[MITgcm-support] Banding, Checkerboarding in KPP Viscosities

[Senja Walberg]

Greetings,

I have been conducting simulations of a lake and using KPP for vertical
mixing. I encountered an issue of noise in derived |grad_h T| fields and
traced it to checkerboarding in KPPdiffT and KPPviscA. Based on a 2004
discussion between Dimitris and Martin I found that the horizontal
checkerboarding could be fixed by toggling on the KPP_SMOOTH_DENS flag in
KPP_OPTIONS.h. However, I am still encountering vertical banding. Attached
is a snapshot of this banding from a simplified lake simulation.
[image: KPPviscA_y_Support.jpg]
The lake is stratified between ~25 degrees at the top and ~10 degrees at
the bottom with the pycnocline at 15 m. I am using dz=0.5 m and dx=100m.

I have tried turning on KPP_SMOOTH_VISC, KPP_SMOOTH_DIFF, and
ALLOW_KPP_VERTICALLY_SMOOTH, as well as increasing both horizontal and
vertical viscosities and diffusivities, and have had no luck getting rid of
this banding.

I'm currently experimenting with changing the vertical resolution to see
what effect that has, but would like to not be constrained to a coarse
vertical grid if there is some fix to this.

Does anybody know what's causing this issue or how it can be resolved?

Regards,
Senja
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.mitgcm.org/pipermail/mitgcm-support/attachments/20200402/74d58513/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: KPPviscA_y_Support.jpg
Type: image/jpeg
Size: 859222 bytes
Desc: not available
URL: <http://mailman.mitgcm.org/pipermail/mitgcm-support/attachments/20200402/74d58513/attachment-0001.jpg>