My Research


Generally, I am interested in causes and consequences of phenotypic variation and integration at the individual, population and community levels. I use physiological and anatomical traits to study trait-environment relationships. Grounded in empirical data, my research gives me the opportunity to introduce aspiring scientists to ecological research in the field and in the lab.


Can traits predict plant species and community response to climate warming?

Mégantic Project Figure for WebpageIn North America, climate warming is expected to cause northward and upward elevational range shifts in plant species and changes in local community composition. While temporal changes in taxonomic diversity and composition have been studied, their consequences for functional diversity and composition (i.e. the community’s set of phenotypic traits) are poorly understood. Yet, functional composition has been shown to be better predictors of community properties and ecosystem functioning than taxonomic composition. Moreover, while research indicates that traits are good predictors of species performance along spatial gradients, traits have seldom been used to predict changes in species relative abundances over time. We propose to: (1) examine how functional diversity has changed over time in response to a warming climate; and (2) assess whether changes in species relative abundances can be predicted by phenotypic traits. We use a unique dataset combining temporal and spatial environmental gradients within the Parc National du Mont Mégantic, to which we are adding new data on key plant functional traits.

Interspecific integration of trait dimensions at local scales: the plant phenotype as an integrated network

Closely integrated groups of traits (i.e. trait dimensions) are used to classify plant phenotypic diversity into plant strategies, but we do not know the degree of interdependence among trait dimensions. To assess how selection has shaped the phenotypic space of plant strategies, we examine whether trait dimensions are independent.We gathered data on locally coexisting tree species and examined the correlation structure of 20 traits reflecting three well-established trait dimensions: the leaf economic spectrum, the wood spectrum, and Corner’s rules.  Across species, the sapling phenotype is not structured into clear trait dimensions. Instead traits from the three commonly recognized trait dimensions are organized into an integrated trait network. Our results indicate that trait dimensions apparent in broad-based interspecific surveys do not hold up among locally coexisting species. Our study indicates that local and global patterns of phenotypic integration differ, questions the evolutionary origins of trait dimensions and calls into question the use of trait dimensions at local scales. Ch2 figure for webpage

Differences in functional trait integration among temperate deciduous tree species: What are the roles of constraints, phylogeny and environmental niche?

General relationships among functional trait have been identified across species, but it remains largely unknown whether there are ubiquitous patterns of functional trait correlation within populations. To assess the roles of constraints, phylogeny and adaptation to the environmental niche on the phenotype, we explore similarities and differences in intrapopulation (static) integration structures in locally coexisting tree species. Static integration of the different species are weak and all distinct from each other. In only three trait pairs are the intrapopulation correlations are shared across all species, suggesting these correlations are driven by constraints. Variation in intrapopulation integration among species is not related to environmental nor phylogenetic differences. The species-specific nature of intrapopulation trait integration suggests flexibility in plant design. Ch3 Figure for webpage

Testing central assumptions of trait-based ecology: Traits are good predictors of plant performance, but only when phenotypic complexity and individual variation are taken into account.

While trait-based plant ecology has taken an increasingly important role in plant ecology, some of its fundamental assumptions remain poorly tested. Here we examine three key assumptions: (1) that traits are good predictors of plant performance, (2) that traits are better indicators of plant performance than species identity, and (3) that environmental filtering operates on traits. We measure relative growth rate (RGR) of individual tree saplings from 24 co-occurring temperate tree species and predict RGR based on plant traits, environment, size, age and species identity. We find that (1) traits explains 53% of variance in RGR, (2) that adding species does not improve the models but instead that species affect RGR through traits, and (3) that the environmental affects RGR largely through their interaction with traits. Together, traits, age and environmental variables Ch4 Figure for webpageexplain 75% of RGR variance, albeit with a large number of variables. These findings largely support the assumptions of trait-based ecology, but suggest that taking phenotypic complexity into account might be necessary to produce accurate predictions.


The intensity, rather than the strategy, of functional trait sampling determines the detectability of community assembly processes in tropical forests

Functional traits variation is often used to disentangle community assembly mechanisms and ecological processes. Inferred trait distributions can be influenced by the choice of functional traits analysed and by the trait-sampling strategy employed in the field. In this work lead by Dr. Bentley, we examine the implications of varying trait-sampling strategy for the detection of community assembly mechanisms. We find that sampling intensity more strongly affected our ability to detect environmental filtering than did sampling strategy. In general, we could detect community-level trait structure by sampling 10-20% of individuals across multiple plots independently of the sampling strategy. Our results indicate that, independently of sampling strategy, sampling intensity cannot be sacrificed to adequately assess assembly processes.

Bentley Figure for website


Changes in plant biodiversity during the Anthropocene: A review across scales (In Press)

In response to human activity, plant communities have undergone dramatic changes in recent centuries. This has lead to widespread species loss and to the current biodiversity crisis. However, a closer look at biodiversity studies at different scales reveal that not all such changes fit with the dominant “biodiversity-crisis” narrative.  We find that at the global scale, future declines in plant species diversity are highly likely given habitat conversion in the tropics. However, to date  few extinctions have been documented for the Anthropocene (<0.1%).  At regional scales, the introduction of non-native species has greatly increased plant species richness in many regions of the world. At the local scale, conversion of primary vegetation to agriculture has decreased plant diversity, while other drivers of change (e.g., climate warming, habitat fragmentation, nitrogen deposition) have highly context-dependent effects, resulting in a range of net changes in biodiversity with a mean close to zero. We find that the biodiversity crisis might be one of composition instead of net richness. These results prompt a reassessment of how conservation goals are defined and justified.

Trait variation and integration across scales: Is the leaf economic spectrum present at local scales? (2011-2016)

Ch1 Figure for webpageAlthough trait-based strategy dimensions such as the leaf economic spectrum (LES) have been identified primarily at global-scales, trait variation at the community scale is often interpreted in the context of trait dimensions. In this paper recently published in Ecography, we argue from several lines of evidence that a research priority should be to determine whether global-scale trait relationships hold at more local scales. We review recent literature assessing trait variation at smaller scales, and then present a case study exploring the relationship between the correlation strength of leaf traits and their similarity in variation structure across ecological scales. We find that the correlation strength between pairs of leaf traits does not predict whether the traits respond similarly to different drivers of variation. Instead, correlation strength only sets an upper bound to the dissimilarity in trait variation structure. With moderate correlation strengths, LES traits largely retain the ability to respond independently to different drivers of phenotypic variation at different scales. Recent literature and our results suggest that LES relationships may not hold at local scales.

Fitness of multidimensional phenotypes in dynamic adaptive landscapes (2015)

Dr. Laughlin and I suggest that our understanding of community assembly has much to gain from taking a multivariate perspective on the process of ecological filtering. We propose that examining the links between intraspecific trait co-variation and plant fitness will help predict species responses to environmental gradients. These ideas are described in this paper in published in TREE.dynamic adaptive landscape

The effects of grazing on foliar trait diversity and niche differentiation in Tibetan alpine meadows (2014-2015)

In this work headed by Dr. Niu, we examine whether functional trait diversity and niche differentiation increases the potential for species coexistence. In phosphorus limited Tibetan alpine meadows, we test whether grazing increases the functional diversity of leaf
traits and niche differentiation. We compare indicators of the fundamental and realized niches (method from DeBello et al. 2012. Ecology) to quantify the effects of grazing on trait divergence and convergence,Niu paper illustrative figure for website which are indicative of niche divergence and competitive exclusion. We find that grazing promotes divergence in leaf phosphorus, but convergence in leaf carbon and nitrogen. Together with the lack of response in multivariate space, these results indicate that coordinated shifts in carbon economic traits allow to maintain foliar function while allowing for niche divergence in response to increased competition. This manuscript was published in Ecosphere.

A global meta-analysis of the relative extent of intraspecific trait variation in plant communities (2013-2015)

Led by Dr. Siefert, this review paper examines the relative extent of intraspecific trait variation (ITV) for community ecology as a function of traits, spatial extent, species richness, growth form and climate. A meta-analysis find that overall, Siefert 2015 figure for webpageITV accounted for 25% of the total trait variation within communities and 32% of the total trait variation among communities on average.The relative amount of ITV decreased with increasing species richness and spatial extent, but did not vary with plant growth form or climate. These results, published in this Ecology Letters paper, provide practical guidelines for when researchers should include ITV in trait-based community and ecosystem studies.


The return of the variance: intraspecific variability in community ecology (2012)

With Dr. Violle as first author, this TREE review paper argues that intraspecific variance has been neglected in community ecology but needs to be explicitly considered to make accurate predictions. We introduce new T-statistics (‘T’ for trait), based on the comparison of intraspecific and interspecific variances of functional traits across organizational levels, to operationally incorporate intraspecific variability into community ecology theory. We show that a focus on the distribution of traits at local and regional scales combined with original analytical tools can provide unique insights into the primary
forces structuring communities. Violle 2012. Figure for Webpage

How do traits vary across ecological scales? A case for trait-based ecology (2007-2009)

For my Master’s research at McGill University, I looked at the variation pattern of two leaf functional traits (traits that affect plant  function, such as photosynthesis, support, etc) across 6 nested ecological scales in the lowland tropical rainforests of Panama. I found that two important traits defining plant function vary as much within species as they do among species. This suggests that plant function may be very variable within species. LMA LDMC % varianc across scales.pngWe also find that despite having very distinct species composition, different plots exhibit similar trait distributions, indicating that a community’s leaf trait distribution is subject to environmental filtering although species composition may not. These results were published in 2010 in Ecology Letters.



Dissertation: Variation and integration of ecophysiological traits across scales in tropical and temperate trees: patterns, drivers and consequences (2011-2015)

The overarching goal of my dissertation is to explore the potential and limits of a trait-based approach to plant ecology. Together, the different studies presented here address two explicit and implicit foundational assumptions underpinning the trait-based approach: (1) that the correlation patterns and biological significance of traits transfer across scales and (2) that the phenotypic complexity of plants can accurately be synthesized into a few meaningful traits to study their ecology. Moreover, the last chapter focuses on a third key assumption: (3) that traits are strong predictors of plant performance (Shipley et al. n.d.). I examine these assumptions by exploring multivariate patterns of phenotypic variation and integration across different ecological scales (e.g., individuals, populations, species) while explicitly considering the phenotypic complexity of trees, both in terms of their multidimensional and integrated nature.

Two themes thus permeate this body of work: scales and phenotypic complexity. Much of what we know about the relationships among key traits comes from species-scale studies. Trait variation at smaller scales are often interpreted in the context of these interspecific relationships, but it is not clear that interspecific patterns observed at global scales apply to smaller scales. Moreover, although plants are complex, integrated organisms with intricate relationships among their traits, single traits are often studied and interpreted without considering the rest of the phenotype. Yet, examining individual traits outside of their phenotypic context might provide limited insight or be misleading.

Dissertation Overview_Figure for webpage

To address these shortcomings, this body of work examines multidimensional patterns of trait variation and correlation across ecological scales. It uses (1) a set of six ecophysiological leaf traits from mature trees in a lowland tropical rainforest, and (2) a set of twenty leaf, root, stem, branch and whole-plant ecophysiological traits from deciduous saplings in a temperate forest.

The combination of our findings point to three main conclusions: (i) local interspecific and intra-population trait integration structures differ from each other and from the global interspecific patterns reported in the literature, such that global-scale interspecific patterns cannot readily be transferred to more local scales; (ii) considering the complexity of the plant phenotype provides better insights into ecological patterns and processes than what we can learn from considering individual or a handful of traits; and (iii) traits strongly affect individual plant performance, although there is no relationship between a species’ trait correlation structure and its environmental niche, which suggests that there are multiple alternative optimal phenotypes in a given environment.

Go to my Publications

Go to my Curriculum Vitae