[MITgcm-devel] KPP, once again

[Martin Losch]

Good suggestion! The nice thing about the GOTM KPP code is that it is  
almost the same (even the similar variable names) as the MITgcm-KPP  
code, except for a few more comments. This may actually help a lot.  
We'll see how far I get.

Martin

On Sep 19, 2006, at 4:59 PM, chris hill wrote:

> Martin,
>
>  Not sure if this is relevant.
>  There is a project called GOTM that provides fairly clean  
> implementations of a number of turbulence schemes that can be  
> linked as a library I believe. No built-in TAF compatibility  
> though. However, if you are going to be hacking on the KPP routines  
> in a serious way and worrying about turbulent mixing under shelf  
> ice it might be worth looking at/googling GOTM.
>
> Chris
>
> Martin Losch wrote:
>> I guess, what I have to do is, to tell the subroutines bldepth and  
>> blmix in kpp_routines.F that the surface is no longer at k=1 but  
>> at k=kTopC. But before I can do that I'll have to spend some time  
>> with these routines, and with the gentlement Large, McWilliams and  
>> Doney, I suppose (o;
>> Martin
>> On Sep 19, 2006, at 4:22 PM, Martin Losch wrote:
>>> Hi Dimitris,
>>>
>>> I agree: setting the "surface"-diffusion coefficients to zero is  
>>> the minimum I'll have to do (and that does not pose any problem,  
>>> does it?). Modifying the surface boundary layer is a little more  
>>> tricky. So far, I have managed to make KPP compute a hbl that  
>>> looks reasonable (although not correct). For now I should  
>>> problably turn it off in the presence of shelfice, but I am just  
>>> starting to look at that.
>>>
>>> Thanks for you input.
>>>
>>> Martin
>>>
>>> On Sep 19, 2006, at 4:06 PM, Dimitris Menemenlis wrote:
>>>
>>>> Hi Martin, without having tried it, it seems that replacing  
>>>> nzmax with klowc should work but that this would not be  
>>>> sufficient for KPP to work with shelfice.  Somehow 1) the  
>>>> surface boundary layer scheme would need to be modified or  
>>>> turned off and 2) the mixing coefficients evaluated between  
>>>> ktopc and ktopc-1 would need to be set to zero to avoid  
>>>> diffusive flux into the ice shelf.  Dimitris
>>>> _______________________________________________
>>>> 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