Evolution to alternative levels of stable diversity leaves areas of niche space unexplored

Abstract

One of the oldest and most persistent questions in ecology and evolution is whether natural communities tend to evolve toward saturation and maximal diversity. Robert MacArthur’s classical theory of niche packing and the theory of adaptive radiations both imply that populations will diversify and fully partition any available niche space. However, the saturation of natural populations is still very much an open area of debate and investigation. Additionally, recent evolutionary theory suggests the existence of alternative evolutionary stable states (ESSs), which implies that some stable communities may not be fully saturated. Using models with classical Lokta-Volterra ecological dynamics and three formulations of evolutionary dynamics (a model using adaptive dynamics, an individual-based model, and a partial differential equation model), we show that following an adaptive radiation, communities can often get stuck in low diversity states when limited by mutations of small phenotypic effect. These low diversity metastable states can also be maintained by limited resources and finite population sizes. When small mutations and finite populations are considered together, it is clear that despite the presence of higher-diversity stable states, natural populations are likely not fully saturating their environment and leaving potential niche space unfilled. Additionally, within-species variation can further reduce community diversity from levels predicted by models that assume species-level homogeneity.Author summaryUnderstanding if and when communities evolve to saturate their local environments is imperative to our understanding of natural populations. Using computer simulations of classical evolutionary models, we study whether adaptive radiations tend to lead toward saturated communities in which no new species can invade or remain trapped in alternative, lower diversity stable states. We show that with asymmetric competition and small effect mutations, evolutionary Red Queen dynamics can trap communities in low diversity metastable states. Moreover, limited resources not only reduces community population sizes, but also reduces community diversity, denying the formation of saturated communities and stabilizing low diversity, non-stationary evolutionary dynamics. Our results are directly relevant to the longstanding questions important to both ecological empiricists and theoreticians on the species packing and saturation of natural environments. Also, by showing the ease evolution can trap communities in low diversity metastable stats, we demonstrate the potential harm in relying solely on ESSs to answer questions of biodiversity.

Ecological diversity exceeds evolutionary diversity in model ecosystems

“…As nicely summarized by Edwards et al . [2018], communities not at an ESS will experience selective pressure towards the ESS [however, see Doebeli & Ispolatov, 2014; Doebeli et al ., 2017; Rubin et al ., 2021, for discussions of non-equilibrium evolutionary dynamics]. In evolutionary time, super-saturated communities will likely experience the extinction of certain species and collapse to the ESS [e.g., 1].…”

“…Asymmetric competition can lead to both more interesting ecological (unstable or cyclic coexistence [Hernández-García et al ., 2009]) and evolutionary dynamics in high dimensional phenotype space (e.g., Red-Queen dynamics [Rubin et al ., 2021] or evolutionary chaos [Doebeli et al ., 2017]), depending on the specific functional form of the competition kernel. While diversity patterns due to asymmetric competition is beyond the scope of this article, while investigating systems with non-stable (Red Queen) evolutionary dynamics we found that randomly assembling super-diverse communities was possible, but exceedingly rare [Rubin et al ., 2021], creating a possible analogy to the difficulties in creating saturated communities with symmetric competition. We have also described a related phenomenon where continual selection can trap communities in low diversity meta-stable states [Rubin et al ., 2021].…”

“…Here we chose to use a Gaussian competition function and a quartic carrying capacity function ( σ α = 0.6, σ K = 1 unless otherwise noted). This system is a commonly used model to study diversification [Doebeli & Ispolatov, 2014; Doebeli et al ., 2017; Rubin et al ., 2021] and is a mathematically simple and structurally stable description of diversification dynamics. Using a higher order carrying capacity (and thus comparatively flat-shaped) function compared to the competition kernel both naturally restricts the viable phenotype space to an area around the origin (roughly between –2 and 2 for the parameters chosen) and guarantees the presence of a branching point at the origin [Baptestini et al ., 2009; Doebeli, 2011], regardless of the choice of σ α .…”

“…As the evolutionary dynamics are not a focus of this paper, and we are using a standard model and technique, the evolutionary dynamics will not be explained in further detail. Please see Doebeli [2011]; Rubin et al . [2021] or the source code for these simulations included in the supplementary materials for information on the adaptive dynamics of this system.…”

“…As the evolutionary dynamics are not a focus of this paper, and we are using a standard model and technique, the evolutionary dynamics will not be explained in further detail. Please see Doebeli [2011]; Rubin et al [2021] or the source code for these simulations included in the supplementary materials for information on the adaptive dynamics of this system. Notably, while we used adaptive dynamics as we felt it was the fastest and most accurate way to determine the ESS of each system, a successive invasion simulation like the one used by D'Andrea et al et al [2018]; Sasaki & Ellner [1995] would deliver exactly the same ESS results.…”