Frontogenesis function should be computed on pressure levels

### What happened?

Walter Hannah (DOE) reported a bug in the computation of the frontogenesis function (computed in dp_coupling; dycore dependent) that was brough to our attention by @chihchen24. Code modification for cam_development provided by Walter Hannah and Jack Chen.

This needs to be fixed for all dynamical cores computing the frontogensis function. @kuanchihwang @JulioTBacmeister @mgduda @mbramberger @mijeong135

@PeterHjortLauritzen will do PR for SE (since we need it ASAP for CESM3)

Also: we need to make same changes in CAM-SIMA for SE and MPAS.

### What are the steps to reproduce the bug?

Run any version of CAM7 or WACCM!

### What CAM tag were you using?

any tag

### What machine were you running CAM on?

CISL machine (e.g. cheyenne)

### What compiler were you using?

Intel

### Path to a case directory, if applicable

_No response_

### Will you be addressing this bug yourself?

Yes, but I will need some help

### Extra info

_No response_