kice in mushy thermodynamics

[dabail10]

I think this is kind of a minor bug. I found that kice (ki) is hardcoded in ice_mushy_physics.F90 and I thought it should be using kice from the icepack_parameters.F90 file. Thoughts?

[eclare108213]

There is a comment in icepack_parameters.F90 acknowledging that mushy doesn't use kice:

                           ! kice is not used for mushy thermo
    kice      = 2.03_dbl_kind    ,&! thermal conductivity of fresh 

BL99 thermo uses kice only for the Maykut and Untersteiner form of conductivity (based on the salinity and temperature profiles), not for the 'bubbly' option, so the comments/documentation for kice need to say that it is specific to the MU71 conductivity.

icepack_mushy_physics.F90 also uses ki as the basis for computing a variable conductivity profile (km):

ki = 2.3_dbl_kind , & ! fresh ice conductivity (W m-1 K-1)

I think it's reasonable for mushy to have its own form (and its own parameter). I would prefer to not have parameters like this buried in the code, though. Also, kice is used in the meltpond schemes, not ki. That might be inconsistent when using mushy thermo? It might come down to an assumption of how saline the ice/snow/meltponds are. This could be a research project.