given by Reference 98, however, has been converted to the Wahoo and Umbrella cases, using a column diameter D of 2,000 feet and a column height C of 1,500 feet for Wahoo; similar parameters for Umbrella are 1,800 feet and 5,000 feet, respectively. The cases fora solid column and for two hollow columns (one with an inner core D, , 67 percent of the outer diameter, and a second with an inner core D, , 85 percent of the outer diameter) are computed for both shots. None of the resultant radial velocity curves compare with those obtained for Wahoo; three such curves representing the three types of columns have been superimposed on the ob- served curve in Figure 3.132. All velocities are plotted relative to zero time instead of relative to the time of base surge formation as suggested by Reference 98. The case for a hollow core, 85 percent of the outer diameter, most closely approximates the radial velocities observed for Umbrella, and three such curves representing the collapse of columns of three different denSities have been superimposed on the observed curve in Figure 3.133. The comparison with these fluid models is, however, poor at best. The collapse of the fluid models starts with the column at rest, a static condition that only approximates the actual situation. The collapsing column, particularly on Wahoo, must have had someinitial radial velocity before collapse as indicated by the throwout plumes. A more extensive study of these phenomena is required before any definite conclusions can be drawn. The time of cessation (TOC) (see Appendix F) is subject to a number of definitions even greater than TOA. TOC may be defined either as the time at which the normalized rate curve becomes essentially horizontal after registering the passage of the main series of dose rate peaks or as that point at which the normalized rate curve drops permanently below 10) r/hr. Both TOC’s have been determined for each station, and these values are presented in Table 3.18. A plot of cessation times defined either way versus distance is badly scattered particularly for the close-in stations where TOC is influenced both by surge development and by. contributions from waterborne radioactive sources. At greater distances, the slope of the data points roughly approximates the reported surface winds. Since the latter definition of the TOC sometimes indicated by the abbreviation “norm 1074” , tS more readily corrected for the effects of waterborne sources, this TOC is used for the study of the surge tails, the postulated diffuse remnants which trail behind the base surge (Section 3.3.2). The length of these tails is computed on the basis of the time difference between the photo-TOC and the rad-TOC, using the official Task Force surface wind speed. The distance of the primary surge photo-boundary Py along each of the station legs has been determined at various times after zero time (Table 3.19). A plot of these distances versus time may be approximated by a straight line for most downwind legs. The slopes of these straight lines are also given in Table 3.19. At approximately 4 minutes after Umbrella, a break occurs in the downwind plots, which probably represents the passage of the downwind surge over the atoll reef (Section 3.3.2). The slopes for Umbrella are, therefore, given both before and after this time. In general, the siopes of all lines are close to the reported surface wind speeds; however, once again there is some evidence of a difference in base surge velocity along specific radii. There is also some indication that the point of maximum dose rate (source center) recedes farther behind the surge front at later times. Since the later visible boundaries are rather diffuse, the postulated recession can only be illustrated by comparing the time at which the dose rate reached 0.1 percent of peak values with time of the peak value (Figures 3.134 and 3.135). The tendency for the time of peak to fail farther behind the time of the first rise in dose rate is, however, so slight that it cannot be conclusively demonstrated with the available data. If this phenomenon is real, it may possibly be explained by the fact that the base surge increases in height with time and thus increases its effective radiating area. The base surge radius has beer. determined by calculating the position of the hypothetical surge center for the time of peak dose rate recorded at a given station and measuring the dis- tance from the station to this center. These measurements have been made, both for the first peak representing the downwind surge transit and for the completion of the upwind surge transit (photo-TOC). These radii are presented in Table 3.20. The measured radii for the first peak are also plotted in Figures 3.136 and 3.137. For Wahoo, the surge radii at time of peak appear 237