Previous: Observation and data analysis Next: Summary Up:Ext. Abst.
3. Gravity wave characteristics
For the analysis of the seasonal variation of gravity wave
activity, monthly mean gravity wave energy densities, and
, are calculated in the stratosphere and the troposphere.
and
are given by
![]() |
![]() |
![]() |
(1) |
![]() |
![]() |
![]() |
(2) |
where and
can be derived using three-dimensional gravity wave spectrum
model suggested by Fritts and VanZandt (1993).
Figure 2 shows time series of monthly mean and
in the stratosphere and the troposphere. Although gravity wave
energy
are derived in the vertically uniform basic state zonal wind
and static stability, the magnitude and tendency of
and
in the stratosphere are almost the same. However,
is quite different from
in the troposphere. The similarity between
and
in the stratosphere suggests that the calculated perturbation
variables in the stratosphere can be referred to as gravity wave
perturbations.
![]() |
In the stratosphere, is much larger in January and November than in the other months.
The gravity wave activity in the stratosphere can be controlled
by the basic state flow and the characteristics of wave sources.
The stationary mountain waves can propagate vertically into the
stratosphere in winter because they do not meet critical level
for the zonal wind structure in the troposphere(Figure 1b). However,
those stationary waves can not propagate into the stratosphere
in summer because of the reversed zonal wind near the tropopause.
For the summer zonal wind structure, non-stationary waves induced
by convective storms may not propagate into the stratosphere,
either. Thus the strong wave activity in the winter stratosphere
can be determined by the combination of the above-mentioned wave
sources and basic state flow conditions.
In the atmosphere, there are many possible sources for gravity
waves besides mountain and convection. Kitamura and Hirota(1989)
showed the relevance of the subtropical jet to the wave activity
through the anaysis of the propagation direction of waves. Their
study suggests that the observed gravity waves in this study are
generated near the subtropical jet region far away form Pohang.
Accordingly, wave propagation characteristics should be estimated
in order to precisely analyze the strong gravity wave activity
in the winter stratosphere. However, the spectral characteristics
and dominant spatial and temporal scales of gravity wave should
be calculated in advance in order to estimate wave propagation
characteristics.
In this study, we calculated the power spectral densities
of the normalized temperature as a function of the vertical wavenumber.
Allen and Vincent (1995) fitted their model spectra into the monthly
mean PSD to obtain several spectral characteristics of gravity
waves. The model spectra used in Allen and Vincent is given by
![]() |
(3) |
where is vertical wavenumber(
),
is the characteristic vertical wavenumber(
), and
is the log-scale spectral slope in the large vertical wavenumber
region.
The characteristic vertical wavenumber() indicates the dominant vertical scale in the observed gravity
wave field because the gravity wave energy is concentrated near
the vertical scale corresponding to
in the area-preserving form of PSD(not shown). Yearly mean
s are
(4.37 km) and
(3.92 km) in the stratosphere and the troposphere, respectively.
The spectral slopes(
) of monthly mean PSDs in the large vertical wavenumber region
are slightly less than -3 except for the stratospheric PSD in
May, July, and August. Yearly mean
are 2.66 and 2.86 in the stratosphere and the troposphere, respectively.
The intrinsic frequency and mean propagation direction of
the wave are estimated using Stoke's parameter method(Eckermann
and Vincent 1989) and Hilbert transform, and mean horizontal and
vertical wavenumber are obtained assuming that the observed perturbation
variables are due to inertia gravity waves. The monthly mean vertical
wave lengths are about 2.94 km and 2.55 km in the stratosphere
and the troposphere, respectively. The estimated monthly mean
horizontal wave lengths are about 430.94 km and 96.59 km in the
stratosphere and the troposphere, respectively. Thus it can be
immediately seen that the horizontal scales of waves are, on average,
200 times as large as the vertical scales of waves in the stratosphere.
The ratio of intrinsic frequency to inertia frequency (
) are about 2.26 with small seasonal variations in the stratosphere.
The intrinsic phase speed and group velocity are written as
![]() |
(4) |
![]() |
(5) |
where
is the intrinsic frequency,
is mean horizontal wavenumber, and
is mean propagation direction.
Figure 3 shows the intrinsic phase velocity and group velocity
vectors in the stratosphere in January and July. In the stratosphere,
the mean direction of intrinsic phase velocities in winter is
mainly toward the northwest, while that in summer is toward the
northeast. That is, the observed gravity waves in the stratosphere
have the anisotropic propagation characteristics. The monthly
mean
s in the stratosphere show an interesting seasonal variation that
there exist the negative
in winter, and positive in summer. In the troposphere, however,
the propagation characteristics do not have the anisotropy that
exists in the stratosphere.
![]() |
The monthly mean shows the seasonal variation opposite to that of
in the stratosphere. Thus we can see that the basic state wind
significantly affects the dominant direction toward which the
gravity wave energy propagates.
In this study, because is not directly observed, the zonal and meridional momentum fluxes
are indirectly estimated in order to examine the interaction between
the observed gravity waves and the large-scale circulation. the
zonal and meridional momentum fluxes are calculated using
![]() |
(6) |
![]() |
(7) |
where
is the spectral average value,
is the Hilber transformed normalized temperature perturbation,
and
.
From the monthly mean zonal and meridional momentum flux
in the stratosphere, we can see a clear seasonal variation of
zonal momentum flux. In the stratosphere, zonal momentum is transported
downward in winter, while that is transported upward in summer.
Because the intrinsic phase velocities in the winter stratosphere
are mainly westward, the downward trasfer of the zonal momentum
should be observed for the gravity waves that propagate their
energy upward. Thus it is expected that the gravity waves will
deposit their negative momentum to the large-scale flow, and accelerate
the large-scale zonal flow westward in the region where the diffusion
or breaking of the gravity waves exist. As a result, the easterly
zonal mean flow in the winter stratosphere in January and November
may be due to the deposition of gravity wave momentum to the large-scale
zonal flow. Non-zero vertical gradient of zonal and meridional
momentum flux in January and November can be clearly seen in the
vertical profiles of monthly mean zonal and meridional momentum
fluxes in the stratosphere. In January, the magnitude of the zonal
momentum flux above z = 20.5 km decreases with the altitude, and
approaches to zero above z = 28 km. In November, the magnitude
of zonal momentum flux decrease rapidly with height in the altitude
range between z = 19.5 km and 22 km. This vertical structure of
the zonal momentum flux in the winter stratosphere indicates that
the gravity waves can accelerate the large-scale zonal wind westward.