Estimating internal dose due to ingestion of radionuclides from Nevada Test Site fallout

The U.S. Department of Energy initiated the Radiation Exposure Review Project to provide a critical reexamination of radiation doses to people resulting from testing nuclear devices at the Nevada Test Site. One part of this effort focused on the dose resulting from the ingestion of contaminated food. The PATHWAY radionuclide transport model was developed to provide estimates of food concentrations for 20 radionuclides for each of 86 test events and 15 agricultural scenarios. These results were then used as input to the Human Ingestion model to provide dose estimates for individuals and populations in 9 western states. The model considered the life-style and age of the people, and accounted for the transport of milk between locations. Estimates of uncertainty were provided for all doses using Monte Carlo simulation techniques. Propagation of uncertainty between the PATHWAY model and the Human Ingestion model required the development of special strategies to ensure that the inherent correlations between concentrations of the radionuclides in foods were handled properly. In addition, the size of the input data base (60 megabytes), the number of cases to consider (over 30,000), and the number of Monte Carlo simulations (over 6 million) required the development of efficient and reliable methods of data access and storage while running simulations concurrently on up to 14 UNIX workstations. The problems encountered in this effort are likely to be typical of any dose reconstruction involving geographically heterogeneous environmental conditions. This paper documents the methods used to disaggregate the system to achieve computation efficiency, the methods used to propagate uncertainty through the model system, and the techniques used to manage data in a distributed computing environment. The radionuclide- and age-specific dose factors used in the analysis are also provided.