[MITgcm-support] Grid-scale noise in a nested simulation with EXF

[Martin Losch]

Hi Xiaozhou,

as you found out, open boundaries are always a drag. I don’t have any good suggestions other than playing with viscosity parameters (you’ll probably see these stipes in the velocity fields very well). In the animations that you sent to me, I can also see grid scale patterns away from the open boundaries so I suspect that your viscosity may be too low.

I have had quite decent experience with use the stevens boundary conditions, because they reduce strong vertical motion near the boundary which is often the cause of problems.

If you suspect the exf package and in particular the exf-interpolation, then you could try for a test to do the interpolation offline, at least you should inspect the interpolated fields that you provide (i.e. air temp, humidity, wind fields, etc, the sensible heat flux will only reflect your ocean surface temperature fields).

For more detailed suggestions you might want to send your namelist files and $PKG_OPTIONS.h files to the list

Martin

> On 3. Jul 2019, at 05:25, Ruan Xiaozhou <saberruan at hotmail.com> wrote:
> 
> Dear MITgcm users,
> 
> I’m running a hydrostatic regional simulation with all obcs, rbcs and exf (JRA55) packages enabled. As I was looking through the output of the temperature field, there was grid-scale noise appearing near the surface (top 3 levels) right after the simulation began. The noise seems to be related to very strong convective events at the surface (it’s gone when I turn off the EXF package). The top layer becomes unstable with super strong cooling which could not be improved with the convective adjustment (e.g. ivdc_kappa). When I diagnosed the EXF variables, similar noise pattern can be found in the sensible heat flux as well. I was wondering if anyone has seen this before? I have checked that all the input files have the correct precision. Any suggestions would be appreciated. Many thanks for your time!
> 
> Best,
> Xiaozhou
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support