Evolutionary lability underlies drought adaptation of Australian shrubs along aridity gradients

Abstract

Leaf drought tolerance traits influence plant survival in water deficit conditions, and these traits are influenced by both the plant's evolutionary history and the environment in which the plant is currently growing. However, due to the substantial phenotypic plasticity in leaf traits, we still do not know to what degree variation in leaf traits is governed by species' phylogenetic history or by their environment. To explore this question, we re-examined a drought tolerance dataset from 37 native Australian shrub species with varying climate origins growing in a common garden located in Melbourne, Australia. We previously measured seven leaf morphophysiological traits, and here, we estimated how phylogenetically conserved these traits are. We quantified phylogeny and the strength of correlation between the morphological traits and physiological traits before and after accounting for shared phylogenetic history. We also evaluated the relationship between species' leaf traits and the climate of their native ranges. We present three main findings: (a) most leaf drought tolerance traits had weak phylogenetic signals, which is consistent with the convergent evolution of these traits. (b) There is weak but consistent coordination between distinct leaf drought tolerance traits, which can be masked due to species' phylogenetic histories. (c) Leaf drought tolerance traits show strong correlations with the climate of species' origins, and this relationship is only weakly impacted by phylogenetic signals. Therefore, the role of phylogeny on the coordination among leaf functional traits and their links to climate were limited. A better understanding of trait-environment relationships might be more pivotal than understanding the evolution of these traits for improving the predictions of species' response to climate change-type drought, especially for shrub species that span substantial aridity gradients.

Keywords: climate change; common garden; ecophysiology; functional traits; phylogenetic niche conservatism; shrubs.

Figure 1

Data visualization of the three morphological traits (leaf size, specific leaf area, and Huber values) mapped along the phylogeny of 37 shrub species. If these traits were phylogenetically conserved, closely related species should share similar bar lengths. By default, data are centered and scaled by trait.

Figure 2

Data visualization of the four physiological traits (π_o, π_tlp, ϵ, and Ψ _mid) mapped along the phylogeny of 37 shrub species. If these traits were phylogenetically conserved, closely related species should share similar bar lengths. By default, data are centered and scaled by trait.

Figure 3

Phylogenetic correlograms for two traits: (A) π_o and (B) π_tlp. The solid bold black line represents the Moran’s I index of autocorrelation, and the dashed black lines represent the lower and upper bounds of the confidence envelope (95%). The horizontal black line indicates the expected value of Moran’s I under the null hypothesis of no phylogenetic autocorrelation. The colored bar shows whether the autocorrelation is significant based on the confidence interval (red) or not (black).

Figure 4

Principal component analysis (PCA) for the seven leaf traits of 37 shrub species. Panels (A, B) are not correcting for phylogeny, while (C, D) are a phylogenetically corrected PCA. The color gradients of the legend in each panel show the contribution of a trait to a given principal component in percentage. The hollow circles labeled with a four-letter species code aside in panels (A, B) represent each plant species (see supp. table 1 for full species names and corresponding species code).

Figure 5

Correlation between leaf drought–tolerant traits after removing the phylogenetic signals. Solid lines represent the linear regressions, and shallow gray bands represent 95% confidence in the lower triangular intervals. The correlation coefficients are given respectively in the graphics above the diagonal. Histograms showing trait value distributions are given in the diagonal. *P < 0.05, ** P < 0.01, *** P < 0.001. N.S., non-significant relationship.

Figure 6

Relationships between morphological traits and the 5th percentile of the aridity index (AI) at the species level: (A–C) linear regression between morphological traits and the 5th percentile of AI; (D–F) phylogenetically independent contrast (PIC) linear regression between morphological traits and the 5th percentile of AI. Solid black lines indicate the linear trends of the 5th percentile of the AI changes of morphological traits, and shallow gray bands represent 95% confidence intervals. N.S., non-significant relationship.

Figure 7

Relationships between physiological traits and the 5th percentile of AI at the species level: (A–D) linear regression between physiological traits and the 5th percentile of AI; (E–H) PIC linear regression between physiological traits and the 5th percentile of AI. Black dashed lines indicate marginally significant linear trends of the 5th percentile of the AI changes of physiological traits, and shallow gray bands represent 95% confidence intervals. N.S., non-significant relationship.