A Bayesian network-based study on the evolution of ecosystem service trade-offs intensity mechanisms in the Nansihu Basin, China

Understanding the mechanisms that drive the trade-off intensity of ecosystem services in basins is critical to achieving sustainable regional development. First, this study quantifies ecosystem services using the InVEST model and then calculates the trade-off intensity using the root mean square error (RMSE). Second, a Bayesian network model is combined to analyze the dynamics of the drivers of trade-off intensity over multiple time periods. Finally, a probabilistic model of Bayesian networks is used to simulate areas within the basin that are more likely to have high trade-off intensities. The results show a significant increase in grain production but a decrease in water conservation, water purification, habitat quality and carbon storage. Bayesian inference identified precipitation, the proportion of cropland and the NDVI as dominant drivers. Spatial optimization simulations highlighted Heze and Zaozhuang city as priority regions, suggesting integrated land–water management strategies to reconcile agricultural and ecological objectives.