Time step vs accumulated graupel: Thompson vs GFS_MP_generic.F90

[ericaligo-NOAA]

In mp_thompson.meta we see graupel defined as:
lwe_thickness_of_graupel_amount.

From mp_thompson.F90, graupel is accumulated during the model integration.
graupel = graupel + max(0.0, delta_graupel_mp/1000.0_kind_phys)

In GFS_MP_generic.meta, we see lwe_thickness_of_graupel_amount assigned to graupel0 and totgrp is said to be the accumulated graupel. As noted above, graupel in mp_thompson.meta and graupel0 in GFS_MP_gneric.meta are actually the accumulated graupel.

[totgrp]
standard_name = accumulated_lwe_thickness_of_graupel_amount
long_name = accumulated graupel precipitation
units = kg m-2
dimensions = (horizontal_loop_extent)
type = real
kind = kind_phys
intent = inout

In GFS_MP_generic.F90:
graupel = frain*graupel0 ! time-step graupel <-------this is actually the accumulated, not time step graupel.
totgrp (i) = totgrp (i) + graupel(i)

Am I correct in that the treatment of graupel in GFS_MP_generic.F90 is not correct since it assumes what's coming in is the one time step value, when it's actually the accumulated value?

Eric