Trait-based approaches to global change ecology: from description to prediction

Abstract

As global change forces species' ranges and abundances into novel configurations, traits-based approaches could allow predictions of community re-assembly. We present a quantitative review of traits-based research globally to (1) evaluate the extent to which this approach has been applied, and (2) evaluate moving from description and to prediction. We highlight the application of traits-based frameworks to describe ecological patterns; terrestrial plant morphology comprises >30% of the literature alone. But fewer than 3% of studies predict ecological effects of global change, mostly in the past five years. While organism size is the most common trait, we identified 2,430 other morphological, physiological, behavioural, and life history traits that mediate environmental filters of species' ranges across ecosystems and taxonomy. Global change studies forecast range shifts from a few physiological or life history traits. Though uncommon, spatially-explicit models constructed from correlated multivariate trait assemblages (or 'syndromes') offer the best chance of predicting shifts under global change scenarios. Moving the field towards trait-based prediction requires (1) matching the scale of trait measurement to the ecological processes, (2) increasing the resolution of environmental gradients along which traits are measured, (3) moving from single to multivariate traits, and (4) accounting for intraspecific trait variation.