Ingenious

	sitemap \| help
	> Advanced Search


you are here: > Home > Read > Understanding our world > Predicting Weather > Just because you can model it

save to my links [ + ] read caption

TOPIC SECTION:

Just because you can model it, doesn’t mean you can predict it

On ‘Black Monday’, 19 October 1987, three days after the Great Storm, the London Stock Market suffered a massive £50.6 billion loss, which was partly triggered by the aftermath of the storm.

Meteorology and economics are both predictive sciences that aim to forecast future events. Governments and businesses like certainty and one way to gain cer

Front view of Philips’s hydraulic model of the British economy, 1949.
Credit: Science Museum/Science & Society Picture Library

tainty is to have a reliable method of forecasting. However, global weather and national economies are huge, highly complex systems, which are changing all the time.

The role model for meteorology and economics has always been astronomy, a science which can be used to predict solar eclipses, comets and even sunspots with great accuracy. For more than a century, meteorology and economics have sought a similar standard of forecasting, using mathemati

Meteorologists can only ‘predict the weather accurately provided it doesn’t do anything unexpected’

cal methods, but without complete success. Lewis Fry Richardson published his Weather Prediction by Numerical Process in 1922, and Jan Tinbergen, who originally studied physics, developed a mathematical (econometric) model of the American economy in the 1930s. The calculations required in both cases are very complex and detailed modelling of the weather or the national economy only became feasible with the introduction of electronic computers after the Second World War. The American-Hungarian mathematician John von Neumann promoted the use of computers in both fields. The dream of modelling the economy and the weather was aided by the systematic collecting of more economic and meteorological data in the 1950s and 1960s, notably the National Income Accounts and meteorological observations from balloons and, from 1964, satellites.

Spurred by the availability of this data, and improved theoretical models, two branches of the British Civil Service –the Met Office and the Treasury – poured resources into the construction of increasingly complex mathematical models in the 1970s and 1980s using ever more powerful computers. By the 1990s, however, the early optimism about a technological breakthrough in both fields had receded. Meteorological forecasts are more accurate than twenty years ago, but meteorologists can only ‘predict the weather accurately provided it doesn’t do anything unexpected’. Much the same is true in economics.

This oil painting ‘Serendip’ by Keith Holmes, 1997-1999, depicts the chaos butterfly in the bottom right-hand corner.
Credit: Science Museum/Science & Society Picture Library

Models can predict the impact of a single factor (for instance a tax cut) tolerably well, but they cannot predict the behaviour of the entire economy accurately.

The new discipline of ‘chaos theory’ seemed to offer a way forward. Chaos theory suggests there is a sensitive dependence in complex systems on the initial conditions – tiny events at the outset can have massive, hitherto unexpected, consequences. This is sometimes called the ‘butterfly effect’ after a talk by Edward Lorenz in 1972 entitled ‘Predictability: Does the Flap of a Butterfly’s Wings in Brazil set off a Tornado in Texas?’. Lorenz had first formulated the concept in 1963 and around the same time Benoit Mandelbrot at IBM suddenly realised that cotton prices also displayed chaotic behaviour. Meteorologists can harness the sensitivity of chaotic systems to the initial conditions by slightly changing these conditions and looking at the total collection (‘ensemble’) of forecasts calculated in this way.

Chaos theory can improve our understanding of the weather and the economy, but it also undermines the earlier vision of accurate forecasting. Unless we know the initial conditions to an impossibly high degree of precision, we cannot predict with certainty what path any complex system will take. The weather and the economy never repeat exactly the same paths again.

Topic section: Playing monopoly with the weather forecasts

Playing monopoly with the weather forecasts

The Meteorological Office has a near-monopoly of weather forecasts in the UK. We can try to predict the weather ourselves by watching the direction of the wind and tapping the barometer or can Piers Corbyn and his sunspot model do any better? > more

Topic section: 'There isn’t a hurricane on the way...'

'There isn’t a hurricane on the way...'

The wind may have uprooted a tree in your garden, but it was only a severe gale, not a hurricane. You may be drenched, but it was only a shower. Weather forecasters use a very specific terminology that baffles most of us. It is all a matter of interpretation. > more


	> Printer friendly version	> Back to top
© NMSI. All rights reserved.		\| terms of use \| sitemap \| contact us \| accessibility \| privacy \| who we are



If your browser is not javascript enabled then click here to Read More. To learn how to javascript enable your browser click here.
DISCOVER: Voices - of people involved Unusual Takes - the unexpected angle REFERENCES LINKS



5 images Related to: Just because you can model it, doesn’t mean you can predict it