[MITgcm-devel] changing offline_exf_seaice

[Jean-Michel Campin]

Hi Gael,

I did some plots (at the time I was working on this set-up) 
of the norm of the oceanic velocity (constant in time here), 
not very noisy, just feeling the cost-line.
I guess this re-enforce my 1rst comment:
> > Would be good to talk about this when you pass by MIT.

Cheers,
Jean-Michel

On Mon, Dec 17, 2012 at 05:04:56PM -0500, Gael Forget wrote:
> Cool. That was it. The results look
> Much simpler when I remove those
> Highly heterogeous ocean velocities.
> I guess those were meant for dyn test.
> Cheers,
> Gael
> 
> Sent from my iPhone
> 
> On Dec 17, 2012, at 15:40, Jean-Michel Campin <jmc at ocean.mit.edu> wrote:
> 
> > Hi Gael,
> > 
> > Would be good to talk about this when you pass by MIT.
> > Did not look at our plots yet (but I did many previously),
> > and likely related to ocean current which influence
> > the exchange coeff between ice and sea-water.
> > 
> > Cheers,
> > Jean-Michel
> > 
> > On Mon, Dec 17, 2012 at 02:06:46PM -0500, Gael Forget wrote:
> >> Hi Jean Michel,
> >> 
> >> the input.thermo experiment is getting there. 
> >> 
> >> I am a little puzzled by the input.thsice experiment results though. 
> >> They turn out much more complicated than expected, given 
> >> our simple experiment design. I dont understand why 
> >> the viscinity of the coast influences the result as it does. 
> >> See e.g. the ocean temperature at day 5 (bottom right 
> >> panel in attached plot). There should be no horizontal 
> >> flux in our simple problem design, right? So where 
> >> is that behavior coming from? Is there an implied 
> >> diffusion of some sort in the thsice thermodynamics?
> >> 
> >> I looked a little at thsice parameters and codes but I dont get it. 
> >> The input.thermo results are much more homogeneous in y.
> >> I must be missing something.
> >> 
> >> Cheers,
> >> Gael
> >> 
> > 
> > 
> >> 
> >> 
> >> On Dec 6, 2012, at 8:41 PM, Jean-Michel Campin wrote:
> >> 
> >>> Hi,
> >>> 
> >>> I would like to change the test experiment "offline_exf_seaice",
> >>> to use a re-entrant channel with idealized costline (based on the
> >>> set-up I put few weeks ago in http://mitgcm.org/~jmc/icedyn_v02.tar.gz)
> >>> If things go well, would like to start the changes tomorrow (Friday)
> >>> with only the forward set-up for now.
> >>> 
> >>> I propose the following forward tests:
> >>> (1) input        : seaice-dynamics + thermodynamics and SST relaxation
> >>>                  (to replace current offline_exf_seaice/input.seaice)
> >>> (2) input.thsice : thermodynamics(pkg/thsice) and SST relaxation
> >>> (3) input.thermo : thermodynamics(pkg/seaice) and SST relaxation
> >>> (4) input.fixedSST : thermodynamics(pkg/seaice) and fixed SST
> >>>                  (to replace current offline_exf_seaice/input.seaicetd)
> >>> (5) input.dyn_lsr  : seaice-dynamics, using LSR
> >>> (6) input.dyn_jfnk : seaice-dynamics, using JFNK
> >>> (7) input.exf_bulk : no seaice, just bulk-formulae
> >>>                  (to replace current offline_exf_seaice/input)
> >>> 
> >>> I think I have 2 & 5 ready to go. I need to put something for (1)
> >>> (otherwise the other will not be tested), and as a temporary test,
> >>> I was going to use thsice and LSR (as in 2 + 5) but this could be changed 
> >>> later on.
> >>> 
> >>> Martin, when you have time, if you can take care of (6), it would be nice.
> >>> One thing we could try is to test both pkg/seaice advection and 
> >>> pkg/thsice advection in 5 & 6 (e.g., thsice in 5 ?).
> >>> 
> >>> Gael, I made several changes to the thermodynamic forcing you sent me,
> >>> so that it fits thsice, it's also a stronger forcing (but once I add 
> >>> the dynamics, it's still dominated by advection vs thermodynamic forcing),
> >>> and I get immediately some places with surface melting and other places
> >>> with melting and freezing from below.
> >>> When you are back, we could talk about (3) and (4), and from there
> >>> it should be easy to get (7) (+ update (1) ?).
> >>> 
> >>> Comments and suggestions are welcome (specially for dir name, since
> >>> it's a pain to changed them with CVS).
> >>> 
> >>> Cheers,
> >>> Jean-Michel
> >>> 
> >>> _______________________________________________
> >>> 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
> > 
> > 
> > _______________________________________________
> > 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