Spatial-temporal dynamics and driving forces of urban vegetation carbon storage and carbon sink in Shanghai megacity

Abstract

Urban vegetation has important ecological functions such as purifying the air, alleviating the heat island effect and improving the quality of the living environment. Accurately assessing the carbon sink (CSK) of urban vegetation, understanding the spatial and temporal variations in carbon storage (CS) and CSK along the urban–rural gradient, and identifying their driving factors are essential for addressing climate change in cities. This study employed machine learning and geospatial statistical analysis to evaluate CS and CSK in Shanghai, a major metropolitan area in China. The results showed that: (1) From 2015 to 2020, the total CS in Shanghai increased by 1.10 Mt (41.98%), while carbon density (CD) rose by 41.54%. A pronounced gradient was observed in CS and CSK, with 75% of CS concentrated in the suburban area (Ring5). (2) Although suburban regions exhibited a larger total CSK than urban areas, it showed high fluctuation, including negative values and cold spots. In contrast, the central urban area (Ring1) demonstrated stable CSK with no negative values. (3) During the study period, the spatial distribution of CS became more homogeneous, with a reduction in the extent of high-value and low-value regions decreased. The area of the low-value region decreased by more than 50%. CSK cold spots were mainly distributed in the south, while hot spots clustered in the north, with higher index values in cold spot regions. (4) In central area (Ring1), soil conditions were the dominant factor affecting both CS and CSK. Meanwhile, the transition zone (Ring2–Ring4) was influenced by an interplay of natural environmental, socioeconomic, and locational conditions, whereas the suburb (Ring5) was predominantly controlled by climatic factors. All driving factors exhibited interactive enhancement effects, suggesting that improving regional carbon sink capacity requires integrated measures. This study demonstrates that remote sensing inversion and spatial analysis offer effective technical support for dynamic urban carbon assessment and can inform the development of differentiated carbon management policies.