Pitfalls of ignoring trait resolution when drawing conclusions about ecological processes

Abstract

Aim Understanding how ecological communities are assembled remains a grand challenge in ecology with direct implications for charting the future of biodiversity. Trait‐based methods have emerged as the leading approach for quantifying functional community structure (convergence, divergence) but their potential for inferring assembly processes rests on accurately measuring functional dissimilarity among community members. Here, we argue that trait resolution (from finest‐resolution continuous measurements to coarsest‐resolution binary categories) remains a critically overlooked methodological variable, even though categorical classification is known to mask functional variability and inflate functional redundancy among species or individuals. Innovation We present the first detailed predictions of trait resolution biases and demonstrate, with simulations, how the distortion of signal strength by increasingly coarse‐resolution traits can fundamentally alter functional structure patterns and the interpretation of causative ecological processes (e.g. abiotic filters, biotic interactions). We show that 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. Main conclusions Inferences of assembly processes must account for trait resolution to ensure robust conclusions, especially for broad‐scale studies of comparative community assembly and biodiversity change. Despite recent improvements in the collection and availability of trait data, great disparities continue to exist among taxa in the number and availability of continuous traits, which are more difficult to acquire for large numbers of species than coarse categorical assignments. Based on our simulations, we urge the consideration of trait resolution in the design and interpretation of community assembly studies and suggest a suite of practical solutions to address the pitfalls of trait resolution biases.

The dimensionality and structure of species trait spaces

“…Conversely, considering redundant or correlated traits, even if meaningless, has no expected impact on dimensionality so can be very neutral in the building of species trait space and the computation of indices. Yet using surrogate traits or traits with a coarse resolution to describe a given dimension of ecological strategy can substantially affect the results (Kohli & Jarzyna, 2021; Loranger et al, 2016).…”

“…The environment can also be crucial in determining the course of multicellular evolution and organismal complexity, with aggregative multicellularity evolving more frequently on land whereas clonal multicellularity is more frequent in water (Fisher et al, 2020). On the other hand, the number of species and trait characteristics are likely to influence the complexity of species trait spaces beyond the type of organism and the environment (Kohli & Jarzyna, 2021; Zhu et al, 2017). Yet the relative importance of these different potential drivers has never been tested across kingdoms and realms for a vast number and diversity of traits and taxa.…”

“…A second key aspect of any species trait space is its robustness to the choice or the omission of traits so its capacity to consistently position species relative to each other whatever the sub‐selection of traits for a given goal (environmental filtering, competitive interactions etc.). This capacity ultimately determines the confidence by which we can estimate metrics like species trait dissimilarity or functional diversity (Carscadden et al, 2017; Kohli & Jarzyna, 2021; Zhu et al, 2017). However, this robustness has been largely overlooked and deserves a dedicated analysis across multiple datasets where the number, completeness, correlation and type of traits cover a broad range of options.…”

AVONET: morphological, ecological and geographical data for all birds

“…Another recurring theme is the idea that functional traits can help us to devise a quantitative framework for understanding and predicting ecological communities (Schleuning et al, 2020; Winemiller et al, 2015). However, unlocking the true potential of functional traits is highly dependent on comprehensive sampling at the species level, whereas coverage remains patchy for all major taxonomic groups, particularly for continuous morphological traits (Cernansky, 2017; Kohli & Jarzyna, 2021).…”

“…This creates problems for analytical approaches that assume coverage is complete, including phylogenetic comparative analyses and evolutionary models. Missing species in partially sampled trait datasets can radically alter the trait structure of communities and the fit of evolutionary models, reducing predictive power and restricting studies to a biased sample of well‐known clades (Kohli & Jarzyna, 2021; Weiss & Ray, 2019). The main obstacle to the completion of species sampling for plant traits is the sheer diversity of plants themselves.…”