Abstract

Against the backdrop of global climate change, enhancing the carbon sink capacity of ecosystems has become one of the key pathways for implementing climate action. This study takes the Western Hills–Yongding River Cultural Belt in Beijing as its study area to systematically investigate the coupling mechanisms between multidimensional spatial pattern factors and the carbon sink capacity of vegetation net primary productivity (NPP). The study employed the morphological spatial pattern analysis (MSPA) method to structurally identify and quantify regional green space landscape elements. By integrating vertical vector data such as building height and canopy height, and incorporating terrain background and human activity intensity indicators, a comprehensive system of driving factors was established, encompassing two-dimensional landscape patterns, three-dimensional spatial structures and multi-dimensional attributes. Based on the Local Climate Zone (LCZ) classification system, a detailed classification of the study area’s land cover was carried out, and eight typical units—encompassing built-up settlements, brownfield sites, forested green spaces and riparian wetlands—were selected as the core objects of analysis. Building on this, multiple non-linear machine learning regression models were constructed to conduct fitting analyses and accuracy validation; the LightGBM model was identified as the optimal fitting model based on the test set coefficient of determination (R2) and root mean square error (RMSE) as core indicators; Furthermore, the SHAP interpretability analysis framework was introduced to systematically reveal the relative importance of each of the driving factors, their positive and negative effects, non-linear response characteristics, and two-factor interaction mechanisms. Based on these findings, category-specific, differentiated spatial regulation strategies for enhancing carbon sinks were proposed, providing scientific support for the optimisation of ecological spaces, the ecological restoration of brownfield sites, and the realisation of carbon sink potential within the study area.

IPC Classification

Keywords