Xin Li ,  Xinjie Liu ,  Leping Liu ,  Pengfei Sun ,  Jiajing Li ,  Xiaoqian Li ,  Zhi Dong ,  Xinchuang Chen

Abstract 

Urbanization reshapes interactions among ecosystem services (ES), making it necessary to clarify how urbanization gradients, ecological mechanisms, together with nonlinear driver responses, jointly regulate ES dynamics to support ecological zoning management, especially in mountainous and hilly regions. Focusing on the central-south Shandong hilly region, we delineated developed (E), developing (I), and rural (R) zones through multidimensional clustering, quantified six major ES, and examined their trade-offs, synergies, and nonlinear drivers using Geographically Weighted Regression and XGBoost-SHAP. ES bundles were further identified using a Self-Organizing Map, and a dual-zone framework was established by integrating urban functional zones and ES bundles, with spatially explicit driver threshold characteristics used to support interpretation and management diagnosis. Results reveal a distinct R > I > E gradient in ES supply and increasing spatial heterogeneity toward urban cores. ES interactions shift from synergy-dominated in E to increasingly complex trade-offs in I and R, with water conservation‑carbon storage being the most urbanization-sensitive pair. Driver regimes differ across zones: socioeconomic factors dominate in E, natural-anthropogenic interactions in I, and biophysical constraints in R. Key nonlinear thresholds include forest cover >10%, temperature ranging from 14 to 14.7 °C, and precipitation >900 mm. Linking these nonlinear response characteristics with ES bundles and functional zones enables precise identification of sensitive areas, providing a scientifically grounded pathway for differentiated and zone-specific ES management in urbanizing mountainous regions.

Paper Linkage:https://doi.org/10.1016/j.ecolind.2026.114707