Professor Malcolm Bolton
Cambridge University Engineering Department
One of the tasks of the Civil Engineering research worker is to develop an understanding of the processes that underlie the various classes of natural disaster that may threaten the civil population or its infrastructure. The test of real understanding is that it should be useful in developing improved strategies or technologies by which hazards can be mitigated. However, the public funding of risk-reduction is a highly contentious matter. The legal, financial and political consequences of disasters may be severe but the costs of preventing them might be even higher. Precisely because there is a lot at stake, the introduction of new ideas or new technologies into an existing consensus may not be welcomed by the professional groups who are being paid to manage the risk.
There is also the philosophical problem of the roles of determinism and chance in setting public or professional policy in relational to rare events of great severity. In the IT era we can arrange to know a great a deal more than was previously possible - does that alter the balance of the argument? Does a focus on statistics (probability of collapse of a building in an earthquake of a given magnitude, probability of a landslide on a given slope in a given rainstorm, etc.) draw attention away from the correct understanding of mechanics in those collapses that do occur?
These issues are explored in relation to two questions - the design of foundations to resist earthquakes, and the stabilisation of loose soil slopes in regions prone to landslides following heavy rain. The discussion will be based around physical models observed in centrifuge tests, as compared and contrasted with the aftermath of real disasters. It will be shown in each case that the actual behaviour seems to contradict the consensus that had emerged unchallenged in the academic literature over the last 25 years. On the other hand, the observed behaviour was not very complicated; a proper description of the mechanisms could have been included in a fourth year undergraduate module.
The colloquium will conclude with contrasting visions of what might be done in response to the two classes of problem that have been introduced. Money for re-engineering existing structures could simply be redirected on the basis of the new understanding - but laws and design codes militate against this. Furthermore, no deterministic evidence at model scale can ever be taken to adjudicate conclusively on the causes of rare, full-scale catastrophes. Model conditions might be different in some subtle but significant way. For example, undiscovered scale distortions might have led to misleading observations, or the soil sample in the tests might not have been representative of the in situ conditions of those rare soils on sites where disaster struck. Perhaps money should be spent on extending the research, before it is spent on the assumption that the research is right.
