Edward N. Lorenz, the Massachusetts Institute of Technology meteorologist whose efforts to use computers to increase the precision of weather forecasts inadvertently led to the discovery of chaos theory and demonstrated that precise long-range forecasts are impossible, died of cancer recently at his home in Cambridge, Mass. He was 90.

Lorenz was perhaps best known for the title of a 1972 paper, "Predictability: Does the Flap of a Butterfly's Wings in Brazil Set Off a Tornado in Texas?" The memorable title pithily summarized the essence of chaos theory -- that small changes in a system can have large and unexpected consequences.

Although the chaos theory was initially applied to weather forecasting, it subsequently found its way into a wide variety of scientific and nonscientific applications, ranging from the geometry of snowflakes to the predictability of which movies will become blockbusters.

His work "profoundly influenced a wide range of basic sciences and brought about one of the most dramatic changes in mankind's view of nature since Sir Isaac Newton," wrote the committee that awarded him the 1991 Kyoto Prize for basic sciences in the field of Earth and planetary sciences.

By showing that there are limits to the predictability of many systems, Lorenz "put the last nail in the coffin of the Cartesian universe and fomented what some have called the third scientific revolution of the 20th century, following on the heels of relativity and quantum physics," said atmospheric scientist Kerry Emanuel of MIT.
Lorenz also was "a perfect gentleman, and through his intelligence, integrity and humility set a very high standard for his and succeeding generations," he added.

One dramatic conclusion of his work is that it is impossible to predict weather more than two or three weeks ahead of time with any certainty.

The roots of chaos theory trace back to at least the late 19th century, when French physicist Henri Poincare discovered to his chagrin that it was not possible to calculate the stability of a celestial system containing more than two bodies -- at least using techniques available at the time.

That was a shock because Newton's laws of gravity and motion promise order and predictability, and Poincare concluded that there must be other equations that would eliminate the problem. In the absence of computers, however, there was little anyone could do to test that thesis.

Cut to 1961, when a young assistant professor of meteorology at MIT was using a primitive Royal McBee LPG-30 computer to study simplistic models of the atmosphere based on a series of 12 differential equations.
After one run, he decided he wanted to study the end of the calculation in greater depth. Rather than running the entire calculation again, he chose a point in the middle of the calculation and entered the previously calculated values for that point. He then went off for a coffee break to avoid the incessant noise from the machine.

When he returned, to his surprise he found that the calculated weather patterns were grossly different from those produced in the first computation. After determining that the Royal McBee hadn't simply blown a vacuum tube, he began looking more closely at the calculation itself.

Ultimately, he concluded that the original calculations had been carried out to six significant figures. To save space, however, the printout he used had rounded each value off to only three places. That subtle difference, less than 1 percent of the original values, was sufficient to drive the system in a completely different direction.
Lorenz worked out the math involved and reported his findings in the Journal of Atmospheric Sciences in a 1963 paper called "Deterministic Nonperiodic Flow."

The first paper fell on deaf ears, however, and he received little attention until his 1972 "butterfly" talk at a meeting of the American Association for the Advancement of Science.

Lorenz later said that he had planned to use a seagull as an illustration, but that an MIT colleague suggested a butterfly would have more effect. He chose Brazil for its alliterative value.

According to the Web of Science online database, Lorenz's original paper has now received at least 4,000 unique citations by subsequent authors, making it one of the most prolifically cited papers of all time.

Edward Norton Lorenz was born May 23, 1917, in West Hartford, Conn. He received his bachelor's degree in mathematics from Dartmouth College in 1938 and a master's degree from Harvard in the same subject in 1940. During the war, he served as a weather forecaster for the U.S. Army Air Corps, earning a master's degree in meteorology from MIT in 1943. After the war, he continued his studies and received a doctorate in 1948.
He spent his entire career at MIT.

In addition to the Kyoto Prize, he also received the Royal Swedish Academy of Sciences' 1983 Crafoord Prize, established to honor fields not eligible for the Nobel Prize.