[MITgcm-support] [EXTERNAL] Banding, Checkerboarding in KPP Viscosities

[Martin Losch]

Hi Senja,

in order not confuse this with the KPP_ESTIMATE_UREF issue (https://github.com/MITgcm/MITgcm/issues/336), let’s move this discussion back to the support list. I am citing you:

> It could be that the issue I'm encountering can't be solved by modifying KPP_ESTIMATE_UREF, of course. KPPdiffT and KPPviscA fields do look noticeably different in the various configurations we've tried (without ESTIMATE_UREF, with it 'fixed' with dbloc imported and work1 not, and with your new improved fix).


Not knowing what’s going on, I would save all available KPP (3D) diagnostics to see which KPP variables are affected. Do the fluxes due to viscosity and the countergradient fluxes compensate the checkerboard pattern? Does reducing the time step do anything?

Is your configuration simple enough so that you can share it?

Martin

> On 6. Apr 2020, at 11:38, Martin Losch <Martin.Losch at awi.de> wrote:
> 
> I think that there are serious issues in the KPP_ESTIMATE_UREF, that should be looked into, before it can be used.
> 
> Martin
> 
>> On 3. Apr 2020, at 16:55, Senja Walberg <senja.w at gmail.com> wrote:
>> 
>> Hello,
>> I wanted to follow up that Matthew's solution of turning on the KPP_ESTIMATE_UREF flag was able to solve my problem, so thanks for that tip.
>> 
>> However, the code did not compile by default with this flag turned on, I've mentioned this on the github for those interested.
>> 
>> Thanks again,
>> Senja
>> _______________________________________________
>> MITgcm-support mailing list
>> MITgcm-support at mitgcm.org
>> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support