Fine-scale variation in the effect of the cushion plant Azorella selago on vascular plants, mosses, hepatics and lichens in the sub-Antarctic

Abstract

QUESTION : Plant–plant interactions can strongly influence community structure and composition. The outcome of these interactions can vary considerably across space and is often linked to environmental conditions, with, for example, a higher prevalence of facilitative interactions typically being observed under greater environmental severity. To date, most studies have documented shifts from competitive to facilitative (or neutral) plant–plant and plant–lichen interactions along gradients of increasing environmental severity from pairwise interspecific interactions. However, few studies have examined if the outcome of these interactions for different taxonomic groups is dependent on environmental conditions across multiple environmental stress gradients. Location: Sub-Antarctic Marion Island. METHODS : We examine community-level variation in the response of four taxa (i.e., vascular plants, hepatics, mosses, and lichens) to an interaction with a long-lived cushion plant species (Azorella selago) that ameliorates microenvironmental conditions, testing how the effect of the cushion plant on the taxa varies along multiple stress gradients at the scale of a landform. RESULTS : Contrary to expectations, even when considering multiple proximate predictor variables, fine-scale spatial variation in the effect of A. selago on the taxa could not be explained. However, the outcome of the interaction with A. selago differed between taxonomic groups, with vascular plants benefitting and the non-vascular taxa experiencing neutral or negative impacts. CONCLUSIONS : This study highlights that the impacts of biotic interactions cannot always be generalized across plant groups, and that it is necessary to consider taxon-specific responses when predicting community-level impacts of biotic interactions. More generally, we demonstrate how complex spatial variation in environmental stressors can be explicitly considered when modelling variation in the outcome of plant–plant interactions.