On this page we list information about the stratospheric chemistry mechanism in GEOS-Chem.
For the influence of stratospheric composition on tropospheric photolysis rates, see the Fast-J page.
Though GEOS-Chem is predominantly used for simulating tropospheric chemistry, it is necessary to deal with species that may have net stratospheric sources and transport to the troposphere (e.g., NOx), or net stratospheric sinks of gases transported from the troposphere (e.g., CO). An updated linearized stratospheric chemistry scheme has been prepared for implementation into GEOS-Chem v9-01-03 to complement the linearized ozone chemistry afforded by Linoz.
Authors and collaborators:
- Lee Murray (Harvard)
New implementation of linearized stratospheric chemistry
An updated linearized stratospheric chemistry scheme has been implementated into GEOS-Chem v9-01-03o. It currently works with GCAP/GEOS4/GEOS5/MERRA/GEOS5.7 simulations at all resolutions, for full chemistry, SOA, dicarbonyl, isoprene, H2/HD, tagged Ox, and tagged CO.
One may turn on the linearized stratospheric chemistry scheme in the chemistry menu of the input.geos file
%%% CHEMISTRY MENU %%% : Turn on Chemistry? : T Use linear. strat. chem?: T => Use Linoz for O3? : T Chemistry Timestep [min]: 60 Read and save CSPEC_FULL: T USE solver coded by KPP : F
At the beginning of a new month, the model reads in the archived 3D monthly mean production rate, P (v/v/s), and loss frequency, k (s-1), for each species. The new concentration at the end of each chemistry time step is then determined by solving
! Simple analytic solution to dM/dt = P - kM over [0,t] if ( k .gt. 0d0 ) then STT(I,J,L,N) = M0 * exp(-k*t) + (P/k)*(1d0-exp(-k*t)) else STT(I,J,L,N) = M0 + P*t endif
This is applied for each pertinent species to every grid box for which the tropospheric chemistry solver is not applied in a given chemistry time step.
Also, monthly mean stratospheric OH fields are read in and used to generated decay for certain species (currently, CH3Br, CHBr3, CH2Br2).
The bromine species are hard-wired to be ignored for now. Bry in the stratosphere is prescribed using monthly mean nighttime and daytime concentrations from the GEOS CCM as in its original (v9-01-03m) implementation. For more information, see the bromine chemistry mechanism wiki page.
Note: Linoz is still the recommended option for stratospheric ozone, and ozone prod/loss rates were only implemented here for completeness. If one turns off Linoz, then Synoz is applied in the code.
Preparation of monthly mean 3D prod/loss rates
The production rates (v/v/s), loss frequencies (s-1) and concentrations (v/v) have been archived from the GMI "Combo" model simulations, which one can think of as having the GEOS-Chem troposphere + the GMI full stratospheric chemistry model. For GEOS4/GEOS-Chem simulations, long-term monthly means were calculated from the "Aura4" GMI monthly concentrations, which used GEOS-4 assimilated met fields for 2004-2006. For GEOS5/ or MERRA/GEOS-Chem simulations, long-term monthly means were determined from the "MERRA" GMI rerun (containing the Synoz bug fix in the GMI model) using MERRA met fields for 2004-2010.
The GMI model outputs as diagnostic the monthly mean rate of each kinetic or photolysis reaction. That output was aggregated to get the total production rate (v/v/s). Photolysis frequencies are also available, but kinetic loss frequencies had to be estimated from the burden/loss rate, to determine the mean loss frequencies (s-1).
Species with linearized stratospheric chemistry
Numbers indicate number of GMI Combo model reactions aggregated, see [reaction details document] for more information
|Species Name||Photolysis Prod||Kinetic Prod||Photolysis Loss||Kinetic Loss|
Additionally available species
The following species from the GMI model have also had their monthly mean production rates (v/v/s), loss frequencies (s-1), and concentrations (v/v) archived for GEOS-Chem. The linearized chemistry will automatically have prod/loss applied if a tracer sharing one of the following names is introduced into the model's STT array of transported tracers and DO_STRAT_CHEM is called.
H, H2, HCOOH, H2O, HO2, MO2, MOH, N, N2O, O, O1D, OH, Br, BrCl, BrO, BrONO2, HBr, HOBr, Cl, Cl2, ClO, Cl2O2, ClONO2, HCl, HOCl, OClO, CH3Br, CH3Cl, CH3CCl3, CCl4, CFCl3, CF2Cl2, CFC113, CFC114, CFC115, HCFC22, HCFC141b, HCFC142b, CF2Br2, CF2ClBr, CF3Br, H24O2, A3O2, ATO2, B3O2, EOH, ETO2, ETP, GCO3, GP, IALD, IAO2, IAP, INO2, INPN, ISN1, ISNP, KO2, MAN2, MAO3, MAOP, MCO3, MRO2, MRP, MVN2, PO2, PP, PRN1, PRPN, R4N1, R4O2, R4P, RA3P, RB3P, RCO3, RCOOH, RIO1, RIO2, ROH, RP, VRO2, VRP.
The bromine species are hard-wired to be ignored for now. Bry in the stratosphere is prescribed using monthly mean nighttime and daytime concentrations from the GEOS CCM, located in the bromine_201205/CCM_stratosphere_Bry data directory.
Also note, the input file may be accessed from anywhere to obtain monthly mean concentrations (e.g., 4D stratospheric OH) for use within the model.
All stratospheric-relevant chemistry routines (including Linoz) have been consolidated into a single driving routine DO_STRAT_CHEM in GeosCore/strat_chem_mod.F90, called from DO_CHEMISTRY in GeosCore/chemistry_mod.F for full chemistry, tagged Ox and H2/HD simulations. The tagged CO simulation has been updated in GeosCore/tagged_co_mod.F.
- GeosCore/strat_chem_mod.F: Source code file with stratospheric chemistry. The Linoz routines are called from here.
- GEOS_NATIVE/strat_chem_201206/gmi.clim.*.nc: Species-specific vertical and horizontal resolution dependent files containing mean production rates and loss frequencies from the GMI combo model in the primary data directory
Note, the input file is in the netCDF file format, and therefore also requires GEOS-Chem be compiled with netCDF.
The original stratospheric chemistry scheme through GEOS-Chem v9-01-02 (v9-01-03n) consisted of:
- A fixed stratosphere-to-tropophere flux of ozone from Synoz (an alternative and recommended linearized ozone chemistry Linoz option introduced in v8-02-04)
And performed a simple linearized chemistry scheme from archived monthly mean production rates, J-values, and OH fields from the 2D stratospheric chemistry model of Hans Schneider and Dylan Jones using GEOS-1 and/or GEOS-STRAT met fields (36 vertical levels) that have since been regridded to subsequent GEOS grid vertical resolutions.
- Fixed monthly zonal mean production rates of NOy in pnoy_200106
- Fixed monthly zonal mean production rates and loss frequencies of CO in pco_lco_200203
- Loss of selected species by reaction with archived monthly zonal mean stratospheric OH fields in stratOH_200203
- Loss of selected species by photolysis using archived monthly zonal mean J-values archived in stratjv_200203
These were variously performed by the subroutines
- DO_UPBDFLX - Called from main.f before transport, calls Synoz/Linoz ozone.
- UPBDFLX_NOY - Called first from DO_UPBDFLX for NOy production, and then again after transport in main.f to repartition NOy into NOx and HNO3.
- schem.f - Called just after SMVGEAR/KPP solves tropospheric chemistry in chemdr.f. Performs
- Photolysis loss
- CO prod/loss
- Reaction with OH
|Species||Photolysis Loss||Loss by OH||Special Treatment|
|Ox||Synoz (P) or Linoz (P&L)|
Coming soon, with STE fluxes.
Murray, L.T., D.J. Jacob, J.A. Logan, R.C. Hudman, and W.J. Koshak, Optimized regional and interannual variability of lightning in a global chemical transport model constrained by LIS/OTD satellite data, submitted to J. Geophys. Res., 2012.
Previous issues that are now resolved
Minor bug in parallel DO loop for Bry concentrations
The BEFORE variable was declared as a 3-D allocatable array at the top of module strat_chem_mod.F90:
REAL*8, ALLOCATABLE :: Before(:,:,:) ! Init atm. state each chem. dt
and was allocated at the start of the GEOS-Chem simulation by routine INIT_STRAT_CHEM. BEFORE was also used in this parallel loop in routine DO_STRAT_CHEM:
!=============================== ! Prescribe Br_y concentrations !=============================== !$OMP PARALLEL DO & !$OMP DEFAULT( SHARED ) & !$OMP PRIVATE( NN, I, J, L, BryDay, BryNight, IJWINDOW ) DO NN=1,6 IF ( GC_Bry_TrID(NN) > 0 ) THEN ! Make note of inital state for determining tendency later BEFORE = STT(:,:,:,GC_Bry_TrID(NN)) ... etc ...
As you can see, the BEFORE array has (I,J,L) scope, but is located within an OpenMP parallel DO loop with (I,J,L,N) scope. Therefore, BEFORE should have been declared in the !$OMP PRIVATE listing, but was not.
To correct this situation we changed BEFORE from an allocatable array to a local array within routine DO_STRAT_CHEM:
! Arrays REAL*8 :: STT0(IIPAR,JJPAR,LLPAR,N_TRACERS) REAL*8 :: BEFORE(IIPAR,JJPAR,LLPAR)
and also added it to the !OMP PRIVATE declaration in this parallel DO loop:
!=============================== ! Prescribe Br_y concentrations !=============================== !$OMP PARALLEL DO & !$OMP DEFAULT( SHARED ) & !$OMP PRIVATE( NN, BEFORE, I, J, L, BryDay, BryNight, IJWINDOW ) DO NN=1,6 IF ( GC_Bry_TrID(NN) > 0 ) THEN ! Make note of inital state for determining tendency later ! NOTE: BEFORE has to be made PRIVATE to the DO loop since ! it only has IJL scope, but the loop is over IJLN! ! (bmy, 8/7/12) BEFORE = STT(:,:,:,GC_Bry_TrID(NN)) ... etc ...
This ensures that the strat-trop exchange diagnostic will be computed properly.
--Bob Y. 17:29, 8 August 2012 (EDT)
Bug fix for Intel Fortran Compiler 12
Prasad Kasibhatla wrote:
- I am experiencing the IFORT 12 compile failure of GEOS-Chem v9-01-03 with nested grid options turned on in define.h. The compile seems to be failing in the Stratospheric chemistry module GeosCore/strat_chem_mod.F90.
Please see this wiki post on our Intel Fortran Compiler wiki page for a full description of the issue.
--Bob Y. 15:27, 20 December 2012 (EST)