Files
Abstract
Ecologists have long been interested in understanding the causal links between environmental stochasticity and population abundance. Most theories developed to explain these dynamics use a deterministic representation of the natural world where substantial changes of a population’s size occur through one of two paths as explained by alternative stable states. Either the system moves into another stable state due to an external push, or the equilibrium population size itself changes as a kinetic parameter changes. However, if the natural world is recast using a stochastic representation, a third path arises: noise-induced transitions (NITs). By taking stochasticity in the natural world into account, the location and/or number of equilibrium population sizes can change. NITs have been demonstrated in chemistry and physics but have largely been ignored in ecology.
This dissertation consists of two combined theoretical/empirical studies to demonstrate NITs in biological populations. First, I demonstrate the destabilizing effects of environmental noise through a noise-induced extinction—as expected under the ecological dogma of stochasticity—using chemostats of the cyanobacteria, Aphanizomenon flos-aquae. Then, I present the surprising result of noise-induced persistence in stochastically harvested populations using microcosms of Saccharomyces cerevisiae. This dissertation is the first step in building a concept map of noise-induced outcomes given different combinations of nonlinear dynamics and types of environmental noise. Building upon these results, we can develop a heuristic understanding of how and when a stochastic rather than deterministic model is a more appropriate description of the natural world.