The eccentric engineer

E&T celebrates the life and career of Lewis Richardson, the founder of modern meteorology.

It has been a poor summer and my experience of building sandcastles in the rain, digging potatoes in the rain and, frankly, doing everything in the rain, is not one I want to repeat. The only consolation is that meteorologists have had an even worse time of it, partly because everyone seems to think the weather is their fault, but mainly because predicting the unsettled weather patterns is rather tricky, and when they predict wrong we blame them for that, too.

The story of modern weather prediction and the extraordinary machine first envisaged for doing it is a rather heroic one. Well into the 20th century, most forecasting was the 'seen that before' principle. Developing weather patterns were compared with previously recorded ones and, if the two looked similar, then the forecast was that a similar sort of thing would happen. Which it sometimes did and sometimes didn't.

However, working in the meteorological office at the remote Eskdalemuir Observatory in 1913 was a scientist called Lewis Richardson. He had a brilliant idea that weather could be predicted by using a mathematical technique he'd invented while studying peat. He'd made his idea public three years previously in his work 'The approximate solution by finite differences of physical problems involving differential equations'.

The run up to World War I was not the best time to have such thoughts, particularly when you're tucked away in Eskdalemuir and certainly not when you're a Quaker and an ardent pacifist. With the outbreak of war, Richardson became a conscientious objector and in 1916, like many Quakers, took on a non-combative role in the Friends' Ambulance Unit attached to the 16th French Infantry Division. It was an extremely dangerous job; indeed, the most decorated NCO, William Harold Coltman VC, DCM & Bar, MM & Bar, was a pacifist stretcher bearer who never fired a shot.

While Richardson was ducking shells, he decided to test if the weather could be predicted mathematically, so he took a day for which the weather in past years was known. The day was 20 May 1910, which had been meticulously studied due to an experiment in which hundreds of weather balloons had been launched across central Europe to record weather patterns.

With just a pencil, paper and a slide rule, Richardson used his numerical process to calculate the weather over one location in southern Germany for six hours. The result was a disaster, his calculations predicting a world record increase in atmospheric pressure when it actually remained fairly level. It was, admittedly, a bad start, but Richardson persevered.

Richardson's work was rewarded with indifference and hostility, his pacifism effectively barring him from most academic posts. Initially, he returned to the Met Office, but when that merged with the Air Ministry in 1920, he felt conscience-bound to resign as he could not work for a military organisation. A lesser man might have been put off by this, but not him, and in 1922, working on the very outskirts of the academic world, he published his great work 'Weather Prediction by Numerical Process'.

He imagined a wonderful machine. A huge, spherical room, with the map of the globe painted on the interior, around which would be tiered galleries filled with desks for 64,000 people. Each person would be a 'computer' responsible for the equation, or part equation, modeling the weather on the part of the globe behind where they sat. Around each desk, small signs would flash the results of each calculation, while in the centre, on a tall podium, the weather conductor - the person responsible for keeping the 64,000 calculations running in perfect harmony - would stand. 

If some areas fell behind, he would shine a blue light on them to tell them to speed up, if others ran ahead, a 'beam of rosy light' would be shone in their direction to tell them to slow down. Around the conductor, four clerks would dispatch the results by pneumatic tube to a telegraph office where they could then be radioed to the waiting world.

Richardson's weather machine was, of course, just a dream. There were no machines that could make so many calculations in tandem and no-one seriously considered putting 64,000 mathematicians in a spherical room for any length of time. But the idea was prescient. Richardson had realised that by modeling weather patterns and running those models, weather could be predicted mathematically far more accurately than by simply using precedent.

When Richardson died in 1956, the electronic computers capable of making his dream a reality were just coming into existence. His numerical process soon became the single most powerful tool in meteorology. It is still not perfect, but before we complain we should remember that without this genius we'd still just be guessing that tomorrow's weather might be a bit like today's - and I certainly hope that's not true! 

Justin Pollard has been shortlisted for the UK Press Gazette Magazine Design & Journalism Columnist of the Year Award for 2008 for his 'Eccentric Engineer' column in E&T.

Recent articles

Info Message

Our sites use cookies to support some functionality, and to collect anonymous user data.

Learn more about IET cookies and how to control them