[MITgcm-devel] depressed Eta and sIceLoad

[Martin Losch]

Thanks! Now I do not have any excuses any more to not learn the  
diagnostics package (o: I am still using taveFreq for everything ...

Another thing about the seaice model:
While looking for possible causes for this depression in eta, I found  
the following part in growth.F, where the freshwater flux is computed:
           EmPmR(I,J,bi,bj)= maskC(I,J,kSurface,bi,bj)*(
      &         EVAP(I,J,bi,bj)-RUNOFF(I,J,bi,bj)
CML     &         EVAP(I,J,bi,bj)*(ONE-AREA(I,J,2,bi,bj))
CML     &         -RUNOFF(I,J,bi,bj)
      &         +SEAICE_SALT(I,J,bi,bj)*0.92 _d 0/SEAICE_deltaTtherm
      &         )
I added the commented lines to test whether it makes a difference to  
suppress evaporation underneath sea-ice, and it does seem to make a  
big difference in my Weddell Sea (part of my global 2x2 model  
config): The WDW of the Weddell Sea is eroded much more slowly (which  
is good) by (erroneous) deep convection. I get the impression that  
with the original code one extracts too much water (adds too much  
salt) to the surface of the ocean underneath ice (evaporation + sea  
ice formation), so that I get too saline surface waters and  
consequently too much convection!

I think that's a bug!
Martin


On 23 Oct 2006, at 17:05, Patrick Heimbach wrote:

>
> Hi Martin,
>
> I implemented some initial diagnostics
> (i.e. using pkg/diagnostics)
> for the seaice package over the weekend.
> It does include HSNOW.
>
> So you could have a look at it.
> As template for the use of diagnostics
> check out labsea/input/
> which has two seaice variables as part
> of the diagnostics list
> (you need to add 'SIhsnow ').
>
> You need code updates in pkg/seaice, pkg/diagnostics
> and model/src.
>
> Hope it works.
>
> -p.
>
>
>
> On Mon, 2006-10-23 at 10:50, Jean-Michel Campin wrote:
>> Hi Martin,
>>
>> My first guess would be the snow-thickness (unfortunately, there was
>> not many diagnostic for this field in pkg/seaice, may be Patrick
>> add one), since in the run I did, there was quiet a lot of snow.
>> Cheers,
>> Jean-Michel
>>
>>
>> On Mon, Oct 23, 2006 at 10:26:09AM +0200, Martin Losch wrote:
>>> Hi Jean-Michel,
>>>
>>> unfortunately I cannot say when, but at sometime the sIceLoad fields
>>> started to cause problems in the height field Eta. With useSeaice
>>> = .true., useRealFreshWaterFlux=.true., exactConserv=.true., and
>>> LINEAR free surface (#undef NONLIN_FRSURF) I get local  
>>> depressions of
>>> ETA underneath seaice  along the coast of -40m and more in 100years.
>>> They go away, when I set useRealFreshWaterFlux = .false. I need to
>>> mention, that I have a global EmPmR imbalance of approximately 10cm
>>> per year, so that my mean Eta is approximately -11m after 100years,
>>> but that wasn't a problem earlier. I attach a figure which shows Eta
>>> and surface velocities averaged over the years 91 to 100 of a 100
>>> year integration. What I find strange is that the strong depression
>>> in Eta does not have any corresponding signal in the surface
>>> velocities. T/S do not have any striking anomalies in these regions
>>> around Antartica, either, that could compensate Eta-gradients of  
>>> more
>>> 10m per 100km. The model does not explode, the mean effective ice
>>> thickness is way below 2m in the areas of Eta-depression.
>>>
>>> Do you have any idea what's going on?
>>>
>>> Martin
>>
>>
>>> _______________________________________________
>>> 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
> -- 
> Dr Patrick Heimbach | heimbach at mit.edu | http://www.mit.edu/~heimbach
> MIT | EAPS, 54-1518 | 77 Massachusetts Ave | Cambridge, MA 02139, USA
> FON: +1-617-253-5259 | FAX: +1-617-253-4464 | SKYPE: patrick.heimbach
>
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel