totalChl_loc mismatch between surface_flux_compute() and interior_tendency_compute()

[mnlevy1981]

In surface_flux_compute(), we compute totalChl_loc by simply summing the surface values of the per-autotroph Chl tracers:

MARBL/src/marbl_surface_flux_mod.F90

Lines 178 to 189 in c23d407

	!-----------------------------------------------------------------------
	! Compute total chlorophyll
	!-----------------------------------------------------------------------
	if (sfo_ind%totalChl_id.ne.0) then
	totalChl_loc = c0
	do auto_ind = 1,autotroph_cnt
	totalChl_loc = totalChl_loc + &
	max(c0, tracers_at_surface(:,marbl_tracer_indices%auto_inds(auto_ind)%Chl_ind))
	end do
	surface_flux_output%sfo(sfo_ind%totalChl_id)%forcing_field = totalChl_loc
	end if

In interior_tendency_compute(), we first enforce a consistency check (which sets Chl to 0 if any of C, P, Fe, or Si are 0):

MARBL/src/marbl_interior_tendency_mod.F90

Lines 809 to 813 in c23d407

	! autotroph consistency check
	call autotroph_zero_consistency_enforce(column_kmt, marbl_tracer_indices, autotroph_local)
	! set totalChl_local
	totalChl_local = sum(autotroph_local%Chl(:,:), dim=1)

This discrepancy in how we treat Chl is coming up now because of differences in MOM and POP: in POP, we were using the surface flux totalChl_loc to compute absorption but MOM wants a 3D Chl field and so I'm updating MARBL to be able to provide the GCM with the interior tendency totalChl_local term.

Do we want the surface chlorophyll values returned to the GCM to match the surface values of the 3D chlorophyll returned to the GCM? If so, I think we have two options:

1. Enforce consistency in the surface, or
2. Compute totalChl_local in the interior prior to calling autotroph_zero_consistency_enforce()