[MITgcm-devel] OBCS_SPONGE vs RBCS

[Menemenlis, Dimitris (3248)]

OK.  I will put in pkg/seaice, same as we discussed
for an eventual seaice_ptracer capability.

Your suggestion about moving obcs to seaice
sounds reasonable.  it would be similar to having
rbcs in seaice.

seaice rbcs is kind of urgent for Michael's work.
do you mind delaying the obcs move for a few
days, until I complete the seaice rbcs change?

D.

On Nov 17, 2010, at 9:06 AM, Jean-Michel Campin wrote:

> Hi Dimitris,
> 
>> On a separate but related topic, I want to implement RBCS for pkg/seaice.
> Good thing.
>> Should the code live in pkg/seaice or in pkg/rbcs?
> I would be tempted to favor the pkg/seaice location (all what is 
> currently in RBCS is for 3-D fields, so would need separated mask anyway), 
> but ptracers is done within pkg/rbcs, so it's not really consistent 
> with what I propose ...
> 
> On similar subject, I was thinking of moving out of pkg/obcs
> the ptracer and seaice parts (in the 1rst step, would keep the
> namelist unchanged, and could see later when parameter files
> would be changed). There are reasons in favor of this move,
> e.g., the pieces in obcs_init_variables.F are not right
> with PTRACERS_Iter0 > 0 ; + splitting obcs_calc would make it
> more flexible on where/when to call it.
> It's not urgent, but it would be a good thing to agree on this 
> (moving or not) before extending OBCS-STEVENS to seaice & ptracers.
> Do others agree ?
> 
> Cheers,
> Jean-Michel