Changes in composition of the dominant eigenvalue in Portal control plots (colored lines) and known timespans of community shift (gray highlights)

Dynamic stability

One classical approach for stability involves studying the species interaction matrix, and determining whether the matrix produces asympotitic stability near an equilibrium. In Ushio et al. 2018, we extended this framework by using empirical dynamic modeling to quantify how the species interaction matrix changes over time. This then allows for estimation of stability as a changing quantity, or “dynamic stability”.

Ushio et al. 2018 describes and demonstrates the approach for a fish community in Maizuru Bay. Currently, I am investigating whether dynamic stability can be a useful guide for identifying abrupt shifts in ecological communities. Specifically, whether the approach can give insight into change in the rodent community at the Portal experimental site, and validating the method against the known shifts identified in Christensen et al. 2018.

Code and data for this project are available on GitHub and initial results were presented at the 2018 ESA meeting in New Orleans.