JUST for a few minutes let me take you away from your worries. Forget your ever-rising cost of living, your concern over world leaders acting strangely or not acting at all, and your anxiety about the political and climatological state of the world. Instead I’ll tell you why the weather on May 20 more than one hundred years ago became so important.
In the grim months of 1917 there was a British ambulance driver working with the French 16th Infantry Division in the trenches near Rheims. He had been in a secure and responsible job of national importance with the UK Met Office, but as a Quaker appalled by what was happening in France, in 1916 he volunteered to join the Friends’ Ambulance Unit.
He had a first-class Cambridge degree and was fascinated by the possibility of mathematics. When he heard about the Titanic disaster in 1912 he devised a method of calculating the distance of an iceberg by timing sound echoes, using a penny whistle, an umbrella (to amplify the echo) and an Isle of Wight pier.
He realised that given sufficient data, the problems of weather forecasting might be attacked mathematically. He took with him into the trenches a mass of reports from weather stations for 7am on May 20, 1910. He used all his spare time constructing equations to produce a forecast for six hours ahead, wanting to compare his result with what actually happened.
At the same time he wrote a book which described how, given enough data and a great many human calculators, it might be possible to forecast the weather by using numerical methods. The whole thing got lost in the battle of Champagne in April 1917, but was miraculously discovered in a heap of coal several months later. An official history of the Met Office confirms this unlikely tale.
The book, Weather Prediction by Numerical Process, came out in 1922. Reviewers were unimpressed because of its difficult maths, and because the author had been totally honest and included his forecast results, which were not very convincing. In fact they were pretty awful. Then there was his picture of the future: numerical forecasting would be carried out in a huge theatre, where 64,000 ‘computers’ (i.e. people, possibly with slide rules) would sit in various positions appropriate to that particular country’s data. Supervisors would issue and collect the data under the control of one director seated in a high tower in the middle of the arena.
In a remarkable vision he wrote: ‘Perhaps some day in the dim future it will be possible to advance the computations faster than the weather advances and at a cost less than the saving to mankind due to the information gained. But that is a dream.’
In 1922 it was certainly a dream. The Manchester Guardian reviewed the book as the ‘rhapsodies of an irresponsible visionary’, but tempered its assessment by saying it was also ‘an attempt to picture a forecast-factory of the future’. The Harvard meteorology professor Alexander McAdie wrote that the book ‘will probably be quickly placed on a library shelf and allowed to rest undisturbed by most of those who purchase it’.
It seems incredible now that the book was hardly mentioned in the meteorological text books for the next 20 years. The experts clearly saw it as an impossible dream and dismissed it from their minds. Curiously, plate tectonics, the theory of continental drift put forward in 1915 by Alfred Wegener, went the same way. Again, the experts laughed – who could imagine whole continents drifting about all over the place? The east coast of South America seemed to fit into the west coast of Africa. That might be curious but was surely accidental. Schoolchildren (myself among them) were still laughing about this theory in the 1940s.
The trouble was that no one could see any mechanism by which continents could float around, and no one could see any kind of mechanism that would take the place of the 64,000 computers. Before the 1940s that word still meant human beings.
But by 1950 our Quaker mathematician had received a forecast produced electronically in America on the primitive ENIAC computer, and by 1954 the Met Office was realising that numerical forecasting showed great promise. A computer was installed by January 1959 and now, of course, they have an array of super-machines producing forecasts for the whole world.
But who was that ambulance driver? I have told you the story of Lewis Fry Richardson, born 1881 in Newcastle upon Tyne, died 1953, the solitary pioneer of Numerical Weather Forecasting, the result of which you probably consult now and again on your phone.
It’s lucky someone looked in that coal heap.
