Probabilistic Versus Deterministic Analysis for Demonstration of Compliance with the Dose Criteria in 10 CFR Part 20, Subpart E

The U.S. Nuclear Regulatory Commission (NRC) promulgated its regulations on the “Radiological Criteria for License Termination” in 10 CFR Part 20, Subpart E. The regulations require that, for the release of sites for unrestricted use, the total effective dose equivalent (TEDE) to an average member of a critical group, resulting from residual radioactivity that is distinguishable from background radiation, does not exceed 0.25 milliSievert per year (mSv/yr), (25 millirem/yr (mrem/yr)) . Subpart E also requires that the residual radioactivity has been reduced to levels that are as low as is reasonably achievable (ALARA). In addition, Subpart E establishes the criteria for license termination with restrictions on future land use, as long as specific conditions are met. Finally, subpart E provides alternate criteria for license termination in unusual situations where the site may exceed the 0.25 mSv/yr (25 mrem/yr) limit, but would not be permitted to exceed 1.0 mSv/yr (100 mrem/yr) or 5.0 mSv/yr (500 mrem/yr), under certain conditions. Licensees, staff, and/or stakeholders may demonstrate compliance with the dose criteria through site-specific dose impact analysis to establish derived concentration guideline levels (DCGLs). Site-specific dose impacts analysis may also be conducted to assess remedial actions or options for site release restrictions based on measured concentrations of residual radioactivity at the site. The NRC is committed to a risk-informed performance based (RIPB) approach in regulatory decision-making and licensing activities. Staff commonly uses probabilistic analysis to assess the significance of risk and uncertainties in dose analyses. The analyst typically considers realistic bounding conditions based on probabilistic distributions of input parameters and evaluate uncertainties in the dose outputs. Licensees and stakeholders currently use two approaches for dose analysis. One approach is based on a deterministic methodology, which typically employs a highly “conservative” single value for each parameter assumed to bound variability and uncertainty in site specific conditions. The second approach is based on a probabilistic methodology which employs sensitivity and uncertainty analysis of input parameters using probabilistic distributions of sensitive parameters. NRC staff have conducted dose analyses for decommissioning sites using both deterministic and probabilistic approaches to review and evaluate licensees’ derived concentration guideline levels (DCGLs) equivalent to the dose criteria in 10 CFR Part 20, Subpart E. Review of licensee’s dose analysis includes assessment of input parameters and evaluation of the dose outputs for the two approaches. Based on staff analyses, and using common codes/models acceptable by most Federal agencies, the deterministic analysis approach can produce unrealistically conservative dose values. Probabilistic dose analyses, however, generally produce more realistic dose results and help define the associated uncertainties. Detailed comparative analyses of the two approaches with an actual example is presented.