[MITgcm-support] adding uniform background flow to all water grid points smoothly.

[Martin Losch]

Hi,

do you want the flow to interact with topography? If so, then the restoring option with RBCS near the boundaries (like a sponge layer) away from the topography (that’s important) appear to be the best solution. I would start from initial conditions at rest (U0=0) and impose this flow near the (open?) boundaries with strong restoring ( tau ~ O(deltaT) or just a bit larger ) and then spin up the model. For slower background flow I would even risk a non-zero U0, but with 10cm/s you might run into cfl issues in the spinup phase.

What do you mean by “dramatic distinct patterns”? The flow is going to interact with topography. I would not use restoring near or at the topography.

Martin

> On 12. Nov 2020, at 18:32, Li Jianing <li.jia.ning at outlook.com> wrote:
> 
> 
> Hi again, the former e-mail is broken and I have
> no idea. So I'm re-sending it from  my phone, hope this works. Here is the message:
> 
> Hi there,
> 
>  I'm trying to add a uniform background flow field (U=0.1m/s) to my 2D (x-z) simulation, which is laterally homogeneous and vertically uniformly stratified. To avoid obvious noise induced by flow-topography interaction, I hope the background flow field firstly grow slowly and then remain stable.
>  I tried to use U_init+OBCS or RBCS to do that, but the results seem not so reasonable. I think the better way is to apply an additional forcing to the rhs of momentum equation (x: -fV; y: fU). So I altered the 'myPa_TendVelV' to fU in '/pkg/mypackage/mypackage_init_varia.F' and enabled this package and run for 12 hours. The result shows a gorwing U field (0 to 0.03m/s or so generally, bigger or smaller values at topography) but didn't reach 0.1 m/s at the end (12hr modeltime).
> So,
>   1. Is this a right way to adding uniform background field, any way to do this easier?
>   2. By RBCS relaxing the flow field also makes U grow gradually from 0 to ~0.1. These show dramatic distinct patterns, which is more reasonable ?
> 
> Thanks.
> Jianing Li
> 
> &&&this is how I modified 'mypackage'&&&
> # mypackage_init_varia.F:
> line 50: myPa_TendVelV (i,j,k,bi,bj) = 4.E-6
> 
> # MYPACKAGE_OPTIONS.h
> line 14: #undef MYPACKAGE_2D_STATE
> 
> #data.mypackage(whole file):
> &MYPACKAGE_PARM01
> myPa_MNC=.FALSE.,
> myPa_applyTendU=.TRUE.,
> myPa_applyTendV=.TRUE.,
>  &
> 
> 
> 发送自 Outlook Mobile
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support