Observed Ecological Communities Are Formed by Species Combinations That Are among the Most Likely to Persist under Changing Environments

Despite the rich biodiversity found in nature, it is unclear to what extent some combinations of interacting species, while conceivable in a given place and time, may never be realized. Yet solving this problem is important for understanding the role of randomness and predictability in the assembly of ecological communities. Here we show that the specific combinations of interacting species that emerge from the ecological dynamics within regional species pools are not all equally likely to be seen; rather, they are among the most likely to persist under changing environments. First, we use niche-based competition matrices and Lotka-Volterra models to demonstrate that realized combinations of interacting species are more likely to persist under random parameter perturbations than the majority of potential combinations with the same number of species that could have been formed from the regional pool. We then corroborate our theoretical results using a 10-year observational study, recording 88 plant-herbivore communities across three different forest successional stages. By inferring and validating plant-mediated communities of competing herbivore species, we find that observed combinations of herbivores have an expected probability of species persistence higher than half of all potential combinations. Our findings open up the opportunity to establish a formal probabilistic and predictive understanding of the composition of ecological communities.

Automated summary

The study found that specific combinations of interacting species emerging from regional species pools are more likely to persist under changing environments. By using niche-based competition matrices and Lotka-Volterra models, it was demonstrated that realized combinations of interacting species are more likely to endure random parameter perturbations compared to potential combinations with the same number of species. This research provides insights into the probability and predictability of the composition of ecological communities.