Abstract |
To evaluate the role of proximity to mobile source emissions on ambient air surrounding residences, statistical analyses using linear regression models were conducted for selected volatile organic compounds, carbonyls, PM2.5 mass, elemental carbon and organic carbon with mobile emission sources. The log transformed ambient air concentration of individual air toxics measured in Elizabeth, NJ during the Relationship of Indoor, Outdoor and Personal Air (RIOPA) study was used as the dependent variable and inverse distance to roadways, gas stations, and point sources and meteorological parameters as the independent variables in the regression models. The home, roadway, point and area sources in and around Elizabeth, NJ were geo-coded using Geographic Information System (GIS) techniques to determine the distance between the homes and potential ambient sources. Meteorological data (wind speed, wind direction, temperature, and atmospheric pressure) were obtained from the NOAA, Weather-Bureau-Army-Navy (WBAN) station in the Newark Liberty International Airport, which is immediately to the north of Elizabeth, and mixing height data from Brookhaven, NY (the closest station to Elizabeth containing that type of data). The meteorological data were averaged over the 48 hour sampling period to provide a single value for each sample. The roads were stratified into six roadway types based on categories used in the EPA Mobile 6 model. Quality assurance steps were taken to confirm the location and each home and location, including direct visits to Elizabeth to verify the address and coordinates. Various regression models (and selection criteria) were used to confirm that repeatable set of associations were obtained. |