[MITgcm-devel] vintage bug in seaice_growth.F ?

[Martin Losch]

Hi Gael,

as usual I am not absolutely positive, but this is my take on this:
- (growthHEFF+growthNeg) is the left over (conductive) heat (in units of ice volume/grid-cell-area) after melting all the ice. It does not matter, if this heat is available as an average over the entire grid cell or just underneath the "area" fraction where there is ice. It's just the amount of heat that can potentially melt ice with volume "growthHEFF+growthNeg". So it's not "effective"
- in line 583, QNETI is this same quantity converted to heat flux units (QI/dt), and it is really only the heat flux underneath the ice, so that when you average over the cell you still have to apply the fractional area factor.
- the whole thing is one of the tricky parts of using HEFF (volume) as a diagnostic variable as opposed to thickness.

Bottom line: I think it is correct the way it is (and it's also the way it was introduced originally in growth.F (see here:
<http://mitgcm.org/viewvc/MITgcm/MITgcm/pkg/seaice/growth.F?hideattic=0&revision=1.1.2.1&view=markup>)

Martin

On Oct 10, 2010, at 5:54 PM, Gael Forget wrote:

> Hi Martin,
> 
> as part of the seaice_growth.F merge process, I stumbled 
> on something that looks odd. Either I am missing something, 
> or it is a bug that dates back to the growth.F era...
> 
> It would be great if you could double-check my reading of  
> seaice_growth.F V1.70 (the last one before my modifs):
> ------------------------------------------------------------
> line 445	FICE is multiplied by area -> GHEFF becomes 'effective'
> line 472	FICE is updated, dividing GHEFF by area -> FICE remains 'NOT effective'
> => I think this is correct.
> ------------------------------------------------------------
> line 569	FICE is multiplied by area -> growthNeg becomes 'effective'
> line 583	QNETI is defined, without area division -> QNETI is 'effective' too
> line 778	as part of the QNETO/QNETI weighted average, 
> 	QNETI, which is already 'effective', is wrongly mutliplied by AREA
> 	QNETO, which is 'NOT effective', is correctly mutliplied by (1-AREA)
> => I think the double multiplication by area is wrong.
> ------------------------------------------------------------
> 
> Can you confirm that I should fix this? 
> Or please tell me what I am missing.
> 
> Cheers,
> Gael
> 
> 
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel