Estimating Exposures to Traffic-Related PM in New York City using a Spatial Regression Model

Abstract

MS7-08 Abstract: Accumulating evidence from recent epidemiologic studies suggests that traffic-related air pollution contributes to the health effects associated with long-term exposure to air pollution. Many of these epidemiologic studies lack adequate methods of evaluating small-scale (intraurban) spatial variations in air pollution. We are developing a regression-mapping model to estimate small-scale ambient diesel exhaust particle (DEP) concentrations in New York City. This model is being developed using a dataset of 100 residential outdoor DEP measurements as the outcome measure and land use and traffic characteristics measured at the same sites as predictors. DEP measurements were obtained as part of the ongoing Columbia Center for Children's Environmental Health Study, in which 100 homes throughout Northern Manhattan and the South Bronx were sampled to assess the health impacts of PM2.5 and DEP. Simultaneous indoor and outdoor samples were collected over 2 48-hour periods over a 2-week period at each home. Samples were collected on 25-mm Teflon filters at 4 L/min and were analyzed for mass using a microbalance and for reflectance using a smoke-stain reflectometer. Using the reflectance results, an absorption coefficient was calculated, which is directly correlated to black carbon content, a surrogate for DEP. Based on a preliminary sample of 66 homes, absorption coefficients (mean ± standard deviation) were 1.38 ± 0.51 and 1.30 ± 0.46 for the outdoor and indoor samples, respectively. Indoor and outdoor measurements were highly correlated (r = 0.79, P < 0.001), indicating a strong impact of outdoor sources on the DEP content of indoor air. Slightly higher absorption coefficients were measured in the outdoor samples than in the indoor samples during warmer months (r = 0.86, P < 0.001), and similar coefficients were measured during colder months (r = 0.74, P < 0.001). Using annual average daily traffic volume estimates obtained from the New York State Department of Transportation, a low but significant correlation (r = 0.32) was found between outdoor absorption coefficients and daily traffic volume estimates. We are also performing a validation of the New York Metropolitan Transportation Council's Best Practice Model (NYMTC BPM), a traffic-forecasting model that is being used to predict traffic volumes, against New York State Department of Transportation official traffic count data. The BPM includes a highway network containing 53,000 links, encompassing most minor arterial and above roadway facilities. This study lends validity to using BPM estimates in traffic studies as well as provides a reliable tool for estimating outdoor exposure to traffic-related pollutants for use in health studies.