Make precipitation part of the working fluid

[trontrytel]

No description provided.

[szy21]

I only looked at changes to grid-mean now and they look good, thanks! Could you double check it still makes sense when we use TD.relative_humidity and TD.vol_vapor_mixing_ratio in the code? They should be mostly in radiation callback and diagnostics.

[szy21]

Do we have an has_condensate for non-equilibrium thermo state?

[trontrytel]

We do. However if would return true for a case where hypothetically q_liq=0 but q_rai!=0. I wanted to avoid that problem, so I do it by hand with the specific precomputed values

[szy21]

Just note that these could be negative.

[trontrytel]

That should be fine. I'll double check just in case but all microphysics functions return zero if one of the tracers is negative. And all the limiters inside ClimaAtmos precipitation are written anticipating tracers can be negative too. No trust!

[trontrytel]

I only looked at changes to grid-mean now and they look good, thanks! Could you double check it still makes sense when we use TD.relative_humidity and TD.vol_vapor_mixing_ratio in the code? They should be mostly in radiation callback and diagnostics.

Water vapor related things should be fine since q_vap = q_tot - q_liq - q_ice from PhasePartition. So whether q_liq/ice contain just the cloud condensate stuff in the equil case or also contain precipitation in the nonequil case the resulting q_vap is fine.

I'm not so sure about some other thermodynamics functions. Like, do we want out potential temperature to now also contain precipitation contribution? This kind of ties back to the buoyancy gradients too.