Computer weather forecasts?
In the USSR as in the west, the possible use of electronic computers for weather forecasting has attracted considerable attention from both scientists and the public. In particular, it seems that the Russians have made considerable progress towards a reasonably successful prediction of a weather map 24 hours in advance.
The atmospheric pressure systems revealed in a weather map are too complicated for mathematical analysis, so that changes in them – which are closely connected with changes in the weather – must at present be forecast by guesswork backed by experience of similar weather systems in the past.
Advertisement
In principle, however, they can be deduced by a process of step-by-step computation. Inevitably, errors are introduced in each successive step as the weather continually changes its patterns, which limits the range in time over which reasonably accurate forecasts can be obtained. The success of the methods adopted is best judged from the accuracy of the eventual forecast.
In the Russian experiment, pressure calculations were made for some 450 points in Europe and western Siberia. The total calculation took about 40 minutes to complete using the very high-speed BESM computer based at the Academy of Sciences of the USSR. It enabled a weather map to be prepared that was in good general agreement with the actual weather readings that were taken 24 hours after those on which the calculations were based.
Particularly well shown was the movement of a depression initially located over the North Sea which moved into a general region of low pressure near Finland. On a small scale, however, agreement with the predictions was less good, and it is clear that though this innovative approach to weather forecasting shows much promise, it is not yet of practical use.
From The New ÐÓ°ÉÔ´´, 30 May 1957