Coupling response mechanisms of land use conflicts and ecosystem health using the GWRF-SHAP model: A structural and functional perspective

Clarifying the coupled response relationship between land use conflicts (LUCs) and ecosystem health (EH) is crucial for regional ecological conservation and sustainable development. While previous studies have primarily focused on the interactions between land use structural conflicts (LUSC) and EH, the role of land use functional conflicts (LUFC) remains underexplored. This study takes Kunming as study area and constructs a comprehensive analytical framework that integrates both structural and functional dimensions of LUCs. The interactions between LUCs and EH were assessed using the coupling coordination degree (CCD) model and the generalized additive model (GAM). Meanwhile, the nonlinear driving mechanisms of CCD were elucidated through an interpretable spatial machine learning model (GWRF-SHAP). The key findings are summarized as follows. (1) LUSC, LUFC and EH exhibited significant spatial heterogeneity. Specifically, from 2010 to 2020, the centers of gravity of both LUSC and LUFC consistently shifted southward, while EH showed an overall fluctuating upward trend. (2) The CCD values between EH and both LUFC and LUSC were generally high, mainly reflecting well or highly coordinated states, with notable north–south spatial differences. The interaction between LUFC and EH followed an inverted U-shaped pattern with two thresholds, whereas LUSC and EH showed a nonlinear negative correlation characterized by a single threshold. (3) Key drivers of CCD included human activity intensity, land use degree, elevation, slope, distance form the city center, and distance form the county center all exhibiting obvious nonlinear characteristics and threshold effects. Additionally, the second-order interaction between human activity intensity and land use degree exerted the strongest influence on CCD. These findings offer deeper insights into the interaction between land use systems and ecosystem, and provide a theoretical foundation for optimizing land resource allocation and coordinating regional ecological management. • Proposed a land use conflicts (LUCs) framework based on structure and function. • Revealed coupling coordination and nonlinear forms of LUCs and ecosystem health (EH). • Identified dominant factors driving the coupling coordination degree of LUCs and EH. • Visualized the nonlinear, threshold and interactive effects of dominant factors.