Pre-disposing factors leading to drought-induced mortality in broad-leaved evergreens
Drought-induced mortality has major impacts on landscapes but is currently challenging to predict. Not only are the mechanisms of mortality still unresolved, but we also have a poor understanding of environmental factors that pre-dispose vegetation to drought mortality. Bioclimatic, process-based and dynamic global vegetation models must incorporate these thresholds in order to improve the accuracy of their predictions. A thorough understanding of the pre-disposing factors leading to drought-induced mortality is needed to predict water and carbon fluxes, and estimate vegetation boundaries in future.
Here we synthesize the literature on biotic (age, plant size, stand density, stand basal area, root depth) and site conditions (microclimate, shade, aspect, soil depth and type, and species region of origin, temperature, atmospheric CO2) drivers of drought-induced mortality, using examples from South Africa and Australia, examining climatic and site-specific variables correlated with mortality. An understanding of mortality in broad-leaved evergreen Australian and South African vegetation will inform our understanding of vegetation from other regions, as convergent strategies to avoid water stress have been observed globally in vegetation from nutrient-limited sites.
Synthesis of the literature on factors pre-disposing plants to drought mortality showed that not all of the pre-conceptions of factors leading to mortality are supported by evidence. Many pre-disposing factors had thresholds which needed to be exceeded before % mortality increased. Further, most relationships between pre-disposing factors to drought and mortality appeared to be non-linear. Stand density and rooting depth were identified as key pre-disposing factors, with strong evidence for a positive correlation between stand density and drought-induced mortality, and a negative correlation between root depth and mortality rates. The relationship between drought mortality and tree age or size had a ‘U shaped’ curve, where young, small trees and older, larger trees died, although many studies reported no effect of tree age or size.
These non-linear thresholds need to be considered in Dynamic Global Vegetation Models and process-based ecophysiological models. Many research gaps still exist, with some plant functional types having few results, including needle-leaved evergreens. Further, more studies are needed which examine the relationship between growth rates, grazing, region of origin, elevated CO2 and mortality rates.