Effects of Built Environments on Travel Behavior and Emissions: A Reexamination by Addressing Methodological Issues

Abstract

Urban transportation researchers have been studying the relationship between land use policy and travel behavior for several decades due to the topic's great importance in public policy-making. Because of the improvements in energy efficiency, large reductions in emissions have been achieved for a given amount of travel. Unfortunately, the rapid growth in total travel distance over the past several decades, especially for light duty vehicles, has reduced the benefits from technological improvements. Therefore, many urban planners have suggested land use planning as an alternative and fundamental way to reduce auto dependency and thereby, transportation emissions. However, several empirical studies about the linkage between built environments and travel behavior produced mixed results. In light of the inconsistent findings, research efforts to reconcile the discrepancy among different studies are required. Several methodological issues are found based on the previous literature and four main challenges are addressed in this study: self-selection, spatial autocorrelation, trip-interdependency, and geographic scale. In addition, two key methodological issues in modeling transportation emissions are found and addressed. First, transportation emissions per person are often estimated by using vehicle miles traveled (VMT) and emissions factors, but these emissions factors do not fully consider variations in travel speed and vehicle characteristics. Second, VMT and emissions factors are associated with travel characteristics, implying that the same methodological challenges existing in the land use-travel behavior analysis can exist in the land use-transportation emissions analysis. This research obtained several important results. First, increasing residential density can reduce VMT and emissions significantly. In addition, the impact of residential density on VMT is higher than that on transportation emissions, indicating that negative externalities such as congestion generated from compact developments should be considered in the land use-transportation emissions analysis. Second, analyses show that the effects of land use factors on VMT and emissions are different according to tour types and geographic scales. These results imply that different land use policies should be implemented according to neighborhoods characteristics. Finally, the sensitivity analyses of built environment factors show that ignoring trip and vehicle characteristics in the emissions calculation can inflate the influences of built environments on emissions.