Nonlinear Impact Analysis of Urban Road Traffic Carbon Emissions Based on the Integration of Gasoline and Electric Vehicles

With the rapid proliferation of electric vehicles (EVs) in China, the landscape of transportation carbon emissions has undergone significant changes. However, research on the impact of the built environment on the carbon emissions of mixed traffic from gasoline and electric vehicles remains sparse. This paper focuses on urban traffic scenarios with a mix of gasoline and electric vehicles, analyzing the spatiotemporal distribution of carbon emissions from both types of vehicles and their nonlinear association with the built environment. Utilizing trajectory data from gasoline-powered and electric taxis in Chengdu, China, we establish segment-level carbon emission estimation models based on the vehicle-specific power of gasoline vehicles and the equivalent energy consumption of electric vehicles. Subsequently, we employ the XGBoost algorithm and SHapley Additive ExPlanation (SHAP) to analyze the nonlinear relationships between 13 built environment variables and vehicle carbon emissions. This paper reveals that most built environment variables exhibit nonlinear relationships with traffic carbon emissions, with five factors—population density, road density, residential density, metro accessibility, and the number of parking lots—having a significant impact on road carbon emissions. Finally, we discuss the carbon reduction benefits of EV adoption and propose policy recommendations for low-carbon initiatives in the transportation field.