Matching the scale of trait measurement to the ecological processes disrupted by global change
Ultimately, trait-based analyses of ecological variance strive for generality in explaining ecological processes, for example size-based relationships in predator-prey interactions (Rice et al. 1993) or thermo-tolerance in species responses to environmental gradients of change (e.g. using naturally occurring climate resilient corals to construct bleaching-resistant nurseries; Morikawa and Palumbi 2019). However, relevance and specificity drive the need to down-scale trait-based approaches to the system and questions, while maintaining scalable analytical frameworks. The resolution at which traits are measured and associated with the constituents of biological communities (i.e. by individual, life stage, population, or species) determines the level of inference that can be achieved in evaluating their role in mediating outcomes across environmental and ecological contexts. We suggest that the scale at which traits should be measured depends on 1) the ecological scale of the research question (individual, population, community, or ecosystem), 2) the focus on ecological patterns or process, 3) the extent to which intra- and inter-specific variation is hypothesized to play an important role in mediating the focal environmental filtering process in question, and 4) the analytical method used to generate inference.
Trait variation can be measured for a community of organisms or at the species-level, as well as at subpopulation to individual level. It must be clear whether intra-specific variation in trait values within species or groups of taxa that share measurable traits is the required currency for predictive modeling. But species’ trait information is typically available to analysts at either broad (species- or metapopulation-) scales or very specific geographic scales, with a rate-limiting middle ground of information on spatially and temporally explicit subpopulations. Where the consequences of species interactions are the focus of investigation, and food web traits are conserved at a species or population level, trait-based linkages between species groups may be most appropriately coded at that level (i.e., one value per species or population). Trait variation can be measured at a subpopulation to individual level, where intra-specific variation in trait values within species or groups of taxa that share measurable traits is the required currency for a model. For example, across large longitudinal or latitudinal gradients of change that span multiple species’ subpopulations. However, traits of interest were typically represented using a single value/level per species in the predictive global change studies identified here (Figure 7), regardless of study type, ecosystem, or taxonomic focus.