Dramatic shift in the drivers of ecosystem service trade-offs across an aridity gradient: Evidence from China's Loess Plateau.

Increased aridity creates challenges for sustainable ecosystem management due to the potential for trade-offs among ecosystem services. However, our understanding of how ecosystem service trade-offs (EST) respond to aridification remains limited. Here, generalized additive models and structural equation modeling were used to explore EST dynamics within an aridity gradient on the Loess Plateau, China. Trade-offs between water yield and both carbon storage and habitat quality showed nonlinear relationships with aridity, first increasing and then decreasing. Interestingly, climatic and human factors mostly indirectly influenced EST via effects on landscape characteristics. In regions with an Aridity Index (AI) value of <0.5, climatic and human factors strongly drove EST; in regions with AI > 0.5, landscape characteristics were most important. Therefore, landscape characteristics acted as the key regulators of EST. Importantly, AI values of ∼0.5 represented a transition point, after which dramatic shifts in EST-driver relationships were observed. As >22 % of the Earth's terrestrial surface is projected to reach this level of aridity by 2100, further research on this boundary (between sub-humid and semi-arid areas) is urgently needed to protect ecosystems from the effects of increasing aridity. This study may serve as a valuable reference for mitigating the potential negative effects of increased aridity on human well-being.