Modeling the formation and aging of secondary organic aerosols in the Los Angeles metropolitan region during the CalNex 2010 field campaign

Several different literature parameterizations for the formation and evolution of urban secondary organic aerosol (SOA) are evaluated using a box model representing the Los Angeles Region during CalNex. The model SOA formed only from the oxidation of VOCs (V-SOA) is insufficient to explain the observed SOA concentrations, even when using SOA parameterizations with multi-generation oxidation that produce much higher yields than have been observed in chamber experiments, or when increasing yields to their upper limit estimates accounting for recently reported losses of vapors to chamber walls. Including SOA from primary semi-volatile and intermediate volatility organic compounds (P-S/IVOCs) following the parameterizations of Robinson et al. (2007), Grieshop et al. (2009), or Pye and Seinfeld (2010) improves model/measurement agreement for mass concentration at shorter photochemical ages (0.5 days). Our results strongly suggest that other precursors besides VOCs are needed to explain the observed SOA concentrations. In contrast, all of the literature P-S/IVOC parameterizations over-predict urban SOA formation at long photochemical ages (3 days) compared to observations from multiple sites, which can lead to problems in regional and global modeling. Sensitivity studies that reduce the IVOC emissions by one-half in the model improve SOA predictions at these long ages. In addition, when IVOC emissions in the Robinson et al. parameterization are constrained using recently reported measurements of these species model/measurement agreement is achieved. The amounts of SOA mass from diesel vehicles, gasoline vehicles, and cooking emissions are estimated to be 16 - 27%, 35 - 61%, and 19 - 35%, respectively, depending on the parameterization used, which is consistent with the observed fossil fraction of urban SOA, 71(±3)%. The percentage of SOA from diesel vehicle emissions is the same, within the estimated uncertainty, as reported in previous work that analyzed the weekly cycles in SOA concentrations (Bahreini et al., 2012; Hayes et al., 2013). However, the modeling work presented here suggests a strong anthropogenic source of modern carbon in urban SOA, possibly cooking emissions, that was not accounted for in those previous studies, and which is higher on weekends.