[MITgcm-devel] thsice/seaice

[Martin Losch]

Hi Jean-Michel,

Dimitris just pointed out that the conversion between the effective  
thickness = normalized volume of seaice (HEFF) to the actual  
thickness of thsice (iceHeight) is problematic as it involves the  
division by the fractional ice area (AREA resp. iceMask):
HEFF  = iceHeight*iceMask
iceHeight = HEFF/AREA

In this way small HEFFs can translate into large iceHeights if AREA  
is very small (which can happen as a result of advection/diffusion).  
Do you think it is possible to use effective thickness within pkg/ 
thsice so that we don't have to make this conversion? The other  
(quick) fix is to impose a minimum fractional area (say, 1e-4) to be  
effectively zero, that is
if (AREA>1e-4); iceHeight=HEFF/AREA

This may actually be partially responsible for the strange results  
that I get when I apply advection to the thsice variables (at least  
when I do something like that, the model runs much longer without  
crashing).

Martin