Update scm_vgrid and UI

This pull request addresses the following issues and discussions:

https://github.com/NCAR/ccpp-scm/discussions/305
https://github.com/NCAR/ccpp-scm/issues/288
https://github.com/NCAR/ccpp-scm/issues/234 (partial -- removes namelist option)
https://github.com/NCAR/ccpp-scm/issues/208 (partial -- last suggestion RE: output location)

Objective 1: update the vertical grid code to match the latest FV3 code and improve vertical coordinate usability

update scm_vgrid.F90 to match changes in FV3's fv_eta.F90
like FV3, put hard-coded ak/bk coefficients in fv_eta.h and bring the file in via include
remove old get_GFS_vgrid code -- not needed anymore since one can read files containing ak/bk coefficients in get_FV3_vgrid
remove "model_name" case configuration namelist variable since it is no longer needed; only get_FV3_vgrid should ever need to be called
remove n_levels from case config files (will use the default value unless one passes in an argument to the run script)
modify the run_scm.py script to take arguments of --levels, --npz_type, and --vert_coord_file
modify the multi_run_scm.py script to take the same arguments and pass them through to run_scm.py
set default npz_type to '' and --levels to 127 in scm_input.F90
set the path to the vertical coordinate date directory (for file reads) in run_scm.py

Implications:

Change the number of vertical levels by specifying them via the --levels (or -l) argument of the run_scm.py script (or multi_run_scm.py script); the default is now 127 levels
Set the --npz_type argument of the run scripts to use any of the npz_type options in scm_vgrid.F90; the default is npz_type = ''
To read in the ak/bk coefficients, set the npz_type command line argument to 'input', put a file with format specified in scm_vgrid.F90 into data/vert_coord_data, and specify the name of the file via the command line argument --vert_coord_file

Objective 2: Since we needed to modify all case configuration files, use this opportunity to clean up all unnecessary case configuration namelist options

remove n_columns option (since only 1 works right now anyway)
remove time_scheme (since we only really need to use Forward-Euler timestepping for forcing)
move dt, n_itt_out, and n_itt_diag to command line arguments for the run script since these aren't really tied to the cases anyway (should make it easier to modify these values on the fly)
set default values for n_itt_out and n_itt_diag in run_scm.py
set default value of dt according to the suite (since this is usually a function of the physics, or should be)
add a maximum value of dt according to the suite (and exit if user tries to use one bigger than the max)
remove 'output' namelist variable (user won't need to change the name of the file if run directories are correctly set up)
move 'case_data_dir' to a run script command line argument; the default is set to data/processed_case_input (where case data is currently stored), but setting the command line argument will tell the run script where to look for the data if it is elsewhere (e.g., in data/DEPHY-SCM)

Objective 3: Clean up suite defaults

remove default_tracers.py and default_namelists.py
add suite_info.py with new 'suite' class and default suite-based information

Objective 4: Add flexibility for output location

separate the build directory and run directory by default
modify the run script to set up a run directory outside the build directory (scm/run by default), controlled by the --run_dir command line argument
modify rt.sh to work with the changes in run_scm.py
add --runtime_mult command line argument to run script to replace running all SCM regression tests for a hard-coded number of seconds. This variable (if less than 1) reduces the runtime for each case, which is useful for regression testing.
require users to set SCM_ROOT environment variable to point to the top level ccpp-scm directory prior to running run_scm.py in order facilitate maximally-flexible run directory (and output) location

NCAR / ccpp-scm

Update scm_vgrid and UI #307