[MITgcm-devel] freezing in the interior

[Chris Hill]

Another reason to put this in its own package.

Chris

On Wed, Apr 20, 2011 at 11:48 AM, Martin Losch <Martin.Losch at awi.de> wrote:
> Hi Dimitris,
>
> is this supposed to work with the adjoint? if so you need to update model_ad_diff.list and probably add some store directive to do_ocean_physics.F. If you do not want it to work with the adjoint then you should exclude the call when allow_autodiff_tamc is defined.
>
> Martin
> On Apr 20, 2011, at 3:59 PM, Dimitris Menemenlis wrote:
>
>> Gael, Steph, JM, I have checked in a first draft of freeze_interior.F
>> subroutine.  It is enabled with flag allowInteriorFreezing in PARM01.
>> Right now all the negative heat comes to the surface level, where it
>> would contribute to ice shelf or sea ice thickness or to cooling the
>> surface level.
>>
>> I like Gael's suggestion of distributing the negative heat throughout
>> the water column as the ice rises.  It's a bit like the inverse of
>> a salt plume.  The question is how much heat such a process would
>> give up on the way up.  Would it use up all the available heat in
>> each level down to the freezing point or would it mostly rise to the
>> surface at the sub-grid scale?
>>
>> Before coding anything more complicated, I want to experiment with
>> this basic first version in our setups, see what happens and
>> correct any obvious bugs.
>>
>> Dimitris
>>
>> On Tue, 2011-04-19 at 15:50 -0700, Gael Forget wrote:
>>> Hi Dimitris,
>>>
>>> I also had thought of doing such a hack, so I am glad you are doing it.
>>>
>>> One thought : given that you will take some negative heat from a typically thick layer at depth,
>>> putting it all at once in the surface layer may not prove very numerically stable; another option
>>> would be to spread that negative heat over the above layers (looping from the bottom layer up?).
>>>
>>> Cheers,
>>> Gael
>>>
>>>
>>>
>>>
>>>
>>> On Apr 19, 2011, at 9:18 AM, Dimitris Menemenlis wrote:
>>>
>>>> Jean-Michel,
>>>> I was thinking of adding to main code as this could also
>>>> be useful for pkg/icefront, and in some circumstances
>>>> for pkg/seaice.
>>>>
>>>> I would follow the
>>>> allowFreezing/freeze_surface.F
>>>> useOldFreezing/freeze.F
>>>> template and add a flag and subroutine in main code say:
>>>> allowFreezingInterior/freeze_interior.F
>>>>
>>>> Is this OK?
>>>>
>>>> On Tue, 2011-04-19 at 05:48 -0700, Jean-Michel Campin wrote:
>>>>> Hi Dimitris,
>>>>>
>>>>> Would this be part of the pkg/shelfice ? I know there is
>>>>> a depth dependent freezing temperature used in this pkg.
>>>>>
>>>>> Cheers,
>>>>> Jean-Michel
>>>>>
>>>>> On Mon, Apr 18, 2011 at 05:32:24PM -0700, Menemenlis, Dimitris (3248) wrote:
>>>>>> MITgcm ice czars, in configurations with ice shelf cavities we often end up with temperatures below the in situ freezing point at mid depth.
>>>>>>
>>>>>> I am thinking of adding a subroutine similar to freeze.F but that will move the excess negative heat from the ocean interior to the surface level (or to the level right below the ice shelf) in order to conserve heat in the model.
>>>>>>
>>>>>> Has anyone else given some thought to, or already written some code to deal with, this mid-depth freezing problem?
>>
>>
>>
>> _______________________________________________
>> 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
>