SHOC-HOMME 3D Turbulence
This page describes how to enable the 3D-turbulence coupling between SHOC and HOMME, along with the main user-facing parameters associated with it.
Enabling 3D Turbulence Coupling Between SHOC and HOMME
EAMxx exposes the 3D-turbulence option through the HOMME control-namelist
parameter ctl_nl::do_3d_turbulence. This is the setting that users should
change.
To enable the feature, run
./atmchange ctl_nl::do_3d_turbulence=true
After changing the setting, rebuild the generated namelists before the next
run in the usual way, for example via case.submit or by re-running the case
setup/build workflow used for your case.
What This Switch Does
When ctl_nl::do_3d_turbulence=true:
- HOMME enables its 3D-turbulence path.
eamxx_buildnml.pymirrors that value into the locked EAMxx parameterhomme::do_3d_turbulence_homme.- During atmosphere-driver initialization, that HOMME-facing flag is copied
into SHOC's internal runtime option
do_3d_turbulence_shoc. - SHOC computes horizontal eddy diffusivities for heat and momentum and passes them back to HOMME.
- HOMME computes horizontal shear components and passes them to SHOC, which uses them to form the 3D shear-production term in the TKE equation.
Users should NOT edit homme::do_3d_turbulence_homme directly. That parameter
is intentionally locked and is maintained automatically from
ctl_nl::do_3d_turbulence.
Related SHOC Parameters
The following SHOC parameters remain user-configurable and are relevant when 3D turbulence is enabled:
shoc::coeff_kh_horiz: horizontal eddy-diffusivity coefficient for heat.shoc::coeff_km_horiz: horizontal eddy-diffusivity coefficient for momentum.shoc::coeff_kh: vertical eddy-diffusivity coefficient for heat.shoc::coeff_km: vertical eddy-diffusivity coefficient for momentum.
For example:
./atmquery shoc::coeff_kh_horiz
./atmquery shoc::coeff_km_horiz
./atmchange shoc::coeff_kh_horiz=0.1
./atmchange shoc::coeff_km_horiz=0.1
The default values of coeff_kh_horiz and coeff_km_horiz are both 0.1.
These parameters are tunable, but early testing suggests that 0.1 provides a
good balance for preserving isotropic and anisotropic turbulence behavior as
resolution changes.
A more exhaustive parameter study is still needed. At present, values much
larger than 1.0 appear to reduce the model's effective resolution
substantially and should generally be avoided.