Dose assessments for SFR 1

Following a review by the Swedish regulatory authorities of the safety analysis of the SFR 1 disposal facility for low and intermediate level waste, SKB has prepared an updated safety analysis, SAR-08. This report presents estimations of annual doses to the most exposed groups from potential radionuclide releases from the SFR 1 repository for a number of calculation cases, selected using a systematic approach for identifying relevant scenarios for the safety analysis. The dose estimates can be used for demonstrating that the long term safety of the repository is in compliance with the regulatory requirements. In particular, the mean values of the annual doses can be used to estimate the expected risks to the most exposed individuals, which can then be compared with the regulatory risk criteria for human health. The conversion from doses to risks is performed in the main report. For one scenario however, where the effects of an earthquake taking place close to the repository are analysed, risk calculations are presented in this report. In addition, prediction of concentrations of radionuclides in environmental media, such as water and soil, are compared with concentration limits suggested by the Erica-project as a base for estimating potential effects on the environment. The assessment of the impact on non-human biota showed that the potential impact is negligible. Committed collective dose for an integration period of 10,000 years for releases occurring during the first thousand years after closure are also calculated. The collective dose commitment was estimated to be 8 manSv. The dose calculations were carried out for a period of 100,000 years, which was sufficient to observe peak doses in all scenarios considered. Releases to the landscape and to a well were considered. The peaks of the mean annual doses from releases to the landscape are associated with C-14 releases to a future lake around year 5,000 AD. In the case of releases to a well, the peak annual doses were obtained around year 4,000 AD, when an increase of groundwater discharges occurs in connection with the ongoing shoreline displacement. The peaks for the well are also dominated by C-14, although a few other radionuclides also contribute. This pattern of the peak doses was observed for calculation cases of the main scenario and less probable scenarios. Moreover, the predicted doses, including peak dose values, for calculation cases of the less probable scenarios were very close to the predictions for calculation cases of the main scenario. Predicted annual doses to the most exposed individuals during the first 1,000 years after the repository closure did not exceed 0.05 μSv per year. Low release rates are predicted for this period, when the recipient for the releases is the sea, and doses per unit release rate are low, as compared to the case with releases to a lake or a mire. Doses from short-lived radionuclides were very low, as these can only be released during the sea period. The doses from actinides were also low, due to effective retention in the engineered barriers and their low inventory. In general, total mean annual individual doses were low during the whole simulation period, with values below 14 μSv per year for the main and less probable scenarios. Uncertainty analyses were carried out using probabilistic methods. These analyses showed that the parameter uncertainty in the peak doses from releases to the landscape are low due to the dominant role of C-14, for which the dose factors and the biosphere release rates have low uncertainty. Sensitivity studies were carried out to identify the near field and biosphere parameters that have the highest contribution to the uncertainty of the peak dose estimates. The parameters with the highest contribution to the uncertainties in estimates of release rates from the near field are the flow uncertainty factors for non-sorbing radionuclides and the distribution coefficients in the construction cement for sorbing radionuclides. The biosphere parameters with the highest effect on the uncertainty of the peak dose estimates for releases to the landscape are the dissolved inorganic carbon content and the net primary production in lakes. In the case of releases to a well, the biosphere parameter with the highest and dominant effect on the uncertainty of dose estimates is the well capacity