[MITgcm-devel] another bug in growth.F ?

[Martin Losch]

I forgot to say, that with the new/current versions of growth etc.  
the effective ice thicknesses have increased by a factor of 2 (in the  
Southern Ocean), to be as large a 3m in a 10-year average. While I  
don't know if that good or bad, it's nevertheless dramatic. (See  
attached figure run34 is with "old" growth etc, run38 with current  
growth, etc. Note that the extend of significant ice and snow (>1cm)  
in run34 is approximately similar to that of run38, but that very  
thin layers of ice and snow O(1e-5) reach very far north, remember  
its a 10year average over all seaons).

Martin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: snow.png
Type: image/png
Size: 77685 bytes
Desc: not available
URL: <http://mitgcm.org/pipermail/mitgcm-devel/attachments/20061128/b574f7f9/attachment.png>
-------------- next part --------------

On 27 Nov 2006, at 16:57, Martin Losch wrote:

> Hi Dimitris,
>
> sorry to pick this up again, but you changes do something strange  
> to the testing of lab_sea: The results become hight machine and  
> compiler dependent. Please have a look at:
> http://mitgcm.org/testing.html
> and in particular at:
> http://mitgcm.org/testing/results/2006_11/tr_rays1_20061126_0/ 
> summary.txt
> (SunOS WS compile f77 vs. g77, also on SunOS)
> http://mitgcm.org/testing/results/2006_11/tr_rays1_20061126_1/ 
> summary.txt
>
> also the pgi compiler on xd1 and the pgi and ifort compiles on  
> faulks appear to have problems (I haven't looked any further). Only  
> the g77 seems to be doing fine. Any idea, what could be the reason?  
> Maybe this is something that could lead you to the adjoint problems  
> you are having with the new growth routines?
>
> Martin
>
> On 16 Nov 2006, at 02:00, Dimitris Menemenlis wrote:
>
>> Jinlun, to make sure we will be talking about same code, please  
>> look at:
>>
>> http://mitgcm.org/cgi-bin/viewcvs.cgi/MITgcm/pkg/seaice/
>>
>> http://mitgcm.org/cgi-bin/viewcvs.cgi/MITgcm/pkg/seaice/growth.F
>>
>> http://mitgcm.org/cgi-bin/viewcvs.cgi/*checkout*/MITgcm/pkg/seaice/ 
>> growth.F?rev=1.32
>> _______________________________________________
>> 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