[MITgcm-devel] conservation // sublimation

[Gael Forget]

Martin,

> If you think of sublimation as melting that followed immediately by evaporation

did you try thinking of it as a direct conversion of ice to water vapor?  ;-)

More seriously, I see no double accounting of the latent heat flux on the ocn/ice side.
In previous emails, there is some confusion regarding signs though. 
May be it would help if we clarified this point first.

So let us say that sublimation occurs -- trimming the ice pack and generating water vapor. This phase change 
requires heat to be taken out of the ocn/ice system, which is the upward latent heat flux that solve4temp computes.
If anything, this ocn/ice heat sink will tend to grow ice.
Not to melt ice, right?

In hypothetical coupled ocn/ice/atm runs, the latent heat loss on the ocn/ice side would have 
to balance out the storage of latent heat on the atm side (namely the water vapor gain).
So, we dont need to add another heat source on the ocn/ice side (which is what you seem 
to argue for, AFAIU), but to account for the upward ice->vapor water flux.	No?

Cheers,
Gael