The operational version of GEM-MACH employs a limited-area configuration with 15-km grid spacing over most of Canada and the U.S. Two size bins and 9 chemical species (SO4, NO3, NH4, EC, pOC, SOA, crustal material, sea-salt and water) are used to represent the size distribution and chemical composition of particulate matter. Air-quality process representations include a 42-species gas-phase chemical mechanism, and additional modules for aqueous chemistry and inorganic and organic heterogeneous chemistry, dry and wet deposition, emissions, and particle microphysics. Anthropogenic emissions files are created using the SMOKE emissions processing system, and biogenic emissions are calculated on-line using BEIS v3.09.
Several process modifications were made to the original model: (1) the gas-phase mechanism was modified to include alpha-pinene and beta caryophyline as representatives of monoterpenes and sesquiterpenes, as well as diagnostically predicted concentrations of intermediate volatility organic compounds (IVOCs), benzene, methacrolein and methyl-vinyl-ketone; (2) the aerosol yields were parameterized based on a 2-product fit to Cal Tech and Carnegie Mellon chamber data (Kroll et al., 2006, Lane et al., 2008, Pathak et al., 2007, Ng et al., 2007) , and the formation of secondary organic aerosols from a simplified IVOC oxidation scheme was added (Stroud et al, 2010); (3) the original Young-Boris gas-phase solver was replaced by a modified version of the RODAS3 chemistry solver of the KPP package (Sandu and Sander, 2006); and (4) the biogenic emissions for pine and spruce in Canada's boreal forests were updated to more representative values (based on leaf biomass measurements in Chen et al. (1996) and Bond-Lamberty et al., (2002)).
The updated model has been reconfigured for nesting: the original 15-km grid covering North America has been reduced to a western regional domain in order to reduce processing time and storage, and three 2.5-km subdomains of interest were selected for nesting: (1) a Georgia Basin subdomain, including the Seattle, Wa. to Vancouver, B.C. corridor; (2) a SW California subdomain (to be used for comparisons to the CalNEX measurement campaign data); and (3) a Canadian oil sands subdomain (for use in forecasting and comparisons to an upcoming EC-sponsored measurement campaign in that region). The setup and testing of these nested domains will be discussed, along with preliminary comparisons to observations and to the operational GEM-MACH15 forecasts.