PhaseTransitions incorrectly removes "ice" qualifier from heat capacity and density

[glwagner]

@simone-silvestri I don't think this is the right change. The problem is that "snow density" and "snow heat capabity" are not single numbers.

These numbers refer to the microscopic density and heat capacity of pure ice. They do not refer to the macroscopic "bulk density" and "bulk heat capacity" of a porous media which includes 1) ice and brine (what we call "sea ice") or 2) ice and air (what we call "snow").

Can we please clear this up? This struct is currently quite confusing now because it does not specify what to what phase the liquid is changing into.

Originally posted by @glwagner in a8852e1

[glwagner]

We should have a design that will generalize to more sophisticated ice and snow models in which the porosity is a prognostic variable. In that case, the "bulk" density and heat capacity will depend on the porosity (eg the mixture of liquid+ice for sea ice, and air+ice for snow).

The change that was made to PhaseTransitions is not future proof. It would have to be reverted in the future in order to represent this microscopic/macroscopic difference.

We need to have a different design for representing bulk properties vs microscopic properties.