Uncovering nonlinear thresholds and functional zoning of ecosystem services: Evidence from the cities along the Yellow River in Shandong, China

Rapid urban expansion along the Yellow River has exerted increasing pressure on ecosystem functions and ecological sustainability, highlighting the need for precise quantification of ecosystem service ( ES ) supply and identification of key driving thresholds for effective ecological management. This study focuses on riparian cities in Shandong Province, evaluating seven critical ES . Utilizing the XGBoost-SHAP framework, we identified dominant drivers and their non-linear threshold effects, while Self-Organizing Map was applied to delineate functional zones. Spatially, Carbon storage ( CS ), Habitat quality ( HQ ), Soil retention ( SR ), Water conservation ( WC ), Water purification ( WP ), and Wind and sand stabilization ( WS ) cluster in central-southern mountainous regions, whereas Food production ( FP ) shows an opposite pattern. Spatiotemporal analysis from 2000 to 2020 reveals increasing trends in CS and FP supply alongside a general decline in other services. The responses of ES to drivers display clear differentiation: vegetation and human activities jointly regulate CS and FP ; topography dominates SR , WC , WP , and WS ; and cumulative anthropogenic disturbance shapes HQ . Furthermore, distinct service bundles and thresholds are observed across zones; for instance, zones B2, B4, and B5 are severely constrained when population density exceeds 833–837 persons/km 2 [810–860]. These findings provide a robust scientific basis for ES -based ecological management and spatial optimization in rapidly urbanizing regions along the Yellow River.