PhD comprehensive seminar | Joe Pharaon, Anticipating critical transitions in coupled human-environment systems

Monday, August 10, 2015 1:00 pm - 1:00 pm EDT (GMT -04:00)

MC 6496


Joe Pharaon, Applied Math, University of Waterloo


Anticipating critical transitions in coupled human-environment systems


In the past few decades, it has become increasingly clear that human activity is the major driving force behind environmental deterioration, causing rapid climate change, habitat loss and deforestation.  Such changes can occur through a critical transition, defined as a dramatic shift in an ecosystem state over a relatively short time period, associated with a bifurcation of a dynamical system.  Researchers have studied how to use the changing statistical properties of stochastic dynamics preceding critical transitions to predict them ahead of time.  It is therefore of utmost importance to study critical transitions in the environment and their early warning signals. However, due to the pervasiveness of human influence, most environmental systems are actually part of a coupled human-environment system, where the human population and the environmental dynamics interact with one another.  Despite this, the impact of human feedback on environmental states on how the early warning signals of critical transitions manifest has not been explored.  In this seminar, I will describe previous research on coupled human-environment system (HES) models, and early warning signals of critical transitions in non-HES models that do not incorporate human feedback.  I will also present a model designed to explore early warning signals for critical transitions in coupled human-environment systems.  The model views human behavior as dynamic, changing according to social and economic norms, and strength of conservation values.  The model is parameterized using data on old-growth forests in the Northwest Pacific of the United States of America, and corresponding changes in conservation values over time. I propose to explore using changes in variance and lag-1 auto-correlation in time series as early warning signals, and to compare how those signals manifest in the coupled HES model, compared to a non-coupled model that lacks human feedback. . I hypothesize that human feedback “mutes” early warning signals, thereby making it more difficult to anticipate a critical transition.  Such research has the potential to improve our ability to predict, and possibly prevent, catastrophic environmental changes.