Physics cleanup for MPAS

[dustinswales]

Description

The GFS based physics parameterizations contained within the CCPP Physics repository have some host-specific, not necessarily dycore-specifc, code to handle the evolution of the inter-physics state. There are two groups of parameterizations to consider, process-split and time-split. Process-split schemes are all presented with the same input state (i.e. order does not matter), whereas Time-split schemes are provided a modified, or updated, state from the previous scheme (i.e. order matters). MPAS has a different physics-2-dynamics coupling than in the FV3 based UFS, which requires that all schemes be written as process-split.

Solution

About half of the existing GFS based parameterizations are written as process-split. For these existing process-split schemes, the accumulation of the instantaneous tendencies is done within the scheme. This will need to be refactored so that the scheme outputs the instantaneous tendency and the host model applies the change in a post interstitial scheme, as either time-split or process-split.
The time-split schemes will also need to be modified to not change the input state and compute/output the instantaneous tendencies instead. These tendencies will be applied as either time-split or process-split within the post interstitial scheme(s).