Discrepency between marbl_settings_mod.F90 and settings_latest.yaml

[mnlevy1981]

As part of #430, I updated the documentation to mention the new unit_system_opt argument to marbl_instance%init(). The init page lists default parameters, both those hard-coded in the Fortran and those provided by settings_latest.yaml. It had been a while since the documentation had been updated, so the new parameters lcheck_forcing, bftt_dz_sum_thres, and zooplankton_settings(1)%basal_respiration_rate_per_day were all added to the list generated by the fortran. The list generated from the YAML included these three new variables, but also changed the reported default value in several other parameters. See the diffs for more detail, but things like

-  caco3_bury_thres_omega_calc = 0.89
+  caco3_bury_thres_omega_calc = 1.0
-  gQ_Si_kSi_thres = 6.0
+  gQ_Si_kSi_thres = 5.0
-  gQsi_max = 0.822
+  gQsi_max = 0.685
-  o2_sf_val_lo_o2 = 2.6
-  parm_CaCO3_gamma = 0.02
+  o2_sf_val_lo_o2 = 3.0
+  parm_CaCO3_gamma = 0.1

only appeared in the "generated from YAML" section. The fortran default for caco3_bury_thres_omega_calc is 0.89, and that's also what is set in settings_cesm2.1.yaml so it's unclear to me why these values didn't propagate to the latest YAML file.

The above is very verbose, but the point of this issue ticket is that I think we want the Fortran defaults and the values provided in settings_latest.yaml to match but there are several parameters where that is not the case. For those, should I update the Fortran or the YAML?

A secondary question: is it concerning that we appear to have altered a handful of settings just for CESM 2.1 and then reverted back to old values for CESM 2.2 and beyond?