Forecasting by the numbers

While employed as an ambulance-driver for the French army during World War I, Lewis Fry Richardson worked on a scheme that's still in use today for predicting weather by computer. Richardson devised seven equations that could be used to extend the current weather into the future. He put his ideas to the test with a grand experiment in 1922. Using data collected across western Europe a decade before – at 4am on May 10,1910 – he calculated what the weather should have been at 10am on that day. It took Richardson six weeks to crunch the numbers for this six-hour forecast and the results were abysmal. The pressure change he predicted for one point in Germany was a hundred times greater than the real thing. As author Frederik Nebeker points out,”It is ironic that this forecast, which is probably the forecast that consumed the greatest amount of labour by a single person, is no doubt one of the least accurate forecasts ever Nonetheless, Richardson's test demonstrated to the world that one could – in principle, at least – carry out calculations and forecast the weather using more than intuition and rules of thumb that were standard in the 1920s. Richardson was a Quaker and conscientious objector, and he brought his mathematical bent to studies of human conflict ranging from arguments to wars. There was a lighter side to him as well. In describing how turbulence occurs, he wrote a classic one-sentence poem:'Big whirls have little whirls that feed on their velocity, and little whirls have lesser whirls and so on to viscosity.”