debris vy the tropopause unless definite proof of such a mochanism was aveilable. 4s e consequence, the techniques described above wore used with” comfidonee up to about six hours, and examined in their relation to the longerange forecast air perticle trajectories for times bey-ad six heures. This employment was later expanded ints o mre detailed and = - formal, technique developed subsequent to the third shot and employed thereafter in the series. (See Incl 2) TU The method of elliptical app UPSHOT, SUPPER series in Nevada tions developed following the (and wed with remarkable success on OLE) was used on C..STLE for «better aprreciation of the dezree of contamination and the extont of the forecast fall-cut aress. Due to yield scaling considerations, amd the unique meteorclogical differences between the Nevada and the Pacific proving grounds, confidence in this methed for the first shot of the C..STLE series was low, In addition, although this method has certain practical and aprealing features, in its delineation 2f a picture of the fall-out pattern on the ground, it is no less restricted to the ground cero wind system and the stability of these winds, than the methods described heretofor. C.\STLE use of the existing fall-cut forecasting methods wes sub stentially as follow: a. Vector solution, This method was apriied by the vectcrial addition of winds from the surface t ooxicum height, all vectcrs normalized to 5,000 feet per hour for convenience in ccnputatione, and «ith the vector lengths ~roportiocnal to the wind speed in imcts. Since tho surface wind and the arcas ccntizucous to GZ in the PFG are essentially at zcro elevation, no correction was necessary for the asc-callod “average fall-out surface" elevation. Winds were normally rlotted for each 2,000 fo7t levela fr-m tw thousand feet to twenty thousand feet and for every five thousand feet levels up to seventy thousand feet. (Dove sevonty th-usand, due to the relatively stable wind directions, ten thousand foot levels were plotted es a normal rule. The 2,000 foot intervals were used in the lower tradewinds primarily to smocth cut the wind wector dlegr=m fcr these levele which are critical fram a close-in fall-out viewpoint. For this purpose, the 2,000 foot weters were normalised by pictting vectors of 4 length twoofifthe of the wind speed im invte (1.6. 2,000/5,000 of « full hour wind vwocter). Ten thousand fect levels were similarly treated, plotting 10,000/5,000 or twice « full hour wind veetor, Im this technique, the 2,000 fort level was aseumed to represent the aversze wind Sctween the warface and 2,000 feet, the 4,000 foot leval was assumed to reprosont the averase wind between 4,000 feet and 2,000 feet, ete. The addition of clesin vectors between the ground sero (initial point of the first voctur) and cach succeesive altitude rrovided the necessary resultant winds from oach lovel, and consequently, the line on the ground on which fall-out shvuld cecur from the levels involved. Computations of time of arrivel of foll-cut end ires of fall-out followed the same pattern as presented in NSM 105-33, ecnsistin basically of dividing the resultant winds into hourly inercments <cerencing