Modeling of in-vehicle human exposure to ambient fine particulate matter

A method for estimating in-vehicle PM2.5 exposure as part of a scenario-based population simulation model is developed and assessed. In existing models, such as the Stochastic Exposure and Dose Simulation model for Particulate Matter (SHEDS-PM), in-vehicle exposure is estimated using linear regression based on area-wide ambient PM2.5 concentration. An alternative modeling approach is explored based on estimation of near-road PM2.5 concentration and an in-vehicle mass balance model. Near-road PM2.5 concentration is estimated using a dispersion model and fixed site monitor (FSMs) data. In-vehicle concentration is estimated based on air exchange rate and filter efficiency. In-vehicle concentration varies with road type, traffic flow, windspeed, stability class, and ventilation. Average in-vehicle exposure is estimated to contribute 10–20 percent of average daily exposure. The contribution of in-vehicle exposure to total daily exposure can be higher for some individuals. Recommendations are made for updating exposure models and implementation of the alternative approach.