To explore how environmental variability may create non-random community structure, we simulated the assembly of model communities under varying levels of environmental variability. We assembled communities by creating a large pool of randomly constructed species, and then added species from this pool sequentially, allowing extinctions of invading and resident species to occur until the community became saturated. Because much current research on community structure focuses on single trophic levels, we constructed species pools consisting only of competitors. To compare with more realistic communities, we also created species pools with multiple trophic levels. For both types of communities, following assembly we calculated a variety of metrics of community structure, and five measures of community stability.
Communities assembled under high environmental variability had fewer species, fewer and weaker interactions among species, and greater evenness in abundance of persisting species. For single trophic-level communities, community size was dictated primarily by competitive exclusion. In contrast, for multiple trophic-level communities, community size was increasingly limited by dynamical instabilities as environmental variability increased.
Differences in community structure resulting from assembly under high environmental variability led to differences in community stability. According to two measures of stability related to population variability – the characteristic return rate to equilibrium and the coefficient of variation in individual species densities – stability increased for communities assembled under high environmental variability. In contrast, three additional measures of stability that are not directly related to population variability showed a variety of patterns, either increasing, decreasing, or remaining constant. Thus, communities assembled in highly variable environments are not necessarily generically more stable.
Our results demonstrate that environmental variability can structure communities and affect their stability properties in non-trivial ways. Thus, when making predictions about the response of communities to future extinctions or environmental degradation, account should be given to the forces responsible for community structure.