[MITgcm-devel] exf runoff and seaice
Jean-Michel Campin
jmc at ocean.mit.edu
Thu May 17 15:01:12 EDT 2007
Dimitris,
> Reason for that is that if SEAICE_EXTERNAL_FORCING is defined, pkg/seaice
> uses EXF_FIELDS.h common blocks rather than the SEAICE_FFIELDS common block
> (see SEAICE_FFIELDS.h). This means that runoff array must exist, even if
> it just filled with zeros, which is why ALLOW_SEAICE is "not" needed
> everywhere.
I understand this SEAICE_EXTERNAL_FORCING issue. But my point was
that the ALLOW_SEAICE is not always there in pkg/exf where it really
needs to be:
in exf_init_fixed.F (few days ago was in exf_readparams.F), I see:
> #ifdef ALLOW_RUNOFF
> if ( runofffile .NE. ' ' .AND. runoffperiod .gt. 0. ) then
> call cal_FullDate ( runoffstartdate1 , runoffstartdate2,
> & date_array ,mythid )
> call cal_TimePassed(modelstartdate,date_array,difftime,mythid)
> call cal_ToSeconds ( difftime, runoffstartdate ,mythid )
> runoffstartdate = modelstart + runoffstartdate
> endif
> #endif
it means that if I don't set ALLOW_RUNOFF, I would be able to read
the runoff file (because the loading part is with define ALLOW_SEAICE),
will be able to store it (again, because define ALLOW_SEAICE in EXF_FIELDS.h),
it will not be correct, since the date_array will not be set,
the exf_monitor will print something, but I have no idea of what forcing
I will get at the end (will not be part of sflux in exf_getforcing.F,
might come back in seaice_growth.F ?, and if I use pkg/thsice ? )
> It is needed, so let's add to seaice_check.F. Shall I or will you take
> care of this. Also what is decision re SEAICE_EXTERNAL_FORCING flag?
> Shall we force pkg/seaice to be useable only in conjunction with pkg/exf?
As I wrote earlier, I would like to have a test experiment with
seaice+thsice checked in, before removing this SEAICE_EXTERNAL_FORCING
option. And I am not yet done, this was the suject of my last remark.
I will remove the define ALLOW_SEAICE in pkg/exf, and I guess will have
to fix what this breaks in pkg/seaice.
Jean-Michel
More information about the MITgcm-devel
mailing list