Resumen
Compact development is often recommended to reduce auto-dependency thereby decreasing related energy consumptions and transportation emissions. However, there could be a non-linear relationship between density and transportation emissions because of a possible non-linear association between density and vehicle miles travelled (VMT); low travel speed due to congestion; and the relationship between neighborhood characteristics and vehicle characteristics (e.g., vehicle type and age). In addition, the self-selection issue can exist in the land use-transportation emissions analysis because transportation emissions are often estimated based on travel behavior. Using the 2006 Puget Sound Regional Council (PSRC) Household Activity survey, the follow-up stated preference survey, the Motor Vehicle Emission Simulator (MOVES) data, and the GIS network data, this study investigates the non-linear effects of densities on CO2 equivalent (CO2e) emissions with the consideration of self-selection. Specifically, quadratic forms of population and employment densities, different population density group indicators, and attitudinal factors are employed in the regression models. The results indicate that people living in denser neighborhoods tend to generate fewer CO2e emissions. However, this effect becomes insignificant as population density reaches a certain level.