What contributes to DTV, the vertical mixing temperature tendency?

[khartig]

I am analyzing the temperature tendency contributions over continental North America and the ice-covered Arctic during winter in CAM5 (CESM1-CAM5, F_1850_CAM5 compset with prescribed SSTs and sea ice, constant pre-industrial CO2), often within the first couple of model layers near the surface. I am trying to understand the vertical mixing parametrizations so that I can identify which variables are driving periods of strong DTV (in particular, strong cooling/negative temperature tendencies). From digging through the technical documentation and source code, there seem to be the following contributions:

• SHFLX, sensible heat flux, is added to DTV at the lowest model level
• UW_kvh, the eddy diffusivity of heat
• the counter-gradient term for heat, which is cgh in the source code and seems to be the non-local turbulent transport but doesn't appear to have a corresponding history output available
• the vertical first and second derivative of dry static energy, which can be computed from T and height

Are there any terms or processes that I am missing? Sometimes strong DTV corresponds with large magnitudes in one or more of these terms, but in other cases I seem to see strong DTV even when all terms are relatively small (or at least when these terms are comparable to their values during other periods of weak DTV). Any pointers or references to additional documentation would be appreciated.