Bromine chemistry mechanism

NOTE: Page under construction!

Overview

In GEOS-Chem v9-01-03, we have added 10 additional bromine species to the standard GEOS-Chem chemistry mechanism. You can revert to the previous mechanism by turning off emissions for these species.

Original implementation

This update was tested in the 1-month benchmark simulation v9-01-03m and approved on 06 Jun 2012.

Justin Parella added bromine chemistry as an extension to the standard model's full chemistry simulation in GEOS-Chem v8-03-01. We have now brought Justin's updates into GEOS-Chem v9-01-03. Important features include:

1. Add 10 bromine tracers: Br2, Br, BrO, HOBr, HBr, BrNO2, BrONO2, CHBr3, CH2Br2, and CH3Br
2. 4 source gases:
   1. CH3Br: PBL concentrations are set to average CMDL measured concentrations prior to each chemistry step.
   2. Very short lived (VSL) source gases:
      • CH2Br2 and CHBr3. These account for > 80% of the VSL organic bromine through the troposphere (WMO, 2007).
      • Emissions are taken from Q. Liang et al. [2010] + a seasonal scaling for CHBr3 at lat > 30N described in J. Parrella et al. [in prep.]
        • 429 Gg CHBr3 / yr; 62 Gg CH2Br2 / yr
   3. Sea-salt debromination:
      • Largely follows Yang et al. [2005], treating the debromination as an emission of Br2, constrained to measured bromide depletion factors.
      • 1.4 Tg Br2 / yr using the Monahan sea-salt production function and coarse aerosol bin range of 0.5 - 10 um.
3. Wet and dry deposition for HBr, HOBr, and BrONO2.

Updates for GEOS-Chem v9-01-03o

Lee Muray and Justin Parrella modified the original implementation of the bromine chemistry. We now revert to prescribing stratospheric bromine concentrations from the CCM_stratosphere_Bry fields, but from within the new stratospheric code.

Lee Murray wrote:

In addition to the problems with bromine that Bob pointed out in his email from last week (e.g., won't work with nested simulations), the bromine code also would not work for GEOS-4 or GCAP as it used some GEOS-5-only met fields, or any type of other chemistry simulation (e.g., SOA) since the tracer numbers were hardwired. I made the following changes to the bromine code in this commit:

1. I made use of IDTBR, IDTBrO, etc., in the stratospheric code to generalize tracer number references for other simulation types
2. The Warwick VSL emission files are now a netCDF file always read in at 2x2.5, and re-gridded to whatever resolution the actual simulation is using REGRID_A2A :#The bromine stratospheric input files have been re-gridded for the GEOS-4 and GCAP vertical resolutions
3. For nested simulations, the stratospheric bromine fields are interpolated from the native coarser resolution to the nested grid, as my new P & k files are
4. I have switched the use of FROCEAN and FRLAND (GEOS-5 met fields) to the helper functions IS_LAND, IS_WATER and IS_ICE to generalize the code to GEOS4/GCAP/other. I left in the old treatment for GEOS_5 for now with preprocessing statements, though the GEOS4/GCAP treatment should be identical and preferable as it is more universal.
5. Only GEOS-5 has cloud ice and liquid water content met fields, which is used by the bromine rate calculation in calcrate.F. Mimicking what we do in wetscav_mod.F, I parameterize those variables as a function of temperature for GEOS-4 and GCAP.

I have extensively tested and validated the global simulations. I have tested the nested simulations, but not extensively.

In my comparisons of v9-01-03n and my most recent commit, I found a few small differences in the in the bromine species in stratosphere (e.g., pages 44-47 of this PDF). Stratospheric bromine fields in GEOS-Chem are a weighted average of the monthly mean day and night concentrations in CCM_stratosphere_Bry/, depending on relative amount of daylight and nighttime in the grid box. In July, the South Pole is entirely night and North Pole is entirely day. We should expect them to match the CCM_stratosphere_Bry/ exactly there, but v9-01-03n does not (v9-01-03n benchmark file is used in this plot). That being said, I do not see anything wrong with the original implementation in SCHEM.F. Regardless, my implementation in strat_chem_mod.F90 does match the input fields exactly as expected, and I am confident it is correct.

--Bob Y. 12:42, 10 July 2012 (EDT)