Changes between Version 11 and Version 12 of ticket/1658/General


Ignore:
Timestamp:
2016-06-10T14:11:08+02:00 (4 years ago)
Author:
frrh
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • ticket/1658/General

    v11 v12  
    103103   Well that now runs OK. (Remember the atmosphere is merely receiving the DMS field, it's not using it anywhere... that's for the code owner to figure out.) Checking the DMS coupling field using EXPOUT we see that at TS=1 we have uniform zero and non zero values evolve at subsequent coupling exchanges. However, looking at the incoming field to the atmos we see a column of an apparent steep gradient in the field below India. This is the point at which the orca grid begins and ends - i.e. this looks like a wrap around issue with the MEDUSA DMS field.  
    104104 
    105   Sure enough, the outgoing DMS field from NEMO has zeros in the first and last columns! This is consistent with the loops in trcbio_medusa.F90 which only operate over columns 2 to 361 and rows 2 to 331. Is that deliberate? One presumes it must be otherwise MEDUSA runs would have bad things going on around the grid edges. If it is all ex3epected, then we probably need to add a halo update before coupling DMS (and CO2 flux).  
     105  Sure enough, the outgoing DMS field from NEMO has zeros in the first and last columns! This is consistent with the loops in trcbio_medusa.F90 which only operate over columns 2 to 361 and rows 2 to 331. Is that deliberate? One presumes it must be otherwise MEDUSA runs would have bad things going on around the grid edges. If it is all expected, then we probably need to add a halo update before coupling DMS (and CO2 flux).  
    106106 
     107  What I've done for now is to modify a working copy of the MEDUSA branch to add an explicit lbc_lnk to DMS_out_cpl as soon as it's populated in trcbio_medusa. So [log:branches/UKMO/dev_r5518_GSI7_GSI8_landice_bitcomp_medusa dev_r5518_GSI7_GSI8_landice_bitcomp_medusa] is safe to go into GO6/GC3 now and the only key_medusa refernces are in oce.F90 where they set up ln_medusa which is used at run time and the allocation of coupling fields if key_oasis3 is defined. As a consequence of this approach, no extra keys are needed in sbccpl.F90 as we can control everything by the RTL ln_medusa.  
     108