49. FORECASTING THE WEATHER

Systematic weather forecasting by technical methods has been practised for about a century. The early success depended on the development of the "electric telegraph" and has throughout gone forward with the improvements in rapid communication, the telegraph, the telephone, radio, teleprinter, radio-teleprinter, facsimile transmission, television. The possibility of forecasting beyond a few hours ahead in changeable weather arises simply because the weather systems, especially depressions and anticyclones, move slowly across the map, admittedly changing their structure as they do so but broadly carrying their typical weather conditions with them.

If, therefore, we devote sufficient funds and effort to obtaining weather messages from a wide area, we may chart the weather on suitable maps, study how the patterns move and change, and predict by extrapolations. Although there are many complications due to such factors as diurnal variations, mountains and valleys, land and sea distributions, man made atmospheric pollution and a host of natural peculiarities in the behaviour of the atmosphere itself, the simple facts about the behaviour of depressions and anticyclones (outside the tropics) have made forecasting possible. At the same time they have made great improvements in forecasting seem almost impossible.

Putting it very simply, if we study any weather map, we find pressure systems which behave reasonably coherently for one day, two days or perhaps, in favourable circumstances, a little longer. However, after that time they disappear, new systems take their place and, until they have begun to take shape, we can do little more than guess their behaviour in detail.

The progress of forecasting this century has been mainly in predicting the weather for a day or two ahead, with ever greater attention to detail. The atmosphere is a fluid in three dimensions and it can be studied thoroughly only if we know the temperatures, pressures, winds, etc., not only near the ground but far into the air above.

Balloon-borne radio equipment, the radio sonde, permitted a revolutionary advance in theory and in forecasting technique. Weather maps for most of the world can now be drawn every day for levels far into the stratosphere; we have the required three-dimensional picture. The theoretician was in this way soon enabled to see how the well-known basic laws controlling all kinds of fluid motion, from ocean currents to water in pipes, had to be adapted to satisfy the ways of the atmosphere. This done, mathematical equations were constructed which, if solved, would enable us to forecast the future from our knowledge of the present.

When the forecaster tries to look beyond a day or two ahead, many of his depressions and anticyclones literally dissolve into thin air before his eyes, others take their place, and his methods of extrapolation or calculation fail. The forecaster can usually say something useful for as long as four or five days ahead if it is sufficient to say that the weather will remain unsettled or generally fine or will tend to become warmer or cooler. Such general persistencies or trends, which may often be discerned, are the basis of what is called medium-range forecasting. By long-range forecasting one implies outlooks for even longer periods, of the order of a month or a season. Here, no attempt at detail is possible.

(from Discovery, August 1961)