Professor Edward Lorenz: Meteorologist whose work on weather prediction led to the discovery of chaos and the 'butterfly effect'

Edward Lorenz was a meteorologist at the Massachusetts Institute of Technology (MIT) whose investigations into predictability of the atmosphere led to the discovery of chaos and "strange attractors", while also changing the way operational forecasts are made. Known for his passion for science, his kindness, and his gift for writing, Lorenz produced papers of clarity and beauty, illustrated with understanding and gentle humour.

He was fascinated by forecasting, and studying the various ways our ignorance can lead to wildly misleading predictions of the future. The idea that a small event can produce large changes in the atmosphere dates back at least to Edgar Allan Poe, who discussed the growing impact a wave of the hand would have on the future. In the 1920s the British astronomer Sir Arthur Eddington focused on a match idly thrown from the window of a train, and in the early 1960s Lorenz considered the flap of a seagull's wings.

The origin of the now popular phrase "butterfly effect" is unclear; the contenders include a 1952 story by Ray Bradbury in which the death of a prehistoric butterfly leads to a very different modern world, as well as the butterfly-like shape of the strange attractor – an odd mathematical object which reveals more and more structure when examined closely – now known as the "Lorenz attractor". For Lorenz himself, the move from seagull to butterfly came about when his friend Philip Meriless, unable to reach him as a meeting deadline passed, submitted the title "Predictability: does the flap of a butterfly's wings in Brazil set off a tornado in Texas?" for a talk Lorenz was to give in 1972.

For more than half a century, Lorenz was a master computer modeller. He did not construct his models to be realistic, but rather to provide a compelling example by capturing the relevant mathematics, if not the relevant physics, of a situation. He had a knack for constructing nonlinear models that were more than fit for this purpose, providing new mathematical insights as well as hitting the targets they were designed to exemplify or contradict. He believed strongly that such stories allowed those with little or no training in mathematics to understand the essence of chaos, and without question his examples proved of great value in clarifying the vision of many who did have such training.

Asked why we could not make better weather forecasts, he wondered why we could make forecasts at all, and developed a clear understanding of the difference between the predictability of a physical system itself, the limits imposed by our lack of precise observations and those imposed by our not knowing exactly which equations to use. In constructing simple mathematical models to illustrate the lessons of predictability, the ways and means by which weather and climate forecasts go wrong, Lorenz developed insights applicable far beyond the earth sciences.

Edward Norton Lorenz was born in 1917 in West Hartford, Connecticut. While an undergraduate in mathematics at Dartmouth College, New Hampshire, in the Thirties, Lorenz took an interest in simple pinball machines which appeared in the local drug stores, and in the arguments that followed as to whether they were illegal games of chance or acceptable games of skill. He obtained a masters degree in mathematics at Harvard, before taking a graduate degree in meteorology from MIT.

After serving in the US Army Air Forces as a weather forecaster during the Second World War he returned to MIT, receiving his doctorate in meteorology in 1948 and joining the MIT faculty, to which he belonged for the rest of his career. His description of chaos in a simple model of convection in the early Sixties sometimes overshadows his wide-ranging contributions to atmospheric science, forecasting and the use of computers in scientific research.

Although trained to work with the equations that "governed" the evolution of the atmosphere, in 1955 he was appointed to lead a project in statistical weather forecasting. He immediately challenged the popular claim that the then recent results by Norman Wiener on linear methods provided an optimal approach to weather forecasting, leading some of his colleagues to view him not as a statistician, but as an infiltrator from the numerical weather prediction camp. The challenge took on a form that Lorenz used throughout his career: he developed a counter-example within the context of a simple nonlinear model. This approach did not question Wiener's mathematical results, but only their application, and proved much simpler than establishing the details of whatever mistake had been made in their application.

Lorenz first developed a simple nonlinear model of the atmosphere that considered only 12 variables and tuned it to display nonperiodic behaviour. He then showed that the linear method did not yield perfect predictions of this behaviour and since perfect predictions could be made by forecasting with the exact equations, he had demonstrated that the linear method could be beaten.

Attempting to examine a particular solution more closely, he famously went to get a cup of coffee after typing in a starting point from the original printout and restarting the computer, returning to find that the computer had generated a very different result from what he had expected. Initially suspecting a hardware fault in the computer, he found instead that the "error" had developed slowly due to the amplification of a slight difference in the starting point: he had typed in numbers as the computer had printed them out, and those values differed slightly from the internal values that the computer had used in the original run, hence he had introduced a small difference between the original run and the second.

He soon recognised the implications this held for weather forecasting: small unavoidable uncertainties in the observations could quickly lead to the loss of predictability, even if the weather model was perfect. He often drew an analogy with the pinball games he had witnessed in Dartmouth, where the placement of a winning pinball required super-human skill, an analogy strengthened by the habit of meteorologists to place small wagers even on their real-time forecasts of the 12 variable model, and then watching as the computer printer slowly revealed the results.

In the late Fifties, Lorenz did not wish the issue of sensitivity to small uncertainties to distract from his main message, which was the inferiority of linear forecasting methods. So he delayed detailed discussion of the effect he had witnessed while looking for a simpler model "in the hopes of being able to demonstrate exactly what was happening". This hope was realised with a model consisting of only three variables, which Lorenz introduced in a 1963 paper, "Deterministic Non-periodic Flow". This remains one of the clearest introductions to chaos and includes a description of how strange attractors are formed and why they are fractal.

In a less widely known paper in 1964, "The problem of deducing the climate from the governing equations", Lorenz moved from uncertainty in observations to uncertainty in model parameters, and showed how uncertain parameters pose significant challenges to determining the statistics which define the Earth's climate. Using an even simpler model, now called the Logistic Map, Lorenz demonstrated how its behaviour – the "climate" of the Logistic Map – changed, sometimes drastically, as model parameters were varied.

He even identified the so-called "period three window" which, a decade later, the mathematicians T.Y. Li and J.A. Yorke would discuss when introducing "chaos" as a technical term. Lorenz stressed that in mathematics, a computer could suggest propositions to be proven, propositions whose very existence was unknown until one saw numerical simulations. Indeed, simulations were to play this role years later in the work of Mitchell Feigenbaum and others.

Lorenz also argued that climate forecasting need not be impossible, even if detailed weather forecasts proved impossible due to chaos. His original paper on chaos is often cited by climate-change nay-sayers, who fail to note that in a paper published in the 1970s, "Climate Predictability", he stressed that chaos "does not indicate the rate of decay of climate predictability". In the same paper, Lorenz touched on a theme he was still developing last year: just how long might we need to run a realistic climate model if we want to be fairly sure we know what its climatological distribution really is?

Operational forecasters have the advantage of seeing their forecasts go wrong, day after day. After day. Observing how various forecasts went wrong, Lorenz used simple models to clarify problems and test potential solutions. As computers got faster, his simple models got a bit more complex, but the goal remained to illustrate and motivate methods rather than to simulate the actual weather.

He also worked with those who ran operational weather models, and often visited the European Centre for Medium-Range Weather Forecasts in Reading. Since 1992, the first line of defence against chaos deployed by operational weather forecasters has been to run not just one simulation per day, but an "ensemble" consisting of a few dozen simulations each initiated from a slightly different starting point, in order to determine how sensitive today's forecast is – Lorenz had pioneered the approach used to select an effective ensemble in the late Sixties.

In the Nineties Lorenz advanced the question of how to best improve a forecast if one was allowed to make one more observation of the atmosphere – how would you determine what to observe? Today, when satellites can produce more data than one has time to consider, these ideas can help determine which data to analyse in real time.

A member of the US National Academy of Sciences, Lorenz was also elected a foreign member of both the Royal Society and the Soviet Union's Academy of Sciences, and received numerous awards, including the Crafoord Prize, awarded jointly with Henry Stommel in 1983, and the Kyoto prize in 1991. Accepting the Kyoto prize, he stressed that a scientist "must be willing to question any commonly accepted explanation that is not logically and clearly stated, rejecting it and seeking a new explanation if it cannot be acceptably restated".

When questioned as to whether the Earth's atmosphere was chaotic, Lorenz would note the difference between what we can know regarding our models, and what we can know regarding the real world. He would often suggest that if our best models of a system were chaotic, then it seemed reasonable to call the system being modelled chaotic, while realising that the best model of tomorrow might prove rather different from those we have today. From the early Sixties onwards, he took the position that, even if the jury is still out when it comes to the Earth's atmosphere, "the most recent evidence seems to favour the seagulls".

Leonard A. Smith

Edward Norton Lorenz, meteorologist: born West Hartford, Connecticut 23 May 1917; staff, Department of Meteorology, Massachusetts Institute of Technology 1948-55, Assistant Professor of Meteorology 1955-62, Professor 1962-87 (Emeritus), Head of Department 1977-81; married 1948 Jane Logan (died 2001; one son, two daughters); died Cambridge, Massachusetts 16 April 2008.

Start your day with The Independent, sign up for daily news emails
Arts and Entertainment
Attenborough with the primates
tvWhy BBC producers didn't want to broadcast Sir David Attenborough's famed Rwandan encounter
News
people
News
Campbell: ‘Sometimes you have to be economical with the truth’
newsFormer spin doctor says MPs should study tactics of leading sports figures like José Mourinho
Sport
football
Have you tried new the Independent Digital Edition apps?
ebooks
ebooksA special investigation by Andy McSmith
Life and Style
Alexander McQueen's AW 2009/10 collection during Paris Fashion Week
fashionMeet the collaborators who helped create the late designer’s notorious spectacles
News
i100
News
Perry says: 'Psychiatrists give help because they need help. You would not be working in mental health if you didn't have a curiosity about how the mind works.'
people
Life and Style
Stepping back in time: The Robshaws endured the privations of the 1950s
food + drinkNew BBC series savours half a century of food in Britain
News
Google celebrates St David's Day 2015
newsWales' patron saint is believed to have lived in the 6th century
Latest stories from i100
Have you tried new the Independent Digital Edition apps?
Independent Dating
and  

By clicking 'Search' you
are agreeing to our
Terms of Use.

iJobs Job Widget
iJobs General

Recruitment Genius: Bookkeeper / Office Co-ordinator

£9 per hour: Recruitment Genius: This role is based within a small family run ...

Recruitment Genius: Designer - Print & Digital

£28000 - £32000 per annum: Recruitment Genius: This Design and marketing agenc...

Recruitment Genius: Quantity Surveyor

£46000 per annum: Recruitment Genius: This property investment firm are lookin...

Recruitment Genius: Telesales / Telemarketing Executive - OTE £30k / £35k plus

£18000 - £35000 per annum: Recruitment Genius: This company specialises provid...

Day In a Page

War with Isis: Fears that the looming battle for Mosul will unleash 'a million refugees'

The battle for Mosul will unleash 'a million refugees'

Aid agencies prepare for vast exodus following planned Iraqi offensive against the Isis-held city, reports Patrick Cockburn
Yvette Cooper: We can't lose the election. There's too much on the line

Yvette Cooper: We can't lose the election. There's too much on the line

The shadow Home Secretary on fighting radical Islam, protecting children, and why anyone in Labour who's thinking beyond May must 'sort themselves out'
A bad week for the Greens: Leader Natalie Bennett's 'car crash' radio interview is followed by Brighton council's failure to set a budget due to infighting

It's not easy being Green

After a bad week in which its leader had a public meltdown and its only city council couldn't agree on a budget vote, what next for the alternative party? It's over to Caroline Lucas to find out
Gorillas nearly missed: BBC producers didn't want to broadcast Sir David Attenborough's famed Rwandan encounter

Gorillas nearly missed

BBC producers didn't want to broadcast Sir David Attenborough's famed Rwandan encounter
Downton Abbey effect sees impoverished Italian nobles inspired to open their doors to paying guests for up to €650 a night

The Downton Abbey effect

Impoverished Italian nobles are opening their doors to paying guests, inspired by the TV drama
China's wild panda numbers have increased by 17% since 2003, new census reveals

China's wild panda numbers on the up

New census reveals 17% since 2003
Barbara Woodward: Britain's first female ambassador to China intends to forge strong links with the growing economic superpower

Our woman in Beijing builds a new relationship

Britain's first female ambassador to China intends to forge strong links with growing economic power
Courage is rare. True humility is even rarer. But the only British soldier to be awarded the Victoria Cross in Afghanistan has both

Courage is rare. True humility is even rarer

Beware of imitations, but the words of the soldier awarded the Victoria Cross were the real thing, says DJ Taylor
Alexander McQueen: The catwalk was a stage for the designer's astonishing and troubling vision

Alexander McQueen's astonishing vision

Ahead of a major retrospective, Alexander Fury talks to the collaborators who helped create the late designer's notorious spectacle
New BBC series savours half a century of food in Britain, from Vesta curries to nouvelle cuisine

Dinner through the decades

A new BBC series challenged Brandon Robshaw and his family to eat their way from the 1950s to the 1990s
Philippa Perry interview: The psychotherapist on McDonald's, fancy specs and meeting Grayson Perry on an evening course

Philippa Perry interview

The psychotherapist on McDonald's, fancy specs and meeting Grayson Perry on an evening course
Bill Granger recipes: Our chef recreates the exoticism of the Indonesian stir-fry

Bill Granger's Indonesian stir-fry recipes

Our chef was inspired by the south-east Asian cuisine he encountered as a teenager
Chelsea vs Tottenham: Harry Kane was at Wembley to see Spurs beat the Blues and win the Capital One Cup - now he's their great hope

Harry Kane interview

The striker was at Wembley to see Spurs beat the Blues and win the Capital One Cup - now he's their great hope
The Last Word: For the good of the game: why on earth don’t we leave Fifa?

Michael Calvin's Last Word

For the good of the game: why on earth don’t we leave Fifa?
HIV pill: Scientists hail discovery of 'game-changer' that cuts the risk of infection among gay men by 86%

Scientists hail daily pill that protects against HIV infection

Breakthrough in battle against global scourge – but will the NHS pay for it?