Comparison of 24-hour vs. Inversion-Period Integrated Samples of PM2.5 and Elemental Carbon in Pittsburgh, PA, USA

Abstract

Background: Pittsburgh, Pennsylvania is home to local industrial sources of air pollution, periods of heavy traffic congestion, a complex terrain of river valleys and hills, and frequent atmospheric subsidence inversions. A two-year monitoring campaign was designed to capture intra-urban seasonal variability in fine particulate matter (PM2.5) and elemental carbon (EC) concentrations. Inversion-focused integrated monitoring was performed during year one (2011-2012), vs. 24-hour integrated sampling during year two (2012-2013). Methods: Spatial distributions of local pollution sources (i.e., traffic, industry) and potential modifiers (i.e., elevation) were explored via geographic information systems (GIS). Sampling locations (n=38) were allocated to capture spatial and source variability across a metropolitan domain (~150 sq. miles). Results: In year one, we found temporally-adjusted mean PM2.5 concentrations of 14.1 ± 3.6 µg/m3 for summer and 12.5 ± 2.4 µg/m3 for winter. In year two, we found slightly lower mean PM2.5 concentrations of 13.8 ± 2.8 µg/m3 for summer and 11.5 ± 3.7 µg/m3 for winter. Mean seasonal EC ranged from 1.3 ± 0.51 abs during winter to 1.80 ± 0.46 abs during summer for year one. Temporal variability, industrial land use, and elevation were significant predictors in land use regression (LUR) modeling, resulting in final pollutant-specific seasonal R2 from 0.58 for PM2.5 to 0.76 for EC during year one, and 0.71 for EC to 0.86 for PM2.5 during year two. Seasonal-specific factor analysis for trace metals will be performed for the four seasons, for source apportionment analyses. Conclusions: The paired 24-hour and inversion-focused stationary monitoring campaigns were used to explore spatial variability and seasonal contrasts in PM2.5 and EC concentrations under inversion conditions, improving source apportionment analyses. Industry and elevation were the stronger identified predictors of spatial variability for both PM2.5 and EC.