Correlation of drought traits and the predictability of osmotic potential at full leaf turgor in vegetation from New Zealand

Abstract

Scientists do not know precisely how severe will be the impact of climate change on species. Evidence suggests that for some species, their future distributions might be jeopardized by local extinctions and drought-induced tree mortality. Thus, we require models capable of estimating drought tolerance across many species. We can approach this goal by assessing functional traits. The trait osmotic potential at full turgor, πO, is potentially a good drought indicator; however, few studies address its importance as a drought-tolerance predictor and it is difficult to measure in the field with accuracy. In this work, we aim to answer the questions: which drought traits correlate with πO?; do morpho-anatomical traits correlate with πO?; and which trees and shrubs are more (or less) vulnerable to drought? To achieve this aim, we assessed physiological and morpho-anatomical traits for 14 native species from New Zealand forests. We included leaf- and wood-related traits, πO, water potential and stomatal conductance. We examined how these traits correlate with πO and sought to generate models to predict πO as a function of other traits. We tested 33 different models and evaluated them using Akaike's information criterion. Unfortunately, none of the morpho-anatomical traits correlated well with πO. Instead, water potential correlated most strongly with πO. None of the models using only morpho-anatomical traits produced plausible results. The model with the best predictive performance incorporated the effects of both morpho-anatomical and physiological traits: water potential and wood saturated water content. Of the species analysed, and based on their πO response, Lophozonia menziesii was considered the most vulnerable to drought stress, whereas Plagianthus regius was the least vulnerable. Our findings imply that it is potentially valuable to keep exploring the use of πO as a drought indicator and that the effort required to measure some physiological traits, such as water potential, may be essential to consider plant drought responses and to predict πO.