Stokes MOST modifications to unresolved shear #306
Conversation
* Add a control of the Stokes enhancement to the turbulent scales contributing to the unresolved shear (Stokes_Vt = 0 for no enhancement; Stokes_Vt = Stokes_Xi for full enhancement); * Make a depth restriction to the turbulent velocity scale contribution to the unresolved shear, consistent with the contribution to diffusivity and viscosity (Vt_layer = 1 use full boundary layer depth; vt_layer = surf_layer_ext only includes the surface layer). Stokes MOST parameters are based on Vt_layer = 1; * Replace divisions with reciprocal multiplications.
| surfV = surfHv / hTot | ||
| surfUs = surfHus / hTot | ||
| surfVs = surfHvs / hTot | ||
| I_hTot = 1./hTot |
There was a problem hiding this comment.
@gustavo-marques While this is better for performance, it changes answers, soI reverted it in the previous PR. Are we okay with answer changes, or should we revert this again?
There was a problem hiding this comment.
Yes, we are okay with answer changes. Thanks!
| @@ -1085,20 +1086,16 @@ subroutine KPP_compute_BLD(CS, G, GV, US, h, Temp, Salt, u, v, tv, uStar, buoyFl | |||
| !$OMP uS_SLD, vS_SLD, uS_SLC, vS_SLC, uSbar_SLD, vSbar_SLD, & | |||
| !$OMP StokesXI, StokesXI_1d, StokesVt_1d, kbl) & | |||
| !$OMP shared(G, GV, CS, US, uStar, h, dz, buoy_scale, buoyFlux, & | |||
| !$OMP Temp, Salt, waves, tv, GoRho, GoRho_Z_L2, u, v, lamult) | |||
| !$OMP Temp, Salt, waves, tv, GoRho, GoRho_Z_L2, u, v, lamult, Vt_layer) | |||
| do j = G%jsc, G%jec | |||
| do i = G%isc, G%iec ; if (G%mask2dT(i,j) > 0.0) then | |||
|
|
|||
| do k=1,GV%ke | |||
| U_H(k) = 0.5 * (u(I,j,k)+u(I-1,j,k)) | |||
| V_H(k) = 0.5 * (v(i,J,k)+v(i,J-1,k)) | |||
There was a problem hiding this comment.
For multiple tests including SMS.TL319_t232.G1850MARBL_JRA and other, I get the following error:
dec1095.hsn.de.hpc.ucar.edu 2093: forrtl: severe (174): SIGSEGV, segmentation fault occurred
dec1095.hsn.de.hpc.ucar.edu 2093: Image PC Routine Line Source
dec1095.hsn.de.hpc.ucar.edu 2093: libpthread-2.31.s 0000147407AC08C0 Unknown Unknown Unknown
dec1095.hsn.de.hpc.ucar.edu 2093: cesm.exe 0000000001683D0D mom_cvmix_kpp_mp_ 1095 MOM_CVMix_KPP.F90
dec1095.hsn.de.hpc.ucar.edu 2093: cesm.exe 0000000000D8BF79 mom_diabatic_driv 1326 MOM_diabatic_driver.F90
dec1095.hsn.de.hpc.ucar.edu 2093: cesm.exe 0000000000D7F76E mom_diabatic_driv 395 MOM_diabatic_driver.F90
dec1095.hsn.de.hpc.ucar.edu 2093: cesm.exe 0000000000C66171 mom_mp_step_mom_t 1608 MOM.F90
dec1095.hsn.de.hpc.ucar.edu 2093: cesm.exe 0000000000C5A936 mom_mp_step_mom_ 956 MOM.F90
dec1095.hsn.de.hpc.ucar.edu 2093: cesm.exe 0000000000C21689 mom_ocean_model_n 638 mom_ocean_model_nuopc.F90
dec1095.hsn.de.hpc.ucar.edu 2093: cesm.exe 0000000000BDC250 mom_cap_mod_mp_mo 1845 mom_cap.F90
There was a problem hiding this comment.
We need to evaluate this in conjunction with CVMix/CVMix-src#97
I apologize for missing this detail in the original description.
There was a problem hiding this comment.
Then we need to update the CVMix submodule with this PR. When you do a git status, I believe you'll have something like:
modified: pkg/CVMix-src (new commits)
If so, can you do:
git add pkg/CVMix-src
git commit -m "update CVMix submodule"
git push
Note that this will not only modify CVMix but also change the CVMix remote from mom-ocean organization to CVMix organization, because I think the mom-ocean fork doesn't have those changes yet. Before submitting a MOM6 candidate PR to mom-ocean, we'll need to push these CVMix changes to the CVMix mom-ocean fork, I think. Comments, @mnlevy1981?
There was a problem hiding this comment.
I've opened mom-ocean/CVMix-src#1 to get these changes to mom-ocean/CVMix-src; I don't know if we can update dev/ncar before that PR is taken care of, because the git add pkg/CVMix-src will tell git modules to look for commit mom-ocean/CVMix-src@5e3d0ce and I don't know if having the open PR is sufficient for finding that external.
|
Once I reverted to the * hTot operations and added an if statement to guard the STOKES_MOST calculation, the results were b4b with the baseline. Here is the path to the pr_mom tests:
|
|
It looks like this is missing the update to the latest CVMix commit |
Add a control of the Stokes enhancement to the turbulent scales contributing to the unresolved shear
(Stokes_Vt = 0 for no enhancement; Stokes_Vt = Stokes_Xi for full enhancement);
Make a depth restriction to the turbulent velocity scale contribution to the unresolved shear, consistent with the contribution to diffusivity and viscosity (Vt_layer = 1 use full boundary layer depth; vt_layer = surf_layer_ext only includes the surface layer). Stokes MOST parameters are based on Vt_layer = 1;