Abstract
The safety goals adopted by the U.S. Nuclear Regulatory Commission (NRC) consist of two qualitative safety goals backed up by two quantitative health objectives (QHOs). The QHOs establish risk limits for severe accidents in terms of their radiological consequences to affected individuals, in particular, the average individual health risks of early fatality and latent cancers from radiation exposure of members of the public living in the vicinity of a nuclear power plant. This paper is devoted to a reexamination of the coverage of the current safety goals as they constrain (or fail to constrain) the total (radiological and nonradiological) risk posed by nuclear power plant operation. Specifically, we suggest the need to address societal consequences. By societal consequences, we mean measures of consequences that reflect the number of people affected and the offsite effects both radiological and nonradiological, not just the individual risks. Recent Level 3 probabilistic risk assessments suggest that given a high likelihood of evacuation of the close-in population before any release occurs the current QHOs are satisfied by large margins, and the experience of an actual severe accident at Fukushima showed that actual human health effects from released radiation were not the dominant consequences, as there were no early fatalities and no measurable increases expected in cancer rates above the baseline rates in the Japanese population. Hence, regardless of accident probability, Fukushima-type accidents with evacuation would satisfy the NRC’s health-related safety goals. However, there were very significant societal costs in that large numbers of people were relocated for long periods and there was substantial property damage and community disruption along with the costs of recovery and decontamination. We argue that, in addition to the risks addressed in the current safety goals, societal risk should also be considered. This paper discusses specific possibilities for a goal and an associated quantitative objective.
Acknowledgments
Work at Idaho National Laboratory (INL) was performed under U.S. Department of Energy Idaho Operations Office contract DE-AC07-05ID14517. We acknowledge useful conversations with numerous colleagues, including V. Bier, M. Corradini, and R. Denning, as well as exceptionally constructive comments from two referees. Opinions expressed in this paper are the opinions of the authors and do not necessarily represent the views of INL.
Notes
a In this paper, Commission refers to the five NRC commissioners and not to the agency as a whole.
b The prompt fatality QHO has been assumed for some time to have a frequency of 5E−07 per year based on data from the 1980s showing that the risk of a prompt fatality in the United States (due to traffic accidents, etc.) was around 5E−04 per year at that time. It is likely less now due to improvements in traffic safety. The latent cancer fatality QHO is 2E−06 per year based on the cancer fatality rate of 2E−3 per year.
c The exception to this is if the accident is an external event such as a large seismic event that could disrupt the road and bridge network, making evacuation difficult.