Edward Lorenz

Edward Lorenz, 90, MIT professor developed chaos theory

Edward Lorenz, an MIT meteorologist whose meticulous attempt to predict the weather through an early computer unraveled into a spectacular failure that he turned into the chaos theory, died at his home in Cambridge yesterday. He was 90.

His development of the chaos theory is considered a seminal moment in 20th-century science. Also called "deterministic chaos," the theory brought about "one of the most dramatic changes in mankind's view of nature since Sir Isaac Newton," said the committee that awarded Dr. Lorenz the 1991 Kyoto Prize for basic sciences.

It was one of many scientific awards that the quiet professor won. There is no Nobel Prize for his specific field of expertise, meteorology.

At a meeting of scientists in 1972, he gave a talk with a title that captured the essence of his ideas: "Predictability: Does the Flap of a Butterfly's Wings in Brazil Set Off a Tornado in Texas?"

The phrase "butterfly effect" has become part of the lexicon of both pop science and pop culture.

His work on the theory began a decade earlier. In the winter of 1961, Dr. Lorenz was trying to determine how accurate a computer could predict long-term weather patterns. He ran one simulation with a computer model, then wanted to extend the forecast, so he added a second simulation, with the same parameters and conditions of the first model. The weather pattern should have seamlessly flowed into the second simulation.

Instead, the trajectories quickly diverged.

The problem: a rounded decimal number. Dr. Lorenz realized that the computer stored numbers to an accuracy of six decimal places but, to save space, printed out results shortened to three decimal places. So, for example, 0.310625 became 0.311. For the second simulation, he had used the shortened figure.

Even this minute discrepancy drastically altered the forecast.

Tiny changes, in effect, could have catastrophic, and often unpredictable, consequences. And they made perfect predictions of weather, even through the emerging power of computers, impossible: Exact measurements of all the conditions could be upset by one small event, such as the flap of a gossamer wing.

The development of this theory changed not only how scientists viewed the prediction of weather, but also had applications in such sciences as fluid dynamics.

In his desire to improve weather forecasts, Dr. Lorenz had developed the theory that, ultimately, undermined flawless predictions.

"For the past 15 years or so I've been interested in predictability," he said in 1983. "I guess most people would consider me a pessimist [since] I think there's a definite limit [in long-term weather prediction] which we can't go beyond."

On the other hand, he added, "I think we still have a long way to go before we reach that limit."

The scientist was incredibly quiet. Getting him to talk was painfully difficult, his colleagues told the Associated Press. He rarely wrote papers with others.

"Of all the geniuses of that era, he was the quietest and most humble and the most kind," said Jerry Mahlman, a longtime friend and a retired federal climate scientist.

Dr. Lorenz was born in West Hartford, Conn., in 1917 and later wrote in a biographical sketch: "As a boy I was always interested in doing things with numbers and was also fascinated by changes in the weather."

He had degrees from Dartmouth College and Harvard University as well as the Massachusetts Institute of Technology, which he joined as part of the meteorology staff in 1948.

He became a professor at MIT in 1962 and was the head of the meteorology department from 1977 to 1981. He retired in 1987.

Dr. Lorenz was an avid hiker and mountain climber, colleagues and family members said.

Dr. Lorenz's wife, Jane, died in 2001. He leaves two daughters, Cheryl, of Eugene, Ore., and Nancy of Roslindale; a son, Edward H. of Grasse, France; and four grandchildren.

Material from the Associated Press was used in this obituary.

Published in The Boston Globe April 17, 2008 Copyright 2008 Globe Newspaper Company.