The conservatism and uncertainty in a regulatory model for predicting derived release limits for aquatic emissions

AECL, Chalk River Laboratories, Chalk River, Ontario K0J 1J0, Canada

Abstract

The conservatism and uncertainty in the Canadian Standards Association (CSA) model for calculating derived release limits (DRLs) for aquatic emissions of radionuclides from nuclear facilities was investigated. The model was run deterministically using the recommended default values for its parameters and its predictions were compared with the distributed doses obtained by running the model stochastically. Median doses from 400 stochastic runs were higher than the deterministic doses predicted using CSA default values for more than half of the radionuclides considered. Thus the CSA model is not always conservative for calculating DRLs for aquatic emissions as it was intended to be. The uncertainty in the predicted ingestion dose was high, with the 95% confidence interval exceeding an order of magnitude for most radionuclides. A sensitivity study revealed that ingestion doses to adults predicted by the CSA model are sensitive primarily to water intake rates, bioaccumulation factors for fish and marine biota, dietary intakes of fish and marine biota, the fraction of consumed food arising from contaminated sources and irrigation rate. To improve DRL models, further research in aquatic exposure pathways should concentrate on reducing the uncertainty in these parameters.