Revealing the driving mechanisms of ecosystem services supply and demand relationships in a coastal city: A geospatial machine learning approach

Driven by human activities and climate change, the contradiction between supply and demand of ecosystem services (ESs) in global coastal areas is becoming increasingly acute, and the sustainable development of coastal cities is facing severe challenges. However, the complex driving mechanism and spatial heterogeneity of the ecosystem services supply and demand relationships (ESSDR) in coastal cities are still unclear, and it is difficult to support refined ecosystem management. Therefore, taking Qingdao as an example, this research employed the G-XGBoost model combined with the SHAP interpretability framework to assess the nonlinear and local spatial relationships between drivers and ESSDR, based on the quantification of six ESs (food production, water yield, marine aquaculture, carbon sequestration, soil retention, and recreational capacity) from 2002 to 2022. The results show that: (1) the comprehensive ecosystem services supply-demand ratio (CESSDR) exhibited significant spatial differentiation, with an overall pattern of southern deficit and northern surplus; (2) The proportion of built-up land (BLP) is the most influential driving factor; and (3) Urban expansion exerts a significant nonlinear threshold effect on ESSDR: when BLP exceeds 30%, the imbalance between ESs supply and demand intensifies significantly. Based on these findings, this study proposes a differentiated zoning management strategy centered on dominant-factor identification and provides scientific support for ecosystem management in coastal areas.