Ecosystems may undergo critical transitions from one stable state to another in response to gradual changes in external drivers including climate and disturbances. This shift in ecological states can be non-linear and abrupt, making it difficult to anticipate in real-world situations. The cold-sensitive woody species Morella cerifera L. has been encroaching into adjacent grassland over the last century in Virginia’s barrier islands. It has been shown that climate warming may induce an abrupt transition from grassland to shrubland through vegetation-microclimate feedbacks. However, it is unclear how such a transition affects the spatio-temporal patterns of vegetation and whether such patterns can serve as leading indicators of critical transitions in barrier island vegetation dynamics. Here, we combine high-resolution imagery with a modelling framework to investigate the spatial patterns of woody patches on Hog Island, Virginia.
Results/Conclusions
We show that the distribution of woody patch sizes in 1978 and 1990 can be described by a power law (R2 > 0.99). Our stochastic cellular automata model suggests that the distribution of vegetation cluster sizes follows a power law under critical minimum temperatures (nearly -16°C) but deviates from a power law in the absence of the positive vegetation-microclimate feedback. Both theoretical analyses and empirical evidence indicate that the spatial pattern of vegetation in coastal ecosystems may follow universal scale-invariant power laws when the system approximates critical transitions, as a result of local positive feedbacks with the physical environment. Therefore the emergence of power law scaling can be considered as an early warning signal of critical grassland-to-woodland transitions in Virginia barrier islands.