Predicting the Unpredictable: The Tumultuous Science of Earthquake Prediction

Seismologist Ian Main on why more lives are saved by preparation than prognostication

June 17, 2010

Despite all the advances in scientific understanding, destructive earthquakes such as the recent ones in Haiti, Chile and China still strike suddenly and without warning. Why is this? Will it always be so? What should we do about them?

In this forensic and engaging overview, Susan Hough presents a frank, entertaining and personal review of the history of ideas, practice, personalities and experience in the science of earthquake prediction. Although Hough is a respected scientist, she takes a journalist's viewpoint here, not shying away from legitimate criticism of those she regards as friends, and taking on the credulous at the edge of, or even beyond, the mainstream scientific community.

Recent images on our TV screens highlight a much bigger issue: earthquakes don't kill people directly, buildings do, particularly heavy masonry structures. But anti-seismic design, based on careful use of known best practice in hazard mitigation and engineering, simply doesn't grab the headlines. It may be what we actually need in earthquake-prone countries, but "it isn't what we want", says Hough - we want prediction.

Sadly, placing narrow bounds on the time, place, depth and magnitude of a future event has proved an elusive goal. Instead, the history of earthquake prediction research is a cautionary tale in the highly complex and quite unreasonable (non-linear) behaviour of the Earth, the dangers of conflating science with advocacy, the difficulties of working at the limits of scientific understanding and uncertainty, and the mismatch between what the wider public wants (and increasingly thinks science is for) and what we can actually do. Hough's book is a triumph in deconstructing what has been a frustrating experience for many.

The only compelling predictive signal, of many examples covered in the book, is the tendency of earthquakes to cluster in time, as aftershocks or as more generally triggered events. The best example is the "predicted" 1975 earthquake in Haicheng, China. Hough provides an illuminating critique of just how messy the process actually was - with large variations in local authority levels of alert, questionable scientific methods of interpretation influenced by local traditions and beliefs, and a political environment unconducive to independent scrutiny. It remains the only "successful" prediction of a major event in terms of evacuation, although the high level of preparedness by the local population, and a cluster of shocks sufficiently frightening in their own right immediately prior to the event, probably played a large part.

One of the strongest aspects of the book is its human dimension, with "only a few heroes" and "no villains", although Hough does identify some cases where over-optimistic prediction statements have been used, consciously or not, to attract funds that might have been better spent on more basic hazard-reduction programmes. A significant minority of researchers have been guilty of a lack of self-criticism, to the point of using methodology known to be flawed in other aspects of predictive science, the most obvious being the lack of appreciation of how data selection, often with the benefit of hindsight, can strongly bias results.

In blunter terms, Richard Feynman once cautioned against scientists "fooling themselves", and Irving Langmuir warned against avoiding a known pattern of behaviour in the "science of things that aren't so". The jury is out on the eventual destination, but earthquake prediction research is quietly moving back to the mainstream by recasting even its name as a question - earthquake predictability - by opening new frontiers in satellite and borehole observation, and by adopting careful "blind" testing of clearly stated falsifiable hypotheses, as used elsewhere in science and medicine. The jury is out on the eventual destination, but in the meantime we really should consider, urgently, what we can actually do if the answer is not what we really want.

Predicting the Unpredictable: The Tumultuous Science of Earthquake Prediction

By Susan Hough
Princeton University Press, 2pp, £16.95
ISBN 9780691138169
Published 13 January 2010

You've reached your article limit.

Register to continue

Registration is free and only takes a moment. Once registered you can read a total of 3 articles each month, plus:

  • Sign up for the editor's highlights
  • Receive World University Rankings news first
  • Get job alerts, shortlist jobs and save job searches
  • Participate in reader discussions and post comments

Have your say

Log in or register to post comments