The impact of low and intermediate-level radioactive waste on humans and the environment over the next one hundred thousand years.

In order to assess the potential radiological risk to humans and the environment from a geological repository for radioactive waste, a safety assessment must be performed. This implies that the release and transfer of radionuclides from the repository into the surface environment are calculated and that the effects in the biosphere are evaluated for an assessment period up to one hundred thousand years according to Swedish regulations. This paper discusses the challenges associated with the modelling of surface ecosystems over such long time scales, using the recently completed assessment for the extension of the existing repository for the low- and intermediate-level nuclear waste (called SFR) in Forsmark, Sweden as an applied example. In the assessment, natural variation and uncertainties in climate during the assessment period were captured by using a set of climate cases, primarily reflecting different expectations on the effects of global warming. Development of the landscape at the site, due to post-glacial isostatic rebound, was modelled, and areas where modelling indicated that radionuclides could discharge into the biosphere were identified. Transfers of surface water and groundwater were described with spatially distributed hydrological models. The projected release of radionuclides from the bedrock was then fed into a biosphere radionuclide transport model, simulating the transport and fate of radionuclides within and between ecosystems in the landscape. Annual doses for human inhabitants were calculated by combining activity concentrations in environmental media (soil, water, air and plants) with assumptions on habits and land-use of future human inhabitants. Similarly, dose rates to representative organisms of non-human biota were calculated from activity concentrations in relevant habitats, following the ERICA methodology. In the main scenario, the calculated risk for humans did not exceed the risk criteria or the screening dose rate for non-human biota, indicating that the repository design is sufficient to protect future populations and the environment. Although the combination of radionuclides, land-uses/habitats, type of most exposed population and area of exposure that contribute most to the total dose shifts over time, the total calculated dose shows limited variability. Significant reductions in the dose only occur during submerged periods and under periglacial climate conditions. As several different water and food pathways were equally important for endpoint results, it is concluded that it would be difficult to represent the biosphere with one or a set of simplified models. Instead, we found that it is important to maintain a diversity of food and water pathways, as key pathways for radionuclide accumulation and exposure partly worked in parallel.