Desertification vulnerability under accelerated dryland expansion

Given accelerated dryland expansion, studies on aridity-driven desertification are of great importance for advancing the understanding of land degradation mechanisms and developing possible mitigation measures. However, how increased aridity drives ecosystems toward desertification and its impacts on desertification vulnerability both in space and over time are largely unclear. Here, we collected eight key ecosystem structure/functioning data (including soil, vegetation and animal) for threshold identification using piecewise linear models, and then linked them to biome shifts to explore the process of AI-driven desertification. The results show that ecosystems along aridity index (AI) gradients experience three vulnerable stages featuring different threshold and biome responses: a low vulnerability stage (AI: 0.5-0.65; mainly featuring a soil fertility loss and a forest decline), a moderate vulnerability stage (AI: 0.65-0.75; mainly featuring accelerated vegetation degradation) and a high vulnerability stage (AI: 0.75-1; mainly featuring rapid overall land degradation). According to further simulations, areas with low to high vulnerability account for 31% of the global terrestrial area, which is projected to increase by 18.8% (21.2%) under the greenhouse gas emission scenario SSP2-4.5 (SSP3-7.0) by the end of the 21st century. Given increased vulnerability and population growth, Africa and East Asia are exposed to great desertification risk and ensuing contradiction between humans and lands. These findings can provide a theoretical and practical reference for ecosystem management and policy development, which are critical for global ecological security and human welfare in areas threatened by desertification.