We would like to have reversible processes as well as pseudo adiabatic processes. A few references:

1. reversible moist-adiabatic process
2. moist-adiabatic lapse rate
3. pseudoadiabatic lapse rate

Upon looking into our current implementation of moist_lapse it appears to come from (2) above. Given that our docs say it's pseudo adiabatic, I would have expected an implementation of (3). (@dopplershift?) I'm unclear on the difference between the two actually, but it looks like it can make up to 0.7 degC difference for a 25 C surface parcel lifted to 100 hPa.

Implementing (1) requires some mixing ratios that we do not yet have #433