Abstract
The spread of invasive species has severe ecological and economic consequences, making it critical to understand the processes driving range expansion. A contemporary challenge in invasion ecology is to understand mechanisms generating observed spatial and temporal patterns of spread, which are inconsistent with classical models that predict rates of spread that are constant through space and time. Using the invasion of North America by the gypsy moth, this dissertation address four questions concerning mechanisms underlying spatiotemporal variability in the spread of invasive species, each representing a separate chapter: 1) Are observed temporal fluctuations in gypsy moth invasion rate driven by high-amplitude population fluctuations in established populations? 2) Does topography, by altering reproductive timing, affect population growth rates in spreading gypsy moth populations? 3) Does the configuration of gypsy moth habitat on the landscape affect gypsy moth mate finding and population growth? And 4) what general patterns of invasion dynamics result from spatial and temporal heterogeneity in environmental conditions?
In the first study, I combined time series analysis with a simulation model to find that yearly fluctuations in the invasion rate of the gypsy moth were driven by cycles in established populations, which affected the number of immigrants arriving to nascent populations near the invasion front via long-distance human-mediated transport. In the second study, I integrated empirical data and simulation models to demonstrate that the growth of nascent gypsy moth populations is strongly slowed by increases in elevation, and slowed more modestly by increases in the elevational variability (hilliness) of landscapes. These reductions in population growth are driven by topographically induced changes in the timing of reproductive maturation that result in mating failure. For example, moths developing at low elevations reach reproductive age before moths at higher elevations, isolating potential mates in time. Mating failure may also result from spatial isolation of potential mates; for example, boundaries between forest and non-forest habitat may act as barriers to movement. In the third study, I evaluated whether mating failure may be increased in landscapes where forest habitat is patchily distributed. Field experiments showed that gypsy moths strongly resist leaving forest patches and that mate-location probabilities decay more quickly in the non-forested matrix than in forested habitats. A simulation model predicted that increased spread rates would accompany increases in the abundance of forest habitat and the connectivity of forest patches, and an empirical analysis of gypsy moth spread rates was consistent with the model predictions. The second and third studies are among the first to show how environmental heterogeneity affects the severity of the Allee effect, a phenomenon causing slow population growth and extinctions in small populations. In the fourth study I used a theoretical model to investigate general responses of invasions to spatial heterogeneity in the severity of the Allee effect. Patterns of spread in landscapes with varying Allee effect severity depended on the spatial configuration of Allee variability, and depended also on an interaction between long-distance dispersal and the severity of the Allee effect.
This dissertation makes several contributions to applied and theoretical invasion ecology. These studies support ongoing efforts to slow the spread of the gypsy moth by improving predictions of risk of future gypsy moth spread. My findings also suggest that range expansion may be slowed by suppression of gypsy moth outbreaks. More generally, this work provides novel insights into two potentially widespread mechanisms causing variations in the severity of Allee effects, and suggests that such variations are an underappreciated source of variability in the dynamics of biological invasions.
