get SCM vertical levels from namelist file instead of .nc simulation file.

[haakon-e]

No description provided.

[haakon-e]

This would imply that scm_path points directly to a namelist file instead of a simulation directory.

[haakon-e]

Two approaches:

1. For non-isotropic domains, a TC namelist file will still define parameters for some function that generates vertical levels. Instead of running a simulation, we can generate these vertical levels using TC functions along with parameters from a namelist file
2. Run a simulation for 1 timestep (linked to Allow setting SCM runtimetf from calibration pipeline. #71), then fetch vertical levels with existing methods.

[ilopezgp]

We need to:

• Read z_scm (grid),
• Namelist (to modify parameters and generate runs),
• Location of TC.jl variables (z_c or z_f) without running a full simulation.

@charleskawczynski had some ideas of how this might be done without even running the simulation for a single time step.

[charleskawczynski]

Can call stuff defined in TC.jl's integration_test/ folder?

[charleskawczynski]

If so, we could use:

sim = Simulation1d(namelist)

for variable_name in keys(sim.io_nt.aux)
    space = sim.io_nt.aux[variable_name].dims[1]
    # One of these will be true:
    # space == "zf"
    # space == "zc"
end

This will be subject to change, but it should work without actually simulating any steps.

[ilopezgp]

Fixed in #87 .