Model Representation of Secondary Organic Aerosol in CMAQv4.7

Abstract

Numerous scientific upgrades to the representation of secondary organic aerosol (SOA) are incorporated into the Community Multiscale Air Quality (CMAQ) modeling system. Additions include several recently identified SOA precursors: benzene, isoprene, and sesquiterpenes; and pathways: in-cloud oxidation of glyoxal and methylglyoxal, particle-phase oligomerization, and acid enhancement of isoprene SOA. NO(x)-dependent aromatic SOA yields are also added along with new empirical measurements of the enthalpies of vaporization and organic mass-to-carbon ratios. For the first time, these SOA precursors, pathways and empirical parameters are included simultaneously in an air quality model for an annual simulation spanning the continental U.S. Comparisons of CMAQ-modeled secondary organic carbon (OC(sec)) with semiempirical estimates screened from 165 routine monitoring sites across the U.S. indicate the new SOA module substantially improves model performance. The most notable improvement occurs in the central and southeastern U.S. where the regionally averaged temporal correlations (r) between modeled and semiempirical OC(sec) increase from 0.5 to 0.8 and 0.3 to 0.8, respectively, when the new SOA module is employed. Wintertime OC(sec) results improve in all regions of the continental U.S. and the seasonal and regional patterns of biogenic SOA are better represented.

Multi-generational oxidation model to simulate secondary organic aerosol in a 3-D air quality model

“…The "Base" SOA model is equivalent to that used in the Community Multiscale Air Quality (CMAQ) model version 4.7 (Carlton et al, 2010). This Base model is representative of current-generation SOA models.…”

“…It should be noted that the experimental data used here to determine the SOM fit parameters are not the same data as used in developing the parameters in the Base model (Carlton et al, 2010). This difference in data sets can be expected to lead to some differences in the resulting simulated SOA concentrations.…”

“…Since the model-predicted OA at Riverside is dominated by POA (∼ 90 %), the O : C is controlled by the O : C of the emitted POA (∼ 0.1-0.2) and is lower than the campaignaveraged O : C of 0.31 inferred from the AMS data. The underprediction (in SOA concentrations and O : C) is typical of predictions in regional (Carlton et al, 2010) and global models (Farina et al, 2010) and arises mostly from an incomplete understanding of the sources and pathways of OA. Numerous factors may contribute to the underprediction of O : C at Riverside, including missing emission sources for SOA precursors, semi-volatile and reactive behavior of POA (Robinson et al, 2007), SOA formation from unspeciated emissions , aqueous production of SOA in cloud, fog and aerosol water (McNeill, 2015) and multi-generational ageing (Donahue et al, 2012b).…”

“…Most box (0-D) and large-scale (3-D) models represent SOA production from the gas-phase oxidation of certain VOCs (large alkanes, aromatics, isoprene and terpenes) to yield 2-4 low-volatility products that partition into the particle phase (Odum et al, 1996;Carlton et al, 2010;Lane et al, 2008). Laboratory chamber data provide the basic information on which these SOA formation models are built.…”

Impact of the 2008 Global Recession on air quality over the United States: Implications for surface ozone levels from changes in NOx emissions

“…This version of CMAQ is configured with the Carbon Bond 2005 chemical mechanism [Yarwood et al, 2005] and AERO5 aerosol module [Carlton et al, 2010]. The lateral boundary conditions used in the simulation are monthly averaged profiles extracted from the 2006 simulation with Harvard University's GEOS-Chem model [Zhang et al, 2011].…”