PAttribution Analysis and Dynamic Prediction of Urban Pollutants from New Energy Vehicle Promotion Based on GNN-Transformer Full Life Cycle Models

With the rapid expansion of the global new energy vehicle (NEV) market, the transportation sector is undergoing a low-carbon transformation. However, the environmental benefits across the entire life cycle of NEVs remain controversial. Although NEVs exhibit near-zero emissions during the usage phase, the high energy consumption in battery production and recycling challenges may undermine their emission reduction benefits. Existing research predominantly focuses on emission reductions during the usage phase, neglecting the dynamic impact of regional energy structures and the spatiotemporal heterogeneity of traffic flows on pollutant emissions. This oversight results in a lack of precise assessment tools for policy formulation. This study proposes a spatiotemporal pollutant prediction model integrating multi-source data and machine learning algorithms. It reveals nonlinear relationships between NEV penetration rates, charging infrastructure density, and pollutant concentrations, while quantitatively analyzing constraints imposed by upstream supply chain factors (e.g., electricity carbon intensity) on emission reductions. Findings indicate that promoting new energy vehicles effectively reduces concentrations of pollutants like CO and NOx, while the clean transition of the power structure during electrification is crucial for mitigating secondary pollutants such as SO ₂ . The paper proposes a differentiated policy strategy: prioritizing public transportation in pollution-sensitive areas during the initial phase, enhancing charging infrastructure in the intermediate phase, and promoting low-carbon private vehicle adoption through carbon trading mechanisms in the later phase. Additionally, it demonstrates the environmental impacts across the entire lifecycle of new energy vehicles and offers region-specific deployment recommendations, providing a decision-making basis for future environmental policies and technological innovations.