The impact of dynamic assignment methods and speed variability on regional vehicle emissions inventories

Accurately estimating mobile source emissions is critical because it determines unbiased transportation conformity and evaluations of emission control measures. However, the traffic activity data used for developing emissions estimates are typically provided by traditional travel demand models with inadequate temporal and spatial resolutions. This study systematically examines the variation of vehicle emissions inventories that can result from using activity data with highly resolved temporal and spatial scales.; Two approaches are considered to improve traffic activity data. The first approach is post-processing link speeds based on available travel demand modeling outcomes and roadway characteristics. Five speed post-processing methods are compared and demonstrated to substantially influence regional emissions inventories. The study supports US EPA's recommendation on using post-processed speeds to improve regional mobile emissions modeling. The second approach is using dynamic simulation-assignment techniques to explicitly generate finely resolved traffic data. The variation of mobile emissions inventories is quantified based on a new modeling framework, which incorporates time-dependent traffic outcomes from a state-of-the-art dynamic simulation-assignment model. Two specific issues are addressed within the new framework. One is the consistency issue of interface between traffic activities and emission factors. The analysis indicates that trip-based data can more sensitively reflect emission changes induced by regional scale travel demand variations; while the link-based approach is more sensitive in measuring emissions effect as results of facility-related projects. The other issue is to specify how finely, resolved traffic data improve the entire regional mobile emissions inventories. The comparison between static and dynamic scenarios suggests that traditional aggregated static data tend to result in larger TOG and lower NOx and CO emissions estimates.; This dissertation study advances vehicle emissions estimation in both the state of the art and the state of the practice. The study addresses the key issues pertaining to the improvement of linkage between traffic activity data and regional vehicle emissions. The study also provides valuable information to help planners assess desirable directions of future modeling enhancement in estimating mobile emissions inventories.