+Add grid_unit_to_L to the ocean_grid_type

[Hallberg-NOAA]

Add the new element grid_unit_to_L to the ocean_grid_type and thedyn_horgrid_type, which can be used to convert the units of the G%geoLat and G%geoLon variables to rescaled horizontal distance units ([L ~> m]) when they are Cartesian coordinates. When Cartesian coordinates are not in use, G%grid_unit_to_L is set to 0.

This new element of the grid type is used to test for inconsistent grids or to eliminate rescaling variables in set_rotation_beta_plane(), initialize_velocity_circular(), DOME_initialize_topography(), DOME_initialize_sponges(), DOME_set_OBC_data(), ISOMIP_initialize_topography(), idealized_hurricane_wind_forcing(), Kelvin_set_OBC_data(), Rossby_front_initialize_velocity(), soliton_initialize_thickness(), and soliton_initialize_velocity(). These are the instances where this new variable could be used and bitwise identical answers are recovered. There are a few other places where they should be used, but where answers would change, and these will be deferred to a subsequent commit.

All answers are bitwise identical, but there are new elements in two transparent and widely used types.

[MJHarrison-GFDL]

The introduction of G%grid_unit_to_L makes sense and I approve. Unfortunate that this requires updating user code - unmerged user code on other forks will require manual updates.

[Hallberg-NOAA]

The use of this variable is not required, so if users do not want to update their code on other forks, they don't have to do anything.

[Hallberg-NOAA]

This was failing the pipeline testing on the first try for one case due to overly aggressive error detection on line 361 of Idealized_Hurricane.F90 , but this was readily corrected with a single-line fix to the logic around an error message that I showed to Theresa (who agreed that it was innocuous if it works), and now it has passed pipeline testing at https://gitlab.gfdl.noaa.gov/ogrp/mom6ci/MOM6/-/pipelines/26079.