Objective Urban ecosystem services play an irreplaceable role in maintaining life support and environmental dynamic balance. Urban green spaces, as an important component of urban ecosystems, are the only element that directly increases carbon sinks and indirectly reduces carbon emissions in cities. While beautifying the environment, they play an extremely important role in ecological benefits and carbon sequestration. Quantifying the carbon sequestration capacity of urban green spaces is of great significance for achieving carbon peaking and carbon neutrality goals and addressing climate change.

Methods Based on Carnegie-Ames-Stanford approach (CASA) model, Landsat 8 TM remote sensing data and meteorological data from 2014, 2017, and 2020 are used to estimate the net primary productivity (NPP) and carbon sequestration capacity of parks in the central urban area of Beijing, and explore the spatial and temporal distribution pattern and change characteristics of carbon sequestration capacity of parks in the central urban area of Beijing. Pearson correlation analysis is used to study the driving relationship between the carbon sequestration capacity of park green space, vegetation index, climate environment and landscape pattern. Additionally, the biomass allometry equation method is used to measure the actual carbon sequestration of 33 typical plant communities in Beijing, in order to verify the accuracy of remote sensing model estimation.

Results The results show that the area of park green space in the central urban area of Beijing increased year by year from 2014 to 2020, with a 10.89% increase in the area of park green space in 2020 compared to 2014. Varying significantly in spatial distribution, park green spaces in the central urban area of Beijing are mostly distributed in Haidian District and Chaoyang District. Due to regional limitations, the area of park green spaces in core areas such as Dongcheng District and Xicheng District is relatively small. The total carbon sequestration capacity of park green spaces in the central urban area of Beijing in 2014, 2017 and 2020 was 5.23×10⁴ tC, 7.09×10⁴ tC, and 7.45×10⁴ tC, respectively. Accordingly, the annual average carbon sequestration capacity per unit area was 2.46 tC/hm², 3.19 tC/hm², and 3.29 tC/hm², respectively, showing an increasing trend year by year. 82.15% of park green spaces have significantly improved their carbon sequestration capacity. There are significant differences in the spatial distribution of carbon sequestration capacity in park green spaces, showing a pattern of “high at the edge area and low in the center area overall”. Haidian District, Chaoyang District and Shijingshan District have higher carbon sequestration capacity. In 2014, park green spaces were mainly characterized by medium to high carbon sequestration capacity, while in 2017 and 2020, they shifted to medium to high carbon sequestration capacity and high carbon sequestration capacity, with large green patches playing an important role in achieving high carbon sequestration capacity. The carbon sequestration capacity of park green spaces is significantly positively correlated (p<0.01) with normalized difference vegetation index (NDVI), leaf area index (LAI), and rainfall, and significantly negatively correlated (p<0.05) with temperature, while having little correlation with wind speed. The carbon sequestration capacity of park green spaces is significantly positively correlated (p<0.01) with patch area (PA), largest patch index (LPI), percentage of like adjacency (PLADJ), patch cohesion index (COHESION), and aggregation index (AI), and significantly negatively correlated (p<0.01) with landscape shape index (LSI), mean perimeter-area ratio (MPAR), and mean patch fractal dimension (MPFD). To verify accuracy of the remote sensing model adopted, the measured and simulated values of carbon sequestration capacity were compared. Specifically, the relative error between the measured and simulated values is 5.64%, the mean square error is less than 0.1, and the fitting accuracy is 0.7314. The estimated result is good, basically meeting the accuracy requirements.

Conclusion The carbon sequestration capacity of urban park green spaces in Beijing significantly improved from 2014 to 2020. The higher the carbon sequestration capacity of green spaces, the higher their vegetation coverage and LAI, and the better their cooling effect. The higher PA, the higher AI, and the simpler the patch shape of park green space patches, the stronger the carbon sequestration capacity to a certain extent. This research may provide ideas for scientific assessment and improvement of the carbon sequestration capacity of urban green space, formulation of climate change response measures, and lay a foundation for assessment of the carbon balance of urban green space ecosystem. The research results can directly support the whole process of planning, construction and management of urban green space, and provide a scientific basis for the optimization and regulation of the layout of urban park green space.