A few remarks on theme 9

Sir James Lighthill †
University College London, Gower Street, London WC1E 6BT, United Kingdom.
Tel.: +44-171-3 87 70 50, Fax: +44-171-3 83 55 19

From one highly important point of view, I agree that all early warning systems for natural disasters share a common characteristic. It is that the populations under threat need to be continually aware of all the potential dangers, and given detailed training on how it will be necessary for them to respond to different categories of warning. Yet from other points of view there are some genuine differences in the relationship between forecasting authorities and the public for different types of natural disaster. I will illustrate them from the types connected with the dynamics of the earth’s fluid envelope (atmosphere, oceans, rivers, groundwater). Inevitably, the public’s approach to any problem takes one form in relation to wholly harmful natural phenomena like coastal flooding (whether by storm surge or tsunami) or like exceptionally strong winds, but quite a different form in relation to flooding of other kinds. Here I have in mind extremely intense rainfall, produced either by monsoons or by tropical cyclones, much of which is enthusiastically welcomed by peoples who without it would be subject to severe drought. Only most infrequently does such rainfall reach those seriously threatening levels which demand public response of an unusual kind. The ambivalent relationship that exists between great rivers and the populations living on and around their flood plains: the general welcome afforded to regular river flooding gives way to general dismay only rarely! And inevitably, this circumstance affects the nature of the early warning problem. Even where two ‘wholly harmful’ natural phenomena are concerned, e.g. extreme winds and the storm surges they generate, those issuing early warnings may still face varying problems in winning public confidence. Recent improvements in localising where an extreme-wind tropical cyclone will make landfall have unfortunately not been matched by improvements in predicting the exact time of landfall; or, more importantly, the phase of the tidal cycle at which landfall will occur. Therefore, forecasters of storm surge intensity are obliged to assume landfall at high tide – the worst case – and true intensity is often far less than predicted if landfall coincides with low tide or mid-tide. Costly evacuations arising from what is subsequently seen to have been a false alarm can have an unfortunate influence on public response to future warnings. My third illustration of how the early warning problem varies as between types of natural disaster is simply related to the amount of warning time available. An early warning system that works well for threats from a tropical cyclone (where public awareness can be built up in successive stages over a period of days as gradually more reliable estimates of landfall location and intensity are made) is hardly suited to the tsunami problem. Indeed, those living in coastal areas threatened by tsunamis need to be trained in making exceedingly rapid responses to the admirable Tsunami Warning System. Yet, notwithstanding all the differences that exist, I have just now come back again to awareness and training! – two essential features which, as I mentioned at the beginning of this paper, must lie at the heart of all early warning systems. Theme 9, in short, offers great potential for highly interesting discussion.