H2SO4 production rate is needed for SO4 nucleation in MAM4 to produce more SO4.
Previously, the H2SO4 production rate (del_h2so4_gasprod) was an order of magnitude too low in CESM-GC compared to CAM-chem. This is because GEOS-Chem does not have a H2SO4 tracer; H2SO4 "production rate" is instead computed as the change in H2SO4 after GEOS-Chem time step.
CESM-GC's coupling to MAM4 simulates the H2SO4 tracer by recording the relative mass contributions of so4_a1, so4_a2, so4_a3, and h2so4 to the GEOS-Chem bulk SO4 tracer. This is the binRatio (for a1~a3) and gasRatio (for h2so4). Because these mass ratios come from MAM4, where sulfate nucleation has just occurred in the previous timestep, h2so4 is very low since it should have been consumed by nucleation. Thus, gasRatio is very low after coming out of the MAM4 timestep, resulting in "change in SO4" * gasRatio = "change in H2SO4" to be abnormally low.
Daniel and I discussed that the appropriate science fix is to use the SO2 + OH -> SO4 rate in GEOS-Chem for the H2SO4 production rate. This is recorded in fullchem_mod.F90, converted to mol/mol production of H2SO4, then stored in State_Chm%H2SO4_PRDR, which is now used here.
The conversion process is
#ifdef MODEL_CESM
! Calculate H2SO4 production rate for coupling to CESM (interface to MAM4 nucleation)
DO F = 1, NFAM
! Determine dummy species index in KPP
KppID = PL_Kpp_Id(F)
! Calculate H2SO4 production rate [mol mol-1] in this timestep (hplin, 1/25/23)
IF ( TRIM(FAM_NAMES(F)) == 'PSO4' ) THEN
! mol/mol = molec cm-3 * g * mol(Air)-1 * kg g-1 * m-3 cm3 / (molec mol-1 * kg m-3)
! = mol/molAir
State_Chm%H2SO4_PRDR(I,J,L) = C(KppID) * AIRMW * 1e-3_fp * 1.0e+6_fp / &
(AVO * State_Met%AIRDEN(I,J,L))
IF ( State_Chm%H2SO4_PRDR(I,J,L) < 0.0d0) THEN
write(*,*) "H2SO4_PRDR negative in fullchem_mod.F90!!", &
I, J, L, "was:", State_Chm%H2SO4_PRDR(I,J,L), " setting to 0.0d0"
State_Chm%H2SO4_PRDR(I,J,L) = 0.0d0
ENDIF
ENDIF
ENDDO
#endif
Updates H2SO4 production rate passed to MAM4 sulfate nucleation to use actual H2SO4 production rate computed by KPP in GEOS-Chem (SO2 + OH). This commit requires GEOS-Chem 14.1.1, specifically the field
State_Chm%H2SO4_PRDR
introduced in https://github.com/geoschem/geos-chem/commit/a79ba0c4cbf335642a4c256f35593bfd461e9a62.The background is as follows:
del_h2so4_gasprod
) was an order of magnitude too low in CESM-GC compared to CAM-chem. This is because GEOS-Chem does not have a H2SO4 tracer; H2SO4 "production rate" is instead computed as the change in H2SO4 after GEOS-Chem time step.so4_a1
,so4_a2
,so4_a3
, andh2so4
to the GEOS-Chem bulkSO4
tracer. This is thebinRatio
(for a1~a3) andgasRatio
(for h2so4). Because these mass ratios come from MAM4, where sulfate nucleation has just occurred in the previous timestep,h2so4
is very low since it should have been consumed by nucleation. Thus,gasRatio
is very low after coming out of the MAM4 timestep, resulting in "change in SO4" * gasRatio = "change in H2SO4" to be abnormally low.SO2 + OH -> SO4
rate in GEOS-Chem for the H2SO4 production rate. This is recorded infullchem_mod.F90
, converted tomol/mol
production ofH2SO4
, then stored inState_Chm%H2SO4_PRDR
, which is now used here.The conversion process is