The dynamics of the lower thermosphere has been recently
reviewed by Fuller-Rowell [1994]. The large scale
circulation of the MLT has been addressed experimentally over the
past quadrennium by chains of medium/high frequency (MF/HF) radars
that can determine wind systems between 
60 and 100 km [e.g.,
Manson et al., 1991a]. The validity of MF/HF measurements
above 
80 km has been questioned due to ambiguities between
the Doppler velocities of the observed density fluctuations and the
background wind [ Hines et al., 1993]. However, an
experimental study by Manson et al. [1991b] has shown
excellent agreement between measurements made by a co-located MF
radar and a Fabry-Perot interferometer (an optical instrument with
sufficiently high spectral resolution to determine winds through
observations of Doppler-shifted atmospheric emissions).
Measurements from incoherent scatter radars have also been used to
extend the range of observations of the MLT into the lower
thermosphere [ Johnson, 1991; Reese et al., 1991;
Johnson and Virdi, 1991; Salah et al., 1991; Morton et
al., 1993a, Salah et al., 1994; Salah, 1994],
providing information on tides and tidal variability. Satellite
measurements of lower thermospheric winds at high latitudes have
been presented by Killeen et al., [1992] from Doppler
observations of O(
S) metastable atomic oxygen emissions. Also
significant new global observations from the UARS are appearing in
the literature, elucidating the zonal mean winds in the equatorial
mesosphere [ Lieberman et al., 1993], the global nature of the
diurnal tide [ Morton et al., 1993b; Hays et al., 1994],
and describing a large and significant þtwo-dayþ planetary wave
[ Wu et al., 1993]. Satellite observations of the two-day
wave complement recent progress made with observations from
ground-based instruments [ Clark et al., 1994] and theoretical
approaches [ Hagan et al., 1993].
The effects of geomagnetic activity in the upper mesosphere and thermosphere have also been investigated using theoretical and observational approaches. Fesen et al., [1991c] studied the effect of auroral activity on mid-latitude semidiurnal tides using the National Center for Atmospheric Research's Thermosphere-Ionosphere General Circulation Model (NCAR TIGCM), a previous version of the TIMEGCM. In a subsequent study, the global effects of geomagnetic activity on thermospheric tides was investigated [ Fesen et al., 1993b]. These studies show that the response of the atmosphere depends on activity level, latitude, and altitude with maximum effects at high altitude and high latitude. Fuller-Rowell et al. [1991] simulated the response of the upper atmosphere to the variable geomagnetic conditions during the first Lower Thermospheric Coupling Study (LTCS-1) with the University College London/Sheffield Thermosphere-Ionosphere Model. The model simulations were successful in predicting the range of observed tidal amplitudes at mid to high latitudes when tidal forcing functions imposed near the lower boundary of the model were increased from the values inferred from linear tidal models. Johnson and Virdi [1991] compared lower thermospheric neutral winds derived from the European Incoherent-Scatter Radar facility (EISCAT) and the Sondrestrom incoherent scatter measurements obtained during LTCS-1. This 5 day interval included a period of enhanced geomagnetic activity. Both radars observed surges in the neutral winds associated with the enhancement and in approximate agreement with NCAR TIGCM simulation results. Johnson and Luhmann [1993] reported observations of highly unusual wave and tidal structures determined from results obtained during periods of continuous MST measurements at Poker Flat which included two solar proton events.
Figure 2 illustrates theoretically-calculated low latitude
zonal-mean zonal and meridional wind profiles as a function of
local time and altitude from 30-500 km. These winds were calculated
by the TIMEGCM of Roble and Ridley [1994] for a solar minimum
case with diurnal tides specified (at the 30 km altitude lower
boundary). The model predicts the vertical propagation of the
diurnal tide with amplitude growth, until dissipation occurs at
90-100 km due to the relatively short vertical wavelength (
30
km) and the effects of cancellation over the height of the source
region. This is seen in the zonal mean meridional winds shown in
Figure 2b. Above 100 km, the self-consistently-calculated
semi-diurnal tides dominate (having longer vertical wavelengths).
Whereas the importance of tides is clear from such a presentation,
the influence of gravity waves is not as obvious since these waves
are basically sub-grid-scale phenomena. The effects of gravity wave
forcing, however, are in fact profound and changes in the
theoretical prescription for the gravity wave forcing process in
the TIMEGCM lead to appreciable changes in the calculated general
circulation [ Roble and Ridley, 1994]. For example, Roble
and Ridley were able to use the gravity wave parameterization of
Fritts and Lu, [1993], to control the extension into the
thermosphere of the zonal-mean zonal wind jets (shown in Figure
2a), while simultaneously controlling the level at which the
diurnal tide was dissipated (Figure 2b). Thus gravity wave
descriptions can now be used in models to simulate both the mean
circulation and tides in the upper mesosphere.