Urban forests play a critical role in mitigating climate change through carbon sequestration, yet their spatial variability at fine scales remains insufficiently understood. Most existing studies focus on city- or regional-level analyses, largely due to limitations in high-resolution data availability. Using 2014 as a representative year with consistent and comprehensive LiDAR and multi-source datasets, this study investigates neighborhood-scale urban forest carbon storage in New York City (NYC). A neighborhood-level analytical framework was developed by integrating LiDAR-derived structural data, remote sensing information, and literature-based variables, and applying Pearson correlation analysis and boosted regression tree (BRT) modeling to identify key drivers. The results reveal that carbon storage vary widely across NYC space. The area of Staten Island has the most significant carbon storage, where over 40% of the neighborhoods are more than 20, 000 T C of carbon. Also, on the other end of the spectrum is Manhattan with over 90% of neighborhoods containing less than 1000 T C of carbon. Further: The total carbon storage (r = 0. 96) strongly correlates to tree number and canopy cover area. The forest metric Treecanopy Percentage of Landscape (TPLAND) strongly associated with carbon density (r > 0. 5), while canopy height is the major factor for per-tree carbon storage (r = 0. 46). Variables in the landscape explain 95. 0% of the variation in total carbon storage, and tree number is the main driver. Soil nitrogen and canopy cover of a forest showed positive correlation with carbon storage, while precipitation (PCP) and Roads Percentage of Landscape (RPLAND) showed negative correlation. This study argues that increasing green space, modifying the built-up and vegetation ratio, and altering spatial configuration can enhance neighborhood-level carbon storage. It gives a scientific basis for urban forest engineering and carbon sink improvement.
Wei et al. (2026) studied this question.