Predicting long and short-term species persistence following invasion of ecological communities

Large ecological communities are underpinned by a complex web of interactions whose exact structure, strength and signs are tremendously elusive. While many large-scale patterns in such communities are well-understood and predictable, the outcomes of processes such as invasions are still difficult to analyze. Given that the assembly history of communities is marked by invasion events at many points in time, it is useful to identify those aspects of invasions that can be reliably predicted even if the exact invasion outcomes cannot be determined. We introduce the notion of proximate uninvadable systems and use these to develop a framework for predicting the structural outcomes of invasion events, i.e., what species are present/absent in the eventual equilibrium. The method is particularly illuminating in large ecological communities and applies even when the invading species is initially abundant. We test this method on a broad class of settings and also demonstrate its robustness against imperfect knowledge of species interactions. Using an example of a large food-web from peri-Alpine lakes, we show how this framework can be applied to systems with fluctuating species abundances. Given that these systems exhibit large fluctuations for prolonged periods of time, we make forecasts for extinction risk in the short term thereby extending the purview of our predictive apparatus.