinctive hydraulic and geochemical features. The main reactive mineral in the till deposits, for the time horizons considered in this work, is calcium carbonate together with minor amounts of clay minerals (e.g. illite). The till deposits forms aquifers with relatively high hydraulic conductivities. In contrast, glacial and post-glacial clays are basically composed of illite with low to very low amounts of calcium carbonate, and containing organic matter-rich layers (gyttja), which can promote reducing conditions in the porewaters. All these clays exhibits relatively low hydraulic conductivity values. Five radionuclides have been selected for conceptualization and qualitative evaluation of retention process: U as an actinide, Se as a redox-sensitive radionuclide, Cs as a monovalent cation, Sr as a divalent cation, and I as an anion radionuclide. Overall, radionuclide retention capacity in the surface systems at Forsmark can be provided by sorption on charged surfaces of clays and oxyhydroxides, co-precipitation with sulphates, sulphides, oxyhydroxides and carbonates, and sorption on organic matter. Two-dimensional coupled hydrogeological and reactive solute transport models have been developed to simulate the geochemical behaviour of U, Cs and Sr. These three radionuclides have been selected for quantitative modelling based on the availability of data and parameters. Two distinct geological and hydrogeological domains have been considered: (1) the till system and (2) the clay system. The first case simulates the intrusion of a radionuclide-bearing fluid from granite bedrock into a relatively dynamic till aquifer. In the second case, the radionuclide bearing fluid interacts with a low permeability, reducing clay layer that is present at the bottom of a discharge zone (such as a lake or the Baltic Sea, for instance), overlying the till deposit. Groundwater flow and transport modelling results highlight the different behaviour of the two simulated systems. The till system constitutes a dynamic aquifer, which reaches the transport steady-state in less than 8 years. On the other hand, the clay system constitutes a relatively low permeability aquitard in which the transport steady state needs hundreds of years to be reached. Reactive transport results indicate that caesium is very strongly retained in the FES of illite in both till and clay systems. Most of the caesium mass entering the system from the deep source is effectively retained in the very close vicinity of the source, independently of the hydrogeological conditions. In the case of uranium, the most effective processes for retention are very different depending on the two considered hydrogeological systems. In the till aquifer, the dissolved uranium is mainly adsorbed onto the charged surfaces of ferrihydrite. It is seen that this dynamic aquifer system still exhibits a uranium retention efficiency of about 50% even after about 100 years of simulation time. On the other hand, the simulated clay system is much more efficient than the till aquifer for uranium retention due to the precipitation of amorphous uranium (IV) oxides. Uranium retention efficiencies higher than 95% are computed for the clay system, even after more than 400 years. As in the case of caesium, most of the uranium mass entering the system from the deep source is effectively retained (in this case precipitated into solid phases) in the vicinity of the source. Finally, model results indicate that strontium is retained by two different geochemical processes: (1) cation exchange within illite interlayers and (2) precipitation into a SrxCa1-xCO3 solid solution. Even though there are two distinct retention mechanisms affecting strontium, the clay system exhibits the lowest efficiency for retention of this radionuclide compared to the other two simulated radionuclides. Cation exchange is the dominating retention process in both the clay and the till system, but the contribution of carbonate solid solution is not negligible; its relative importance depends on the initial pool of carbonate minerals in the system. The overall efficiency of both systems to retain strontium shows a clear decrease with time. However, the efficiency for strontium retention of both systems is still higher than 60% after simulation times of hundreds of years. From this study it can be concluded that the near-surface system at Forsmark constitutes a geochemically reactive barrier able to retain radionuclides by several key processes. Such retention mechanisms produce an effective overall retardation of the migration of the studied radionuclides