Core A : Exposure Core Director : Petros Koutrakis Co - Director : Brent Coull

– Exposure Core (Core A) Current understanding of air pollution effects on human health has been limited by the inability of available analytical methods to assess exposure to ambient multi-pollutant mixtures, which encompass both physico-chemical and spatio-temporal complexity. To address this problem, the Exposure Core will develop and employ new methodologies to assess short-and long-term exposures to individual pollutants, sources and mixtures, permitting us to assess pollution effects on human health outcomes for our New England Cohorts (Projects 2-4) and National Study (Project 5). In addition, these methodologies will be made available to researchers outside the Center. This Core unites our research agenda, bringing all human Projects together to address a common set of exposure assessment-driven hypotheses. Accordingly, we have designed the Projects such that each of the three Cohorts and the National Study will assess the effects of the same six exposure metrics (short-and long-term exposures to individual pollutants, sources and mixtures) on cognitive and vascular health and inflammation, birth weight and growth, cause-specific hospitalization and mortality. Exposures to individual pollutants will be based on direct measurements from local, regional and national networks combined with exposure models. Exposures to sources will be calculated using source apportionment techniques. Exposures to mixtures will be characterized using temporal clustering methodologies to group days with similar pollutant concentration profiles and using spatial clustering methodologies to identify groups of geographical locations, e.g. zip codes or counties, with similar pollutant concentration profiles. Our approach to exposure assessment will employ: 1) the use of air pollution data collected by local and regional monitoring sites; 2) the use of satellite Aerosol Optical Depth (AOD) data to estimate spatial pattern of PM 2.5 levels; 3) the development of clustering algorithms to identify profiles of air pollution mixtures; and 4) the use of state-of-the-art spatiotemporal and long-term spatial exposure models. To assess exposures within the New England Region we will characterize both temporal and spatial patterns of pollution. We will assess temporal patterns of exposures within the Region using data from our Boston Supersite combined with spatial patterns analyzed using spatial networks, satellite data and spatiotemporal statistical models. We will also assess exposures to pollutant sources and mixtures using this information. This investigation will also include populations across the US, for which our exposure assessments will be based on data from national monitoring networks. Time series analysis will be based on nearest-monitor air pollution data. In addition to assessing …