Butterfly

The Butterfly Effect

Oh, yeah!

This is the part of chaos about which most people hear: it has been mentioned in Jurassic Park, Dilbert comics, and even in a recent Venus Williams DMX endorsement commercial. This is the idea that "a butterfly stirring the air today in [Beijing] can transform storm systems next month in New York"1 This statement, although somewhat weird, actually provides and interesting explanation of one of the most fundamental behaviors of chaotic systems: sensitivity.

Boo-hoo

Not that kind of sensitivity. Sensitivity is basically the idea that small differences in initial conditions have drastically different effects on the system over the long run: basically, it is a generalized rule for the Butterfly Effect.

This kind of sensitivity can be seen in weather systems where it was first discovered by Edward Lorenz.

Teacher, that kid is giving me a history lesson

Well, it would help to understand how this discovery came to be, right? If you're more interested in the concepts, skip to the section although the history reinforces sensitivity. Actually, the discovery of sensitivity in chaotic systems actually has an interesting story. In 1960, MIT professor Edward Lorenz designed his own weather toy using a set of relatively simple Newtonian physics equations. His main goal was to understand weather systems in order to be able to predict conditions better.

One day, Lorenz decided that he wanted to go back a certain part more carefully. Computers back then were slow, so Lorenz decided to take a short cut. Instead of starting his "weather machine" from the beginning and wait for it to repeat its pattern on the day that he wanted, the professor entered in data closer to the storm to start his experiment.

Initially, things were looking pretty good. Lorenz then saw the pattern start to repeat itself. It was what happened next that was startling. Slowly, the path of his original weather pattern split from the weather pattern of his second run. After a while, the patterns became completely different.

How did that happen?

You see, Lorenz had his computer printouts stop after the fourth decimal place. In reality, however, the computers held data to several more. During his initial run, Lorenz had unrounded values. By using the rounded values, Lorenz had seen the sensivity on initial conditions that marks system. I guess rounding to the fourth decimal place may have represented that Beijing butterfly flapping its wings!


Last updated 8/8/99