Pervasive Loss of Arctic-Boreal Vegetation Resilience from Climate Change and Disturbance

Arctic-boreal ecosystems have experienced extensive vegetation change during recent decades. Rapidly changing ­climate and more frequent disturbances cause large uncertainties in future vegetation distribution and carbon sink strength. A major challenge for accurately predicting Arctic-boreal vegetation change is to characterize vegetation resilience, i.e., when and where vegetation becomes vulnerable to climate fluctuations and how fast it recovers after disturbances. However, besides previous studies focusing on post-fire recovery at limited field sites, resilience change resulting from climate trends and disturbances has not been systematically quantified at large scales in Arctic-boreal ecosystems. Here, we map vegetation resilience using temporal autocorrelation of greenness at a 300 m resolution across the core domain of NASA Arctic-Boreal Vulnerability Experiment (ABoVE). We plan to leverage a Bayesian dynamic linear model to estimate time-varying autocorrelation using submonthly time series of greenness dada from satellite remote sensing, including data from Landsat 7. Our initial results in Alaska show that vegetation resilience has increased only in 25% of Alaska but significantly reduced in 62% of the area. Notably, greenness trends did not explain resilience changes, suggesting that greening does not enhance resilience. Instead, reduced resilience is primarily detected at locations with warm temperature and high temperature variability. Next, we categorize resilience change before and after different land cover changes and disturbance types, including fire, insect outbreak, and logging. Our findings highlight pervasive loss of vegetation resilience in the Arctic-boreal region despite widespread greening trends, suggesting that vegetation is more vulnerable to and less capable of recovering from perturbation with warming. The derived resilience maps and the identified control factors will provide a benchmark facilitating further observational and modeling research for better prediction of Arctic-boreal vegetation dynamics.