Modeling Household VMT from Accessibility

Abstract

To meet goals for climate emissions and vehicle miles traveled (VMT), transportation agencies must understand how household VMT differs predictably across demographic contexts, location, and especially the influence of the transportation and land use system. To aid MnDOT in understanding variation in household VMT, we combine high quality individual travel behavior survey day records with spatially matched residential access to opportunity data calculated for three modes at variable travel time thresholds. Specifically different types of accessibility are characterized as local (bike access to jobs within 20 minutes), transit (walk plus transit access to jobs between 10 and 40 minutes of travel), and regional (auto access to jobs traveling between 20 and 60 minutes). These complementary access variables significantly predicted household VMT, in the full context of other known drivers such as the number of workers, vehicle availability, income, and spatial autocorrelation. Transit and local access impacts are negative, and small, although meaningful VMT reductions could be possible with large interventions. Regional auto access is more influential, and positively predictive of VMT, meaning reductions to regional auto access provide the largest theoretical level to reducing household VMT in the region, absent other changes to household demographics and land use.