Research

Interests:

- Chaotic/nonlinear ecological systems

- Evaluation of ecosystem models

- Nonparametric modeling

- Ecosystem stability/hysteresis/regime shifts

- Pattern formation

Current research:

I recently started a postdoctoral position at Oregon State University, where I will be building models to analyze oyster population dynamics, and comparing the model dynamics to available time series.

My ongoing research with my postdoctoral advisor at William & Mary focuses on using the topology (spatial patterns) of spatially extended populations to understand population dynamics, and perhaps even predict critical transitions in populations, such as extinction events. Topological changes in the population may provide early warning signs of an impending critical transition or extinction event. We use betti numbers to quantify spatial patterns in population distributions produced from models. We are also using computational topology to analyze time series of spatial data, such as GIS images.

Past research:

Dissertation: My dissertation focused on understanding the dynamics of a spatially extended chaotic population in a unidirectional flow. This is a good approximation for populations living in rivers, streams, or marine environments - anywhere with a strong dominant current! These populations exhibit interesting and varied responses to perturbations and changes in system parameters. Such responses are important to understand and anticipate if we want to successfully manage or conserve aquatic/marine species.

Masters thesis: My masters thesis focused on comparing the dynamics of ecosystem models with the dynamics of the populations that are being modeled. In this way we can assess how well a model is capturing the dynamics of a system. It's particularly important to assess fish population models, since the models are used to inform management decisions. This was accomplished by evaluating presence of nonlinearity in time series of fish populations, and comparing that with the presence of nonlinearity in model output time series (we evaluated presence of nonlinearity by performing lagged coordinate state space reconstruction on all of the time series). We found that while most fish time series were nonlinear, most model output time series were not. This means that models of fish populations are probably overestimating the stability of the system, which is not good for managing these populations in a sustainable way!