Modeling and Evaluating Short-Term On-Road PM2.5 E mission Factor Using Different Traffic Data Sources

摘要

Estimating short-term PM2.5 emission factor from the traffic activity is a key step in studying the advert health impacts of the on-road PM2.5 pollutants. When Motor Vehicle Emission Simulator (MOVES) is used for this application, the emission factor of an analysis interval is estimated based on traffic parameters observed in that interval. Since the duration of the interval is short (e.g., usually less than 10 minutes), the emission effects of traffic activity in one interval can last to later intervals. However, such historical influences are not considered in the MOVES model. To overcome this challenge facing off the short-term PM2.5 related analysis with use of MOVES, this study develops a modeling methodology to take into account the historical traffic activity with capability of quantitatively generating PM2.5 emission factor that is consistent with the field observation. Seven traffic data sources currently utilized in Ohio, including per vehicle record data sources and aggregated traffic data sources, are applied to test the presented model. After comparing the modeling results and MOVES results with the ground-truth PM2.5 measurements, it shows that the modeling results are at least 40% more accurate than the MOVES results. The accuracy improvement brought by the model can be universally observed when each of the tested traffic datasets is utilized. Unlike current practices in which costly and time-consuming traffic data monitoring campaigns are needed, the proposed methodology makes it possible to accurately estimate short-term PM2.5 emission factor using the existing data sources. Because of its adaptability to existing data sources, the presented work may be very useful for researchers and practitioners trying to develop comprehensive control strategies in urban air quality management as long as a prevailing traffic data source is available.