Exploring the dynamic relationships and mechanisms driving a long-term sequence of ecosystem services in mountains based on ecosystem service bundles

• Four distinct ecosystem service bundles were identified using a long-term regional assessment. • Ecosystem service relationships showed significant spatiotemporal heterogeneity across the region and four bundles. • Key drivers and their nonlinear threshold effects were quantified for both the region and bundles. • Function-oriented management strategies were proposed based on bundle characteristics. Mountain ecosystems are typical fragile ecological systems that face severe challenges in providing ecosystem services (ES) due to the dual pressures of global climate change and intensified human activities. However, there is limited research on the long-term spatiotemporal dynamics and driving mechanisms of ES in mountainous regions. The Wuling Mountain Area (WMA) is a key ecological function zone in central China. It plays a critical role in maintaining regional ecological security and promoting sustainable development by providing a stable supply of ES. Using long-term data from 2000 to 2020, this study first assessed the spatiotemporal characteristics and variations for four key ES in WMA: habitat quality (HQ), net primary productivity (NPP), soil conservation (SC), and water yield (WY). The K-means clustering method was then used to identify ecosystem service bundles (ESB). A Pearson correlation analysis and an eXtreme Gradient Boosting (XGBoost) model were applied to explore trade-offs and synergies among ES and their driving mechanisms, respectively. The results indicate the following. (1) ES in WMA had significant spatial heterogeneity. HQ showed a slight downward trend ( slope = -0.002), while NPP ( slope = 2.19), SC ( slope = 1.17), and WY ( slope = 5.39) all showed increasing trends. (2) There were four types of ESB: Core Ecological Bundle (CEB, 28.20 %), Water Abundance Bundle (WAB, 21.66 %), Balanced Production Bundle (BPB, 28.16 %), and Ecological Transition Bundle (ETB, 21.98 %). Each had distinct dominant services and spatial distributions. (3) Relationships among ES varied dynamically, with HQ generally showing trade-offs with NPP, SC, and WY. In contrast, NPP-SC, NPP-WY, and SC-WY mostly displayed synergistic relationships. These relationships also varied significantly across different ESB. (4) Both natural and socio-economic drivers influenced ES changes, often with nonlinear threshold effects. For example, HQ declined in CEB when the proportion of cropland area exceeded 14 %. Based on these findings, the study proposes different management strategies for different ESB. This provides scientific support for functional zoning and sustainable development in the WMA.