The Butterfly Effect
why it's impossible to predict weather
One day in 1961, Edward Lorenz
was working in his office, entering data into a newfangled computer program designed to simulate weather patterns. The simulation was a repeat of one he’d run already, but this time he rounded off one of his 12 variables from .506127 to .506. Then he left his office to grab some coffee while the computer crunched the numbers.
When he came back, though, it was clear that something was very, very wrong. That tiny change in his data led to a drastic transformation, completely changing two months of simulated weather. Instead of small changes leading to other small changes, Lorenz realized that small changes could have huge consequences He published his findings in 1963, and the idea came to be known as “sensitive dependence on initial conditions” in scientific circles.
Embrace the Chaos
The butterfly effect gave rise to something called chaos theory. It centers on hard-to-predict phenomena like animal populations, stock prices, and even human behavior.
Chaos may sound like it’s out of the realm of mathematics — if it’s unpredictable, where do you even start? — but everything in the universe is governed by rules, even if we’re not aware of exactly what they are.
One of the most famous illustrations of this came from Lorenz, who plotted a graph of solutions to equations representing the motion of a gas. The results looked aptly enough, like a butterfly. That graph highlighted how chaos always has it limits.
But when it comes to chaos theory, even our best equations can’t always nail 100 percent accuracy. That’s especially true of the weather. While a butterfly’s wings can’t actually cause a tornado, other small quirks in the atmosphere, like the exact location of individual clouds, can have big effects that we can’t predict.
As Lorenz wrote in his pivotal 1963 paper, when his results are “applied to the atmosphere … they indicate that prediction of the sufficiently distant future is impossible by any method, unless the present conditions are known exactly. In view of the inevitable inaccuracy and incompleteness of weather observations, precise very-long-range forecasting would seem to be non-existent.” Fifty years later, and that hasn’t changed.