Monday, November 26, 2018, 12:15PM - 1:15PM
University of Regina
SEEC room S228 (Sievers Room)
Robert May observed that the simple deterministic equations ecologists use to model populations could, under certain circumstances, lead to intricately – and infinitely – complex behaviour. This observation kick-started the study of complexity in ecological systems, and the rich patterns of behaviour that complexity theory can explain. One particular avenue of interest is that of critical transitions, wherein a small perturbation to an ecosystem may lead to a catastrophic shift to a new regime of behaviour, a shift that is difficult to reverse. Such transitions have been identified in a range of systems: in transitions from grassland to desert; from clear-water to turbid states in shallow lakes; in the climate system over the Pleistocene. Recent work suggests that many of these transitions are proceeded by phenomena to warn of impending change. Broad evidence for critical transitions is however generally lacking; we simply don’t have available the time series to study ecosystems in the detail required. But maybe we do; perhaps lake sediment records offer one way to access time series of sufficient length and detail needed. In order to realize this goal we need to overcome some inherent problems with palaeo time series. In this talk I’ll outline what these problems are and why they’re an issue, and then I’ll show how we’ve been using generalized additive models – statistical models that fit wiggly functions to data – to develop models for generic resilience indicators from palaeoenvironmental time series.
Free and open to the public.