The seasonal gravity wave activities were examined though the calculation of mean gravity wave energy densities,
and 
. was directly calculated from the observed perturbation. while was estimated assuming the observed perturbations to be due to
the gravity waves. In the stratosphere, and were almost the same in their magnitude and seasonal variation.
This similarity suggests that the calculated perturbation variables
in the stratosphere can be considered to be due to gravity waves.
The strong gravity wave activity in the stratosphere appeared
in January and November. Because the gravity wave activity depends
on wave sources and mean flow conditions, further information
on the wave sources was needed. In this study, the characteristics
of wave sources were estimated through the caculation of the group
velocity 
and intrinsic phase velocity
. Through the nonlinear fitting, we can estimate spatial scales of gravity wave. The vertical gravity wave scales of 4.37 km and 3.92 km were dominant in the stratosphere and the troposphere, respectively. The spectral slopes were slightly less than -3 except for the stratospheric PSD in May, July, and August, and the spectral slopes in the stratosphere were usually less than those in the troposphere.
Mean intrinsic frequencies(
) of gravity waves were calculated using Stoke's parameter methods.
In the stratosphere, the calculated
s were about 2.26 times as large as the inertial frequency at
Pohang, Korea, and any significant seasonal variation of
was not found. Mean horizontal scale of gravity waves was estimated using the dispersion relation for the inertia gravity waves, mean vertical wavelength in PSD, and the estimated intrinsic frequency. The aspect ratio of horizonal scale to vertial scale of gravity waves was about 200 in the stratosphere. This indicates that the observed wave motion is almost horizontal. That is, the large portion of wave energies can be explained by horizontal kinetic energy.
The intrinsic phase velocities showed a clear seasonal variation in the stratosphere. Westward propagation was dominant in winter, while there exist a weak eastward progation in summer. For the gravity waves propagating their energy upward, such a anisotropy of the wave propagation in the stratosphere suggests that the downward and upward transfer of zonal momentum will be observed in winter and summer, respectively. As we expected, The momentum flux estimated in this study showed the seasonal variation similar to that of the intrinsic phase velocity. The significant vertical gradient of zonal momentum flux was seen in the stratosphere in January and November when the
and were much stronger compared to those in the other months. Thus
the revesal of zonal wind near z = 33 km in January and November
is thought to be greatly associated with the activity of gravity
waves and the vertical structure of zonal momentum flux.