Add abiotic tracers

[mnlevy1981]

Port the abio_dic_dic14_mod.F90 module from POP to MARBL. We should be able to run with the two abiotic tracers without turning on what the base (non-isotopic) MARBL tracers.

[mnlevy1981]

Currently we are using same atmospheric CO2 forcing from GCM for biotic and abiotic tracers; while the majority of runs will likely use the same value for both, we do want to support running both modules with different CO2 forcings (Keith L gave the example of using emissions-driven CO2 for base biotic tracers in a coupled run but something different for abiotic)

[mnlevy1981]

If abio_on = .true. and base_biotic_on = .false., then we need autotroph_cnt and zooplankton_cnt set to 0 (may as well set max_grazer_prey_cnt as well, for completeness) so that setup_local_tracers() doesn't try to set tracers that don't exist. This issue is cropping up now because I just moved the call to setup_local_tracers() out of the if (base_biotic_on) block.

[mnlevy1981]

commits 6f552e9 and a8bb2a7 reverted some changes in settings YAML files that could have been caught by a regression test for MARBL_generate_settings.py that runs for each settings JSON file and makes sure the resulting settings file matches a baseline (in this case, there are expected changes but the YAML file also introduced several unexpected changes). I think that's a simple addition to run_test_suite.sh but if it turns into something more complicated I'll open a separate issue / PR for it

[mnlevy1981]

The tests that fail in 5a1f06e pass on my laptop, but fail on cheyenne (with gnu) -- I'll investigate further on that machine, it's not clear what changed in that commit to cause problems.