Sensitivity analysis for power industry radionuclide air stack emissions leukemia incidence risk comparison models*

The authors developed mathematical models to quantify the impact of accidents on the U.S. nuclear power industry radionuclide air emissions incidence risk. Additionally, sensitivity analysis was performed on the terms in those models. The number of accidents of a given size needed to equilibrate the nuclear power industry leukemia incidence risk with the coal power industry leukemia incidence risk under normal operating conditions was calculated. We evaluated an accident’s impact on the total leukemia incidence risk comparison done using all of the six types of postulated dose response curves. An overlapping plot of the number of nuclear accidents required to equilibrate industry risks versus accident magnitude enabled the comparisons of models. Sensitivity analysis on the developed models for the current mix of U.S. coal and nuclear power plants was used to verify model limitations. Sensitivity analysis also showed that the models with cell killing terms gave meaningless numbers for large dose accidents and that when both the linear and quadratic terms axe present in the dose-response curves, the linear term dominates the quadratic term by a factor of 10 until the dose exceeds 110,000 mrem (1.10 Sv). Air emission leukemia incidence risk projections to the year 2006 were obtained by including plants due to go online by 2000. The application of these models provided an approach toward developing a methodology for identification of the relative risk from power generation alternatives.