The Detection of Tipping Points in Complex Dynamical Systems
|Speaker: Dr. Michael Harre, Complex Systems Research Group, Sydney University|
|When: Tuesday 12th Aug 2014, 1.00pm-2.00pm|
|Location: Biology tearoom E8A 280|
Abstract: “Tipping points” are the sudden sharp changes in a system's dynamics, for example market crashes in finance, species extinctions in ecosystems, epileptic seizures in brain dynamics and abrupt changes in the global climate. There are two important questions we often want to be able to answer about these changes: How close is the system to a tipping point? and can we navigate around a tipping point so that changes, if we cannot reverse them, at least occur more smoothly over a longer period of time? These are key questions because it is not always so important where we end up but it is sometimes the path we take to get there that matters. Taking climate change as an example a sea level rise of 2 to 4 meters, shifting weather patterns reducing crop outputs or the loss of species diversity are more problematic if they occur so quickly that we cannot adapt to the new environmental conditions. Taken over thousands of years the geographical distribution of a culture can change as can the location of their resources. Likewise over millions of years species die out while others can adapt and fill their vacated ecological niches. When these changes occur over mere decades or centuries though, the loss of biodiversity, the moving or restructuring of our cities, the changing of our crop locations and the infrastructure that supports them will cost us dearly in terms of our economic and social wellbeing.
This talk is based on two recent articles in which my collaborators and I have argued that there are sound theoretical reasons to believe there are general techniques for detecting tipping points and nudging system parameters in order to avoid the most catastrophic effects of a system's collapse. These results will be discussed in terms of recent reviews of the different types of tipping points that have been proposed in climate models.
Bio: Dr. Michael Harre is a researcher with the Complex Systems Research Group within the Faculty of Engineering at Sydney University. His main research interests are in the data-driven analysis of coupled dynamical systems, bifurcations and Catastrophe Theory. He and his collaborators have recently applied techniques from information theory to the problem of predicting abrupt changes in one of the most well studied models in physics, the 2-D Ising model, as well as detecting behavioural transitions in expert skills and the dynamics of financial markets. His current goal is to broaden these results to general non-linear transitions in complex dynamical systems.