Attribution of Land Surface Albedo Changes in China over the Last 40 Years

Surface albedo plays a key role in regulating land-atmosphere energy exchange, yet its spatiotemporal variability and underlying driving mechanisms remain inadequately quantified. This study first developed a machine learning model to reconstruct land surface albedo across China from 1980 to 2020, and conducted model experiments to separate the contributions of three primary drivers (i.e., land cover change, vegetation dynamics, and climate change) to albedo variations. Results show that the machine learning model can reproduce surface albedo with high accuracy. Over the past four decades, the mean albedo across the study area decreased by 0.0101, corresponding to a linear trend of -0.0003 yr-1. Attribution analysis indicates that climate change was the dominant driver over 55.30% of the land area, followed by vegetation dynamics (24.26%) and land cover change (20.44%). Climate forcing, through its control over snow cover, temperature, precipitation, and soil moisture, primarily governed both interannual fluctuations and long-term trends in albedo. In contrast, large-scale afforestation and ecological restoration led to substantial albedo decreases, particularly in southern and southwestern China. Sensitivity analysis further reveals strong spatial heterogeneity in albedo responses to leaf area index (LAI), with pronounced negative sensitivities in arid regions and weak or even positive effects in humid zones. Our findings highlight the dominant role of climate variability in shaping albedo dynamics, while demonstrating how large-scale ecological restoration and vegetation greening modulate surface energy balance under ongoing climate change.