Original Article

Journal of Exposure Science and Environmental Epidemiology (2013) 23, 232–240; doi:10.1038/jes.2012.125; published online 30 January 2013

Intra-urban spatial variability in wintertime street-level concentrations of multiple combustion-related air pollutants: The New York City Community Air Survey (NYCCAS)

Jane E Clougherty1, Iyad Kheirbek2, Holger M Eisl3, Zev Ross4, Grant Pezeshki2, John E Gorczynski3, Sarah Johnson2, Steven Markowitz3, Daniel Kass2 and Thomas Matte2

  1. 1Department of Environmental and Occupational Health, University of Pittsburgh, Graduate School of Public Health, Pittsburgh, Pennsylvania, USA
  2. 2New York City Department of Health and Mental Hygiene, New York, New York, USA
  3. 3Center for the Biology of Natural Systems, Queens College, Flushing, New York, USA
  4. 4ZevRoss Spatial Analysis, Ithaca, New York, USA

Correspondence: Dr. Jane E. Clougherty, Department of Environmental and Occupational Health, University of Pittsburgh, Graduate School of Public Health, Bridgeside Point I 100 Technology Drive, Suite 350, Pittsburgh, PA 15219, USA. Tel.: +1 412 624 7494. Fax: +1 412 624 3040. E-mail: jcloughe@pitt.edu

Received 16 December 2011; Revised 14 October 2012; Accepted 30 October 2012
Advance online publication 30 January 2013

Top

Abstract

Although intra-urban air pollution differs by season, few monitoring networks provide adequate geographic density and year-round coverage to fully characterize seasonal patterns. Here, we report winter intra-urban monitoring and land-use regression (LUR) results from the New York City Community Air Survey (NYCCAS). Two-week integrated samples of fine particles (PM2.5), black carbon (BC), nitrogen oxides (NOx) and sulfur dioxide (SO2) were collected at 155 city-wide street-level locations during winter 2008–2009. Sites were selected using stratified random sampling, randomized across sampling sessions to minimize spatio-temporal confounding. LUR was used to identify GIS-based source indicators associated with higher concentrations. Prediction surfaces were produced using kriging with external drift. Each pollutant varied twofold or more across sites, with higher concentrations near midtown Manhattan. All pollutants were positively correlated, particularly PM2.5 and BC (Spearman’s r=0.84). Density of oil-burning boilers, total and truck traffic density, and temporality explained 84% of PM2.5 variation. Densities of total traffic, truck traffic, oil-burning boilers and industrial space, with temporality, explained 65% of BC variation. Temporality, built space, bus route location, and traffic density described 67% of nitrogen dioxide variation. Residual oil-burning units, nighttime population and temporality explained 77% of SO2 variation. Spatial variation in combustion-related pollutants in New York City was strongly associated with oil-burning and traffic density. Chronic exposure disparities and unique local sources can be identified through year-round saturation monitoring.

Keywords:

land-use regression (LUR); urban air pollution; fine particles (PM2.5); black carbon (BC); nitrogen dioxide (NO2); sulfur dioxide (SO2)