Assessment of off-site radiological hazards at the Oak Ridge K-25 Site : an emergency preparedness perspective

The objectives of this study were to perform a parameter uncertainty analysis and a model intercomparison for the Oak Ridge K-25 Site Emergency Preparedness Department. In the uncertainty analysis, 97% subjective confidence intervals were computed for 50-year committed effective dose equivalent (CEDE) per unit activity for four off-site distances for the U-234, U-235, and Tc-99 isotopes, which present the greatest radiological hazards at the Site in terms of accidental atmospheric releases. Normalized CEDEs were considered to bypass the need for source terms, which contribute major uncertainties to model prediction. The goal of the model intercomparison was to compare two atmospheric radiological release models, RASCAL Version 2.1 and HOTSPOT 6.5, and to provide guidance for limits of model use in the emergency preparedness arena. Subjective probability distributions were determined using data provided by experts, for the meteorological input parameters to the models. The models were run 120 times using a source term of 1 mCi of U-235, using 120 6-tuples generated by Latin Hypercube Sampling. Cumulative density functions (cdfs) for normalized CEDE calculated by the models were generated for U-235 for the four off-site distances of interest, and were used to obtain the desired confidence intervals. By applying ratios of dose equivalent factors, confidence intervals were also obtained for 1 mCi releases of U-234 and Tc-99. Considering a 1 kilogram source term of each isotope, it was seen that a 1 kg release of U-234 would produce 10,000 times the CEDE of a 1 kg release of U-235 or Tc-99, likely exceeding the Protective Action Guide of 1 rem for initiating protective actions. A careful evaluation of source terms must be made to adequately assess the hazards of these radionuclides. RASCAL 2.1 predicts CEDE 23% higher than HOTSPOT. While both are useful models, HOTSPOT is recommended for field work.