Plant Functional Traits Show Non-Linear Response to Grazing

• Autores: Arne Saatkamp, Christine Römermann, Thierry Dutoit
• Localización: Folia geobotánica: A journal of plant ecology and systematics, ISSN-e 1874-9348, ISSN 1211-9520, Vol. 45, Nº 3, 2010, págs. 239-252
• Idioma: inglés
• Enlaces
  • Texto completo (pdf)
• Resumen

  • The trait-based approach aims to detect functional patterns in vegetation beyond specific sites or taxa. In most cases, plant traits are assumed to be linearly related to environmental gradients such as grazing intensity. To generalize results beyond specific sites, it is important to know to which extent environment-trait relationships are non-linear. Non-linearity can be a source of inconsistency among different studies according to length and studied portion of a gradient. In this study, we test if and to what extent traits relate non-linearly to a grazing gradient using data from a grassland-matorral interface in the Mediterranean rangeland of �La Crau� (SE France). Because quantification of grazing by itinerant sheep flocks is difficult, we use a marker plant, Phillyrea angustifolia, and several independent pasture indicators to estimate grazing intensity. First, we related traits to grazing using a multivariate three-table ordination method (RLQ), which assumes linear reactions of traits. To evaluate the importance of non-linearity, we applied generalized additive models (GAMs) to our data. This revealed that a third of traits studied here showed non-linear relationships to grazing. These non-linear responses cover a large spectrum including seed mass, life form, phenology dispersal- and leaf traits. The high part of non-linear relations compromises a general assumption of linear trait-environment relationships. Future works should therefore more often consider non-linear relationships using methods with no constraints on shape of response, e.g., GAM in the analysis of functional trait studies. In this way, non-linear relationships can reveal new aspects of species and community response to global change and deepen our understanding of trait-environment relations.