AAAR 31st Annual Conference
October 8-12, 2012
Hyatt Regency Minneapolis
Minneapolis, Minnesota, USA
Abstract View
Refined Estimates of Ambient PM2.5 Exposure: Validation and Refinement of a Mechanistic Indoor Transport Model
NATASHA HODAS, Qing Yu Meng, Melissa M. Lunden, Barbara Turpin, Rutgers University
Abstract Number: 322 Working Group: Aerosol Exposure
Abstract Use of central-site concentrations to estimate ambient PM$_(2.5) exposure is a source of error in air-pollution epidemiology. To reduce this error, computationally inexpensive methods to model indoor concentrations of ambient PM$_(2.5) are needed, since people spend most of their time indoors. We validated a method to predict indoor concentrations of ambient PM$_(2.5) using measurements from the Relationships of Indoor, Outdoor, and Personal Air (RIOPA) study. A mass balance model was used to calculate indoor concentrations of sulfate and elemental carbon (EC) for homes in California, Texas, and New Jersey using measured outdoor PM$_(2.5) species concentrations and air exchange rates. Depositional losses were estimated using published particle-size-resolved deposition rate measurements and species size distributions. We performed a multiple linear regression analysis (MLR) to determine if human activity variables, which were recorded with questionnaires, explained the variance in the difference between modeled and measured indoor PM$_(2.5) concentrations. We then refined the model to account for the factors selected by the MLR analysis. Generalizations about species-resolved size distributions based on knowledge of particle sources/formation mechanisms provided reasonable estimates of indoor particle losses, regardless of geographic region. Opening windows and using central air conditioning or heating explained 30% of the variance in the difference between measured and modeled indoor concentrations. Both the initial and refined models captured the distribution of measured indoor sulfate, suggesting either could be applied to estimate exposure at the population level. Only the refined model, which accounts for human activities, captured indoor sulfate at the individual-home level. Modeled indoor concentrations of EC from both models were similar, suggesting that these human activities minimally impact the penetration and persistence of EC. Based on these results, we provide guidance regarding measurements and human activity data most critically needed to facilitate the prediction of refined exposure surrogates in large, prospective epidemiological studies.