UC Berkeley

The ongoing climate change is driving significant changes in vegetation composition and species range shifts in many ecosystems worldwide. For example, climate warming has reduced the frequency of extreme low temperature events, which may facilitate woody plant encroachment in woodland-grassland ecotones worldwide where freezing stress inhibits the growth of cold-sensitive woody plants. In addition, global warming and increasing rainfall variability can significantly affect community dynamics and ecosystem functioning especially in drylands where plant growth is primarily constrained by water availability. Both CAM plant expansion in desert ecosystems and woody plant encroachment in cold ecotones have been empirically observed, the underlying mechanisms, however, still remain unclear. A more mechanistic understanding of how global climate change and local positive microclimate feedbacks will affect plant community composition, species range expansion and resilience of these ecosystems is still lacking.

In this dissertation work, I first conducted greenhouse and growth chamber experiments to investigate the responses of two CAM-grass communities in arid ecosystems across the southwestern United States and Chihuahuan desert to asymmetric warming and increasing rainfall variability. In addition, I integrated lab experiments and field observations with process-based modelling frameworks to examine the microclimate warming effects of woody canopies, investigate the physiological cold intolerance of woody species, and quantitatively evaluate to what extent climate warming may trigger critical transitions from grassland to woodland in Northern American coastal ecosystems. Furthermore, I used high-resolution imagery data and stochastic cellular automata models to investigate the spatial patterning of woody patches on Hog Island (Virginia) and its associations with critical transitions in vegetation dynamics.

I found that increasing rainfall variability can enable deep-rooted grasses to gain competitive advantages over shallow-rooted CAM plants through increasing deep soil water availability under current temperature scenario. However, the competitive advantage is likely to shift from grasses to CAM plants due to drought-induced grass mortality under asymmetric warming. I also found that abrupt transitions from grassland to woodland may occur in coastal woodland-grassland ecotones including mangrove-salt marsh ecotones along the Atlantic coast of Florida when the minimum nocturnal temperature exceeds a critical threshold. Such critical transitions may be induced by positive vegetation-microclimate feedbacks whereby woody plants create a local warming effect through modifying the surface energy balance. Furthermore, the spatial patterns of woody vegetation on Hog Island exhibit signs of critical phenomena as evidenced by the emergence of power law distribution of woody patch size in some specific years.

These findings suggest that the ongoing climate change may facilitate CAM plant expansion in Northern American desert ecosystems and woody plant encroachment in Northern American coastal ecosystems where the latitudinal limits of woody plants are majorly constrained by freezing stress. The results also provide novel empirical and theoretical evidence of whether the observed scale-invariant vegetation patterning may be considered as a general early warning signal of critical transitions from grassland to woodland in coastal and potentially other ecosystems. Overall, this dissertation highlights the important role of environmental drivers and microclimate feedbacks in driving transitions in plant community composition, species range shifts, and ecosystem structure and functioning in desert and coastal ecosystems.