How the built environment influences driving: Insights from global positioning system data.

摘要

The sprawling low-density car-dependent urban developments in many metropolitan areas in the United States have contributed to severe transportation consequences in the last five decades. These urban developments demand intensive automobile travel which exacerbate the nation's oil dependency and increase greenhouse gas (GHG) emissions which in turn contribute to global warming. While automobile travel patterns have been related to the built environment in current literature, few studies have made the direct connections between the built environment and vehicle fuel consumption and emissions. This dissertation establishes a methodology for understanding the relationships between specific attributes of the built environment, people's driving behavior, and the associated vehicle fuel consumption and emissions.;This dissertation applies a disaggregated analysis scheme, through which an individual driver's travel behavior and travel outcomes are related to the built environment. In addition to the built environment near drivers' home and work places, this dissertation provides detailed examinations on the urban corridors along drivers' commuting routes, an important and yet understudied urban space. A rich global positioning systems (GPS) dataset collected from 73 automobile drivers over 30 days on a second-by-second basis in the Detroit metropolitan region is used to quantify driving behaviors and to estimate fuel consumption and major tailpipe emissions. Multivariate statistical techniques are applied to test the influences of the built environment on driving outcomes, controlling for other factors.;The results of this dissertation demonstrate that built environment features near home and work locations do not have significant associations with total vehicle miles traveled (VMT) and total fuel consumption and emissions on non-work travel. Rather, the influences of built environment along commuting routes on these travel outcomes are statistically significant. Denser and more diverse non-work destination choices are associated with lower levels of driving, less fuel consumption and less air pollution. This research also indicates that denser and more diverse land-use patterns near drivers' homes lead to lower vehicle fuel efficiency with higher emissions per mile.