Date Published: March 28, 2018
Publisher: Public Library of Science
Author(s): Michaël Belluau, Bill Shipley, RunGuo Zang.
Species’ habitat affinities along environmental gradients should be determined by a combination of physiological (hard) and morpho-anatomical (soft) traits. Using a gradient of soil water availability, we address three questions: How well can we predict habitat affinities from hard traits, from soft traits, and from a combination of the two? How well can we predict species’ physiological responses to drought (hard traits) from their soft traits? Can we model a causal sequence as soft traits → hard traits → species distributions?
We chose 25 species of herbaceous dicots whose affinities for soil moisture have already been linked to 5 physiological traits (stomatal conductance and net photosynthesis measured at soil field capacity, water use efficiency, stomatal conductance and soil water potential measured when leaves begin to wilt). Under controlled conditions in soils at field capacity, we measured five soft traits (leaf dry matter content, specific leaf area, leaf nitrogen content, stomatal area, specific root length).
Soft traits alone were poor predictors (R2 = 0.129) while hard traits explained 48% of species habitat affinities. Moreover, hard traits were significantly related to combinations of soft traits. From a priori biological knowledge and hypothesized ecological links we built a path model showing a sequential pattern soft traits → hard traits → species distributions and accounting for 59.6% (p = 0.782) of habitat wetness.
Both direct and indirect causal relationships existed between soft traits, hard traits and species’ habitat preferences. The poor predictive abilities of soft traits alone were due to the existence of antagonistic and synergistic direct and indirect effects of soft traits on habitat preferences mediated by the hard traits. To obtain a more realistic model applicable to a population level, it has to be tested in an experiment including species competition for water supply.
Variation in soil water availability is one of the major environmental gradients along which plant species are differentially distributed  and is the environmental gradient of interest in this study. Presumably, the realized hydrological niches of such differentially distributed species are determined by both biotic interactions within the community and by trade-offs among functional traits involved in the water economy of a plant. It so then one should be able to predict both whole-plant performance and species relative abundances along environmental gradients given the relevant functional traits, but which traits should be used? Choosing the right traits is a key challenge in trait-based ecology and one that has not yet been completely met [2,3].
From Belluau and Shipley (2017)  we already know that the five physiological (hard) traits can predict the habitat affinities of our species along a gradient of soil water availability and that a plausible multivariate causal hypothesis linking them is consistent with both the observed data a pre-existing knowledge. However, these traits are difficult and time-consuming to measure and some of them require that the plant already be wilting before measurements are taken. This limits their usefulness in large-scale trait-based ecology. We expected, given the hierarchy of traits described by Chapin III, Autumn and Pugnaire (1993) and Silvertown et al. (2015) [4,5] that (i) soft (morpho-anatomical) traits would be more loosely linked to habitat affinities than the hard traits; that (ii) the soft traits would, none the less, link to the hard traits; and that (iii) we should be able to express the links as soft→hard→habitat.