Travel Determinants and Multi-scale Transferability of National Activity Patterns to Local Populations.

摘要

The ability to transfer national travel patterns to a local population is of interest when trying to assemble models in a timely manner for national security applications. At the core of this research are questions about the connection between travel behavior and land use, urban form, and accessibility. These questions include determining to what degree persons with similar demographics but different home locations (i.e., different surveys) travel and identify a set of landform and accessibility measures that best distinguish between areas across the entire country.;The 2001 National Household Travel Survey (NHTS) participants are divided into persons, travel, and land use categories using Latent Class Cluster Analysis (LCCA). These are compared to two sets of LCCA clusters constructed for two local travel surveys. In a comparison of the average travel characteristics for the LCCA clusters, the majority of person clusters had t statistics for the survey sets that are larger than the associated person-land use group t statistics. Dividing the person clusters into smaller groups based on where they lived increased the probability that the activity-travel measures were statistically identical.;Then, survey participant and estimated population distributions were examined. For each survey set, three sets of chi-square statistics were estimated to compare the survey distributions for three cluster combinations: person, land use, and both person and land use categories. The distributions for these cluster combinations were compared using survey participant counts and estimated survey populations. For each of these comparisons, the chi-square statistics demonstrated the NHTS as a whole did not have the same distributions as either of the local surveys.;The results of this research show that in fact people with different geographic residence locations do travel differently, even if they share the same sociodemographic characteristics. However, the inclusion of land use characteristics does improve the ability to match average activity-travel behavior characteristics. The land form and accessibility measures available nationally that provided the most descriptive power were a combination of accessibility measures that are defined as a collection of these zones of overall accessibility.