ATMOSPHERIC MERCURY - 29 September 2006 The CMAQ version 4.6 treatments for mercury are essentially the same as for the version 4.5.1 release. Version 4.6 contains mercury-specific treatments for gaseous chemistry, aqueous chemistry, aerosol behavior and deposition. The changes from version 4.5.1 relate to gaseous chemistry and aerosol behavior. The gaseous chemistry mechanism for mercury in version 4.6 is a modification of the CB05 mechanism (see CB05_NOTES) instead of the CB-IV mechanism used in version 4.5.1. As with the previous version, four gas-phase chemical reactions involving mercury have been added. Aerosol behavior is treated in version 4.6 using an adaptation of the AE4 option instead of the AE3 option used in version 4.5.1. Four mercury species were added to the model: elemental mercury gas (HG), divalent gaseous mercury (HGIIGAS), I-mode aerosol mercury (APHGI), and J- mode aerosol mercury (APHGJ). At this time, only the EBI solver has been applied for the gaseous chemistry mechanism including mercury. Simulation of mercury also requires a treatment of aqueous chemistry with a special version of the AQCHEM routine and simulation of chlorine using the capability first added with CMAQ version 4.5 and updated through version 4.6. The cloud chemistry mechanism in AQCHEM has been modified to include six new chemical reactions and 14 sorption/de-sorption reactions involving mercury. Some of the aqueous chemical reactions of mercury are relatively fast compared to the non-mercury reactions previously in the model and the iterative solution criteria used to solve the aqueous chemical system were modified to maintain numerical stability and accuracy. Solution of the aqueous chemical model with mercury added requires a significant increase in the number of calculations. However, cloud chemistry was previously a small fraction of the total computational requirement and the increase in CPU time for the entire model calculation is normally 10 to 20% depending on the fraction of finite volumes where cloud water is present. An earlier experimental implementation of mercury treatments in CMAQ is described in Bullock and Brehme (2002). As with CMAQ version 4.5.1, version 4.6 contains updates in a number of areas to improve the underlying science and to address comments from peer review. The updates in mercury chemistry are as follows: (1) the elemental mercury reaction with hydrogen peroxide assumes the formation of 100% divalent gaseous mercury rather than 100% aerosol mercury; (2) the elemental mercury reaction with ozone assumes the formation of 50% divalent gaseous mercury and 50% aerosol mercury rather than 100% aerosol mercury; (3) the elemental mercury reaction with hydroxyl radical assumes the formation of 50% divalent gaseous mercury and 50% aerosol mercury rather than 100% aerosol mercury; and (4) the rate constant for the HG + OH reaction was lowered slightly to 7.7 x 10**(-14) cm**(3) molecules**(-1) s**(-1) based on the lower range of the kinetic rate constant estimated by Pal and Ariya (2004). Dry deposition of elemental mercury gas was also added, with deposition velocity values provided from MCIP using a heuristic approach involving estimates of surface reactivity, stomatal resistance, and mesophyll resistance factors, and accepted values for the molecular diffusivity of atomic mercury. It should be noted that this dry deposition of elemental mercury can lead to unrealistic depletion of that species if natural mercury emissions and re-emission of anthropogenic mercury are not also accounted for. CMAQv4.6 mercury model development was conducted using biogenic mercury emission information derived from a special version of the Biogenic Emissions Inventory System (BEIS) described in Lin et al. (2005). To run the mercury version of CMAQ, you will need to generate emissions files using SMOKE with the correct ancillary files (see Appendix A below for a description of the mercury emission processing). Initial and boundary condition files should be produced by the ICON and BCON processors with the same mechanism as the CCTM. We have outlined below the procedure for building and running ICON and BCON using a modified set of "profile" data. ICON: You will need to compile ICON to use the "cb05hg_ae4_aq" mechanism. The following is a "patch" style difference in the bldit.icon.pgf script: 42c42 < set APPL = e1a --- > set APPL = h1a 64,66c64,65 < #set ModMech = ( module radm2_to_cb4 $Revision; ) < set ModMech = ( module radm2_to_cb05 $Revision; ) < #set ModMech = ( module radm2_to_saprc99 $Revision; ) --- > #set ModMech = ( module radm2_to_cb4hg $Revision; ) > set ModMech = ( module radm2_to_cb05hg $Revision; ) 71,75c70,71 < #set Mechanism = cb4_ae3_aq < #set Mechanism = cb4_ae4_aq < set Mechanism = cb05_ae4_aq < #set Mechanism = saprc99_ae3_aq < #set Mechanism = saprc99_ae4_aq --- > #set Mechanism = cb4hg_ae3_aq > set Mechanism = cb05hg_ae4_aq Likewise you will need to modify run.icon to access the appropriate profile data. The following is a "patch" style difference in run.icon: 7c7 < # Usage: run.icon >&! icon_e1a.log & # --- > # Usage: run.icon >&! icon_h1a.log & # 22,23c22,23 < set APPL = cb05 < set CFG = e1a --- > set APPL = cb05hg > set CFG = h1a 80c80 < setenv IC_PROFILE $M3DATA/raw/icon/ic_profile_v7.dat --- > setenv IC_PROFILE $M3DATA/raw/icon/ic_profile_v7h.dat BCON: You will also need to compile BCON to use the "cb05hg_ae4_aq" mechanism. The following is a "patch" style difference in the bldit.bcon.pgf script: 42c42 < set APPL = e1a --- > set APPL = h1a 62,64c62,63 < #set ModMech = ( module radm2_to_cb4 $Revision; ) < set ModMech = ( module radm2_to_cb05 $Revision; ) < #set ModMech = ( module radm2_to_saprc99 $Revision; ) --- > #set ModMech = ( module radm2_to_cb4hg $Revision; ) > set ModMech = ( module radm2_to_cb05hg $Revision; ) 69,73c68,69 < #set Mechanism = cb4_ae3_aq < #set Mechanism = cb4_ae4_aq < set Mechanism = cb05_ae4_aq < #set Mechanism = saprc99_ae3_aq < #set Mechanism = saprc99_ae4_aq --- > #set Mechanism = cb4hg_ae3_aq > set Mechanism = cb05hg_ae4_aq You will need to modify run.bcon to use the appropriate profile data. The following is a "patch" style difference in run.bcon: 7c7 < # Usage: run.bcon >&! bcon_e1a.log & # --- > # Usage: run.bcon >&! bcon_h1a.log & # 22,23c22,23 < set APPL = cb05 < set CFG = e1a --- > set APPL = cb05hg > set CFG = h1a 82c82 < setenv BC_PROFILE $M3DATA/raw/bcon/bc_profile_v7.dat --- > setenv BC_PROFILE $M3DATA/raw/bcon/bc_profile_v7h.dat CCTM: In order to build the mercury version of the CCTM executable program, several modules and a different chemical mechanism must be specified in the model build script. You can modify the bldit.cctm.linux script as follows and build the mercury version of the CCTM: 43c43 < set APPL = e3a --- > set APPL = h3a 94,95c94,95 < #set ModVdiff = ( module eddy $Revision; ) < set ModVdiff = ( module acm2 $Revision; ) --- > #set ModVdiff = ( module eddy_hg $Revision; ) > set ModVdiff = ( module acm2_hg $Revision; ) 107,109c107,108 < #set ModChem = ( module ebi_cb4 $Revision; ) < set ModChem = ( module ebi_cb05 $Revision; ) < #set ModChem = ( module ebi_saprc99 $Revision; ) --- > #set ModChem = ( module ebi_cb4hg $Revision; ) > set ModChem = ( module ebi_cb05hg $Revision; ) 112,113c111,112 < #set ModAero = ( module aero3 $Revision; ) < set ModAero = ( module aero4 $Revision; ) --- > #set ModAero = ( module aero3_hg $Revision; ) > set ModAero = ( module aero4_hg $Revision; ) 116c115 < set ModAdepv = ( module aero_depv2 $Revision; ) --- > set ModAdepv = ( module aero_depv2_hg $Revision; ) 119,120c118 < #set ModCloud = ( module cloud_radm $Revision; ) < set ModCloud = ( module cloud_acm $Revision; ) --- > set ModCloud = ( module cloud_acm_hg $Revision; ) 130,134c128,129 < #set Mechanism = cb4_ae3_aq < #set Mechanism = cb4_ae4_aq < set Mechanism = cb05_ae4_aq < #set Mechanism = saprc99_ae3_aq < #set Mechanism = saprc99_ae4_aq --- > #set Mechanism = cb4hg_ae3_aq > set Mechanism = cb05hg_ae4_aq Next, modify the run script as follows (you will need to generate your own emissions files): 7c7 < # Usage: run.cctm >&! cctm_e3a.log & # --- > # Usage: run.cctm >&! cctm_h3a.log & # 22,23c22,23 < set APPL = e3a < set CFG = e3a --- > set APPL = h3a > set CFG = h3a 113,114c113,114 < set EMISpath = $M3DATA/emis/2001 < set EMISfile = emis3d.2001_ah.us36b.20010722.wndw.ncf --- > set EMISpath = $M3DATA/emis/2001hg > set EMISfile = emis3d.2001_ah.hg.us36b.20010722.wndw.ncf 122c122 < set GC_ICfile = ICON_cb05_M_36_2001_profile --- > set GC_ICfile = ICON_cb05hg_M_36_2001_profile 125c125 < set GC_BCfile = BCON_cb05_M_36_2001_profile --- > set GC_BCfile = BCON_cb05hg_M_36_2001_profile Note: You can change the default scripts by using the Unix "patch" utility. Cut the indented section listed above into a file, say "mod." Then type, for example, "patch run.cctm mod." References Bullock, O. R. and K. A. Brehme (2002). Atmospheric mercury simulation using the CMAQ model: formulation description and analysis of wet deposition results. Atmos. Environ. 36, 2135-2146. Lin, C.-J., S. E. Lindberg, T. C. Ho, C. Jang (2005). Development of a processor in BEIS3 for estimating vegetative mercury emissions in the continental United States. Atmos. Environ. 39, 7529-7540. Pal, B. and P. A. Ariya (2004). Gas-phase HO-initiated reactions of elemental mercury: kinetics, product studies, and atmospheric implications. Environ. Sci. Technol. 38, 5555-5566. ------------------------------------------------------------------------ Appendix A. Clean Air Mercury Rule (CAMR) Mercury Emissions Used as Input Ancillary Files Used with SMOKE for Processing the Mercury Inventories During emissions modeling, the ancillary data are combined with the emissions inventory data to convert the inventories into the gridded, hourly resolution, and chemical species needed by CMAQ to simulate mercury. The ancillary files that are different from the ones used for non-mercury modeling are: (1) the inventory table, which is used to read in the pollutant codes from the NEI and convert to SMOKE inventory pollutants, (2) the chemical speciation data used to convert the raw mercury emissions into the species needed by the CMAQ model. (1) Ancillary file for reading in the NEI Inventory We developed the SMOKE input file called the “inventory table” to read in the various species of mercury reported in the 1999 NEI and convert them to either the model species needed by CMAQ or to an unspeciated mercury placeholder called “HGSUM” that was speciated through the SMOKE Spcmat program using speciation profiles. The fields shown in the table below provide the same information as was included in the inventory table. The “adjustment factor” in this table allows SMOKE to adjust the mass of those inventory compounds containing elements in addition to mercury, to account for only the mercury portion of the compound. SMOKE NAME NEI CAS Adjustment Factor Inventory Pollutant Name Inventory Pollutant Code ---------- -------- ----------------- ------------------------------ ---------------------------- PHGI 202 1 Particulate Divalent Mercury Particulate Divalent Mercury HGIIGAS 201 1 Gaseous Divalent Mercury Gaseous Divalent Mercury HG 200 1 Elemental Gaseous Mercury Elemental Gaseous Mercury HGIIGAS 7487947 0.7388 Gaseous Divalent Mercury Mercuric chloride HGSUM 199 1 Mercury Compounds, unspeciated Mercury & Compounds HGSUM 22967926 1 Mercury Compounds, unspeciated MERCURY (ORGANIC) HGSUM 62384 0.5957 Mercury Compounds, unspeciated MERCURY ACETATO PHEN HGSUM 7439976 1 Mercury Compounds, unspeciated Mercury HGSUM 12 1 Mercury Compounds, unspeciated Mercury & Compounds (2) Ancillary Files for Chemical speciation The mercury speciation factors are used by SMOKE to convert any unspeciated mercury (i.e., HGSUM mentioned above) into the CMAQ model species: particulate divalent mercury (PHGI), gaseous divalent mercury (HGIIGAS) and elemental mercury (HG). Speciation profiles were assigned to the inventory sources using MACT codes and SCC codes. We preferentially used MACT-based profile assignments over SCC-based profile assignments. The global default speciation profile (applied for records that neither matched to MACT or SCC-based profiles) use 20% PHGI, 30% HGIIGAS, and 50% HG. In addition, we corrected the speciation profile data file provided with SMOKE version 2.1 to use a molecular weight of 200.59 for HGIIGAS. The profile data used for MACT and SCC codes are those shown in Appendix B of the Clean Air Mercury Rule Technical Support Document for Emissions (CAMR, 2005). Reference: CAMR (2005). Emissions inventory and emissions processing for the Clean Air Mercury Rule (CAMR). U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Emissions Monitoring and Analysis Division, Research Triangle Park, NC, available on-line at http://www.epa.gov/ttn/atw/utility/emiss_inv_oar-2002-0056-6129.pdf