Improving trait data resolution to realize the promise of the functional trait revolution

Brooks A. Kohli and Marta A. Jarzyna, Evolution, Ecology and Organismal Biology, The Ohio State University, Columbus, OH, Brooks A. Kohli, Biology and Chemistry, Morehead State University, Morehead, KY, Marta A. Jarzyna, Translational Data Analytics Institute, The Ohio State University, Columbus, OH

Background/Question/Methods
Understanding how ecological communities are assembled is critical for identifying general rules in ecology and effectively tracking and predicting global biodiversity change. Trait-based methods are now widely used to quantify community structure and diversity patterns, fueled by a wave of efforts to compile and share huge quantities of species-level functional trait data across the tree of life. However, the most well-sampled and widely used traits are coarse functional categories while fine-resolution (continuous) traits with high species coverage are few. Because categorical trait classification inflates the perception of functional redundancy among taxa, reliance on coarse resolution traits may introduce predictable biases and mislead conclusions about the causes of community structure. To date, the direction and magnitude of these potential biases are unknown. We present the first detailed predictions of trait resolution biases and conduct simulations to assess the role of trait resolution in correctly identifying community assembly processes (e.g., abiotic filters, biotic interactions). We further investigate how the number of traits, pool size, community size, and functional diversity index mitigate or intensify the bias associated with trait data resolution.
Results/Conclusions
We show that increasing trait coarseness can obscure the leading processes of community assembly. Coarser trait data impart different impacts on the signals of divergence and convergence, implying that the role of biotic interactions may be underestimated when using coarser traits. Furthermore, in some systems, coarser traits may overestimate the strength of trait convergence, leading to erroneous support for abiotic processes as the primary drivers of community assembly or change. Based on our simulations, we urge the consideration of trait resolution in the design and interpretation of community assembly and biodiversity change studies. For example, comparisons among studies and taxa should account for possible biases if functional diversity was quantified using traits of differing resolution. Our results suggest that modest gains in trait resolution and number can help avoid methodological biases, as convergence and divergence in all diversity indices are most obscured when using binary categorical traits (the coarsest possible resolution). However, in light of similar effects from the other major facet of trait resolution, namely individual-level trait measurements versus species-level means, we advocate for the continued development and collection of individual-level continuous traits to achieve the most robust solution to the problem of trait resolution bias.