Spatial structure and heterogeneous drivers of city-level CO2 emission in China: Insights from geographically weighted spatiotemporal decomposition method

Abstract

Anthropogenic carbon dioxide (CO 2 ) emissions in China have risen persistently, making effective mitigation policies critical for achieving carbon peak and neutrality. Such policies must be regionally tailored, given the large disparities in emission patterns, socioeconomic drivers, and decarbonization potentials across China. Most studies of differentiated emission mitigation concentrate on strategy design but neglect the rationality and accuracy of spatial management units, serving as the spatial objects for policy implementation. The scientific identification of spatial management units is the key to improving the effectiveness of mitigation measures. This study applies the Geographically Weighted Rotation Empirical Orthogonal Function (GWREOF) to anthropogenic CO 2 emissions across 337 Chinese cities from 1990 to 2020, identifying 20 geographically contiguous spatial modes that characterize the spatial structure of China’s CO 2 emissions. These 20 modes are then used to delineate 20 spatially continuous management units for anthropogenic CO 2 emissions (i.e. carbon emission zones). We then analyze the disparities of carbon emission, carbon balance capacity, economic output and energy intensity among different zones and establish the indicator matrix to divide them into eight types. The results show that mitigation priorities differ by zone types: Economically Developed High-Emission Zones and Low-Carbon Sustaining Medium-Emission Zones should prioritize energy decarbonization and low‑carbon tertiary industries. Resource-Dependent High-Emission and Transition-Pressured Medium-Emission Types require stringent coal controls and ecological restoration. Low-Carbon Developing Medium-Emission Type needs green manufacturing upgrades. Energy Optimization Medium-Emission Type should improve coal structures in heating and industry, and Ecologically Prioritized Low-Emission Type should monetize clean energy and forest sinks through compensation. Combining spatially‑based zoning reflecting emission heterogeneity with factor‑based classification reflecting driver heterogeneity, this research improves targeted and systematic governance of anthropogenic CO 2 emissions in China at the national level.