Relative Sensitivities of Existing and Novel Model Parameters in Atmospheric Tritium Dose Estimates

A sensitivity and uncertainty analysis of the parameters used to calculate offsite doses from routine atmospheric tritium releases has been performed. The tests were conducted using MAXIGASP, a site-specific version of the U.S. Nuclear Regulatory Commission code GASPAR. The analyses of MAXIGASP were conducted on the version of the code currently in use and on a modified version which incorporates a number of novel parameters. In this application, measured tritium oxide (HTO) release rate and calculated downwind concentration data for the Savannah River Site (SRS) provided the release information required as model input. Parameter assignments for MAXIGASP were made based on Latin Hypercube sampling of each variable over its expected range. Site-specific ranges were used when available. Parameter sensitivity was evaluated based on relative values of rank-transformed correlation coefficients and sensitivity indices. The most sensitive parameter in the current version of MAXIGASP appeared to be the breathing rate. The most sensitive parameters in the modified version were, in descending order of relative importance, the tritium quality factor, the effective half-life of tritium, the breathing rate, and the produce consumption rate. The resulting dose equivalents to a given individual at the site perimeter resulting from a 1.23 x 1010 MBq (3.33 x 105Ci) HTO release were estimated as 1.7 ± 1.0µSv (0.17 ± 0.10 mrem) and 1.6 ± 1.6 µSv (0.16 ± 0.16 mrem) for the current and modified versions of MAXIGASP, respectively. These values represent means and 2s standard deviations at a subjective confidence level of 95% for nominal HTO annual release conditions at the SRS.