Recognition of the importance of gravity waves for upper
atmosphere structures has led to numerous experimental and
theoretical efforts. Important experimental tools for the
measurement of individual gravity waves and gravity wave spectra
include lidar [e.g., Bills and Gardner, 1993; She et
al., 1991], radar [e.g., Fritts and VanZandt, 1993],
satellite optical observations [ Mende et al., 1994] and
all-sky imagery from the ground [e.g., Taylor and Hill,
1991]. Signatures of ``breaking'' gravity waves at mesopause
altitudes (
90 km) are of particular interest. These waves are
generated at lower altitudes due to tropospheric effects, such as
the passage of weather fronts and intense pressure regions. The
waves propagate upwards and grow in amplitude, until they reach a
critical level near the mesopause where they ``break,'' depositing
energy and momentum.
Figure 3 [ Fritts et al., 1993] shows a photograph of a
noctilucent cloud at 
86 km altitude, together with a
numerical model calculation of breaking gravity waves near the
mesopause. Noctilucent clouds are visible from the ground typically
only near twilight during late summer months at high latitudes. The
clouds, which provide a screen on which the atmospheric waves in
this region can be imaged, often exhibit wave structures that have
multiple characteristic scales. The smallest scale waves (meters to
kilometers) are thought to be the most important for forcing the
mesosphere/lower thermosphere. The ongoing work is motivated by
the desire to develop a quantitative description of gravity wave
spectra and occurrence frequencies, which will lead to a greater
understanding of energy transport and coupling between regions.
Theoretical advances in our understanding of gravity wave production and dissipation have also taken place recently, leading perhaps most significantly to methods for treatment in general circulation codes [ Fritts and Lu, 1993]. Figure 4, taken from the recent review by Gardner [1994] shows a schematic ``m-spectrum'' of gravity waves. Current models for the saturation (dissipation process) include the cascade model of Dewan [1994], the Doppler spreading theory of Hines [1991], the wave-induced diffusion theory of Gardner [1994], and the spectral collocation code of Fritts [1994] and Palmer et al., [1994].