AAAR 33rd Annual Conference
October 20 - October 24, 2014
Rosen Shingle Creek
Orlando, Florida, USA
Abstract View
Modeling the Spatial Variability of Traffic Related Pollutants and the Contribution of High Emitting Vehicles in Pittsburgh, PA
YI TAN, Timothy Dallmann, Allen Robinson, Albert A. Presto, Carnegie Mellon University
Abstract Number: 578 Working Group: Urban Aerosols
Abstract The highly time-resolved data from a recent mobile monitoring campaign in Pittsburgh, PA showed substantial spatial variability of traffic related pollutants. High emitting vehicles caused short-duration plume events, and contributed a disproportionately large fraction of the near road exposures of particle-bound polycyclic aromatic hydrocarbons (PAHs) and BC.
In this study, we developed a two-layer statistical model to predict the spatial variability of PAHs and BC. The two-layer model was based on highly time-resolved data from 42 Phase I sites. The plume layer provided a simplified estimation on the near road decay of pollutant plumes, and the base layer modeled the spatial patterns of background PAHs and BC using the traditional land use regression (LUR) method. When the two layers were combined, the model successfully predicted the mean concentrations of PAHs and BC. We validated our model using measurements collected from 36 Phase II sites which covered a broader area than Phase I sites. The two-layer model reasonably predicted concentrations of PAHs and BC and the impact of high emitting vehicles at Phase II sites. Traditional single layer LUR models were also developed using mean concentrations from the 42 Phase I sites. The two-layer model performed better than the traditional LUR model in capturing the spatial variability of PAHs and BC.
The two-layer model found that around 20% residential units in the study domain were considerably impacted by PAH plumes from high emitting vehicles (i.e., plumes contributed more than 40% of the total PAH exposure). High emitting vehicles had smaller impacts on BC exposures, contributing to less than 20% of the total BC exposure at ~90% residential units. BC exposures were more driven by regional emissions such as coke production.