[MITgcm-devel] changes in OBCs pkg

[Martin Losch]

Hi Jean-Michel,

I cannot see, why useOBCSprescribe should introduce large vertical 
velocities at the open boundaries, other than by prescribing inadequate 
horizontal velocities. In exp4, I moved the calculation of zonal 
velocities U, as in
>               OBSu(I,K,bi,bj)=Uinflow
>      &       *cos(2.*PI*futureTime*recip_TimeScale)
>      &       *max(futureTime*recip_TimeScale,1.0 _d 0)
where recip_TimeScale = 0./2000. (=zero), into gendata.m. The file 
OBzonalU.bin and OBmeridU.bin contain fields of constant UinFlow=0.25 
as in the late obcs_calc.F.
If there are no OB-files the OB values are not touched (have a look in 
obcs_external_fields_load.F), in particular not the w-velocities, which 
are only set if the non-hydrostatic code is used. The strong vertical 
velocities come about because there is this strong gradient (du/dx 
along east-west boundaries and du/dy along north-south boundaries) in 
the beginning. You could get rid of this by initializing the entire 
u-field to Uinflow=0.25.

So in summary, I cannot understand, why useOBCSprescribe should 
introduce strong vertical velocities near the boundaries which haven't 
been there before. I do acknowledge that there are always problems with 
vertical velocities near the boundaires. What do you think?

I don't know anything about the exactConserv=T behavior.

I'll try to reproduce the behavior that you observe, but all of this 
doesn't say anything about your project of changing OBCS, as long as it 
is not motivated soley by the above issue I don't mind.

Martin

PS When i did these changes, the results did not change by very much, 
in fact by so little, that they still passed the test in testreport. 
That suggests to me that this problem was around before.

On Dec 7, 2005, at 2:24 AM, Jean-Michel Campin wrote:

> Hi,
>
> I am going to change the OBCs pkg,
> to always reset eta & wVel (to zero or to prescribed values)
> at the OBs.
>
> I realized that exp4 is now giving very large wVel
> in the OB regions (in 2 points), 10 times bigger than
> the max in the interior. This was not the case when
> using the old exp4/code/obcs_calc.F
> and it happens because of useOBCSprescribe.
> Similar problem with eta when exactConserv=T in
> the present set-up (but was OK before).
>
> Any comments ?
>
> Jean-Michel
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel