Translate linear EOS between MPAS-Ocean and Omega

[cbegeman]

Fixes #406

Checklist

• Testing comment in the PR documents testing used to verify the changes

[cbegeman]

Testing:

For polaris/tasks/ocean/single_column/cvmix/forward.yaml:

ocean:
  debug:
    config_disable_vel_explicit_bottom_drag: True
  eos:
    config_eos_type: linear
  eos_linear:
    config_eos_linear_alpha: 0.2
    config_eos_linear_beta: 0.01
    config_eos_linear_Tref: 5.0

Output for quiet=False in ocean_model_step:

WARNING: No Omega mapping found for these MPASO options:
    Eos/Linear/config_eos_linear_Tref

  Tendencies: BottomDragTendencyEnable = False
  Eos: EosType = linear
  Eos: Linear: DRhoDT = -0.2
  Eos: Linear: DRhoDS = 0.01

[xylar]

@cbegeman, let's chat about this one at our meeting today.

My feeling would be that we should aim to make this PR have bit-for-bit changes unless there is a strong reason not to. That would mean that the new default values would be:

# Equation of state -drho/dT, defaults to mpas-ocean default
eos_linear_alpha = 0.2

# Equation of state drho/dS, defaults to mpas-ocean default
eos_linear_beta = 0.8

# Equation of state constant term, designed to match mpas-ocean's default linear EOS
# 1000.0 + 0.2 * 5.0 - 0.8 * 35 = 973.0
eos_linear_rhoref = 973.0

# Equation of state reference temperature, defaults to zero because omega doesn't have this term
eos_linear_Tref = 0.0

# Equation of state reference salinity, defaults to zero because omega doesn't have this term
eos_linear_Sref = 0.0

Then, I would think we would want to update the single_column/cvmix test to stick with eos_linear_Tref = 0.0:

ocean:
  eos:
    config_eos_type: linear
  eos_linear:
    config_eos_linear_alpha: 0.2
    config_eos_linear_beta: 0.01
    # 1000.0 + 0.2 * 5 - 0.01 * 35
    eos_linear_rhoref: 1000.65

It's okay if this PR is meant to be non-bit-for-bit but I think we want to document that more clearly if so.

[cbegeman]

@xylar Those are good suggestions. I committed them and will test for BFB-ness