**Dynamical Coupling of the Lower and Middle Atmosphere: Historical Background to Current Research***

*Invited review paper presented at the SCOSTEP Ninth Quadrennial Solar-Terrestrial Physics Symposium, Uppsala, Sweden, August 4-8, 1997.

The role of both gravity waves and planetary waves in coupling the circulation in the middle atmosphere with that in the troposphere is now well appreciated. The present paper reviews the history of the study of middle atmospheric dynamics and its coupling with the lower atmosphere. The emphasis is on early developments, principally those before the mid-1970's.

The circulation in the stratosphere, mesosphere and lower thermosphere is observed to have significant variability with a range of periods from minutes all the way up to several years and over a comparably wide range of spatial scales. The current standard view of atmospheric dynamics (e.g., as embodied in the monograph of Andrews et al., 1987) attributes the fluctuations in the middle atmosphere largely to the effects of mechanical coupling with the lower atmosphere. Indeed the field of middle atmospheric dynamics has to a very large degree been concerned with elucidation of the nature of the various wave motions that are responsible for this vertical coupling.

The present brief review aims to document the history of the development of our ideas on the dynamical coupling between the lower and middle atmosphere. There seems to be very little published treating any aspect of the history of this subject in a systematic way, exceptions being Hines' annotations of a collection of reprinted papers on middle and upper atmospheric dynamics (Hines, 1974), and the personal memoirs published by Lindzen (1987) and Hines (1989).

The survey presented here will be affected by the personal biases of the author and the limitations of space. The focus will be on earlier work (particularly the crucial pioneering developments in the 1950's and 1960's), since much of the history of the rapid advances made over the last three decades is easily accessible in the references given in standard texts and the current literature. I will concentrate on the waves which are thought to couple the lower atmosphere with the middle atmosphere (stratosphere, mesosphere and lower thermosphere). I will deal principally with neutral motions, only briefly touching on F-layer ionization effects. I will not provide detailed explanations of the observations and theories mentioned. Given that the audience for this paper is expected to include both meteorologists and aeronomers, it is likely that all readers will encounter some unfamiliar concepts. Up-to-date textbooks on dynamical meteorology include those of Holton (1992) and Pedlosky (1987), while upper atmospheric physics is discussed at a fairly elementary level in Rees (1989). The reader will also no doubt notice that attention has been restricted to the western scientific literature. In some instances somewhat parallel developments may have occurred in the active scientific community of the Soviet Union. Finally I will concentrate on the development of our knowledge of how upward-propagating waves excited in the lower atmosphere act to force the middle atmospheric circulation. Of course, there is reason to believe that downward-propagating waves can lead to at least a modest stratospheric influence on the troposphere (e.g., Kirkwood and Derome, 1977; Boville, 1984), but the history of this aspect of atmospheric coupling will only be considered in passing.

The paper is organized as follows. Section 2 introduces the very early history of the study of gravity waves in the atmosphere. Section 3 discusses the first observations of high-frequency variability in the middle and upper atmosphere. Section 4 describes the initial identification of this variability as gravity waves propagating up from the lower atmosphere, and then briefly reviews more recent developments in observations and theory of gravity waves and equatorial waves. Section 5 discusses the history of the gravity wave-mean flow interaction problem and its important application to the theory of the stratospheric quasi-biennial oscillation. Section 6 discusses early observations and theoretical models of large-scale planetary waves in the extratropics. Section 7 considers the early work on planetary wave-mean flow interaction and models of stratospheric sudden warmings. Conclusions are briefly summarized in Section 8. The organization of this paper reflects the rather separate paths of developments in the study of high-frequency (frequencies higher than the Coriolis parameter) and low-frequency (approximately geostrophic) motions. At least until the late 1960's there was a tendency for the higher frequency motions to be the purview of aeronomers, while the lower frequency aspects of atmospheric variability were studied by meteorologists (with a strong focus on the deterministic weather forecasting problem). As the field of middle atmospheric science matured it became clear that both low-frequency and high-frequency waves significantly coupled the circulation of the middle atmosphere to that in the lower atmosphere, and that complete models of the middle atmosphere require an adequate treatment of both gravity wave and planetary waves effects.

2. Gravity Waves in the Atmosphere

While surface gravity waves have been studied for centuries, Rayleigh (1883) is apparently the first paper to discuss internal gravity waves in a continuous stratification. It is interesting that Rayleigh's motivation was meteorological, specifically to explain observations of wavelike perturbations in stratus clouds (Jevons, 1857). Rayleigh derived the dispersion relation for linear waves in an incompressible fluid with constant stratification. It is certainly noteworthy that Rayleigh included the exponential decrease of mean density with height characteristic of the atmosphere in his equations (essentially a version of what today would be termed the "anelastic system", e.g. Ogura and Phillips, 1962). Love (1891) covered somewhat similar ground with an analysis of waves in a continuously-stratified fluid with a free surface (relevant for oceanographic applications). Rayleigh (1890) returned to this subject, now considering a compressible atmosphere, but one in which both the mean state and perturbations are assumed to be isothermal (in the spirit of Isaac Newton's original attempt to explain the speed of sound). This work was motivated more by a desire to examine the global resonance properties of the atmosphere than from interest in small-scale waves. Rayleigh obtained a solution for a horizontally-propagating free wave - essentially his version of the familiar "Lamb wave" (see below). Of course, as Rayleigh himself noted, given his assumption of isothermal perturbations, the results he obtained cannot be quantitatively correct.

It is finally in the work of Lamb (1910) that the problem of linear, adiabatic internal gravity waves in a compressible atmosphere is treated completely. Lamb considered two cases: one in which the mean state is isothermal, and one with a mean state characterized by a constant temperature decrease with height (and hence an effective "top" of the atmosphere at finite height). He found the solutions corresponding to vertically-propagating waves and also noted the possibility of a purely horizontally-propagating solution with energy density trapped against the ground, what is now referred to as the Lamb wave.

Some other early work on gravity waves in the atmosphere focussed on explaining the quasi-periodic oscillations sometimes seen in surface microbarograph records. Vaisala (1925) and Brunt (1927) independently derived the formula for the buoyancy frequency (corresponding to the highest frequency internal gravity waves that can be supported) in a fully compressible atmosphere. This is now usually referred to by meteorologists as the Brunt-Vaisala frequency. Brunt (1927) noted that under realistic (tropospheric) conditions the buoyancy frequency corresponds to a period of the order of several minutes, roughly comparable to that some of the mircobarograph oscillations that had been reported. Further early work on high-frequency pressure fluctuations in the lower atmosphere was reported by Johnson (1929).

A striking aspect of all of these early papers on internal gravity waves is the absence of any expressed interest in the fate of waves propagating upward out of the lower atmosphere. The growth of plane wave amplitudes by a factor proportional to the square root of the inverse of the mean density is apparent in the solutions presented by Rayleigh (1883), Love (1891) and Lamb (1910). Rayleigh does not remark on this aspect of his solution, Love mentions his belief that the effects of viscosity will restrain the growth of waves with height, while Lamb more reasonably notes that "owing to the indefinite increase of amplitude as the waves travel upward... the results are not in all respects to be interpreted too literally". The idea that the upward propagating waves could produce interesting (and ultimately observable) effects in the upper atmosphere seems not to have occurred to these early researchers.

While the subject of gravity waves remained out of the mainstream of meteorological (and aerological) research until about 1960, there were two special cases that did receive some attention. One of these was atmospheric tidal motion, i.e. the response of the atmosphere to diurnal thermal and gravitational forcing. Weekes and Wilkes (1947) and Wilkes (1949) developed the theory of the tidal oscillations of a compressible atmosphere, and effectively showed that the response to a periodic forcing could take the form of a global-scale internal gravity wave (although significantly modified by planetary rotation). While the conclusion of these early papers (i.e. that global resonance plays an important role in determining the amplitude of the tidal response) turned out to be incorrect (e.g., Chapman and Lindzen, 1970), the basic mathematical development of the linear tidal theory was useful and was to be influential for later developments in the gravity wave theory of the upper atmosphere (Hines, 1989, remarks that in the late 1950's he was familiar with the concept of exponential growth of amplitude with height because of his earlier reading of these papers on tides). The other subject that received some attention in the years after World War II was the gravity wave response generated by flow over isolated topography (e.g., Queney, 1948; Scorer 1949; 1954). It is noteworthy that tides and topographic waves are the two classes of waves for which it is relatively straightforward to quantify the wave excitation. Perhaps understandably, there was much less interest in studying gravity waves from sources that were more difficult to characterize.

3. Initial Gravity Wave Observations in the Middle and Upper Atmosphere

The region of the atmosphere above the tropopause was essentially terra incognita before the 20th century. Two key developments near the turn of the century changed this. One was the discovery from in situ balloon measurements of the global temperature inversion that we now know as the stratosphere. The other was the observation of long range radio propagation by Marconi, which was followed by a recognition that the upper atmosphere must be ionized (Heaviside, 1902, Kennelly, 1902). The practical importance of radio propagation led to a great deal of research on the ionosphere. In particular, conductivity variations associated with the diurnal cycle and changes in solar activity were extensively studied. Interest in higher frequency changes in the ionosphere date at least as far back as the paper of Pierce and Mimno (1940), who interpreted the quasi-periodic fading of a trans-Atlantic shortwave radio broadcast every few minutes as indicating the presence of horizontally-propagating waves distorting the ionization in the F-region. Early observations of these "travelling ionospheric disturbances" as they came to be known were made by Munro (1950, 1958) and Heisler (1958), among others.

The first observations of the wind at mesospheric and thermospheric heights were obtained by exploiting the ionized trails left by meteors incident on the atmosphere. These techniques have been most useful in the height range roughly 85-110 km sometimes called the "meteor region". Manning et al. (1950) and Elford and Robertson (1953) derived vertical profiles of the wind using measured Doppler shifts of radar returns from the meteor trails. Liller and Whipple (1954) used time-lapse photography to obtain winds from the distortion of long-lived visible trails. Although limited to just a handful of meteor events, the Liller and Whipple results were quite influential, due to their rather unambiguous interpretation as indicators of neutral winds and because of their very high vertical resolution (~200 m). The distortions appeared to be predominantly in the horizontal, and thus were analyzed to determine the horizontal wind as a function of height. Apparent in their wind profiles is a complicated superposition of oscillations of vertical wavelengths varying from ~1-20 km. The fluctuations were found to have amplitudes of the order of several m-s-1 to several 10's of m-s-1. Through the 1950's a debate raged on the interpretation of the meteor radar observations. Eshelman and Manning (1954) believed the radar data could be explained by the relatively large-scale deformations of the meteor trails seen in the optical Liller and Whipple observations. Booker (1956) and Booker and Cohen (1956) acknowledged that the data indicted the presence of the large-scale fluctuations, but contended that these are part of a three-dimensional isotropic turbulent cascade. In hindsight it is easy to raise objections to Booker's theory. In particular, he imagined the meteor zone as a remarkably active region with vertical winds as strong as the observed horizontal winds (~25 m-s-1) and eddies with scales ~1 km having turnover times ~100 sec. The viscous dissipation of kinetic energy was estimated by Booker to be ~25 W-kg-1. As pointed out by Hines (1989), this is equivalent to a heating rate of ~2000oC-day-1, which is not plausible, since radiative cooling rates in this region are reliably estimated to be at most of order 10oC-day-1.

An important development was the first application of radar observations to determine the temporal and horizontal scales of the meteor level wind variations (Greenhow and Neufeld, 1959). In particular, Greenhow and Neufeld determined the decorrelation scales for the wind variations seen above their radar at Jodrell Bank in southern England to be ~2 hours, ~6 km in the vertical and ~150 km in the horizontal. Such strong spatial anisotropy and long timescales were inconsistent with a simple turbulent cascade. Greenhow and Neufeld noted the tendency for both spatial and temporal autocorrelations of the wind in space or time to become strongly negative at sufficiently large separations, suggesting the presence of coherent wavelike oscillations.

4. Development of Gravity Wave Theory of the High-Frequency Fluctuations in the Middle and Upper Atmosphere

Through the 1950's observational evidence of high frequency variations in the circulation above 80 km accumulated, but near the end of the decade many basic issues were still somewhat unclear. The relation of the wind fluctuations in the meteor zone to the ionospheric disturbances higher up, the degree to which the meteor zone wind fluctuations could be accounted for by tides and the possible role of turbulence in producing these fluctuations were all debated. The possibility that non-tidal gravity waves could be involved in explaining travelling ionospheric disturbances had been raised by Martyn (1950), but he subsequently withdrew his suggestion (Martyn, 1955).

The notion that internal gravity waves could account for the high-frequency variations observed in the upper atmosphere was revived by Hines (1959) and explicated at length (including reference to the then just-available Greenhow and Neufeld observations) in Hines (1960). In his 1960 paper Hines derived the plane-wave solutions for the linear perturbations about a motionless, isothermal basic state atmosphere, ignoring rotation and the sphericity of the earth, but including the full effects of compressibility. He found that the solutions fell into two classes: high-frequency acoustic waves (modified somewhat by the effects of gravity) and lower-frequency internal gravity waves. Hines showed that the dispersion relation he derived was consistent with the Greenhow and Neufeld (1959) observations of meteor level wind variations. He noted that waves with periods ~100 minutes (i.e. much longer than the Brunt-Vaisala period) would be associated with predominantly horizontal wind fluctuations, again in agreement with the most straightforward interpretation of the Liller and Whipple observations. He also noted the solutions indicated that wind amplitude of the waves should grow as the mean density decreases. He found some evidence for the expected amplitude growth with height in the optical and radar meteor wind observations then available (although these spanned a relatively small range of mean density). More importantly perhaps, Hines grasped the significance of the expected amplitude growth with height for the overall understanding of the atmospheric gravity wave field. Referring to the winds at meteor levels Hines wrote: "Disturbances associated with wind and weather disturbances in the lower atmosphere... may be expected to generate at their boundaries atmospheric waves which propagate away... in view of the decrease of density between the ground and the 90-km level .. it is evident that amplitudes of plane internal gravity waves could increase over the same height range by a factor of 700. Accordingly the observed upper atmospheric winds could be produced by a process of wave generation in the lower atmosphere whose associated oscillatory motions there would be only a few cm per second ... It does not seem unlikely that wave amplitudes of this order would in fact be produced in the troposphere quite commonly." Hines seems to have been the first researcher to understand the enormous implications of the wave amplitude growth with height for understanding upper atmospheric circulation.

Hines went on to explain travelling ionospheric disturbances as phenomena caused by deformation of the F-layer ionization during the passage of internal gravity waves. Observations that had become available (e.g., Munro, 1958) suggested that the perturbations of the F-layer ionization generally travelled downward. Hines realized that this downward phase progression was consistent with the upward energy propagation associated with gravity waves excited in the lower atmosphere. One additional aspect that needed to be explained was the observation that the travelling ionospheric disturbances appeared to have characteristic horizontal wavelengths and horizontal phase speeds that were larger than those dominant in the meteor wind data. Hines (1960) showed that the action of molecular viscosity would absorb many components of the spectrum of upward-propagating waves before they could reach F-layer heights. The waves with low frequency and low horizontal phase speed would be quenched, leaving the F-region to be dominated by the gravity waves with characteristics matching the observed travelling ionospheric disturbances.

The gravity wave theory for upper atmospheric motions initially aroused considerable controversy, particularly among some meteorologists who simply refused to believe that gravity waves (other than topographic waves) could play a significant role in the atmosphere, or who were confused by the counter-intuitive alignment of group and phase velocities of gravity waves (see Hines, 1989). Other researchers tried to explain the meteor level wind observations purely with astronomically-forced tidal oscillations (see Hines, 1966, for a critique of this view). Within a few years, however, it become generally accepted that the high-frequency wind variations seen in the upper reaches of the atmosphere reflect the presence of vertically-propagating gravity waves which are largely excited in the troposphere. Of course, there are genuinely turbulent motions present in the middle and upper atmosphere as well. Indeed at the mesospheric and lower thermospheric levels the atmosphere is thought to support a kind of spectral cascade with energy fed in at relatively large scales by gravity wave propagation from below, then some nonlinear transfer to smaller scales, until at sufficiently small scales the flow approximates isotropic turbulence. The details of the nonlinear energy transfer, the degree to which the motions at various scales are adequately described by linear theory, the degree of spatial and temporal intermittency, and the role of small-scale turbulence in dissipating kinetic energy are still very much open questions.

Once the importance of gravity waves in the middle and upper atmosphere was established, attention turned to more detailed modelling studies of the propagation, dissipation and nonlinear breakdown of gravity waves in mean states with height-varying wind and stratification (e.g., Freidman, 1966; Booker and Bretherton, 1967; Hines and Reddy, 1967, Hodges, 1976, 1969; Lindzen, 1970). The decade after Hines' original paper on gravity waves also saw the basic development of the theory of planetary-scale internal waves, particularly vertically-propagating equatorial waves trapped in low latitudes (Matsuno, 1966; Lindzen, 1967, 1971, 1972).

5. Effects of Gravity Waves on the Mean Flow

Consideration of the effects of gravity waves on the mean flow seems to have begun with a concern about the drag exerted by topography on the atmosphere. Starr (1948) noted that the momentum transfer between the atmosphere and the surface could be thought of as being the sum of a viscous drag and a force due to upstream/downstream differences in pressure across topographic features (what meteorologists refer to as "mountain torque"). Sawyer (1959) argued that in stably-stratified conditions the mountain torque might actually affect the mean flow at heights quite remote from the surface. The flow over topography generates gravity waves which radiate upward, and the driving of the mean flow then depends on the divergence of the Reynolds stress associated with the waves. Rather remarkably, Sawyer was motivated by a concern for the parameterization of the effects of topographic gravity waves in large-scale numerical simulation models of the atmosphere (then in their very earliest stages of development, of course!). Sawyer realized that the actual Reynolds stress profile would depend in a complicated way on the mean flow and dissipation, but suggested a simple ad hoc procedure for numerical models (simply assuming that the stress decayed uniformly with height through the troposphere).

The role of gravity waves in forcing the mean circulation was clarified by the classic work of Eliassen and Palm (1960). They were also mainly interested in topographic wave drag, but their results are applicable to gravity waves with nonzero phase speeds as well. Eliassen and Palm first considered linear, two-dimensional, vertically-propagating gravity waves in a time-invariant, but vertically-varying, mean flow and stratification. They showed that for steady waves outside regions of forcing and dissipation, and in the absence of critical levels (where the intrinsic horizontal phase speed of the wave is zero) the Reynolds stress is always constant. Thus wave transience, wave dissipation, or wave forcing are needed to cause mean flow driving. Waves can be seen as catalysts transferring mean momentum from the earth's surface (for topographic waves) or (more generally) from a wave forcing region to the level where the waves are dissipated.

As noted by Eliassen and Palm, the steady, inviscid, adiabatic, linear equations governing gravity waves have singularities at critical levels where the Doppler-shifted phase speed goes to zero. The behavior of gravity waves near critical layers was studied by Booker and Bretherton (1967) and Hines and Reddy (1967). It seemed that with some even small dissipation the critical level would act as an absorber for gravity waves and that the mean flow forcing from a single monochromatic wave would be concentrated in a thin region around the critical level, if such a level exists.

Lindzen and Holton (1968) used the results of these studies of gravity wave critical level behavior to advance a theory for the quasi-biennial oscillation (QBO) of the tropical stratosphere. The QBO had been discovered in long records of near-equatorial zonal wind observations by Reed et al. (1961) and Veryard and Ebdon (1961). The existence of the QBO was initially quite puzzling since the QBO appeared to be nearly periodic but was apparently not directly forced by astronomical variations. Also difficult to understand was the quite consistent downward propagation of the wind reversals through the middle and lower stratosphere. Early attempts at explanations involving extraterrestrial influence (Shapiro and Ward, 1962; Probert-Jones, 1964) were rather clearly unsatisfactory. Lindzen and Holton (1968) successfully explained the key features of the QBO by invoking the mean flow effects of a continuous spectrum of gravity waves forced in the troposphere, propagating into the stratosphere. In particular, they constructed a simple model with a mean flow (representing the equatorial zonally-averaged zonal wind) which is a function of height and time, and which is forced with the Reynolds stress divergence associated with each component of the wave spectrum. The eastward (westward) propagating waves force eastward (westward) mean flow accelerations a the critical levels where they are absorbed. The effects of these waves are somewhat self-limiting, however, as the production (say) of an eastward mean shear region with winds ranging between 0 and c will effectively shield higher regions from the effects of all waves with easward phase velocities less than c. Lindzen and Holton's numerical results showed that the presence of the self-limiting eastward and westward mean flow driving mechanisms could lead to an oscillatory mean wind, with the wind reversals descending with time (in impressive agreement with observations). Holton and Lindzen (1972) constructed a somewhat revised version of their model in which the mean flow accelerations were assumed to result from the interaction with discrete monochromatic eastward and westward propagating equatorial planetary-scale waves. Holton and Lindzen chose parameters for these waves roughly consistent with prominent equatorial waves that had been observed in radiosonde data (Wallace and Kousky, 1968: Maruyama et al., 1967), and their model produced a mean flow oscillation very similar to the real QBO in terms of amplitude, period and vertical phase structure.

The possible role of gravity waves in the extratropical zonal mean circulation was raised by Hines (1972) who used some observed case studies of waves near the mesopause to estimate that the associated Reynolds stress divergence could (at least on occasion) be of the order of several hundred m-s-1-day-1. As Hines (1989) remarks, at the time these values seemed to many meteorologists too large to be credible, and Hines' suggestion was not followed up directly by the community for some time. However, there was already evidence that a large dynamical drag was needed in the mesosphere. Early work of Murgatroyd and Singleton (1961) on the radiative balance in the middle atmosphere suggested that there must be strong sinking in the winter hemisphere and rising in the summer, and that continuity would require a significant flow from the summer into the winter mesosphere. Groves (1969) examined the available rocketsonde observations and found some support for the existence of a strong (>10m-s-1) poleward flow in winter upper mesosphere. Some kind of dynamical drag would then be needed to balance the Coriolis torque associated with the mean meridional circulation. In particular, strong westward drag is needed in the mesosphere of the winter hemisphere which should then act as a brake on the eastward polar night jet, and, similarly, eastward drag is needed on the westward jet in the summer hemisphere. The results from the zonally-averaged numerical model of Holton and Wehrbein (1980, a generalization of the work of Leovy, 1964) showed clearly that a very large eddy driving of the mean flow (of the order of 10's of m-s-1-day-1 in the extratropical upper mesosphere) was in fact needed to explain the observed zonal-mean wind and temperature structure in the stratosphere and mesosphere. Haurwitz (1961) had suggested a strong effective turbulent viscosity in the mesosphere might account for the needed drag. Houghton (1978) raised the possibility that the Reynolds stress from gravity waves might in fact balance the Coriolis torque. Lindzen (1981) and Matsuno (1982) showed that the sign of the required drag could be explained plausibly if one assumes that a broad spectrum of eastward and westward propagating waves emerges into the stratosphere from the troposphere. Selective absorption of the waves depending on their intrinsic phase speed then should lead to a dominance of westward propagating waves in the winter mesosphere. If these waves are absorbed in the mesosphere (perhaps by nonlinear breakdown) then they will generate the westward mean flow driving that seems needed to explain the observed circulation. A similar process should occur in the summer hemisphere. The significance of the gravity wave driving of the mean flow is now generally acknowledged and much current effort is directed at understanding the details of gravity wave forcing, propagation, and dissipation so that these wave effects can be incorporated into models of the global circulation of the middle atmosphere.

6. Extratropical Planetary Waves

While the basic physics of internal gravity waves is fairly intuitive, the fact that the large-scale horizontal gradient of potential vorticity can act as a restoring force for wavelike flow perturbations was not clearly appreciated until the theoretical work of Rossby (1940) and Haurwitz (1940) who treated linear waves in a purely barotropic (two-dimensional horizontal) fluid. Both Rossby and Haurwitz showed that the barotropic system could support transversely-polarized wave motions if the variation of the planetary vorticity with latitude is taken into account. Rossby also made the important contribution of demonstrating that the full spherical system could be usefully approximated in midlatitudes by a planar geometry as long as the variation of planetary vorticity with the meridional coordinate is included (the "beta-plane" approximation). The development of the quasi-geostrophic theory by Charney (1947, 1948) allowed a clear distinction to be made between the large-scale, low-frequency planetary-wave variations (characterized by a balance between pressure gradient and Coriolis forces) and the higher frequency gravity waves. The quasi-geostrophic theory also provided a formalism to deal with three-dimensional planetary waves. Charney and Eliassen (1949) performed a simple calculation of the mid-tropospheric planetary wave pattern forced by the surface wind blowing over topography. This work was extended by Smagorinsky (1953) to include the effects of zonal asymmetries in stationary heating.

While middle atmospheric applications were certainly not a major focus of the initial work with the quasi-geostrophic system, Charney realized that his equations supported vertically-propagating planetary waves. Charney (1949) made an estimate that the group velocity of linear vertically-propagating planetary waves in a resting mean state should be ~5 km per day. He then used this to estimate that a midlatitude surface weather forecast would be unaffected by poor (or nonexistent) input data in the stratosphere for at least 48 hours (interestingly the result of Charney's simple calculation is roughly consistent with more recent estimates of predictability influence, e.g., Smagorinsky, 1967). Ooyama (1958) applied the same analysis to explore the downward-propagation of planetary wave pulses that he supposed might be generated in the upper atmosphere by transient solar forcing.

The late 1950's turned out to be a crucial time for the observational and theoretical developments that led to the understanding of the significance of planetary waves in the middle atmosphere. Indeed in many ways the situation paralleled that for gravity waves at the same time, although the important observational advances for planetary waves came not from development of new techniques, but rather from the expansion and research application of the worldwide balloon sounding network established for weather forecasting. The first meridional sections of the zonal wind and temperature extending into the stratosphere were constructed in the mid-1950's (e.g., Kochanski, 1955) and the discovery of the strong eastward winter polar vortex and summertime stratospheric westward jet dates from this period. In the late 1950's and early 1960's papers also began to investigate the behavior of planetary-scale zonal asymmetries using meteorological analyses (i.e. horizontal maps of geopotential, wind and temperature) based on the operational radiosonde network. Some of the early work employed only North American sector analyses, but during and after the International Geophysical Year (1957-58) it became possible to plot fairly credible stratospheric analyses for the entire extratropical Northern Hemisphere. Some of these early observational studies include Austin and Krawitz (1956), Palmer (1959), Wexler (1959), Julian et al. (1959), Boville (1960), Hare (1960) and Boville et al. (1961). By the beginning of the 1960's a basic knowledge of the synoptic meteorology of the extratropical stratosphere had been achieved. It was understood that the summer stratosphere was much less disturbed than the winter stratosphere and that the winter vortex was typically distorted by very large-scale quasi-stationary zonal asymmetries and also (at times) by transient waves which could have somewhat smaller horizontal scales (but generally still much larger-scale than typical tropospheric synoptic waves). The large-scale stratospheric quasi-stationary waves were apparently direct upward extensions of tropospheric waves (although with considerable phase variation with height, so that an Aleutian anticyclone dominated at midstratospheric levels, above the familiar Aleutian low at the earth's surface).

A key development was the theoretical treatment of stationary quasi-geostrophic wave propagation in the famous paper of Charney and Drazin (1961; hereinafter CD). They were initially motivated by a consideration of the energy balance in the upper atmosphere. They noted that astronomers had explained the extreme warmth of the solar corona by invoking the mechanical heating by breaking acoustic waves forced far below in the convective region of the sun. This raised the question of why the meteorological motions in the earth's lower atmosphere did not radiate enough energy in the form of vertically-propagating waves to produce a terrestrial corona. Interestingly they referred to Hines' then recent work on gravity waves (particularly Hines emphasis on the implications of the expected growth of wave amplitude with height) as a motivation for their own inquiry into the propagation of large-scale quasi-geostrophic planetary waves.

CD examined the propagation of quasi-geostrophic stationary waves in a simple beta-plane geometry and in a mean flow with vertical, but no horizontal, shear. They showed that the propagation of stationary waves that might be forced by flow over topography or zonal asymmetries in heating will be very strongly affected by the mean flow in the stratosphere. In particular, stationary waves will be vertically-trapped, rather than propagating, in mean westward winds or in mean eastward winds that are too strong. The trapping criterion for eastward winds is dependent on the horizontal scale of the waves. CD showed for realistic wintertime mean flows only zonal wavenumbers one and two would propagate. Thus CD predicted the time-mean flow in the stratosphere ought to be very different from that in the troposphere, with essentially no zonal asymmetries in summer and only very large-scale stationary waves in winter. It is not clear how much CD were influenced by observational work being conducted at around the same time. They do not refer to any of the observational studies mentioned earlier in this section, although they do reference Murray (1960), who in turn refers to many of the then recent observational papers on large-scale circulation in the stratosphere. CD mentioned that their predictions for the character of the large-scale, low-frequency circulation in the stratosphere were borne out by their own inspection of stratospheric daily analyses from the US Weather Bureau and the Free University of Berlin, but gave no details.

Returning to their original motivation, CD noted that their results showed that much of the energy in large-scale waves forced in the troposphere should be trapped in the lowermost reaches of the stratosphere in both summer and winter and thus precluded from reaching the upper atmosphere. Their paper is a somewhat incomplete treatment of the "terrestrial corona" problem, ignoring gravity, equatorial and acoustic waves and not considering some nonlinear effects that would also be expected to limit planetary wave energy propagation of to high levels (e.g., Lindzen and Schoeberl, 1982; McIntrye and Palmer, 1984), but it did focus the interest of meteorologists on vertically-propagating planetary waves (and their effects on the global-scale circulation; see next section).

This initial theoretical work on vertically propagating quasi-stationary planetary waves in the middle atmosphere was followed over the next two decades by increasingly sophisticated calculations of the linear wave response to stationary forcing. Dickinson (1968, 1969) extended the theoretical quasi-geostrophic models of stationary planetary waves to include both vertical and horizontal shear. One crucial issue addressed was the behavior of planetary waves in the presence of critical surfaces (which for stationary waves occur whenever the mean wind is zero). Matsuno (1970) calculated the wave response to zonal wave one and two stationary forcing using a numerical solution of the quasi-geostrophic equations linearized about a realistic mean flow (which varied in both height and latitude). His results accorded quite well with more detailed observations of the three-dimensional structure of the Northern Hemisphere winter stratospheric stationary wave field then becoming available (e.g., Hirota and Sato, 1969). Matsuno's work was followed by many more sophisticated numerical calculations of the stationary waves (e.g., Schoeberl and Geller, 1977). A particular focus on later work was on the nonlinear behavior expected for planetary waves near their critical surfaces and the associated irreversible mixing (e.g., Warn and Warn, 1978; McIntyre and Palmer, 1984; Haynes, 1985).

7. Mean Flow Effects of Planetary Waves and the Stratospheric Polar Warming Phenomenon

The daily stratospheric balloon data that became available in the 1950's revealed the occasional occurrence of a remarkably rapid warming of the Northern Hemisphere winter polar stratosphere (Scherhag, 1952). This sudden warming phenomenon was studied using the increasingly complete synoptic maps that became available by Teweles and Finger (1958), Teweles (1958), Craig and Hering (1959) and Reed et al. (1963), among others. These studies documented the typical development of a very strong perturbation to the polar vortex with the vortex becoming either completely destroyed or very much displaced from the pole in a matter of a few days. Rather like the tropical QBO wind reversals, there is generally a clear downward propagation of the mean flow effects in a sudden warming. This initially suggested to some scientists that extraterrestrial forcing might be involved (e.g., Scherhag, 1952; Palmer, 1959).

The rapidity of the growth of zonal asymmetries in warming events led Fleagle (1958), Murray (1960) and Charney and Stern (1962) to consider the possibility that in situ instabilities could account for the sudden warming phenomenon. These studies were less than successful in explaining the observed features of the warmings. In 1971 Matsuno attempted a very different explanation of the sudden warming phenomenon in terms of the interaction of the mean flow in the stratosphere with the large-scale planetary waves forced in the troposphere.

Important background for Matsuno's work was provided by the careful consideration of the effects of planetary waves on the mean flow given by CD. In particular, CD realized that in a rotating system waves could force the mean flow both by Reynolds stress divergence and also from the divergence of eddy heat fluxes, since these heat fluxes would force mean meridional circulations which in turn would drive mean zonal flow accelerations via the Coriolis torque. They showed that the acceleration of an adiabatic zonal-mean flow forced by the eddy fluxes computed for steady, linear, adiabatic quasi-geostrophic planetary waves was zero. This is the quasi-geostrophic analogue of the Eliassen and Palm (1961) result for two-dimensional gravity waves. Matsuno (1971) was led by the CD results to focus on wave transience and critical level interaction as possibly crucial elements in generating a sudden warming. Matsuno also noted that he was influenced by the success of Lindzen and Holton's (1968) in modelling the QBO with wave critical level interactions.

Matsuno (1971) constructed a hemispheric, time-dependent, quasi-geostrophic model of the circulation above the tropopause and forced it with an imposed mid-latitude wave disturbance at the lower boundary with rapidly growing amplitude. This was meant to simulate some (not necessarily well understood) amplification of the tropospheric stationary wave field, possibly associated with the formation of a tropospheric blocking situation. Assuming reasonable wave amplitudes and rates of wave growth, Matsuno was able to simulate a rather realistic sudden warming in his model. He attributed the mean flow driving in his simulation to critical level absorption.

Matsuno's paper was followed by other studies of the sudden warming phenomenon using somewhat similar mechanically-forced numerical models (e.g., Geisler, 1974; Holton, 1976; Hsu and Holton, 1977; Schoeberl, 1978) . This work helped refine Matsuno's conception by emphasizing the roles of wave transience, dissipation and nonlinear interaction among planetary scale waves in the development of the flow during a warming. The early work on modelling of stratospheric sudden warmings was reviewed by Holton (1980).

The success of models incorporating wave effects on the mean flow in explaining both the QBO and stratospheric sudden warmings helped motivate some further basic theoretical work on wave-mean flow interaction. Particularly notable were efforts in the mid-1970's to generalize the "non-interaction" (or better "non-acceleration") theorems of Eliassen and Palm (1961) and Charney and Drazin (1961) to the full three-dimensional primitive equation system (Holton, 1974; Andrews and McIntrye, 1976; Boyd, 1976). The paper of Andrews and McIntyre (1976), which has been extremely influential in later work, also serves as an excellent summary of, and endpoint for, the early "heroic" period of middle atmospheric dynamics.

The study of the circulation in the middle atmosphere made enormous strides in the 1950's and 1960's. The understanding of the middle atmosphere as a region largely forced by mechanical stresses from the lower atmosphere and the elucidation of the details of this coupling were advanced by a remarkable array of observational and theoretical investigations. Much of this work was motivated by purely abstract scientific concerns. The supposed lack of practical interest in the stratosphere and mesosphere had even led these regions of the atmosphere to be referred to facetiously as the "ignorosphere". In retrospect, the existence of this pure scientific curiosity was fortunate for mankind, since it allowed a basic understanding of stratospheric circulation to be brought to bear on the issue of anthropogenic influence on the ozone layer, when this environmental concern arose quite unexpectedly in the early 1970's (Molina and Rowland, 1974).

Today there are still many uncertainties remaining in our understanding of middle atmospheric circulation, even on some rather basic aspects. For example, the relative roles of tropospheric forcing and internal middle atmospheric instability in generating synoptic variability in the stratosphere and mesosphere is still a somewhat open issue (e.g., O'Neill and Pope, 1988). However, the basic picture of the middle atmospheric circulation that emerged in the 1960's still forms the foundations of the field and underlies current investigations in such topical areas as the role of the middle atmosphere in anthropogenic climate change.

Acknowledgments. The author thanks P.G. Drazin for helpful correspondence and R.A. Vincent, K. Bryan and N.-C. Lau for valuable comments on the manuscript.

Andrews, D.G., J.R. Holton and C.B. Leovy, 1987: Middle Atmosphere Dynamics, Academic Press, New York, 489 pp.

Andrews, D.G. and M.E. McIntyre, 1976: Planetary waves in horizontal and vertical shear: the generalized Eliassen-Palm relation and the mean zonal acceleration. J. Atmos. Sci., 33, 2031-2048.

Austin, J.M. and L. Krawitz, 1956: Fifty millibar patterns and their relationship to tropospheric changes. J. Meteorol., 13, 152-159.

Booker, H.G., 1956: Turbulence in the ionosphere with applications to meteor-trails, radio-star scintillation, auroral radar echoes and other phenomena. J. Geophys. Res., 61, 673-705.

Booker, H.G. and R. Cohen, 1956: A theory of long-duration meteor echoes based on atmospheric turbulence with experimental confirmation. J. Geophys. Res., 61, 707-733.

Booker, J.R. and F.P. Bretherton, 1967: The critical layer for internal gravity waves in a shear flow. J. Fluid Mech., 27, 513-539.

Boville, B.W., 1960: The Aleutian stratospheric anticyclone. J. Meteorol., 17, 329-336.

Boville, B.W., C.V. Wilson and F.K. Hare, 1961: Baroclinic waves of the polar-night vortex. J. Meteorol., 18, 567-580.

Boville, B.A., 1984: The influence of the polar night jet on the tropospheric circulation of a GCM.. J. Atmos. Sci., 41, 1132-1142.

Boyd, J.P., 1976: Noninteraction of waves with the zonally averaged flow on a spherical earth and the interrelationships of eddy fluxes of energy, heat and momentum. J. Atmos. Sci., 33, 2285-2258.

Brunt, D., 1927: The period of simple vertical oscillations in the atmosphere. Quart. J. Roy. Meteorol. Soc., 53, 30-32.

Chapman, S. and R.S. Lindzen, 1970: Atmospheric Tides. Dordrecht Publisher, 200 pp.

Charney, J.G., 1947: The dynamics of long waves in a baroclinic westerly current. J. Meteorol., 4, 135-162.

Charney, J.G., 1948: On the scales of atmospheric motions. Geofys. Pub., 17, no. 2, 1-17.

Charney, J.G., 1949: On a physical basis for numerical prediction of large-scale motions in the atmosphere. J. Meteorol. 6, 371-385.

Charney, J.G. and P.G. Drazin, 1961: Propagation of planetary-scale disturbances from the lower into the upper atmosphere. J. Geophys. Res., 66, 83-109.

Charney, J.G. and A. Eliassen, 1949: A numerical method for predicting the perturbations of the midlatitude westerlies. Tellus, 1, 38-54.

Charney, J.G. and M.E. Stern, 1962: On the stability of internal baroclinic jets in a rotating atmosphere. J. Atmos. Sci., 19, 159-172.

Craig, R.A. and W.S. Hering, 1959: The stratospheric warming of January-February 1957. J. Meteorol. 16, 91-107.

Dickinson, R.E., 1968: Planetary Rossby waves propagating vertically through weak westerly wind wave guides. J. Atmos. Sci., 25, 984-1002.

Dickinson, R.E., 1969: Vertical propagation of planetary Rossby waves through an atmosphere with Newtonian cooling. J. Geophys. Res., 74, 929-938.

Eliassen, A. and E. Palm, 1961: On the transfer of energy in stationary mountain waves. Geofys. Pub., 22, no. 3, 1-23.

Elford, W.G. and D.S. Robertson, 1953: Measurements of winds in the upper atmosphere by means of drifting meteor trails II. J. Atmos. Terr. Phys., 4, 271-284.

Eshelman, V.R. and L.A. Manning, 1954: Radio communication by scattering from meteoritic ionization. Proc. Inst. Radio Eng., 42, 530-536.

Fleagle, R.G., 1958: Inferences concerning the dynamics of the mesosphere. J. Geophys. Res., 63, 137-145.

Freidman, J.P., 1966: Propagation of internal gravity waves in a thermally stratified atmosphere. J. Geophys. Res., 71, 1033-1054.

Geisler, J.E., 1974, Numerical model simulation of the sudden stratospheric warming mechanism. J. Geophys. Res., 79, 4989-4999.

Greenhow, J.S. and Neufeld, E.L., 1959: Measurements of turbulence in the 80 to 100 km region from the radio echo observations of meteors.29-2133.

Groves, G.V., 1969: Wind models from 60 to 130 km altitude for different months and latitudes. J. British Interplanetary Soc., 22, 285-307.

Hare, F.K., 1960: The disturbed circulation of the Arctic stratosphere. J. Meteorol., 17, 36-51.

Haurwitz, B., 1940: The motion of atmospheric disturbances on a spherical earth. J. Marine Res., 3, 254-267.

Haurwitz, B., 1961: Frictional effects and the meridional circulation in the mesosphere. J. Geophys. Res., 66, 2381-2391.

Haynes, P.H., 1985: Nonlinear instability of a Rossby wave critical layer. J. Fluid Mech., 161, 493-511.

Heaviside, O., 1902: Telegraphy I: Theory. Encyclopedia Brittanica (10th ed.) 33, 213-218.

Heisler, L.H., 1958: Anomalies in ionosonde records due to travelling ionospheric disturbances. Australian J. Phys., 11, 79-90.

Hines, C.O., 1959: Turbulence at meteor heights. J. Geophys. Res., 64, 939-940.

Hines, C.O., 1960: Internal atmospheric gravity waves at ionospheric heights. Can. J. Phys., 38, 1441-1481.

Hines, C.O., 1966: On the analysis and interpretation of winds observed at heights of 85 to 135 kilometers: a rebuttal. J. Geophys. Res., 71, 1461-1475.

Hines, C.O., 1972: Momentum deposition by atmospheric waves, and its effect on thermospheric circulation. Space Research, XII, 1157-1161.

Hines, C.O., 1974: The Upper Atmosphere in Motion. American Geophysical Union, 1022 pp.

Hines, C.O., 1989: Earlier days of gravity waves revisited. Pure. Appl. Geophys. 130, 151-170.

Hines, C.O. and C.A. Reddy, 1967: On the propagation of gravity waves through regions of wind shear. J. Geophys. Res., 72, 1015-1032.

Hirota, I. and Y. Sato, 1969: Periodic variations of the winter stratosphere circulation and intermittent vertical propagation of planetary waves. J. Meteorol. Soc. Japan, 47, 390-402.

Hodges, R.R., 1967: Generation of turbulence in the upper atmosphere by internal gravity waves. J. Geophys. Res., 72, 3455-3458.

Hodges, R.R., 1969: Eddy diffusion coefficients due to instabilities in internal gravity waves. J. Geophys. Res., 74, 4087-4090.

Holton, J.R., 1974: Forcing of mean flows by stationary waves. J. Atmos. Sci., 31, 942-945.

Holton, J.R., 1976: Semi-spectral numerical model for wave-mean flow interactions in the stratosphere: application to sudden stratospheric warmings. J. Atmos. Sci., 33, 1639-1649.

Holton, J.R., 1980: The dynamics of sudden stratospheric warmings. Ann. Rev. Earth Planet. Sci, 8, 169-190.

Holton, J.R., 1992: An Introduction to Dynamic Meteorology, Third edition. Academic Press, 511 pp.

Holton, J.R., and R.S. Lindzen, 1972: An updated theory for the quasi-biennial cycle of the tropical stratosphere. J. Atmos. Sci., 29, 1076-1080.

Holton, J.R. and W.M. Wehrbein, 1980: Numerical model of the zonal mean circulation of the middle atmosphere. Pure Appl. Geophys., 118, 284-306.

Houghton, J.T., 1978: The stratosphere and mesosphere.. Quart. J. Roy. Meteorol., 104, 1-29..

Hsu, C.P. and J.R. Holton, 1977: Stratospheric sudden warming: the role of wave-wave interactions. Collection of extended summaries of contributions presented at CMUA sessions, IAGA/IAMAP Joint Assembly 22, Seattle. pp. 48.1-48.6.

Jevons, W.S., 1857: On the cirrous form of clouds. Phil. Mag., 4th series, 14, 22-35.

Johnson, N., 1929: Atmospheric oscillations shown by the microbarograph. Quart. J. Roy. Meteor. Soc., 55, 19-30.

Julian, P., L. Karwitz and H.A. Panofsky, 1959: The relation between height patterns at 500 mb and 100 mb. Mon. Weath. Rev., 87, 251-256.

Kennelly, A.E., 1902: On the elevation of the electrically conducting strata of the Earth's atmosphere. Elec. World and Eng., 39, 473.

Kirkwood, E. and J. Derome, 1977: Some effects of the upper boundary condition and vertical resolution on modeling forced stationary planetary waves. Mon. Weath. Rev., 105, 1239-1251.

Kochanski, A., 1955: Cross sections of the mean zonal flow and temperature along 80W. J. Meteorol., 12, 95-106.

Lamb, H., 1910: On the theory of waves propagated vertically in the atmosphere. Proc. London Math. Society, 7, 122-141.

Leovy, C.B., 1964: Simple models of thermally driven mesospheric circulation. J. Atmos. Sci., 21, 327-341.

Liller, W. and F.L. Whipple, 1954: High-altitude winds by meteor train photography. J. Atmos. Phys. Special Supplement, 1, 112-130.

Lindzen, R.S., 1967: Planetary waves on beta-planes. Mon. Weath. Rev., 95, 441-451.

Lindzen, R.S., 1970: Internal gravity waves in atmospheres with realistic dissipation and temperature. 1. Mathematical development and propagation into the thermosphere. Geophys. Fluid Dyn., 1, 303-355.

Lindzen, R.S., 1971: Equatorial planetary waves in shear I. J. Atmos. Sci, 28, 609-622.

Lindzen, R.S., 1972: Equatorial planetary waves in shear II. J. Atmos. Sci, 29, 1452-1463.

Lindzen, R. S., 1981: Turbulence and stress owing to gravity wave and tidal breakdown. J. Geophys. Res., 64, 9707-9714.

Lindzen, R.S., 1987: On the development of the theory of the QBO. Bull. Amer. Met. Soc., 68, 329-337.

Lindzen, R.S., and J.R. Holton, 1968: A theory of the quasi-biennial oscillation. J. Atmos. Sci, 25, 1095-1107.

Lindzen, R.S. and M.R. Schoeberl, 1982: Note on the limits of Rossby wave amplitudes. J. Atmos. Sci., 39, 1171-1174.

Love, A.E.H., 1891: Wave motion in a heterogeneous heavy liquid, Proc. Lond. Math. Soc. 22, 307-316.

McIntyre, M.E. and T.N. Palmer, 1984: The "surf zone" in the stratosphere. J. Atmos. Terr. Phys., 46, 825-849.

Manning, L.A. , O.G. Villard and A.M. Peterson, 1950: Meteoric echo study of upper atmospheric winds. Proc. Inst. Radio Eng., 38, 877-883.

Martyn, D.F., 1950: Cellular atmospheric waves. Proc. Roy. Soc. London, A201, 216-234.

Martyn, D.F., 1955: Interpretation of observed F2 'winds' as ionization drifts associated with magnetic variations. The Physics of the Ionosphere, Physical Society, London, pp. 161-165.

Matsuno, T., 1966: Quasi-geostrophic motions in the equatorial area. J. Meteorol. Soc. Japan, 44, 25-42.

Matsuno, T., 1970: Vertical propagation of stationary planetary waves in the winter Northern Hemisphere. J. Atmos. Sci., 27, 871-883.

Matsuno, T., 1971: A dynamical model of the stratospheric sudden warming. J. Atmos. Sci., 28, 1479-1494.

Matsuno, T., 1982: Quasi-one-dimensional model of the middle atmosphere circulation interacting with internal gravity waves J. Meteorol. Soc. Japan, 60, 215-226.

Molina, M.J. and F.S. Rowland, 1974: Stratospheric sink for chlorofluoromethanes: chlorine atom-catalysed destruction of ozone. Nature, 249, 810-812.

Munro, G.H., 1950: Traveling disturbances in the ionosphere. Proc. Roy. Soc. London, A202, 208-223.

Munro, G.H., 1958: Traveling ionospheric disturbances in the F-region. Australian J. Phys., 11, 91-112.

Murgatroyd, R.J. and F. Singleton, 1961: Possible meridional circulations in the stratosphere and mesosphere. Quart. J. Roy. Meteorol. Soc., 87, 125-135.

Murray, F.W., 1960: Dynamic stability in the stratosphere. J. Geophys. Res., 65, 3273-3305.

Ogura, Y. and N.A. Phillips, 1962: A scale analysis of deep and shallow convection in the atmosphere. J. Atmos. Sci, 19, 173-179.

O'Neill, A. and V.D. Pope, 1988: Simulations of linear and nonlinear disturbances in the stratosphere. Quart. J. Roy. Meteorol. Soc., 114, 1063-1110.

Ooyama, K., 1958: On the vertical propagation of a disturbance through the mesosphere. New York University Dept. of Meteorology and Oceanography Report. 69 pp.

Palmer, C.E., 1959: The stratospheric polar vortex in winter. J. Geophys. Res., 64, 749-764.

Pedlosky, J., 1987: Geophysical Fluid Dynamics. Springer-Verlag, 710 pp.

Pierce, J.A. and H.R. Minmo, 1940: The reception of radio echoes from distant ionospheric irregularities. Phys. Rev., 57, 95-105.

Probert-Jones, J.R., 1964: An analysis of fluctuations in the tropical stratospheric wind. Quart. J. Roy. Meteorol. Soc., 90, 15-26.

Queney, P., 1948: The problem of airflow over mountains: A summary of theoretical studies. Bull Amer. Meteorol. Soc., 29, 16-26.

Rayleigh, Lord, 1883: Investigation of the character of the equilibrium of an incompressible heavy fluid of variable density. Proc. London Math. Society, 14, 170-177.

Rayleigh, Lord, 1890: On the vibrations of an atmosphere. Phil. Mag., 5th series, 29, 173-180.

Reed, R.J., W.J. Campbell, L.A. Rasmussen and D.G. Rogers, 1961: Evidence of a downward-propagating annual wind reversal in the equatorial stratosphere. J. Geophys. Res., 66, 813-818.

Rees, M.H., 1989: Physics and Chemistry of the Upper Atmosphere. Cambridge University Press, 289 pp.

Rossby, C.-G., 1940: Planetary flow patterns in the atmosphere. Quart. J. Roy. Meteorol. Soc., 66, 68-87.

Sawyer, J.S., 1959: The introduction of the effects of topography into methods of numerical forecasting. Quart. J. Roy. Meteorol. Soc., 85, 31-43.

Schoeberl, M.R., 1978: Stratospheric warmings: observations and theory. Rev. Geophys. Space Phys., 16, 521-538.

Schoeberl, M.R. and M.A. Geller, 1977: A calculation of the structure of planetary waves in winter. J. Atmos. Sci, 34, 1235-1255.

Scorer, R.S., 1949: Theory of waves in the lee of mountains. Quart. J. Roy. Meteorol. Soc., 75, 41-56.

Scorer, R.S., 1954: Theory of airflow over mountains, III: Airstream characteristics. Quart. J. Roy. Meteorol. Soc., 80, 417-428.

Shapiro, R. and F. Ward, 1962: A neglected cycle in sunspot numbers? J. Atmos. Sci, 19, 506-508.

Smagorinsky, J., 1953: The dynamical influence of large-scale heat sources and sinks on the quasi-stationary mean motions of the atmosphere. Quart. J. Roy. Meteor. Soc., 79, 342-366.

Smagorinsky, J., 1967: The role of numerical modelling. Bull. Amer. Meteorol. Soc., 48, 89-93.

Starr., V.S., 1948: An essay on the general circulation of the earth's atmosphere. J. Meteorol., 5, 39-43.

Teweles, S. and F.G. Finger, 1958: An abrupt change in stratospheric circulation beginning in mid-January 1958. Mon. Wea. Rev., 86, 23-28.

Vaisala, V., 1925: Uber die Wirkung der Winschankungen auf die Pilotbeobachtungen. Soc. Sci. Fenn, Commentat. Phys.-Math, 2, 19-37.

Veryard, R.G. and R.A. Ebdon, 1961: Fluctuations in tropical stratospheric winds. Meteorol. Magazine, 90, 127-143.

Wallace, J.M. and V.E. Kousky, 1968: Observational evidence for Kelvin waves in the tropical stratosphere. J. Atmos. Sci, 25, 900-907.

Warn, T. and H. Warn, 1978: The evolution of a nonlinear critical level. Stud. Appl. Math, 59, 37-71.

Weekes, K., and M.V. Wilkes, 1947: Atmospheric oscillations and the resonance theory. Proc. Roy. Soc. A192, 80-99.

Wilkes, M.V., 1949: Oscillations of the Earth's Atmosphere. Cambridge University Press, 250 pp.

Wexler, H., 1959: Seasonal and other changes in the Antarctic atmosphere. Quart. J. Roy. Meteor. Soc., 85, 196-208.

Yanai, M. and T. Maruyama, 1966: Stratospheric wave disturbances propagating over the equatorial Pacific. J. Metoerol. Soc. Japan, 44, 291-294.