Mapping Source-Specific Air Pollution Exposures Using Positive Matrix Factorization Applied to Multipollutant Mobile Monitoring in Seattle, WA

Mobile monitoring
strategies are increasingly used to provide fine
spatial estimates of multiple air pollutant concentrations. This study
demonstrates a novel approach using positive matrix factorization
(PMF) applied to multipollutant mobile monitoring data to assess source-specific
air pollution exposures and to estimate associated emission factors.
Data were collected from one-year mobile monitoring, with an average
of 26 repeated measures of size-resolved particle number counts (PNC),
PM2.5, BC, NO2, and CO2 at 309 sites
in Seattle from 2019 to 2020. PMF was used to characterize underlying
source-related factors. The sources associated with these six factors
included emissions from aviation, diesel trucks, gasoline/hybrid vehicles,
oil combustion, wood combustion, and accumulation mode aerosols. Fuel-based
emission factors for three transportation-related sources were also
estimated. This study reveals that PNC of ultrafine particles with
size <18, 18–42, and 42–178 nm was dominated by features
associated with aircraft, diesel trucks, and both oil and wood combustion.
Gasoline and hybrid vehicles contributed the most to CO2 and NO2 concentrations. This approach can also be extended
to other metropolitan areas, enhancing the exposure assessment in
epidemiology studies.