[MITgcm-devel] idemix

[Jean-Michel Campin]

Hi Martin,

Just my personal view:
1) since it's coded, would be nice to have these pieces of code somewhere.
2) will not be used with KPP since KPP is already doing something
 for internal-wave induced mixing (correct me if I am wrong) and, 
 up to now, the 2 most used vertical mixing pkg are kpp & ggl90,
 so, it will not be a serious problem to add this idmix in ggl90 pkg.
 We could agree that it should be off by default.
3) But in the future, there might be an issue if one wants to use it
 with an other (new ?) vertical mixing scheme.
 A new "idmix" pkg would also involve more memory (I guess) and
 more coding effort (since you have it in pkg/ggl90).

What do others think ?

Cheers,
Jean-Michel

On Wed, Sep 04, 2013 at 11:32:14AM +0200, Martin Losch wrote:
> Hi there,
> 
> There's a new model/parameterization for mixing by internal tides based on a Garret-Munk spectrum <http://journals.ametsoc.org/doi/abs/10.1175/JPO-D-12-0207.1>
> 
> Carsten Eden has implemented this IDEMIX model by Olbers et al as part of the GGL90 package (one extra file and some modifications to about 8 files of GGL90), and he gave me the code. I think technically, it's possible to massage the code so that everything is OK (i.e. does not interfere with other parts of GGL90), but the question is, do we want to include this in the main repository, and if so how? just as part of GGL90 may be a pity, because it could also be useful elsewhere.
> 
> Or just an example/verification experiment in MITgcm_contrib with the modified files in "../code" at the risk that it is outdated very quickly?
> What do you think?
> 
> Martin
> 
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel