Tare complex with electronic gages. There was a definite and consistent variation iu the data between the two lines. It is recognized that a moisture-laden air will attenuate pressures in thc blast wave, simply because blast energy will be lost by an amount proportional to that which is necessary to evaporate the suspended water droplets or rain in the path of the shock. Studies on the problem by the two projects concerned indicated that a moderate shower could contribute sufficient water content to the air to account for the deviation in the pressuredistance curves of the two blast lines (described in the Project 1.24 and 1.2b reports, see Appendix). 2.3.5 Comparison with the 2W Theory. It was anticipated that sufficient data would be obtained from Castle to allow a quantitative comparison to be made, for surface bursts, with the ideal case. Theoretically, such a burst over a perfectly reflecting plane should act like one of twice ita yield in free air. Data from previous surface bursts, Jangle Surface and Ivy Mike, did not entirely confirm this theory. The question was the value of the reflection factor —of necessity between 1 and 2. From Castle data, it appeared to be certainly less than 2—probably between 1.6 and 2. The difficulty, and the reason a more~definite figure cannot be assigned, lies with the determination of yield of the multi-stage devices; fireball and time-of-arrival methods used to estimate yield involve the 2W assumption. A method independentof this assumption is necessary. Unfortunately, only radiochemical analysis, which determines only the fission yield of a device, satisfies this restriction. Figure 2.4 shows a pressure-distance plot of all the surface overpressures scaled to 1 kt at standard sea~level conditions, along with similar data from Jangle Surface and Ivy Mike, compared to the 1W and 2W free-air composite curves. All measured data were scaled to 1 kt at sea-level conditions. The solid line represents a composite pressure-distance curve for a 1-kt surface burst based primarily on Castle measure- ments. Yields used for data reduction were based on a radius-time history of the fire~ ball (involving the 2W assumption)!. All arrival-time data are compared in Figure 2.5 on a similar basis. There were no apparent effects due to refraction observed during Operation Castle. In fact, Figure 2.4 indicates that overpressures at long ranges fall closer to the 2W freeair curve than do overpressures at closer ranges. 2.4 ABOVE-SURFACE MEASUREMENTS The results of Ivy King confirmed the scaling laws for free-air pressures up to a yield of 540 kt. Data obtained from the Mike event, however, were confined to the low- pressure region. There was reason to suspect that for high yields, an altitude correction must be made for propagation vertically through a nonhomogeneous atmosphere. Castle, then, presented an opportunity to document pressures in the air above megaton-yield surface shots. These phenomena include a definition or delineation of the shock from a surface burst as it propagates through the low levels of the atmosphere out to long ranges. 2.4.1 Pressures. The smoke-rocket and direct-shock photography techniques were used for pressure-distance determination in the air and along the surface. In general, 1On Redwing, considerable data was obtained from two land-surface bursts, one a kiloton burst of medium yield determined by radiochemical analysis. burst curve was drawn from the data— it scaled about 1.6W. A composite land~surface