UPPER-ATMOSPHERE WINDS AND THEIR INTERPRETATION .I. EVIDENCE FOR STRONG NONLINEARITY OF HORIZONTAL FLOW ABOVE 80 KM

This paper discusses a homogeneous collection of observational material comprising thirty horizontal-wind profiles, twenty-four of which are published here for the first time, derived from photographs of artificial vapor trails ejected by rockets over Wallops Island, Virginia (38dgN) between 1959 and 1964. The observations extend over the height range 65-225 km and are essentially complete in the range 80-140 km on which the pesent discussion concentrates. They have a height resolution of about 0.1 km at 100km and about 1 km at 130km. All the horizontal-wind profiles have a stratified structure which is especially marked between 80 and 130 km. Changes of the vertical gradient of the horizontal wind tend to be concentrated in narrow layers, some of which are less than 0.1 km in width. Between these layers the vertical gradient of the horizontal wind usually changes slowly in both magnitude and direction. The mean magnitude of the horizontal wind between 80 and 130 km is closely related to the form of the profile. Profiles that have a regular spiral form tend to have large mean wind-speeds; irregular profiles, small mean wind-speeds. In all the profites the wind vector shows a strong tendency to rotate in a clockwise direction with increasing height. This tendency manifests itself even more strongly in profiles of the residual wind (the difference between the actual horizontal wind at a given level and the mean wind in a height interval of 25 km centered on that level). There are indications that the mean amplitude of the residual-wind profile is related to the form of the smoothed-wind profile. Since the pioneering experimental work of Greenhow and Neufeld, it has been customary to decompose the horizontal wind into a prevailing component and a nonprevailing component, and the nonprevailing component into a periodic part and an irregular or noise-like part. The periodic part of the nonprevailing component, which Greenhow and Neufeld further decomposed into components with periods of 24, 12, and 8 hr, has been identified with the atmospheric tidal oscillation. The irregular part has been attributed by some authors to turbulence, by others to a superposition of internal gravity waves, and by still others to various combinations of these two types of motion. Considerations of phase coherence, as evidenced in both the spatial and temporal variations of the horizontal wind, lead us to reject the assumption underlying all these views: that the nonprevailing component can be separated into two physically distinct parts. Other theoretical and observational arguments are offered to support this conclusion. The structure of the horizontal-wind profiles indicates that the wind system is strongly nonlinear. A possible physical interpretation of the profiles proceeds from the conjecture that the narrow transition layers represent gravity shocks-physical surfaces of discontinuity that bear the same relation to internal gravity waves as do compression shocks to acoustic waves. In a first approximation the temporal development of the horizontal-wind profile is determined by the motions of these surfaces in horizontal-velocity space and in height. It is suggested that the aperiodic temporal variation of the nonprevailing component of the horizontal wind may be caused by coupling between the prevailing and nonprevailing components. The former is known to exhibit irregular day-to-day fluctuations at meteor heights. © 1968.