Modeling grassland landscape degradation and policy responses in the Gannan Plateau: a coupled ERI–BRT–ABM framework

Climate change and intensified human activities are leading to the degradation of alpine grasslands, landscape fragmentation, and soil erosion, threatening the ecological security and sustainable development of the Yellow River Basin. To address the limitations of existing research, which often focuses on static risk descriptions and lacks simulations of behavioral responses and policy interventions, this study systematically assesses the spatiotemporal evolution and driving mechanisms of landscape ecological risk (LER) in the Gannan Plateau from 1980 to 2020, and quantifies the long-term impacts of different grazing policies on grassland landscape degradation and restoration. We used the Landscape Ecological Risk Index (ERI) to depict the risk pattern, employed the Boosted Regression Tree (BRT) to identify the relative importance of factors, and constructed an Agent-Based Model (ABM) in NetLogo for grassland landscape ecological risk scenario simulation and policy response based on key factors, conducting multi-scenario simulations to test the effects of policy combinations. The results show that the ecological risk in the study area exhibits a spatial gradient of “high in the northeast and low in the southwest”, with medium and low risks predominating; slope and GDP are the dominant factors. Multi-scenario simulations show that a single policy is insufficient to curb grassland degradation, while a comprehensive strategy combining “Seasonal grazing bans + Spatial grazing prohibition zones + Livestock quantity limits” is optimal in terms of both ecological and cost-effectiveness. This study proposes a “Risk Diagnosis - Driving Mechanism Analysis - Dynamic Simulation - Policy Optimization” framework, providing a scientific basis for differentiated and collaborative management in the alpine pastoral areas of the Yellow River Basin, and has theoretical and practical value for the construction of ecological security barriers and the sustainable utilization of grasslands. • Integrated ERI–BRT–ABM framework links risk diagnosis to policy response simulation. • Slope and GDP are the top drivers of landscape ecological risk in Gannan. • Simulated multi-policy scenarios to assess grassland landscape management and degradation control. • Seasonal bans + spatial zones + livestock limits best prevent grassland landscape degradation.