Urban Travel Demand Modeling

Current emissions factor models (such as MOBILE5a) are unable to directly account for the mobility impacts of TCMs; they require estimates of mobility changes to be estimated by regional travel demand models as a pre-requisite to estimating the effects of TCMs on mobile-source emissions. Regional travel demand models are able to evaluate the impact of some TCMs on travel mode through changes in relevant travel times/costs of alternative modes. However, they are unable to determine the potential impact on time-of-day changes in travel due to peak-period TCMs (light rail/commuter rail service, HOV lanes, This is because time-of-day split is not explicitly modeled and is instead obtained by applying fixed time-of-day factors at the end of the traffic assignment step (Bhat, 1998 has recently shown that peak period HOV provision or peak period transit service improvements does substantially impact time-of-day choice, particularly for non-work trips which comprise as much as one-half to two-thirds of peak period travel in metropolitan areas). Existing regional demand models are also unable to accommodate a) changes in number of trips due to the TCMs because the trip generation step of regional demand model is insensitive to level-of-service, b) changes in trip destination due to TCMs that affect non-auto mode level-of-service (because auto travel time is the only impedance measure in the trip distribution step), c) changes to non-motorized modes of travel (since non-motorized modes are not included in mode choice analysis), and d) changes in vehicle speeds due to traffic system management strategies. The end result is that the existing regional model may not provide reliable effects of TCMs on link volumes by time-of-day (which, in turn, determines VMT impacts and average speeds by time-of-day, two mobility-associated inputs to emissions modeling). This project aims to make several refinements to the urban travel demand modeling process so it is more sensitive to TCMs.