Address inconsistencies in various standard names

Grab bag for to-be-renamed standard names in one single sweep

## PUMAS-related

1. RK and `set_surface_coupling_vars` uses `lwe_snow_and_cloud_ice_precipitation_rate_at_surface_due_to_microphysics` for large-scale snow; PUMAS uses `lwe_large_scale_snowfall_rate_at_surface`

2. Both `effective_radius_of_stratiform_cloud_ice_particle` and `effective_radius_of_stratiform_cloud_ice_crystal` exist and point to `REI` -- should be addressed in PUMAS r3 then revisited

3. `do_hetfrz_classnuc` is currently `do_heterogeneous_ice_nucleation` but should be `do_heterogeneous_freezing_by_classical_nucleation_theory`

## RRTMGP-related

1. `in_cloud_ice_water_path` is inconsistent with `in_cloud_liquid_water_path_for_radiation`

2. `stratiform_in_cloud_graupel_water_path` might need revisiting?

Start by locating the RK and set_surface_coupling_vars uses, then compare the PUMAS-related names with the proposed replacements. Review the RRTMGP-related names and the note about PUMAS r3 before deciding which inconsistencies are in scope. Done means the agreed standard names are applied consistently, with the two explicitly deferred or uncertain items resolved.

Written by the indexing model from the issue text.