Aerosol Chemistry and Physics - Aerosols: Measurement, Dosimetry, and Health Effects - page 134

species also contribute to airborne PM. Secondary organic aerosol (SOA) contributions typically
follow seasonal patterns of ozone concentrations. SOA forms when ozone or OH reacts with hydro-
carbon gases from combustion and emissions from vegetation, leading to high-molecular-weight,
oxygenated compounds that condense on available particles.
Throughout the United States source apportionment is achieved by applying computer models
to ambient PM 2.5 and meteorological data collected at receptor sites in order to quantify the
contributions of the dominant sources. The studies published by Srivastava et al. (2007) and Lane
et al. (2007) provide useful introductions to source apportionment by showing how the predictions
of two very different approaches compare when applied to the same data sets. Sarnat et al. (2008)
found that associations of cardiovascular and respiratory morbidity in Atlanta, Georgia with sources
of PM 2.5 agreed for three different approaches to source apportionment. Using 2 years of speciated
PM data and data from emergency room visits they found that PM 2.5 from mobile and biomass
sources was associated with both cardiovascular and respiratory visits, but secondary particles were
associated only with respiratory visits.
Computer models based on multivariate factor analysis such as positive matrix factorization
(PMF) search for patterns in concentrations measured at receptor sites before associating them with
known PM sources (often guided by emission inventories such as cited in Lane et al., 2007). An
example of this approach applied to Seattle, Washington, showed that wood smoke was the heaviest
contributor to PM 2.5 in 2003-2004 (Wu et al., 2007).
The chemical mass balance (CMB) approach hunts for the vestiges of known molecular markers
of individual sources in ambient PM and then reconstructs contributions of each source at receptor
sites. CMB models are useful because a large fraction of ambient ine carbonaceous mass has not
been traced to individual compounds. For the carbonaceous constituents of PM 2.5 , CMB models in
the United States start with detailed emission factors for molecular markers of individual sources
based on measurements made in California by Rogge et al. (1991, 1997a,b) and Schauer et al.
(1999a,b, 2001, 2002b,c). Soon after these detailed source characterizations became available,
Schauer et al. (2002a) constructed and applied a CMB model by Zheng et al. (2002) to a smog
episode in southern California to derive the contributions of 11 unique sources from measured
concentrations of volatile, semi-volatile, and particulate organic pollutants. They found that vehicle
exhaust and SOAs were the most important sources of particles, among others in that episode. CMB
models continue to evolve as more molecular markers become available for source apportionment
in other regions of the United States with different mixes of local pollutants; for example, new
models address agricultural burning in eastern Washington (Jimenez et al., 2007), soil-derived
markers of crops grown in the Central Valley, California (Rogge et al., 2007), variability in wood
smoke emissions and other emissions from biomass (Robinson et al., 2006a,b), and pollutants from
cooking that often contribute to ambient PM 2.5 (Robinson et al., 2006c). Updating CMB models
with new markers of SOAs derived from various gas-phase precursors (Kleindienst et al., 2007)
has shown that SOA in southern California differs from SOA in the midwest and eastern regions of
the United States (Stone et al., 2009). Researchers have also recognized that CMB models should
incorporate oxidation of molecular markers between emission and measurement at receptor sites
(Robinson et al., 2006a; Roy et al., 2011).
6.1.4 d iFFerences between i ndoor and o utdoor e nvironMents
From a physical chemist's perspective, a building is a leaky reaction vessel with active sur-
faces. Table 6.2 compares some key physical parameters outdoors and in a typical home in the
United States.
Whereas most outdoor atmospheric processes are not constrained by partitions or macroscopic
surfaces, the indoor environment is deined by walls and covered with building materials and
furnishings. Many of these substances can contribute to the air quality by adsorbing or emitting
compounds that are chemically active. Indoor activities also generate gases and particles, leading