SECTION 4 UNCERTAINTY ANALYSIS The uncertainty in calculated film badge doses is estimated from the underlying parameters. Not only is the uncertainty in the mean film badge dose determined, but also the distribution in dose about the mean is estimated for typical personnel. The basic uncertainties in the topside environment include radiation intensities on deck, the positions of personnel (hence their exposure) on deck, the time spent on deck, and the shielding from fallout afforded to those below. Uncertainties in the radiation environment below due to ship contamination are dominated by assumed buildup and decay rates of the radioactive material accumulated on the ship's hull and interior salt water systems. Intensity levels on deck are determined from shipboard radiological survey data, supplemented at late times by decay rates measured on Bikini Atoll. Individual meter readings on deck, where available, are taken as accurate, their inherent error having a negligible influence on the overall uncertainty in dose. Average on-deck intensity as a function of time is taken as accurate; the power law interpolation in time between surveys Closely approximates fission product decay at the times after burst considered. Power law fitting is less accurate during fallout deposition and decontamination; however, the influence of this uncertainty is minimized because the typical crew- member was below during these intervals. Overall, error in on-deck intensity is small compared to the uncertainty associated with crew position in the non-uniform radiation environment. The significant variation in on-deck intensities following fallout deposition focuses attention on the positioning of the crew relgtive to those intensities. Specific data on crew positioning are lacking; however, the crew size and the variety of duties performed suggest that the crew was, on the average, randomly positioned on deck and therefore randomly exposed to each reported intensity. The uncertainty in dose resulting from these assumptions cannot be directly quantified, except by considering unrealistic extremes. However, an indication is provided by the assumption that, for each interval topside, personnel remained in the same general deck area but were 150