Make ttgw in gravity wave drag parameterizations consistent

During the review of https://github.com/ESCOMP/atmospheric_physics/pull/292

Some gravity wave parameterizations use `cpairv` to convert to temperature tendency, some use `cpair`, and some do not do this conversion, when outputting `ttgw`.
```fortran
! Convert T tendency to K s-1
! FIXME: some places use cpairv (e.g., orographic) but cpair is used here. hplin 8/14/25
ttrdg = ttrdg / cpair
```

Work has to be done to make these consistent.

When changing from `cpair` to `cpairv`, there will be answer changes for the temperature tendency diagnostics for both CAM and CAM-SIMA.