[MITgcm-support] OBCS_PRESCRIBE_READ

[Jean-Michel Campin]

Hi Sonya,

Regarding time specification of prescribed OBC, it follows
the same "logic" than surface forcing:
1) without EXF (useEXF=F):
  - constant in time (only 1 time record needed in OBC files)
    when main parameter periodicExternalForcing=False
  - linear time interpolation between 2 consecutive time records
    (separated by a time interval = externForcingPeriod); 
    and the time serie repeats itself over a time = externForcingCycle
2) with EXF (useEXF=T):
   OBC file names are still read from "data.obcs" but time
   specifications are read-in from "data.exf", namelist "EXF_NML_OBCS"
   (can take a look at verification/seaice_obcs/input/data.exf )
   and follow pkg/exf conventions.
   In particular, setting the corresponding period (e.g., 
    obcsNperiod=0., for Northern OBC files) to zero should allow 
   to impose constant-in-time OBCS, that would be read-in only once.
   This feature (i.e., obcs period = 0 with EXF) is currently 
   tested in verification/obcs_ctrl/input_ad/ so there is a good
   chance that it's not broken.
    
Cheers,
Jean-Michel

On Tue, Oct 07, 2014 at 04:00:11PM -0400, Sonya Legg wrote:
> Dear MITgcm-ers,
> 
> I want to run MITgcm with prescribed open boundary conditions, with
> values read in from a file. I see that useOBCSprescribe=.TRUE. is
> the way to do this, setting the arrays for the input variables in
> the files OB[N/S/E/W][t/s/u/v] etc. I notice that as currently set
> up, the input files have dimensions of N(horizontal)XNrX
> time-levels. I want to have constant (in time) boundary forcing
> (i.e. timelevels = 1). I'm having trouble figuring out how I give
> set the time information for the forcing - something in
> exf_readparms.F and namelist exf_nml_obcs?
> 
> Thanks for any pointers
> 
> Sonya
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support