If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Use of plant functional traits, rather than species identities, to generalize complex community dynamics and predict effects of environmental changes has been referred to as a ‘Holy Grail’ in ecology. Particularly in species-rich communities or in large-scale vegetation models, this approach could prove very powerful in the quest for general rules associating biota and environmental conditions. One framework designed to predict vegetation responses to environmental change factors and changes in important ecosystem functions simultaneously is an effect–response framework. In this framework, plants can be classified in terms of their response to environmental factors (via response traits) and in terms of their effects on ecosystem properties (via effect traits) (Chapin et al., 2000; Lavorel & Garnier, 2002) (Fig. 1). As such, it describes an approach to scale up from individuals to communities and ecosystems in the context of environmental change predictions. While this framework is highly relevant and has the potential to advance how we approach questions of vegetation change, empirical tests are rare. In this issue of New Phytologist, Gross et al. (pp. 652–662) take an important step forward by testing this framework for water availability.
‘This striking disparity between conceptual appeal and empirical application probably indicates uncertainty in how to translate the framework into empirically robust tests.’
The revitalization of the functional traits concept
Classification of plants according to functional traits has a long tradition in plant ecology (Raunkiaer, 1937; Grime, 1979), regardless of whether the traits are called life forms, strategies, syndromes, or functional types. Common to all trait classifications has been the search for a functional description of the vegetation, based on attributes that show a common response to the environment, independent of phylogeny. In parallel to the development regarding functional response, classification of vegetation based on the effects on ecosystem processes has an equally long history (Jenny, 1941). Both investigations are based on the reasonable premise that physiological and demographic constraints, together with trade-offs in life-history, should result in predictable changes in trait representation across environmental gradients.
These investigations took largely separate trajectories until the last decade, when several conceptual advances concerning the links between the effects of environmental changes on vegetation and vegetation effects on ecosystem functioning occurred. Chapin et al. (2000) proposed a conceptual framework where modifications of species composition resulting from environmental change translate into modifications of ecosystem functioning via changes in the representation of species traits. Lavorel & Garnier (2002) expanded upon the framework to articulate the environmental response and ecosystem effects through varying degrees of overlap between relevant traits. As one metric of their importance, these two papers have received almost 450 and 170 citations, respectively, at the time of writing this commentary.
While plant functional-type approaches, and in particular the concept of effect and response traits, have been garnering broad interest over the past decade, empirical tests of the concepts have been slower to appear. For instance, of the papers citing Lavorel & Garnier (2002), only 10% measured both effect and response traits. A literature search for ‘effect traits’ brought up only five relevant experimental studies. This striking disparity between conceptual appeal and empirical application probably indicates uncertainty in how to translate the framework into empirically robust tests.
Testing the effect-and-response framework
Are understanding and incorporating functional traits a key to successful prediction of the effects of environmental change, or are they an appealing conceptual way of looking at change with less potential for empirical application? The work presented by Gross et al. is an important demonstration of how the framework can be tested. The authors were able to determine how land-use change in subalpine grasslands affected the community distribution of plant traits related to water use and, in a separate experiment, to determine how the vegetation associated with each land use affected water availability. The authors identified a suite of traits that mediated plant community response and soil water availability across this environmental gradient, validating their model with relationships at other subalpine grassland sites. Here we use their paper as an example that empirical application of the functional effect-and-response framework can be accomplished, aiming to point to ways that other research could follow this lead.
Measuring functional response
A survey of papers that cite Lavorel & Garnier (2002) indicate that investigating functional response is about four times as common as investigating functional effect. The primary approach in identifying response traits is to evaluate how environmental change may alter trait representation via community composition. Environmental change factors can either be simulated in experimental manipulations or evaluated using a correlative approach across an environmental gradient. Gross et al. used 12 grassland sites of different land-use histories, which formed a gradient in water availability and several other covarying environmental factors. Experimental manipulations (e.g. Engelhardt, 2006; Bret-Harte et al., 2008) are less confounded by other site variables, but the results may indicate transient or timelag effects and may not as clearly follow natural gradients. Probably a duel approach is most robust.
To scale up from species-level traits to a community-level response, Gross et al. aggregated aboveground traits using a mass-ratio approach, in which the trait of each species was weighed by the relative abundance of the species at each site. This approach has been utilized by other studies which found that community-aggregated functional traits were correlated with ecosystem process (Garnier et al., 2004; Vile et al., 2006). For belowground traits, Gross et al. measured aggregate traits directly by measuring root traits at each site and bypassing species identification entirely. This approach blurs the line between traditional community and ecosystem measurements, having the advantage that is it much less time intensive than species-by-species measures. Additionally, these measurements take into account the nonadditive effects of species interactions, while the mass-ratio approach assumes additivity (Suding et al., 2008).
Measuring functional effect
To investigate community effect on soil moisture, Gross et al. compared soil moisture in plots at each site where they had removed all vegetation to plots where vegetation was intact. This removal manipulation allowed the separation of plant effects from land-use effects on soil moisture. Removal manipulations provide a strong test of community effects through comparison between manipulated and unmanipulated plots. Similarly, the effects of individual trait groups on ecosystem functioning can be evaluated by selective removals (Cross & Harte, 2007; Bret-Harte et al., 2008) or in single-species monocultures (Eviner, 2004; Engelhardt, 2006; Pontes et al., 2007). While these are all reasonable methods to estimate ecosystem effect, the total community-removal approach by Gross et al. has the advantage that it is relatively simple to conduct and interpret at the community level. Several other studies have simultaneously identified changes in trait groups and ecosystem function along a gradient (Garnier et al., 2004; Quetier et al., 2007). These studies do not directly separate site from species effects on ecosystem functions, providing weaker tests of the framework.
Relating response to effect
A major distinguishing factor of the response-and-effect framework is that it simultaneously incorporates both changes in community structure as a result of the response and effects of these changes on ecosystem function. Gross et al. demonstrate that community-aggregated response traits translate the effects of these changed plant communities on soil water availability. In this case, without information on both response and effect at the community level, impacts of land-use change on soil moisture would be underestimated.
Species that respond to environmental change may share effect traits, leading to shifts in ecosystem processes, or alternatively may vary in traits so that the effect is less than expected, or even in the opposite direction (Suding et al., 2008). While Lavorel & Garnier (2002) emphasize several scenarios where effect and response traits may differ, Gross et al. focused on traits that are related to both response and effect (which they refer to as response–effect traits). Gross et al. assume that the same traits important to the response to an environmental change factor are those that are critical in determining the effect of that factor on ecosystem function. While assuming that response and effects traits are one and the same would simplify relationships, we do not know how often this correlation should occur, or what situations or habitats may be particularly prone to a particular scenario of trait linkages. This correlation is supported for succession by the findings of Garnier et al. (2004), for grazing by Blanco et al. (2007) and for water drawn down in macrophytes by Engelhardt (2006). In contrast, other work has found that response and effect traits did not strongly overlap (Cross & Harte, 2007; Bret-Harte et al., 2008). Further understanding of response-trait linkages is critical for successful application of the framework.
Using functional response and effect traits as ‘common currency’ has great potential to scale up from individuals to communities and ecosystems in the context of environmental change predictions. Given the relative popularity of the conceptual framework but the paucity of the empirical tests, it is essential that methods are developed for application of these frameworks. While the research need is immense, studies such as those by Gross et al. substantiate the importance of community-centered scaling via functional traits to provide more relevant information on the large, but often neglected, role of community dynamics in environmental change.