Primary Research Focuses

Complex Feedbacks in Ecological and Evolutionary processes

My research program uses models to tease apart the affects of eco-evolutionary processes on species invasions and range expansions. Previous projects have investigated how native plants might evolve in response to high density invasive plants, and how mechanisms like gene surfing and spatial sorting affect variance in range expansion speed for Tribolium castaneum (flour beetles). Future work will use similar methods on other species beyond flour beetles, in order to understand what ecological and genetic factors differences in expansion speed variance. Another future project is to develop individual-based models with more than one species to explore how evolution in competitive ability affects spatial sorting and gene surfing.

Early Warning Signals for Multi-species Communities

When passing tipping point, a species may suddenly go extinct, a grassland may suddenly become desert, or lake water may change from clear to cloudy as the population of certain microbes explodes. Tipping points are threshold values of slowly changing variables at which a system changes drastically. As human pressures on the environment increase over time, the idea that stable systems may be approaching tipping points is of increasing concern. My research program explored how to improve early warning signals (EWS) of tipping points, by determining why they occur unequally for species in a multispecies system (Patterson et al 2021). This led to guidelines about which species managers should observe in order to detect approaching tipping points. In the future, my research program will expand on this work by better developing theory on where random variation, or noise, originates in complex systems. Understanding how noise inputs in a system differ among species is another unexplored dimension of the questions of which species is best to monitor for EWS.

Adaption of sparse Bayesian methods for time series data

With the increasing quantity of big data in ecology, from satellites to modern sensors, we need new approaches to turn this data into predictive models. Unfortunately, many popular machine learning techniques are treated as “black box” approaches in ecology- able to make predictions with large amounts of data at the cost of model interpretability. Sparse modeling and regularization offer another way to deal with large data sets with many parameters. My research program is exploring the use of sparse modeling and Bayesian regularization for use with big data sets and time series data. In the future, my research program will use sparse approaches and available data on invasive species to determine which factors predict when a species will become invasive.

See Publications