} (3) Frequent post-shot wind forveaste should be made daring the interval H to H plus 2h hours in order that the actual orientation of f-ll-cut pattern may be more accurately analysed in conjunction with the post= sot weather and radsafe cbeervaticas. (4) The frequency and aceuracy of pre-shot weather forecasts wore adequate for realistic planning and fall-out forecasting following the m iification discussed in paragraph la (4) above. (5) Continued research and stuty of tropical artecrology is saquired in order to improve long-range wind forecasts. (6) The eaployment of constant altitude balloons as a forecasting tool, alttough nct successful on CASTIZE due to operational, supnly and mechanical difficulties, is still considered a valuable adjunct to the ‘duather/Radeafe forecasting techniques. (7) Rawinsonde balloons frequently “Leed" and burst at the cold tropical, tropopause. As @ consequence, critical wind observations were scastimes terminated prior to reaching the naximm significant wind levels. (8) Base to the teniency of the winds to be westerly at points to the north of ground sero, additional raviasonde sites in the northwest and northeast quadrants would have been extremely useful. Ce Pall-out forecasting | (2) Assumptions: (a) As a general rule, the aseumptions of 24 hours and 500 nautieal miles as the upper limite wigdin wiich significant fall-out could oGeur were valid, Although air eont mination from water surface shots extencs the tims to approximately 24 to 30 hourw, the sercscl-type clead from such shots preeluded significant fall-out levels at such lata times. It appears probabTan Bckevey, on the basis of avadlable evikenesr that .@ “seine affast.” 90 I