Résumé
Comparative functional ecology seeks to understand why and how ecological systems and their components operate differently across environments. Although traditionally used in (semi)-natural situations, its concepts and methods could certainly apply to address key issues in the large variety of agricultural systems encountered across the world. In this review, we present major advances in comparative plant functional ecology that were made possible over the last two decades by the rapid development of a trait-based approach to plant functioning and prospects to apply it in agricultural situations. The strength of this approach is that it enables us to assess the interactions between organisms and their environment simultaneously on a large number of species, a prerequisite to address questions relative to species distribution, community assembly and ecosystem functioning. The trait concept will be first defined, before presenting a conceptual framework to understand the effects of environmental factors on plant community structure and ecosystem properties via plant traits. We will then argue that leading dimensions of variation among species can be captured by some selected traits and show that a combination of three easily measured traits-specific leaf area (the ratio of leaf area to leaf dry mass), plant height and seed mass-enables us to assess how different species use their resources, interact with neighbours and disperse in time and space. The use of traits to address central questions in community ecology will be reviewed next. It will be shown that traits allow us to (1) understand how plant species are sorted according to the nature of environmental gradients, (2) evaluate the relative importance of habitat filtering and limiting similarity in the process of community assembly and (3) quantify two main components of community functional structure, namely, community-weighted means of traits and community functional divergence. The relative impacts of these two components on ecosystem properties will then be discussed in the case of several components of primary productivity, litter decomposition, soil water content and carbon sequestration. There is strong support for the biomass ratio hypothesis, which states that the extent to which the traits of a species affect those ecosystem properties depends on the abundance of this species in the community. Assessing the role of functional divergence among species on ecosystem properties will require major methodological breakthroughs, both in terms of metrics and statistical procedures to be used. In agricultural situations, we show that trait-based approaches have been successfully developed to assess the impacts of management practices on (1) the agronomic value of grasslands and (2) the functional composition and structure of crop weed communities and how these could affect the functioning of the crop. Applications in forestry are still poorly developed, especially in temperate regions where the number of species in managed forest remains relatively low. The last decades of research have led to the constitution of large data sets of plant traits, which remain poorly compatible and accessible. Recent advances in the field of ecoinformatics suggest that major progress could be achieved in this area by using improved metadata standards and advancing trait domain ontologies.
Finally, concluding remarks, unanswered questions and directions for research using the functional approach to biodiversity made possible by the use of traits will be discussed in the contexts of ecological and agronomical systems. The latter indeed cover a wide range of environmental conditions and biological diversity, and the prospect for reducing environmental impacts in highly productive, low-diversity systems will certainly imply improving our skills for the management of more diverse systems prone to a trait-based approach as reviewed here.