Transport tracer species aoa*, nh*, St80_25, and StOX did not have molecular
weights defined because they're not traditional chemical species. This causes
issues with unit conversions in the code and results in negative and/or
zero SpeciesConc output. To avoid this, we define MW_g: 1.0 for those
species.

Signed-off-by: Melissa Sulprizio <mpayer@seas.harvard.edu>

1. Move initialization of mask arrays into loop over species.
2. Make sure to do calculations for maintaining the mixing ratio only in the defined region (where Mask is 1)
3. Fix typo in the logic for determining masks of latitude zones.

Signed-off-by: Melissa Sulprizio <mpayer@seas.harvard.edu>

The stOX tracer is defined in GMAO's TR_GridComp but is not commonly utilized
and isn't needed in GEOS-Chem. It is removed here to avoid confusion. The
fields needed for stOX (GLOBAL_O3 and O3_LOSS) are also removed from
tracer_mod.F90 and TransportTracer configuration files.

Signed-off-by: Melissa Sulprizio <mpayer@seas.harvard.edu>

Instead of multiplying the source or sink application by the Mask array,
the code now checks if Mask > 0 before applying the source or sink.
Multiplying by the Mask array may have inadvertently set species concentrations
to zero instead of not implying the source or sink.

Signed-off-by: Melissa Sulprizio <mpayer@seas.harvard.edu>

The Budget diagnostics collection is now needed for TransportTracer benchmarking.
It is now disabled by default to avoid saving out unnecessary data.

Signed-off-by: Melissa Sulprizio <mpayer@seas.harvard.edu>