Global warming caused by greenhouse gas emissions poses a significant challenge to the sustainable development of ecosystems and human society. It is crucial to conduct a comprehensive analysis of spatiotemporal dynamics and the underlying factors influencing CO₂ emissions at a finer spatial scale to advance strategies for mitigating CO₂ emissions. This study integrates energy consumption data, population grid data, and nighttime light data from 2000 to 2020 to construct a comprehensive evaluation system for estimating CO₂ emissions of prefecture-level cities (hereinafter referred to as cities) in China. On this basis, we introduced the Exploratory Spatial-Temporal Data Analysis (ESTDA) method to systematically reveal the spatiotemporal patterns of per capita CO₂ emissions in Chinese cities. Finally, an improved STIRPAT model is employed to analyze the influencing factors of per capita CO₂ emissions. A panel regression model is adopted to examine the relationship between per capita CO₂ emissions and population, urbanization, industrial structure, fixed investment assets, and total import and export volume with the panel data of 284 prefecture-level cities in China spanning from 2005 to 2020. The results indicate that:
1. There are huge regional differences in per capita CO₂ emissions among Chinese cities. Notably, northern cities generally exhibit higher per capita CO₂ emissions compared to southern cities. Moreover, certain provincial capital cities and independent plan cities display higher per capita CO₂ emissions than their surrounding cities.
2. From 2000 to 2020, the spatiotemporal dynamics of per capita CO₂ emissions in various cities demonstrated overall stability with localized variations. This stability is evidenced by the 85% spatiotemporal cohesion rate of per capita CO₂ emissions from 2000 to 2020, indicating a dominant status of no correlation pattern shift. Local dynamics are reflected in the fact that the spatial correlation structure of per capita CO₂ emissions in resource-based cities and some economically developed regions has changed. From the three subtypes of spatiotemporal transitions, Type 1 (0.095)>Type 3 (0.074)>Type 2 (0.060), indicating that some resource-based cities have embarked on a low-carbon transformation development trend in China.
3. The panel data regression results reveal an inverted U-shaped relationship between economic growth and per capita CO₂ emissions at the prefecture-level city scale. I nitially, per capita CO₂ emissions increase with economic growth, first increasing and then decreasing. Per capita CO₂ emissions are positively correlated with population size, the proportion of the secondary industry’s value added to GDP, the proportion of fixed investment assets to GDP, and the proportion of total import and export value in GDP. Conversely, per capita CO₂ emissions. are negatively correlated with urbanization level and the proportion of the tertiary industry’s value added to GDP.