Distribution Dynamics of Park Green Spaces in China and Their Influencing Factors: Evidence from 1760 County Seats

Urban park green spaces are key infrastructure for improving the quality of the living environment, enhancing residents’ well-being, and strengthening ecological resilience. Targeting 1760 county seats in China, this study combines the stock and incremental synergy analysis matrix, exploratory spatial data analysis, and explainable machine learning (EML) methods (SHAP) to systematically reveal the spatiotemporal patterns, spatial association characteristics, and nonlinear paths of influencing factors regarding the spatial configuration of park green spaces in county seats from 2015 to 2024. The low-stock expansion zone is dominant and concentrated in the western region and non-core urban agglomerations. The high-stock expansion zone is of the dominant type, concentrated in the eastern coastal areas and core urban agglomerations. The low-stock contraction zone and high-stock contraction zone are, respectively, marginalized lock-in and degradation risk types, requiring priority intervention. The dynamics of the park green space configuration, including the stock, increment, and their synergy, all exhibit significant positive spatial autocorrelation. The influences of the nine factors exhibit four major characteristics: directional mixing, intensity gradation, path nonlinearity, and regional heterogeneity. The threshold effect is widespread, and the inflection point value varies depending on the factor and region type. This study constructs a nonlinear and interpretable analytical paradigm and, based on empirical results, proposes a new governance framework of “zoning–grading–staging–synergy”. It provides large-sample empirical evidence and theoretical support for the transformation of small-town park green spaces from “sectoral management” to “spatial governance”, offering significant policy value toward achieving the precise distribution and equitable sharing of regional green space resources.