|
DPRI Kyoto and Cabot Institute |
Over the last two days a group from the Cabot Institute has been holding a workshop with colleagues from Kyoto University’s Disaster Prevention Research Institute (or DPRI) on the topic of probabilistic hazard analysis. On the face of it Japan and the UK are very similar: highly urbanised and complex island societies with high population densities and therefore the potential for serious disruption if natural hazards occur. Mind you, the earthquake, tsunami and volcano hazards do put Japan in a different league when it comes to potential impacts. In both countries, robust hazard analysis, planning and decision making is therefore essential to protecting society. Both countries have a lot to learn from each other, and our recent paper on lessons for the UK from the Fukushima disaster is a case in point.
Cabot members Wendy Larner, Colin Taylor, Susanna Jenkins, Jeremy Phillips, Katsu Goda, Philippa Bayleyand myself (Paul Bates) spent two days working with around 30 Japanese colleagues, with Skype presentations from the UK delivered by WillyAspinall, Jonty Rougier and Tamsin Edwards. A full programme of the meeting is on our website, and includes pdfs of the presentations for download. We learned a huge amount about hazard research in Japan and have hopefully begun a large number of research collaborations that will be important for Bristol University for many years. Our profound thanks go to our Japanese hosts Prof. Eiichi Nakahita and Prof. Hirokazu Tatano, and to the Director of DPRI, Prof. Nakashima. The photos here give a flavour of the wonderful time we had.
Possibly the most important theme to emerge from the workshop was that whilst probabilistic analysis of hazards (where we give the chance of an event occurring rather than a definite yes/no prediction) is now commonplace in science, there is still a major issue in educating decision makers, governments and the public in how to use such forecasts to take decisions. Indeed the Daily Mail in the UK has recently been giving the Met Office a hard time for wanting (very sensibly) to move to a probabilistic forecast of rainfall. This shouldn’t be such a big problem, but the fact that it is tells us an awful lot. Intrinsically people deal with probability information all the time: betting and insurance, for example, are both examples of probabilistic contracts that are well understood by the public. So why do we resist being told about other risks in a similar way. My gut feeling is that it is to do with the question of responsibility. A probabilistic forecast of risk forces the decision maker (be they Ministers, civil servants or the public) to deal with uncertainty in predictions, whilst insistence on a deterministic forecast puts the responsibility for this onto the scientists who can then be blamed if things go wrong.