Trait Number and Coding Shape Estimates of Functional Diversity

ABSTRACT: Functional diversity estimates increasingly inform ecological research, yet how methodological choices such as trait number and coding affect common metrics remains poorly quantified. Here, we systematically evaluate the effects of trait number and coding strategy on functional diversity metrics using benthic macroinvertebrate traits. Across 14 functional diversity indices representing functional richness, evenness, dispersion, and redundancy, we showed that metric responses to trait number are highly facet-dependent. Functional richness and evenness indices were particularly sensitive to trait number, whereas dispersion and redundancy metrics were comparatively stable. Estimation uncertainty was minimized at intermediate trait number, indicating a potential balance between functional space resolution and statistical robustness. Despite broad consistency between binary and fuzzy coding approaches for several metrics, redundancy metrics showed substantial divergence between coding schemes. These results demonstrate that hidden methodological decisions can substantially influence functional diversity indices. Our study provides a quantitative framework for evaluating metric robustness and reveals that widely used indices differ fundamentally in their response to dimensionality—a mathematical behaviour that must be understood before ecological interpretation. We recommend reporting trait number alongside functional diversity values and exercising caution when comparing communities assessed with different trait sets, particularly for richness and evenness metrics.