| Predicting Chronic Fine and Coarse Particulate Exposures Using Spatiotemporal Models for the Northeastern and Midwestern United States Jeff D. Yanosky,1 Christopher J. Paciorek,2 and Helen H. Suh1 1Exposure, Epidemiology and Risk Program, Department of Environmental Health, and 2Department of Biostatistics, Harvard School of Public Health, Boston, Massachusetts, USA Abstract
Background: Chronic epidemiologic studies of particulate matter (PM) are limited by the lack of monitoring data, relying instead on citywide ambient concentrations to estimate exposures. This method ignores within-city spatial gradients and restricts studies to areas with nearby monitoring data. This lack of data is particularly restrictive for fine particles (PM with aerodynamic diameter < 2.5 µm ; PM2.5) and coarse particles (PM with aerodynamic diameter 2.5–10 µm ; PM10–2.5) , for which monitoring is limited before 1999. To address these limitations, we developed spatiotemporal models to predict monthly outdoor PM2.5 and PM10–2.5 concentrations for the northeastern and midwestern United States. Methods: For PM2.5, we developed models for two periods: 1988–1998 and 1999–2002. Both models included smooth spatial and regression terms of geographic information system–based and meteorologic predictors. To compensate for sparse monitoring data, the pre-1999 model also included predicted PM10 (PM with aerodynamic diameter < 10 µm) and extinction coefficients (km–1) . PM10–2.5 levels were estimated as the difference in monthly predicted PM10 and PM2.5, with predicted PM10 from our previously developed PM10 model. Results: Predictive performance for PM2.5 was strong (cross-validation R2 = 0.77 and 0.69 for post-1999 and pre-1999 PM2.5 models, respectively) with high precision (2.2 and 2.7 µg/m3, respectively) . Models performed well irrespective of population density and season. Predictive performance for PM10–2.5 was weaker (cross-validation R2 = 0.39) with lower precision (5.5 µg/m3) . PM10–2.5 levels exhibited greater local spatial variability than PM10 or PM2.5, suggesting that PM2.5 measurements at ambient monitoring sites are more representative for surrounding populations than for PM10 and especially PM10–2.5. Conclusions: We provide semiempirical models to predict spatially and temporally resolved long-term average outdoor concentrations of PM2.5 and PM10–2.5 for estimating exposures of populations living in the northeastern and midwestern United States. Key words: air pollution, extinction coefficient, fine particulate matter, geographic information system, generalized additive mixed models, geostatistics, spatial smoothing, spatiotemporal modeling, visual range. Environ Health Perspect 117:522–529 (2009) . doi:10.1289/ehp.11692 available via http://dx.doi.org/ [Online 19 November 2008]
Address correspondence to J.D. Yanosky, Harvard School of Public Health, 401 Park Dr., Boston, MA 02215 USA. Telephone: (617) 384-8836. Fax: (617) 384-8859. E-mail: jyanosky@hsph.harvard.edu Supplemental Material is available online at http://www.ehponline.org/members/2008/11692/suppl.pdf This research has been funded wholly or in part by the U.S. Environmental Protection Agency (EPA) through grant 83054501-0 to Harvard University. This research has not been subjected to the U.S. EPA’s required peer and policy review and therefore does not necessarily reflect the views of the agency, and no official endorsement should be inferred. The authors declare they have no competing financial interests. Received 12 May 2008 ; accepted 19 November 2008.
The full version of this article is available for free in HTML or PDF formats. | 
