[MITgcm-support] Step change in OBC

[Martin Losch]

Hi Olivier,

I don’t have a solution, but do you start from rest in both cases?

Also, the “step change” in the model may be different from what you expect. If you have only two time records in your input files that are supposed to be 5days apart (at t=0 and t=5days), then the model linearly interpolates between these:

flx(t) = flx(0)*(1-t/5days) + flt(5days)*t/5days.

After 5 days, the model will ask for new data at t=10days, etc. So it’s more of a ramped up flux rather than step change.

If you what a real (instantaneous) step change I’d run the model with constant forcing flx(0) for 5 days (obcsSperiod=0) and then restart from a pickup with a different forcing file containing only flx(5days) and again obcsSperiod=0. I suspect that this will also cause some numerical challenges.

Martin

> On 10. Feb 2021, at 23:25, Olivier Marchal <omarchal at whoi.edu> wrote:
> 
> Hello MITgcm-ers,
> 
> I am trying to set up a step change in the OBC (at the southern boundary) for a flow problem with no surface forcing. A sponge layer is prescribed at the southern boundary in order to absorb an outflow at that boundary. An inflow is prescribed along the same boundary but at another location than the outflow. The desired step change is the magnitude of the inflow (from 0 to 50 Sv).
> 
> Here is the current set-up:
> 
> 1) code/packages.conf includes gfd, cal, exf, and obcs
> 
> 2) /input/data.pkg:
> 
> &PACKAGES
> useCAL=.TRUE.,
> useEXF=.TRUE.,
> useOBCS=.TRUE.,
> &
> 3) input/data.cal:
> 
> &CAL_NML
> The Calendar='model',
> startDate_1=0000101,
> startDate_2=000000,
> &
> 4) input/data.exf:
> 
> &EXF_NML_01
> &
> &EXF_NML_02
> &
> &EXF_NML_03
> &
> &EXF_NML_OBCS
> obcsSstartdate1=0000101,
> obcsSstartdate2=000000,
> obcsSperiod=432000., (i.e., step change desired after 5 days of integration)
> &
> 5) input/data.obcs:
> 
> &OBCS_PARM01
> ...
> useOBCSprescribe=.TRUE.,
> useOBCSsponge=.TRUE.,
> OBCSsponge_S=.TRUE.,
> ...
> OBCSuFile = 'uSouth.bin',
> OBCSvFile = 'vSouth.bin',
> OBCStFile = 'tSouth.bin',
> OBCSsFile = 'sSouth.bin'
> &
> 
> &OBCS_PARM03
> spongeThickness=10,
> Urelaxobcsinner=43200.,
> Urelaxobcsbound=8640.,
> Vrelaxobcsinner=43200.,
> Vrelaxobcsbound=8640.,
> &
> The .bin files in input/data.obcs contain fields for two time slices produced from input/gendata.m.
> 
> Here is the issue: the run blows up after 1 iteration ("SOLUTION IS HEADING OUT OF BOUNDS"). Use of a ridiculously small time step (deltaT well below CFL criterion) yields a similar result. If no step change is prescribed (inflow fixed at 50 Sv from the start of the run), then the run goes fine.
> 
> Could anyone help me understand this result? 
> 
> Thanks for any pointers,
> 
> Olivier.
> 
> 
> 
> 
> 
> 
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 1665 bytes
Desc: not available
URL: <http://mailman.mitgcm.org/pipermail/mitgcm-support/attachments/20210211/b0cca9aa/attachment.p7s>