[MITgcm-devel] tices in global_ocean.cs32x15

[Martin Losch]

Hi Jean-Michel,

as usual, you were I right, I had been chasing ghosts …

Martin

On Dec 10, 2014, at 8:41 PM, Jean-Michel Campin <jmc at ocean.mit.edu> wrote:

> Hi Martin,
> 
> When I run this test, I am getting the expected warning:
> ** WARNING ** SEAICE_CHECK_PICKUP: restart from single category Tice (copied to TICES)
> ** WARNING ** SEAICE_CHECK_PICKUP: Will get only an approximated Restart
> which make me think that TICES is OK (I am assuming that the only seaice-surface 
> temperature read from pickup has been repeated 7 times in array TICES, with the logical
> flag "doMapTices" beeing switched to True in S/R seaice_read_pickup.F).
> 
> So what you described below does not correspond to this.
> 
> Are you sure you are running with the right "pickup_seaice.0000072000.meta"
> in your run dir ?
> 
> Cheers,
> Jean-Michel
> 
> On Wed, Dec 10, 2014 at 05:40:15PM +0100, Martin Losch wrote:
>> Hi Jean-Michel,
>> 
>> while looking for the problem that I introduced I stumble over this:
>> 
>> global_ocean.cs32x15 uses SEAICEmultDim = 7 and pickupStrictlyMatch = .FALSE., so the model assumes that the pickup files are OK and just reads them. nbFields = -1, and doMapTices = .FALSE. in seaice_read_pickup.F, but the pickup_seaice does not contain 7 seaice TICES fields as far as I can see. So that TICES(i,j,2,bi,bj) already only contains zeros and you get a division by zero in seaice_solve4temp (when computing mm_log10pi).
>> 
>> I guess it does not matter, because the 1/0 that are produced are overwritten very soon, because it’s an iteration, but I would how this can go undetected and not influence the results. Should we do anything about this?
>> 
>> Martin
>> 
>> 
>> _______________________________________________
>> MITgcm-devel mailing list
>> MITgcm-devel at mitgcm.org
>> http://mitgcm.org/mailman/listinfo/mitgcm-devel
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel