Rethinking dynamic interactions between commuting public transit modal share and land use patterns

Public transit share in commuting has declined in the past decade. While extensive studies have investigated transit ridership decline based on cross-city panel data, dynamic and time-dependent interactions between land use patterns and the public transit share in commuting have still not been well studied. Hence, this study uses a dynamic Gaussian Bayesian network (DGBN) framework to examine this evolving relationship. The results reveal that a mixed land use and a balanced employment-residential distribution positively cause the increasing public transit share rate in commuting. Additionally, the historical public transit share in commuting positively influences future land use patterns, particularly fostering the expansion of commercial, residential, and retail areas. The dual results illustrate the dynamic interaction between the public transit share in commuting and land use pattern at the census tract level and point out the possibility of realizing the promotion of long-term growth of transit ridership through land use planning.