Test Case 1: Mountain-triggered breaking gravity waves

Description

This test examines vertically propagating gravity waves induced by a mountain. Due to the steepness of the topography, gravity waves are directed upwards at a slight angle off the vertical. The model top in this test is approximately 40 km high, with the domain split vertically into a troposphere (~0 km -- 20 km altitude) and a stratosphere (20 km -- 40 km). Temperature reduces with height in the troposphere, but increases with height in the stratosphere (see the US standard atmosphere). This variation in temperature profile is mimicked in this test with a lapse rate of -5 K/km in the troposphere and +5 K/km in the stratosphere. The temperature also varies in the horizontal coordinate (longitude, latitude) due to the mountain, which sits at 20 degrees North. A vertical slice of temperature in the Northern hemisphere is shown in the bottom left figure.

A key parameter of atmospheric flows is the Brunt–Väisälä frequency, N, which governs static stability. Due to variations in lapse rate between the troposphere and stratosphere, there is a sharp transition in N at the tropopause (the boundary between the troposphere and stratosphere); see the below, middle, figure. This large gradient in N causes the gravity waves to break at the tropopause. An example of this is shown in the below right figure with the SE dycore. The horizontal divergence is shown as a diagnostic of gravity wave propagation, along with the potential temperature in contours.

This test draws inspiration from Skamarock et al. (2019), which investigated the impact of vertical resolution in dynamical cores, with one example being for gravity wave breaking. Our test case uses a different topography and temperature profile, but similarly allows for the comparison of different vertical resolutions.

Modifications: All dynamical cores

1.  Three vertical grids of varying resolution are available: L88 (default), L120, L207, where L denotes the number of levels. These correspond to maximum vertical spacings of 800 m, 400 m, and 200 m respectively. How does increasing the vertical resolution impact the simulated breaking of the gravity waves in the stratosphere?

To change the number of vertical levels, make the following xmlchange:

./xmlchange --file env_build.xml --id CAM_CONFIG_OPTS --val "--phys held_suarez --analytic_ic --nlev=xx", where xx in {88,120,207}.

This will require making a new build, with:

case.build --clean-all

case.build

2. Modify the lapse rates in the troposphere and stratosphere.

3. Change the height of the tropopause; this is currently 20 km.

4. Modify the stratospheric lapse rate. 

Modifications: Dycore specific

SE:

See the SE namelist options here.

1. Change the timestepping scheme.

2. Examine the impact of using an alternate vertical hyperviscosity profile.

3. Examine diffusion settings, e.g. hyperviscosity on p or horizontal divergence.

4. Use alternative vertical remapping schemes (e.g. se_vert_remap_T)

5. change the shape or peak height of the topography

FV3:

See the FV3 namelist options here.

1. Modify the transport settings.

2. Modify the diffusion settings, such as the order of divergence damping, nord.

3. change the shape or peak height of the topography

MPAS:

See the MPAS namelist options here.

1. MPAS contains many diffusion options, including 3D options. Because MPAS is designed for weather prediction at very high resolutions, it contains diffusion options not present in other cores. 

user_nl_cam changes

After changing CAM_CONFIG_OPTS so that CAM builds with the correct number of levels, edit the file user_nl_cam in your case directory.

• If you are running MPAS, replace L88 the line ncdata = "/glade/derecho/scratch/owhughes/MPAS_GRIDS/L88.nc" with the appropriate number of levels.

• If you aren't running MPAS, then the line to change will read ncdata = "/glade/u/home/owhughes/vertical_grids/dcmip_vcoords_L88.nc"

References:

Skamarock et al. (2019): "Vertical Resolution Requirements in Atmospheric Simulation"