Define or disambiguate "nonphysics"

[gold2718]

In standard names, such as tendency_of_water_vapor_mixing_ratio_wrt_moist_air_due_to_nonphysics, the term, nonphysics has no definition. I feel like we should either have a definition (e.g., nonphysics == dycore) or break up the term into a set that does have specific definitions.

[grantfirl]

In the UFS, the term "nonphysics" is used in this context to refer to tendencies that are added to the state outside of the CCPP physics. This includes advection from the dycore, of course, but also potentially other tendencies due to coupling, vertical coordinate remapping, DA, etc. This is mainly used for understanding the relative contribution to the change in the state of physics versus "the rest of the model" whatever "rest of the model" means for one's particular application.

[grantfirl]

Note: Unfortunately, there are even caveats to the term "nonphysics" within the UFS. When using the GFDL microphysics, some of the scheme is more tightly coupled to the dycore, the so-called "fast physics". So, even the "nonphysics" tendencies includes part of the microphysics tendencies. So, it's not altogether useful ATM, even for the purpose stated above.