Does accounting for within-individual trait variation matter for measuring functional diversity?

Abstract

Trait-based approaches to disentangle assembly processes in ecological communities typically rely on average trait values obtained from the literature or databases. Recently, ecologists have shown growing concern in accounting for intraspecific variation in trait-based metrics. Besides intraspecific variation, plants and animals exhibit functional trait variation within the same individual (within-individual variation), which may exceed intraspecific variation and constitute a functional trait itself, influencing ecosystem functioning through individual performance variation. However, the role of within-individual variation in functional diversity (FD) remains unexplored. Here, we used sequential foraging observations of four bird species to quantify the magnitudes of interspecific, intraspecific and within-individual functional trait (foraging item, maneuver and stratum) variation. Then, we estimated functional richness using different hierarchical levels of increasing data resolution: (1) average trait values (based on our own literature search, on a global dataset, on the first observation of each foraging sequence and on complete foraging sequences), (2) average trait values plus intraspecific trait variation, and (3) average trait values plus intraspecific and within-individual trait variation. We also performed a series of simulations accounting for different levels of within-individual trait variation. For the empirical data, both intraspecific and within-individual variation accounted for more than 84% of total variation in functional traits, and for one trait (foraging maneuver) within-individual variation accounted for 84.31%. Although all FD metrics showed significant positive correlations, their magnitude consistently decreased when intraspecific and within-individual variations were taken into account (Pearson's correlations from 0.99 to 0.28). Simulations also showed that not accounting for within-individual variation strikingly underestimated functional richness, even at the lowest levels (<5%) of within-individual variation. Our results reveal that within-individual variation may represent a major source of functional trait variation. Overall, the inclusion of within-individual variation in trait-based approaches would improve our understanding and use of FD estimators and determine to what extent it matters for assembly processes.