Relationship between motor vehicle emissions of hazardous pollutants, roadway proximity, and ambient concentrations in Portland, Oregon

Previous methodologies for modeling hazardous air pollutant emissions for onroad mobile sources are based on using spatial surrogates to allocate county level emissions to grid cells. A disadvantage of this process is that it spreads onroad emissions throughout a grid cell instead of along actual road locations. Recent air quality modeling in Portland, Oregon, using the CALPUFF dispersion model assigned emissions to specific roadway links. The resulting data were used to develop a regression model to approximate the CALPUFF predicted concentrations, determine the impacts of roadway proximity on ambient concentrations of three hazardous air pollutants, benzene, 1,3-butadiene, and diesel PM, and to estimate the zone of influence around roadways. Independent variables in the model included emission rates and traffic volumes for individual roadway links, distance and direction between roadway links and receptors, and distributions of wind speeds and directions. Dependent variables were derived from simulated annual average pollutant concentrations from motor vehicles at modeled receptor locations, predicted using CALPUFF. The regression model had limited capability to predict CALPUFF concentrations with an R-squared value of about 0.6. The model indicated the zone of influence around a roadway as between 200 and 400 m. The results support the thesis that in order to capture localized impacts of hazardous air pollutants in a dispersion model, there is a need to include individual roadway links. (C) 2004 Elsevier Ltd. All rights reserved.