o PROlS URBAN SCALE VARIABILITY OF PM2 5 COMPONENTS Carvin D. Stevens, Ron Williams, Alan Vette and Paul Jones U.S. EPA, Research Triangle Park, NC Disclaimer: "Although this work was reviewed by EPA and approved for publication, it may not necessarily reflect official Agency policy." BACKGROUND Introduction An objective of the Detroit Exposure and Aerosol Research Study (DEARS) is to determine the associations between concentrations measured at central site monitors and outdoor residential, indoor residential and personal exposures for selected air toxics, PM constituents, and PM from specific sources. PURPOSE The focus of this poster is to demonstrate that a central site monitor adequately represents ambient concentrations of PM2 5 mass and its secondary constituents over a county-level scale in an urban area (Fig.l). Primary aerosol components, however, are much more spatially variable due to local source contributions in the Detroit urban air shed and more spatially resolved measurements are necessary. MATERIALS AND METHODS Materials and methods Residential outdoor and central site monitoring are being perfonned during the DEARS to assess the variability in composition and concentrations of fine particulate matter across the urban air shed. Participants are monitored for 5 days each in the summer and winter. Their residences are located in seven different environmental monitoring areas (EMA) across a 1400km2 area. The sources potentially impacting each EMA vary from industrial and mobile source categories including coal combustion, coke production, iron and steel production and oil refineries. Sample collections were performed with personal environmental monitors (PEM) using Teflon and quartz filters. Gravimetric analysis were used to detennine the PM2 5 mass. The elemental concentrations (Si, Mn, S, Ni, Cu, Zn, As, Pb, Se, Sr, Cr, Ca, & Fe) were determined using XRF analysis. Concentrations of OC and EC were measured using the thermal-optical reflectance method on the quartz filters. Fig.l, Environmental Monitoring Areas (EMAs) DEARS Study Sites (Emphasis on Proximity to Sources) Exposure Measurement Areas (EMA) 1-Industrial 2-Industrial 3-Diesel 4-T raffic/ Industrial 5-Industrial 6-Highway 7-Regional Table. 1. Species Formation and Sources. Species P/S Sources Comments Sulfate s Fossil fuel combustion Primary component ofPM25 (eastern U.S) Nitrate s Fossil fuel combustion Primary component ofPM25 (western U.S.) Metals p Smelters, soil, incinerators Manganese, nickel, copper, zinc, etc Crustal p Soil, Coal fired boiler Silicon, calcium, iron, titanium Elemental Carbon p Fossil fuel combustion Inert and small size long transport Organic Carbon S/P Fossil fuel combustion A primary component of PM25 P = primary, local sources S = secondary, regional sources fig.2. PM2.5 Composition-Summer ,rtni, ~ EMA1 HEM A3 ~ EMA4 ~ EMA6 ~ EMA7 ¦ Allen Park fig.3. PM2.5 Compositions-Winter ACKNOWLEDGEMENTS I would like to acknowledge Anne Rea, Carry Croghan and RTI International for their contribution in the collection and analyses of the data. RESULTS Preliminary data analysis indicated that the concentrations of PM2 5 mass were fairly consistent across the air shed, varying no more than 10% across an area of 1400km2. However, the composition of PM2 5 in each EMA was highly variable for the primary components (Table I) when compared to the central site at Allen Park. The differences were as great as 36% for the metals. The PM2 5 mass composition was determined to be seasonally-dependent (Figs. 2 & 3). The particle bound nitrate was approximately 45% of the total ambient PM2 5 mass concentration during the winter, and only 7% during the summer. The percentages of the PM2 5 composed of the secondary aerosol components (sulfates and nitrates) were highly correlated and statistically significant across the urban air shed for the summer and winter seasons (Table 2). CONCLUSIONS These data suggest that a central site monitor may adequately represent the spatial distribution of secondaiy components (Nitrates & Sulfates) of PM2 5, but not adequately represent the primary components (primary OC, EC, metals, crustal, etc.) contributed by local sources. Additional source monitoring will be needed with the inclusion of survey, activity, source apportionment and meteorological results to provide better estimates for modeling spatial distributions and exposures to these pollutants across the air shed. Table.2. Pearson correlation coefficients between PM2.5 components measured at the various EMAs and Allen Park during Summer 2004 Species EMA 1 3 4 6 7 OC 0.558 0.476 0.537 0.652 0.718 N03 0.897 0.856 0.893 0.832 0.958 so4 0.968 0.978 0.975 0.983 0.983 Table.3. Pearson correlation coefficients between PM2.5 components measured at the various EMAs and Alien Park during Winter 2005 Species EMA 1 3 4 6 7 OC 0.677 0.755 0.500 0.754 0.571 N03 0.885 0.997 0.785 0.940 0.971 so4 0.946 0.975 0.989 0.988 0.984 ------- |