Assess the need for incl. topo vars in restart files

Low priority, but I wanted to get it on the docket.

In the FV dycore, PHIS is both written and read in the restart files, but it is my understanding from Brian E. that for a restart run, this is over written by the value in the bnd_topo file, which is a namelist var. PHIS is not in SE restarts, but (like FV) SGH and SGH30 variables are in the restarts. It is not clear whether these SGH vars are being read in or used from the restarts. All the ridge topo variables for Julio's OGW are not in any of the restarts, and so are read from bnd_topo in all cases.

I'd like to have a discussion on whether the topo vars should be read in from restarts or read from namelist bnd_topo for restart runs. If it is decided that topo vars should be read from bnd_topo in all cases, then I would advocate for removing topo vars from the restarts.

This issue is inspired by confusion on how to couple CISM to CAM, in the scenario where we want changes in ice sheet topography to be reflected in the CAM topo fields.

@PeterHjortLauritzen
@Katetc