Future vegetation greening on permafrost dynamics

Both field and satellite observations found a significant greening earth under global warming since the early 1980s, even surprisingly at the permafrost regions. However, the holistic effect of vegetation greening on permafrost at global scale remain unclear. This study employs a well-trained ensemble machine learning model, developed using extensive ground temperature measurements, high-quality climate, topographic, soil, and leaf area index (LAI) data, to assess the effect of vegetation greening on permafrost ground temperature and extent. Our model results show that vegetation greening on the permafrost will lead to a warming in mean annual ground temperature (MAGT) by 0.02-0.01 °C, a reduction in the area of permafrost by 6.7-2.5×10⁴ km² compared to the scenario without LAI increasing, under four shared socioeconomic pathways by the end of this century. This found indicate that the impact of vegetation greening on permafrost at global scale is negligible. However, the impact at regional or local scales is substantial under both SSP1-2.6 and SSP5-8.5, with the heating effect in most shrub-tundra areas averaging 0.2±0.1°C (potentially reaching up to 0.8°C). In contrast, the effect in grassland tundra areas is predominantly cooling, with an average of 0.1±0.05°C (potentially reaching up to 0.6°C), and only occurring under SSP5-8.5. Such multi-scale effects will have implications for climate, the carbon cycle, and projection of permafrost dynamics.