revisions to snowfall melting and accumulation from @tanyasmirnova

[SamuelTrahanNOAA]

This is an urgent minor change to RUC LSM from @tanyasmirnova.

This includes:

1. A change in the diagnostic of the snow depth accumulation. The old diagnostic included snow melting, but in the new version it includes only the snow fall with variable density. This was decided after consultations with people from the Weather Service.
2. The old diagnostics that included snow melting in the computation of snowfallacc had a bug in the computation of snow melt. Snow melt and snow sublimation should not be multiplied by snowfrac. It is also only the diagnostic variables.
3. The change to uncomment snow_mosaic=0 for fresh snow changes the model performance. This will help RRFS with the deficient snow depth on the ground.

Issues closed:

• fixes RUC LSM snow depth accumulation diagnostic should only include snow fall with variable density #143
• fixes RUC LSM snow melting and sublimation should not be multiplied by snow fraction #144
• fixes turn off separate treatment of snow covered and snow-free portions of the grid cell in RUC LSM when fresh snow falls on the snow-free ground #145

[dustinswales]

Looks good.