Neutral or non‐neutral communities: temporal dynamics provide the answer

Abstract

Neutral models for community dynamics are based on the assumption that temporal variation is only influenced by demographic stochasticity, ignoring the influence of fluctuations in the environment. Demographic and environmental stochasticity are known to have very different effects on the population dynamics of single species. Communities, however, are usually analyzed with emphasis on the shape of the species abundance distribution rather than on their dynamical characteristics. Accordingly, examination of the validity of the neutral model has typically been based on statistical analyses of the shape of the species abundance distribution or form of the species–area relationships. Here we illustrate that both demographic and environmental noise must be included to obtain realistic description of community dynamics. We do this by a simple approach comparing two different dynamic models both leading to Fisher's log series species abundance distribution: the neutral model with only demographic noise, and a limiting form of an approximate gamma model most strongly affected by environmental noise. While keeping Fisher's diversity parameter α constant, the turnover rates for observable species become much greater and average life‐times of species much smaller in the gamma model than in the neutral model. In practice, the neutral model tends to produce unrealistically small turnover rates for large populations. This illustrates that realistic assumptions about the stochastic components of the underlying population dynamics of single species are important for correctly characterizing dynamical changes in species composition of communities. This approach can provide a powerful tool for identifying the principal factors affecting the structure of communities.

Spatial covariation of competing species in a fluctuating environment

“…Thus, we approximate the density in the noise term by the equilibrium density, Q i , and study relative deviations from this density, z i ðxÞ ¼ n i ðxÞ=Q i À 1. It has previously been shown that approximate results based on a small noise assumption and linearization in models similar to this one are remarkably accurate under moderate and even fairly large population density fluctuations (Engen et al 2017). The linearized version of Eq.…”

Demographic stochasticity alters expected outcomes in experimental and simulated non‐neutral communities

“…Of the few studies that do consider finer-scale outcomes such as abundances over time, most examine the effects of demographic stochasticity alongside other factors such as dispersal (Alexander et al 2012), priority effects (Mertz et al 1976, Fukami 2004 or density manipulations (Svensson et al 2018). These approaches are incomplete because even with fine experimental control, the effects of demographic stochasticity are difficult to isolate from other drivers of variability without ample time-series data (Vellend et al 2014, Engen et al 2017, and the application of detailed stochastic population models (Melbourne and Hastings 2008).…”

On the proportional abundance of species: Integrating population genetics and community ecology

“…We show that when b(x), d(x) and c(x) have a particular functional form (see Eq. (12)(13)(14)) the invariant distribution is a beta distribution (Eq. (15)).…”

“…The recent development of neutral approaches to the study of ecological systems [7][8][9][10] have provided a renewed emphasis upon the value of theory and stochasticity in ecology [11][12][13][14] and a locus for the further integration of genetical and ecological theories 15,16 . By merging the mathematical and statistical tools developed by population geneticists with the neutrality approach, neutral theory in ecology allows us to better understand the factors affecting the abundance and distribution of species [15][16][17][18][19] .…”