Importance of the study context in community assembly processes: a quantitative synthesis of forest bird communities

Species composition is constrained by two upper-level processes in ecological contexts where the dispersion of organisms is not severely limited, namely selection and ecological drift. This intuitive framework has motivated a constant flow of empirical models for linking the species matrix to the local environmental descriptors, in which the environment rarely explains more than 30-40% of the variation in species composition. In most cases, researchers only approximate the environmental axes that drive fitness differences between species, as the list of measured descriptors reflect both logistical constraints and hypothesis-driven questions. Moreover, contextual factors, such as the species pool size (SPS) and the spatial extent of the sampled area, could moderate species-environment associations through sampling effects and dispersal limitations. This study's objective was to quantify the influence of contextual factors (i.e., related to the circumstances in which the study was conducted) on the species-environment association strength on the basis of a synthesis of 156 models of forest bird communities. Our results reveal that factors related to the SPS and the number of independent environmental axes studied affect our capacity to detect selection, whereas spatial factors such as the study's spatial extent and latitude are less important determinants. The study context explains almost a third of the observed variation in the strength of the species-environment association. We conclude that strong species-environment associations can be found for properly designed studies of forest bird communities, which raises the question of whether ecologists have underestimated the importance of selection in community assembly processes.