Electric vehicles and the built environment: Are the relationships different from conventional vehicles?

A large body of research shows that the built environment significantly shapes household vehicle miles traveled (VMT), yet little is known about how these effects may differ between electric vehicles (EVs) and conventional vehicles. As EV adoption accelerates, understanding these dynamics is increasingly important for sustainable transportation planning. This study uses data from the 2023 Utah Household Travel Survey to compare how built environmental characteristics influence VMT across vehicle types. A two-stage hurdle framework, implemented with Gradient Boosting Decision Trees (GBDT), is applied to address zero-inflated outcomes and capture nonlinear relationships. Results show that built environment attributes play a larger role in EV adoption and usage, whereas socioeconomic factors remain more influential for conventional vehicles. In particular, income strongly affects EV adoption but has little effect on EV usage, while it remains an important determinant of both adoption and usage for conventional vehicles. The analysis also reveals notable nonlinear effects of the built environment on VMT. For example, the marginal influence of land-use entropy on EV adoption plateaus at a threshold of 0.85, and the effect of transit stop density on EV usage declines sharply before stabilizing at approximately 20 stops per square mile. For planning and policy implementation, these findings highlight the importance of aligning EV infrastructure with local land use conditions and expanding adoption incentives to reach broader households. Overall, the study provides empirical evidence to inform integrated land use and transportation strategies that facilitate sustainable transportation systems.