Rethinking the nature of intraspecific variability and its consequences on species coexistence

Abstract

Intraspecific variability (IV) has been proposed as a new track to explain species coexistence. Previous studies generally assumed, although often non-explicitely, that IV results from intrinsic differences between conspecific individuals. Under this hypothesis, IV either (i) blurred species differences, thus promoting transient or unstable coexistence, (ii) disproportionately advantaged the strongest competitor, thus hindering coexistence, or (iii) promoted coexistence. Based on a body of evidence, we here emphasize that IV does not necessarily imply inherent differences among conspecific individuals, nor species niches overlap. Conspecific individuals can also differ in their measured attributes due to differences in the micro-environment they thrive in. If this source of variation is predominant, which we suggest it is in many cases given the variation of the environment in numerous dimensions at fine scales and the observed spatial structure of IV, it implies that conspecific individuals respond more similarly to environmental variation than heterospecific individuals, thereby concentrating competition within species -- a necessary condition for species coexistence. Importantly, as the number of environmental dimensions that are typically observed and characterized in the field is generally much lower than the actual number of environmental dimensions that influence individual attributes, a great part of observed IV can often be misinterpreted as random variation across individuals while being environmentally-driven, with radically different consequences for community dynamics. We call for new studies exploring observed IV as an outcome of species-specific responses to high-dimensional environmental variations that can lead to inversions of species hierarchy in space and time promoting stable coexistence.

Beyond variance: simple random distributions are not a good proxy for intraspecific variability in systems with environmental structure

“…Structuring variation: a first step towards accurate representation of multidimensionality Our virtual experiment corresponds to an extreme case in which conspecific individuals have exactly the same response to environmental variation, which is unlikely to be the case under the joint effect of environmental and genetic variation in the field. However, while the intraspecific variation that is added as a noise around species means is not structured in space and time, IV, whether it is environmentally-or genetically-driven or both, is actually highly likely to be structured in space and time (Girard-Tercieux et al 2022). This structure could appear when IV results from spatially-structured environmental variables or from limited dispersion or local adaptation (Marrot et al 2021;Schmitt et al 2021;Westerband et al 2021).…”

“…the ones simulated with the Perfect knowledge model of individual performance). Thus using independent random draws is not a relevant approach to represent intraspecific variability that is often environmentally-driven (Girard-Tercieux et al 2022), and progress must be made to take into account the environment-species interactions in models.…”

“…These studies have led to contrasting results however, letting the debate unresolved: IV could either (i) blur species differences, thus promoting transient or unstable coexistence (Vieilledent et al 2010; Crawford et al 2019), (ii) disproportionately advantage the strongest competitor, thus hindering coexistence (Courbaud et al 2012; Hart et al 2016), or (iii) promote coexistence in specific spatial configurations (Uriarte and Menge 2018). While a unifying framework differentiating whether IV affects niche traits or hierarchical traits has been recently proposed to explain these discrepancies (Stump et al 2021), a major assumption usually made in modeling studies, namely that IV is unstructured in space or time and can be represented by independent random draws around species-specific mean parameters, remains largely undiscussed (Girard-Tercieux et al 2022).…”

“…Differences in attributes among conspecific individuals can result from differences in unobserved environmental dimensions, and do not necessarily mean that conspecific individuals substantially differ in their response to the environment. While it is widely accepted that environmentally-driven structured IV (sIV) is ubiquitous in natural communities (Nicotra et al 2010), the consequences of its substitution by unstructured IV (uIV) on species coexistence and community dynamics remain to be thoroughly tested in models (Clark 2010; Girard-Tercieux et al 2022).…”

“…In particular, high-dimensional (and often unobserved) variation of the environment can lead to large observed IV. For instance, in a highly controlled clonal experiment, IV in tree growth within clones was larger than genetically-driven IV between clones (Girard-Tercieux et al 2022). Differences in attributes among conspecific individuals can result from differences in unobserved environmental dimensions, and do not necessarily mean that conspecific individuals substantially differ in their response to the environment.…”