By Patricia Sullivan

Washington Post Staff Writer

Thursday, April 17, 2008

Edward N. Lorenz, 90, a meteorologist who laid the groundwork for chaos theory, memorably asking whether the flap of a butterfly's wings in Brazil can set off a tornado in Texas, died of cancer April 16 at his home in Cambridge, Mass. He was an emeritus professor at the Massachusetts Institute of Technology.

At MIT, Dr. Lorenz accidentally discovered how small differences in the early stages of a dynamic system, such as the weather, can trigger such huge changes in later stages that the result is unpredictable and essentially random.

At the time, Dr. Lorenz was studying why it's so hard to accurately forecast the weather, but the implications of his work go far beyond meteorology.

The new science of chaos fundamentally changed the way researchers address topics from the geometry of snowflakes to the predictability of which movies will become blockbusters. The butterfly effect became a popular way of describing unpredictability, most recently in "An Inconvenient Truth" (2006), the Academy Award-winning documentary with former Vice President Al Gore.

It also "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.

Yet Dr. Lorenz's 1962 paper on chaos theory was largely ignored for years. A decade later, when he gave a talk about predictability, with a title asking the famous butterfly question, the scientific establishment was ready to consider the idea. Other scientists who had been working on similar questions swarmed to the field, and one by one, certain assumptions of science began to falter.

"When I first heard this [butterfly effect] idea, I thought it very clever but it couldn't be literally true," said James Gleick, a science writer and author of "Chaos: Making a New Science" (1987), which explored Dr. Lorenz's work. "But it is literally true. . . . Complex dynamical systems, if they are chaotic, never repeat themselves. They are capable of an infinite variety of behavior."

This means that simple systems can result in complex behavior and that the slightest change in underlying causes can make the result unpredictable.

Chaos theory -- also known as the science of nonlinearity, the science of complexity, the science of random recurrent behavior or the science of turbulence and discord -- has thus been called the third great scientific revolution of the 20th century, along with relativity and quantum physics.

Edward Norton Lorenz was born May 23, 1917, in West Hartford, Conn., and graduated from Dartmouth College. He received a master's degree in mathematics in 1940 from Harvard University and served as a weather forecaster for the Army Air Forces during World War II.

In 1948, he received a doctorate in meteorology from MIT and joined its faculty. He remained there the rest of his career.

In 1961, he was using a primitive computer to model weather forecasts, which led to his most renowned work.

Using 12 equations, such as the relationship between air pressure and wind speed, he ran the model and found exactly what he sought. But taking a shortcut on the next run, he found that a tiny decimal point change led to a significant error.

Rather than ignore the response, which peers had considered an anomaly, Dr. Lorenz realized measurement is not perfect. If temperature, pressure or humidity measurements were off by a hundredth of a percent, the rainfall he expected in Las Vegas on Thursday could show up as a snowstorm in Beijing a week later. A computerized model of how the "butterfly effect" works can be found at the Exploratorium's Web site.

"He had the ability to make important connections between atmospheric phenomena and simple theoretical models," said Isaac Held, senior research scientist in the National Oceanic and Atmospheric Administration's Geophysical Fluid Dynamics Laboratory. "He taught us that complexity can follow from very simple underlying rules. His study of the limits to the predictability of weather initiated an entire new field of chaotic dynamics."

Dr. Lorenz was elected to the National Academy of Sciences in 1975, and in 1983, he and oceanographer Henry Stommel were jointly awarded the $50,000 Crafoord Prize by the Royal Swedish Academy of Sciences, an honor established to recognize fields not eligible for Nobel Prizes.

Dr. Lorenz was known as shy and humble. He enjoyed hiking and cross-country skiing and sought out mountain trails near every scientific meeting he attended, one of his daughters said.

His wife, Jane Logan Lorenz, died in 2001.

Survivors include three children, Nancy Lorenz of Roslindale, Mass., Edward Lorenz of Grasse, France, and Cheryl Lorenz of Eugene, Ore.; and four grandchildren.

Chaos theory helped shape Dr. Lorenz's conclusions as he worked to calculate long-term forecasts. Given the problems of input, the meteorologist determined that it's impossible to accurately predict weather beyond two or three weeks with a reasonable degree of accuracy.

© 2008 The Washington Post Company