| dc.description.abstract | In this study, the community multiscale air quality (CMAQ), with modifications to track precursor-specific SOA, was applied to model SOA and organic tracer formation from aromatic compounds, isoprene, monoterpenes, and sesquiterpenes. The model predicted aromatic and monoterpene SOA showed strong correlations with the measured daily corresponding organic tracers, which indicates that the tracer-method is a good approach to evaluate model predictions in precursor-specific SOA. However, the tracer-to-SOA ratios (fSOA (italicized)) derived from the modeling results show large variation based on different SOA components considered, and the fSOA (italicized) values showed significant difference from those determined in chamber experiments due to the difference between chamber conditions and ambient atmosphere. The fSOA (italicized) in the ambient air can be assessed by the modified CMAQ model with abilities to simulate organic tracers and SOA simultaneously.
The modeled aromatic SOA tracer, 2,3-dihydroxy-4-oxopentanoic acid (DHOPA), agree well with the field measurements (MFB = 0.15; R = 0.8), and approximately two-thirds of it is from the oxidation of toluene. The modeled fSOA (italicized) shows a strong dependence on the OA loading when only semivolatile aromatic SOA components are included, while this dependence becomes weaker when non-volatile oligomers and dicarbonyl SOA products are considered. To predict total aromatic SOA, a constant fSOA (italicized) of 0.002 is determined, and the common-used chamber-determined fSOA (italicized) value of 0.004 could lead to an underestimation of SOA by a factor of 2.
The isoprene-SOA scheme in the CMAQ model is expanded to simulate the unique isoprene tracers 2-methyltetrols (2-MT) and 2-methylglyceric acid (2-MG) by treating them as semivolatile species and including a non-heterogeneous formation pathway. The modeled fSOA (italicized) of sum of 2-MT and 2-MG in the total isoprene-SOA varies gently, between 0.01-0.02 in polluted regions, suggesting that the chamber-determined fSOA (italicized) of 0.063 may lead to large underestimations of overall isoprene SOA.
The monoterpene (MT) and sesquiterpene (SQT) SOA was simulated by the CMAQ model with five explicit and one lumped MT species and SQT, and the contribution from each oxidation pathway was tracked in the MT SOA formation. Three MT tracers (pinic acid, PA; pinonic acid, PNA; and 3-methyl-1,2,3-butanetricarboxylic acid, MBTCA) and one SQT tracer (β-caryophyllinic acid, BCARYA) were modeled to assess the fSOA(italicized) values to estimate MT and SQT SOA. The fSOA (italicized) shows significant OA dependence, suggesting that using a constant fSOA (italicized) could lead to large errors in estimating terpene SOA. Instead, power-law equations directly link the tracer concentrations to the corresponding SOA concentrations were proposed and lead to good SOA estimations. | |
