An XGBoost-SHAP analysis of the driving factors of carbon emissions in China’s first-tier cities

To implement China's strategic goal of "tailored and categorized approaches" for carbon reduction and enhance urban resilience under climate change pressures, systematic analysis is urgently needed to develop targeted urban emission reduction pathways. Utilizing panel data from 19 Chinese first-tier cities from 2002 to 2023, this study employs the XGBoost-SHAP interpretable machine learning model to investigate the driving effects of eight factors on carbon emissions:economic development level (PGDP), population size (POP), industrial structure (IS), technological innovation (TI), energy intensity (EI), urban form (D), public transportation (PT), and new digital infrastructure (DI). The K-means clustering algorithm is used to categorize the 19 cities into five types, enabling an in-depth analysis of the heterogeneous characteristics of carbon emission drivers across different city types. The main findings are as follows: (1) Population size (POP), energy intensity (EI), technological innovation (TI), public transportation (PT), and economic development level (PGDP) are significant factors influencing carbon emissions in first-tier cities, while the overall impact of new digital infrastructure (DI) remains ambiguous due to its dual role in increasing energy consumption and enabling energy-saving reforms. (2) The influence of individual drivers on carbon emissions exhibits significant heterogeneity across city types. Energy intensity (EI) has a substantial impact on carbon emissions in all five city types, whereas the effects of population size (POP), technological innovation (TI), and public transportation (PT) vary considerably depending on the city type. Based on the findings, this study proposes policy recommendations focusing on systematic governance of key elements, differentiated emission reduction strategies integrating resilience-building, and the establishment of a collaborative governance system to facilitate urban green transformation and enhance comprehensive resilience.