[Mitgcm-support] Re: sea-ice code

[mitgcm-support at dev.mitgcm.org]

Dimitris Menemenlis wrote:
 >>>What do you think about moving toward passing state and grid information
 >>>around via argument lists internally for these, increasingly large,
 >>>ancillary packages?
 >>
 >
 > Patrick, what would be impact on adjoint generation?

Could be a bit (or quite) tricky.
E.g. the model state and some additiona active variables
need to be visible at a level as high as the_main_loop
(e.g. for the checkpointing).
That means, as soon as you add a new field to the state
(as has recently happened with the pressure field)
you may have to modify all the necessary argument lists.
Not very nice...

> My only concern, in the short term, is that this rearrangement might slow down
> progress.  I need to have something up and running and debugged ASAP so that
> Ralf can take a look at it in the January, February time frame (before the
> upcoming birth of their child).
 >
> P.S. Has anyone written a flux coupler for the MIT GCM.  For 1/4-deg
> integration we need to interpolate forcing fields on the fly.

That's supposed to happen within the new ESMF, or is it?

-p.


-- 
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
Patrick Heimbach     Massachusetts Institute of Technology
FON: +1/617/253-5259                    EAPS, Room 54-1518
FAX: +1/617/253-4464               77 Massachusetts Avenue
mailto:heimbach at mit.edu                 Cambridge MA 02139
http://www.mit.edu/~heimbach/                       U.S.A.