Exploring the impact of spatial structure on carbon emissions in Chinese urban agglomerations: Insights into polycentric and compact development patterns

Abstract

Urban agglomerations, as key spatial units, play a pivotal role in fostering economic expansion and reducing carbon emissions (CEs). However, the influence of complex spatial structures within urban agglomerations on CEs remains inadequately explored, thus impeding the progression toward sustainable city. By integrating with multi-source data, this study explores the polycentric and compact spatial structure of urban agglomerations. Applying machine learning approaches, including random forest and SHapley Additive exPlanations model, our research analyzes the heterogeneity, nonlinear characteristics, and interaction effects of spatial structure on CEs. The results indicate that the inverted primacy, effective mesh size, and patch cohesion index are key indicators influencing CEs, and exert heterogeneous impacts on CEs of urban agglomerations with different spatial structure. There are marginal effects of spatial structure on CEs, with thresholds for positive and negative influences varying across different urban agglomerations. When the effective mesh size is less than 500 and the patch cohesion index is at 98 and 99, it can effectively inhibit CEs. Spatial structure indicators interact to influence the intensity and direction of CEs. Our study framework provides new insights into optimizing spatial structures and promoting sustainable development in urban agglomerations.