Spatial variation in inversion-focused vs 24-h integrated samples of [PM.sub.2.5] and black carbon across Pittsburgh, PA

• Published in: Journal of exposure science & environmental epidemiology 2016-07, p.365
• Main Authors: Tunno, Brett J, Michanowicz, Drew R, Shmool, Jessie L.C, Kinnee, Ellen, Cambal, Leah, Tripathy, Sheila, Gillooly, Sara, Roper, Courtney, Chubb, Lauren, Clougherty, Jane E
• Format: Article
• Language: English
• Subjects: Air pollution; Distribution; Environmental health; Health aspects; Particles
• ISSN: 1559-0631
• DOI: 10.1038/jes.2015.14

A growing literature explores intra-urban variation in pollution concentrations. Few studies, however, have examined spatial variation during "peak" hours of the day (e.g., rush hours, inversion conditions), which may have strong bearing for source identification and epidemiological analyses. We aimed to capture "peak" spatial variation across a region of complex terrain, legacy industry, and frequent atmospheric inversions. We hypothesized stronger spatial contrast in concentrations during hours prone to atmospheric inversions and heavy traffic, and designed a 2-year monitoring campaign to capture spatial variation in fine particles ([PM.sub.2.5]) and black carbon (BC). Inversion-focused integrated monitoring (0600-1100 hours) was performed during year 1 (2011-2012) and compared with 1-week 24-h integrated results from year 2 (2012-2013). To allocate sampling sites, we explored spatial distributions in key sources (i.e., traffic, industry) and potential modifiers (i.e., elevation) in geographic information systems (GIS), and allocated 37 sites for spatial and source variability across the metropolitan domain (~388 [km.sup.2]). Land use regression (LUR) models were developed and compared by pollutant, season, and sampling method. As expected, we found stronger spatial contrasts in [PM.sub.2.5] and BC using inversion-focused sampling, suggesting greater differences in peak exposures across urban areas than is captured by most integrated saturation campaigns. Temporal variability, commercial and industrial land use, [PM.sub.2.5] emissions, and elevation were significant predictors, but did not more strongly predict concentrations during peak hours. Journal of Exposure Science and Environmental Epidemiology (2016) 26, 365-376; doi:10.1038/jes.2015.14; published online 29 April 2015 Keywords: 24-h integrated sampling; black carbon (BC); fine particulate matter ([PM.sub.2.5]); inversion-focused sampling; land use regression (LUR)