The simulation considers an initially monochromatic gravity wave (GW) with
horizontal wavelength 120 km and vertical wavelength 24 km. The computational
domain extends from 100 to 250 km in the vertical, 120 km in the streamwise
and 120 km in the lateral direction and 200 x 200 x 256 mesh points were used.
The mesh is uniform in the x and y directions, giving spacings of 600 m. The
mesh is stretched in the vertical direction, with a minimum spacing of
127 m at the lower boundary and a maximum spacing of 1611 m a the upper
boundary. A Realistic background thermodynamic state based on Vadas and
Fritts 2006, 2007 is used. The GW has a period of 1525 sec and a horizontal
phase speed of 78.7 m/s at the lower boundary. The simulation begins
with a x-directed uniform wind of Uo=78.7 m/s, and no GW perturbation.
The GW is introduced gradually via forcing at the lower boundary such that
its non-dimensional amplitude (at the lower boundary) grows from zero to
0.15 over a period of 0.4 wave periods. From this point on, the GW amplitude
is held fixed at the lower boundary. The wave propagates upward, ultimately
filling the domain. Some energy is lost due to viscous dissipation and the
remainder exits through the radiative upper boundary. As the wave approaches
steady state, instability emerges in the altitude range ~130-185 km, at a time
of ~4.5 wave periods. The instability leads to turbulence, which descends
somewhat to the altitude range ~120-180 km. Currently the simulation has
been run out to 5.5 wave periods.
A 100-frame time series of 3D data has been taken during the interval
t=5430-8400 sec (3.5-5.5 wave periods). This interval includes the
approach to steady state, instability, and turbulence. The data is sampled
on a uniform grid with spacing 1 km in all three directions. The full
domain is used in x and y, whereas the vertical direction is limited to
the sub-domain between 105 and 225 km. The spacing in time is 30 sec.
The time series data files are contained in a 3.3 GB tar file and can be
downloaded by clicking the link below. On a unix (or mac) system issue
'tar -xf time_series.tar' to extract the individual data files from the tar
archive. The data files contains u(Nx,Ny,Nz,n), where Nx, Ny, and Nz are
the number of points in x, y, and z, and n=1:5 is for u', v', w', theta', p'.
In this case Nx=120, Ny=120, Nz=120. The background.state file provides
state(Nz,n) where n=1:5 is for rho_bar, p_bar, T_bar, theta_bar, N^2.
All quantities are dimensional in standard S.I. units.
A sample fortran program which illustrates how to read the data is also
provided below.
