The application of plant functional traits into modelling tree mortality globally

Physiological drought, resulting from climate warming and changes in precipitation patterns may be accelerating the rates of tree mortality globally, however, our understanding of how drought and mortality are related across large spatial-temporal scales is still limited. Modelling approaches, such as global-scale vegetation models, are an important tool to explore tree mortality across space and time. However, a necessary step to capturing drought-induced mortality events in model simulations is appropriate representation of relevant plant physiological and morphological traits and their trade-offs within these large-scale models.

We used 40 plant traits from leaf, branch, stem and roots to analyse functional trade-offs across woody species globally, with a particular focus on the role of plant hydraulics in determining species mortality outcomes. We tested the inter-relationships of these traits structured by their linkage to core aspects of plant function. These analyses generated a key subset of traits relating to functional strategies of resource acquisition, growth rate, water conductivity and hydraulic vulnerability. We then assessed these functional strategies correlated to geographic metrics of plant climatic distribution.

The results indicated that plant water-use was negatively correlated with the leaf-level resource capture, but had no correlation with leaf structure. Moreover, we observed resource capture was positively correlated with growth rates and water conductivity of stems, but not with leaf conductivity. Interestingly, hydraulic vulnerability traits were negatively correlated with resource acquisition and plant water flow. We found clear linkages between water flow and hydraulic vulnerability traits, and climatic drivers relating to aridity and plant distributions.

The work identified a continuum of functional strategies which can be discretised into plant functional types, suitable for explanation in drought-induced tree mortality in a global vegetation model. This is a crucial step towards making more reliable large-scale simulations of drought-driven tree mortality within global vegetation and Earth-system models.