Seasonal species interactions minimize the impact of species turnover on the likelihood of community persistence

Many of the observed species interactions embedded in ecological communities are not permanent, but are characterized by temporal changes that are observed along with abiotic and biotic variations. While work has been done describing and quantifying these changes, little is known about their consequences for species coexistence. Here, we investigate the extent to which changes of species composition impacts the likelihood of persistence of the predator-prey community in the highly seasonal Bialowieza Primeval Forest (NE Poland), and the extent to which seasonal changes of species interactions (predator diet) modulate the expected impact. This likelihood is estimated extending recent developments on the study of structural stability in ecological communities. We find that the observed species turnover strongly varies the likelihood of community persistence between summer and winter. Importantly, we demonstrate that the observed seasonal interaction changes minimize the variation in the likelihood of persistence associated with species turnover across the year. We find that these community dynamics can be explained as the coupling of individual species to their environment by minimizing both the variation in persistence conditions and the interaction changes between seasons. Our results provide a homeostatic explanation for seasonal species interactions, and suggest that monitoring the association of interactions changes with the level of variation in community dynamics can provide a good indicator of the response of species to environmental pressures.

💡 Research Summary

This study investigates how seasonal changes in species interactions mitigate the impact of species turnover on the likelihood of community persistence in a predator‑prey network from the Białowieża Primeval Forest in northeastern Poland. The authors compiled presence‑absence data and diet information for predators (carnivores and raptors) and their prey across two distinct seasons—summer (April‑September) and winter (October‑March). In total, 21 predators and 128 prey were recorded for summer, and 17 predators and 127 prey for winter; all winter predators were also present in summer, while a subset of prey species was seasonal. The resulting interaction matrices (γ) capture which predator consumes which prey, with interaction strengths parameterized as γ₀ divided by the number of predators sharing a prey (dₖ) raised to a resource‑partition exponent δ.

To assess community persistence, the authors adopt a structural‑stability framework. They model the dynamics with a generalized Lotka‑Volterra consumer‑resource system, where predator mortality (mᵢ) and prey intrinsic growth (αₖ) are the only free parameters once γ is fixed. Positive, globally stable equilibria exist if the vector of mortality and growth rates lies within the feasibility domain defined by linear equations linking m and α to γ and equilibrium biomasses. The size of this domain, denoted Ω(γ), is interpreted as the probability that a randomly drawn