Sustainable restoration in sandy lands is challenged by the divergent responses of grassland functional types to disturbance

The effective management of regional grassland degradation requires a deep understanding of how different grassland types respond to climate change and human disturbance over different spatial and temporal scales. This study examined the dynamics and underlying driving mechanisms of various grassland types in the Otindag Sandy Land over 36 years (1986-2021) by integrating time-series Landsat images and extensive climate and socio-environmental data. We found that the dynamics and driving factors of different grassland types varied significantly. In low-cover grasslands, the area of desert grassland fluctuated with a decreasing trend after the breakpoint, while that of sandy grassland increased continuously without any breakpoints. In high-cover grasslands, the area of typical grassland only increased after the breakpoint, whereas that of temperate savanna grasslands first decreased and then increased insignificantly. Additionally, the desert grassland area was significantly negatively correlated with annual precipitation (AP), whereas the sandy grassland area was negatively correlated with annual mean wind speed (AMWS). The dynamics of typical grasslands were jointly influenced by climate and human factors, with temperate savannas mainly driven by the AP and AMWS. Moreover, high-cover grasslands were more sensitive to human activities than low-cover grasslands, and the most significant driving factors changed before and after the breakpoint in the area dynamics. These findings provide valuable insights into the monitoring and conservation of sandy land ecosystems.