Could improving public transport accessibility reduce road traffic carbon dioxide emissions? A simulation-based counterfactual analysis

Improving public transport accessibility (PTA) has been considered as an effective measure for promoting sustainable urban development. Based on the grid-level data in Nanjing, China, this paper explores the spatially heterogeneous effects of PTA on road traffic CO2 emissions using a geographically weighted random forest (GWRF) model. A simulation-based counterfactual analysis framework is further proposed to predict the intervention effects of improving PTA. Two kinds of practical interventions, adding facilities and increasing service frequency, are considered in our counterfactual prediction. The effects of improving PTA across different areas are predicted and compared. Our results indicate that the GWRF model with a properly tuned bandwidth outperforms the conventional random forest model. The results of counterfactual analysis show that improving PTA could achieve greater environmental benefits in suburb areas. With the process of urbanization in Nanjing, the population and economy has grown rapidly in suburb areas. Therefore, it is reasonable to improve public transport services in these areas. Based on our findings, PTA has potential to make a significant contribution to sustainable development of urban transportation. Moreover, our findings with respect to heterogeneous effects likely improve the efficiency of local transport policies that target such areas, helping achieve greater environmental benefits.