Predicting 'tipping points' in humans and ecosystems
An international group of scientists looking at studies of critical thresholds in different complex systems such as humans, ecosystems and financial markets has concluded that regardless of the details, the dynamics of each system near its 'tipping point' share generic properties. Their review of early-warning signal modelling, published in the journal Nature, opens up new opportunities for connecting work on tipping-point phenomena across several disciplines. The work was funded in part by a European Young Investigator Award, a scheme run through the Sixth Framework Programme (FP6) to attract outstanding young scientists in all research domains from any country in the world to create their own research teams at European research centres. These awards have effectively been replaced by starting grants through the European Research Council (ERC). Predicting when a system will make a critical shift, triggering an asthma attack or a market crash for example, is tricky. Accurate models to predict thresholds in most complex systems are incipient, as in most cases scientists do not have a full understanding of all the relevant mechanisms and feedbacks. In this latest review, researchers in Germany, the Netherlands, Spain and the US examined early-warning signals in widely different systems to see whether they shared similar features. They looked at similarities between epileptic seizures and the end of glacial periods; desertification and asthma attacks; and climatic transitions and ecosystems. 'It's increasingly clear that many complex systems have critical thresholds -'tipping points' - at which these systems shift abruptly from one state to another,' the study reads. The team observed that some early-warning signals have a generic character, which, they say, suggests that these transitions may be somehow related to 'bifurcations', where universal laws of dynamical systems govern the pattern. In short, such signals may provide valuable information on whether the probability of a major event is increasing. Early warnings of abrupt transitions in human systems show patterns similar to those seen in nature, for example. 'In the case of asthma, it has been shown that human lungs can display a self-organised pattern of bronchoconstriction [when the airways narrow] that might be the prelude to dangerous respiratory failure, and which resembles the pattern formation in collapsing desert vegetation,' the study reads. Picking up early-warning signals in complex natural systems is challenging. But results from recent studies of 8 abrupt climate-change episodes (e.g. the 'greenhouse-icehouse' transition roughly 34 million years ago) show an increased correlation between reconstructed climate dynamics and critical transitions in climate. Including financial markets in the mix is difficult because discovering predictability in this field, the authors explain, leads to its elimination. Profit can be made from predictability, so patterns are quickly annihilated. However, there are early-warning signals that can be picked up using tools such as the 'fear index' (which measures volatility) or by examining patterns embodied in options prices, for example. 'These are compelling insights into the transitions in human and natural systems,' stated Henry Gholz, programme director in the Division of Environmental Biology at the National Science Foundation (NSF) in the US. 'The information comes at a critical time [...] when Earth's (and our) fragility have been highlighted by global financial collapses, debates over health-care reform and concern about rapid change in climate and ecological systems.' 'In systems in which we can observe transitions repeatedly, such as lakes, ranges or fields, and such as human physiology, we may discover where the thresholds are,' the authors say. A key issue for practical application, the study concludes, 'is the question of whether a signal can be detected sufficiently early for action to be taken to prevent a transition or to prepare for one'. More work is needed, they say, to find out how robust these signals are in various mathematical models.
Countries
Germany, Spain, Netherlands, United States