372 MACHTA 10 3 =p No) ‘» 5 . 6 Fig. 2—Estimated upper tropothe Sept. 1, 1961, U.S.S.R. test. Shading indicates dates of first appearance air filters. of debris on surface D, © 6 7-8 spheric trajectory of debris from ‘ iN 9-10 | 3° 72—\ 7 °¢ 9-10 eo ° eete fe ot . Ss GROSS BETA NETWORK ePTIME OF ARRIVAL IN SEPTEMBER 1961 5-6 C= 7-8 (1D 9-10 lay of several days between the predicted arrival at 30,000 ft and at ground level. This delay results, inour opinion, from time to mix down- ward or for particulates to settle out. The literature contains innumerable examples of trajectories of nuclear clouds obtained by many techniques. Of these methods the ap- plication of isentropic analysis is one of the two most valuable improvements over past Weather Bureau procedures. Isentropic analysis permits the tracking of air parcels undergoing adiabatic vertical motions, A dramatic case in which the isentropic trajectories appear to explain better the observations of ground-level air concentrations has been given by Reiter.‘ In the mean, the isentropic surfaces slope downward from north to south in the troposphere; thus trajectories moving southward will be brought closer to the ground or man’s environment, whereas those moving northward rise away from the ground. A second, less important, improvement is the computation of tra- jectories by an electronic computer rather than by hand. An illustration of a computer forecast currently in operational usage by the Weather Bureau appears in Fig. 3 with starting points at the Nevada Test Site (NTS). In practice, it is felt that the additional wind information and knowledge of local topographic conditions permit meteorologists at NTS to start computation of the trajectories more correctly than can be done by machine, After 6 hr the machine trajectories are used as the basis for continuing the forecasts to 30 hr. Figure 4, taken from the Weather Bureau analysis of trajectories for STROBE (the use of constant-level