"Die Politik", said Otto, Prince von Bismarck, "ist keine exacte Wissenschaft." "Politics is not an exact science." The same is often said of meteorology - and for good reasons. Although weather forecasts have improved steadily in quality in recent years, to the extent that total failures nowadays are relatively rare, we are still a long way from being able to anticipate the elements' every whim with total confidence. Absolute perfection remains a distant dream towards which we strive. The computer models used for weather forecasting describe the behaviour of the atmosphere in terms of mathematical equations. Observations describing the present state of the weather are fed into the machine, and the equations applied successively at thousands of different points throughout the "model" atmosphere. In this way a new "picture" of the weather for some future time is calculated. The computer edges its way forward, hour by hour in, say, 20-minute steps, until a forecast chart for two days or so ahead appears upon the screen.
There are a number of factors in this operation which place infallibility just beyond the meteorologist's grasp. There are the uncertainties in assessing the existing state of the atmosphere. The picture is built up from observations of temperature, pressure and humidity made at places perhaps 100 miles apart - or in some parts of the world more than 10 times that distance. The computer's impression of the atmosphere is more a Monet than a Caravaggio; its snapshot has a fuzziness - and some hidden quirk may grow to catch the model unawares. There are also limits to our knowledge of the processes which shape our weather, and the ways in which we can express them. The mathematical equations used in the computer to describe the behaviour of the atmosphere are not exact; necessarily, they are approximations - and future developments predicted by the machine may well be under- or over-estimated, or as happened to the weekend forecasts, it may miss them altogether.
The computer models, unfortunately, are to some extent victims of their own success. In recent years they have become so accurate and so reliable that it would be a brave forecaster indeed who would spurn their advice, and produce a forecast based on more traditional ideas and methods. But as we have seen, they are not infallible. Yet, insofar as Irish forecasters can gain any consolation from the recent dramatic and unfortunate events, it is that weather forecasts nowadays are normally so good that the world at large reacts with shock and unbelief when, just now and then, predictions turn out to be completely wrong.