Historical Doses from Tritiated Water and Tritiated Hydrogen Gas Released to the Atmosphere from Lawrence Livermore National Laboratory (LLNL) Part 1. Description of Tritium Dose Model (DCART) for Routine Releases from LLNL

Abstract

DCART (Doses from Chronic Atmospheric Releases of Tritium) is a spreadsheet model developed at Lawrence Livermore National Laboratory (LLNL) that calculates doses from inhalation of tritiated hydrogen gas (HT), inhalation and skin absorption of tritiated water (HTO), and ingestion of HTO and organically bound tritium (OBT) to adult, child (age 10), and infant (age 6 months to 1 year) from routine atmospheric releases of HT and HTO. DCART is a deterministic model that, when coupled to the risk assessment software Crystal Ball ® , predicts doses with a 95% confidence interval. The equations used by DCART are described and all distributions on parameter values are presented. DCART has been tested against the results of other models and several sets of observations in the Tritium Working Groups of the International Atomic Energy Agency's programs, Biosphere Modelling and Assessment and Environmental Modeling for Radiation Safety. The version of DCART described here has been modified to include parameter values and distributions specific to conditions at LLNL. In future work, DCART will be used to reconstruct dose to the hypothetical maximally exposed individual from annual routine releases of HTO and HT from all LLNL facilities and from the Sandia National Laboratory's Tritium Research Laboratory over the last fifty years.iv This page intentionally left blank.v

Historical Doses from Tritiated Water and Tritiated Hydrogen Gas Released to the Atmosphere from Lawrence Livermore National Laboratory (LLNL). Part 2. LLNL Annual Site-specific Data, 1953 - 2005

“…The high uncertainty exists due to many simplifying assumptions in the model. The choice of CAP88-PC was based on familiarity and its known tendency to slightly overestimate air concentrations at VIS (Peterson 2004a). Although the dose reconstruction will attempt to predict accurate doses with a 95% confidence interval, if there is any bias, it should be slightly conservative.…”

“…Analysis of concentrations in the LLNL swimming pool (Table 19) In DCART (Peterson 2004a), tritium concentrations in the LLNL swimming pool can be estimated when unavailable from the empirical relationship between air moisture tritium concentrations at VIS, air moisture tritium concentrations at the air tritium sampler (POOL) next to the pool, and pool water concentrations. For years with measured concentrations, the mean annual pool water concentrations can be directly input to DCART.…”

“…As explained in Part 1 (Peterson 2004a), soil concentrations are calibrated from air concentrations, and the precipitation pathway has been disabled in DCART because it is not needed to calculate concentrations in food or ingestion dose. Nevertheless, the precipitation equations can be used to determine concentrations in rainfall.…”

“…The uncertainty distribution on these numbers is assumed triangular, with the estimate being the maximum, because it is unlikely that it actually rained throughout all 15-minute periods; the best estimate and minimum are assumed 85% and 50% of the maximum, respectively. Table A1, Part 1 of this LLNL Dose Reconstruction (Peterson 2004a). Because this equation was derived from fractions that represent the maximum time it could have rained, but because the real fraction may be either higher or lower than the best fit to the linear regression, a triangular distribution with maximum, best estimate, and minimum at 1.1, 0.9, and 0.6 times the calculated value, respectively, has been assigned.…”

“…• Part 1 describes DCART, the stochastic model that will be used for the proposed dose reconstruction from routine emissions; it has been published (Peterson 2004a). Corrections to some parameter values and some clarification may be found in Appendix A.…”