Reduced-form air quality dispersion modeling for urban scale traffic-related pollutants

Air quality dispersion models are widely used to assess impacts of emission sources; however, simulating emissions from vehicles in urban are as is both data- and computationally intensive due to the thousands of road segments in urban networks and the spatial resolution needed to characterize traffic-related air pollutants. To address these limitations, a reduced-form air dispersion model (RFM) was developed to efficiently simulate dispersion of traffic emissions while maintaining the accuracy and performance of traditional modeling approaches. The RFM assumes that the spatial relationship, defined as a source-to-receptor transfer matrix derived from a dispersion model, is spatially invariant at the urban scale and thus may be repositioned and used repeatedly to model the entire road network. This approach is developed using traditional dispersion models (RLINE and CALPUFF), the large road network in Detroit, Michigan (9701 links), and a dense receptor network on 50 m centers. Optimized spatial resolution settings specific to Detroit were established, and meteorological homogeneity across Detroit was confirmed prior to applying the RFM. Simulations conducted for a subset of the area verified that the performance of the RFM was comparable to that of the original models. The approach was then applied to the full road network to produce high resolution annual and seasonal maps of PM_2.5 concentrations from on-road sources. Comparisons to previous studies and ambient monitoring confirm the method's validity. RFMs effectively address the computational burden that normally renders modeling at the urban scale infeasible, and they provide a tool and data to support air quality assessment and management for this important source type.