Voveness of the cell sear riwetok to a position southeast of Lajuro gave strseg socthe westerly flow in the local area on 28 Pebruary end west socuthe Tm een etre ee ve f incividual cells drifting eastrard. i west flow on 1 Larch, replacing the light variable winds wich existed on 23 = 25 February. Contimed southward notlor end aprreach of following cell from west brought west to -ombrest flow on 2 and 3 March. (4) &t 50,000 feet, the slow mvement of the clockwise cim culation from a position northwest of Eniwetok to south of : LMejuro changed the flow gradually froma north northrest te west southwest, After 1 Marsh this cireulstion continued its movenent to the southeast and weakened, vith flow in the local area becoming west northwesterly. (5) at 60,00 feet, no clearly defined trend was presext. a “Winds were suite light and erratic at this level, with easterly flow at higher levels. These variable winds persisted through BRAVO plus 3 days. ad. : Differences between forecast and observed winds: In zegard to the differences between forecast and observed winds, reference is made to a report of Proiest 4.5, JTF THE==, by Pacmer, ilier, and Stopinski. This repost stcted that studies of the observa- iN tional errors in upzer wind czservations teen by GO/l ezsimem inéicated thet for wind speeds above 10 imots, arpreximatelr 1f per cent of the observations varied ty core than 20 degrees, eva: om ie when the chservetions were rade by several ciflerem Grows sts, but with the same 2irberne unit. re cifferences Setweer forecast poesia w be