Characterizing tree trait variance over spatiotemporal scales

Abstract

Beyond the study of the mean, functional ecology lacks a concise characterization of trait variance patterns across spatiotemporal scales. Traits are measured in different ways, using different metrics, and at different spatial (and rarely temporal) scales. This study expands on previous research by applying a ubiquitous and widely used empirical model-Taylor's Power Law-to functional trait variance with the goal of identifying general patterns of trait variance scaling (the behavior of trait variance across scales). We compiled data on tree seedling communities monitored over 10 years across 213 2 m² plots and functional trait data from a subtropical forest in Puerto Rico. We examined trait-based Taylor's Power Law at nested spatial and temporal scales. The scaling of variance with the mean was idiosyncratic across traits suggesting that the drivers of variation are likely to differ across traits that may make variance scaling theory elusive. However, slopes varied more in space than through time, suggesting that spatial environmental variability may have a larger role in driving trait variance than temporal variability. Empirical models that characterize taxonomic patterns across spatiotemporal scales, like Taylor's Power Law, can provide an insight into the scaling of functional traits, a necessary next step toward a more predictive trait-based ecology.

Keywords: Puerto Rico; Taylor's Power Law; seed mass; seedling censuses; specific leaf area; wood specific gravity.