I think is just an exageration to show how chaotic some systems like the atmosphere are. If you had such a cumule of casualities it could happen but I guess it doesnt.

@PabloSaudiBombsYemen, except it's not an exaggeration. In a chaotic system, there is no limit to how small the difference can be in the starting points for two completely different trajectories.

@jameslarge infinitely many completely different trajectories originating from various flaps of a butterfly do not necessarily include in them the also uncountably infinite subset of tornadoes that can occur. Chaotic system - divergently different outcomes? Yes. But diverging to tornadoes? Needs to be proven.

why would it not be true? just using my intuition, it makes sense to me. curious why you're so skeptical.

@jameslarge - There is a limit of sorts in that the flap of a butterfly wing pretty quickly gets buried in quantum noise.

All these storms we're having aren't due to climate change, they're due to butterflies! Forget lowering emissions, let's just kill them damn flappy mites.

I think the most common misinterpretation here is that people think the flap of the wings will be somehow amplified to a tornado. It is however more realistic to think of a chaotic system that is at the edge between positiv and negative feedback, and it is so near to that edge that even the perturbation due to a butterfly will push it into one direction while without it, it would have gone to the other.

Cross-site duplicate: earthscience.stackexchange.com/q/245/6

I believe it was first proposed regarding weather forecasts. The scientist (Lorenz?) had a model running, and started a new model using numbers from the first. These numbers were however rounded rather than an exact match, so after a while the two "identical" models showed very different weather. The scientist realized that even if we covered Earth with sensors every meter, a butterfly taking flight between sensors, would be enough to skew the result - so that the weather predicted by the model would not come to pass, but rather some other weather. A storm another place than expected.

I think it's true, but that it's expressed in a way that tends to get people who don't understand it over-excited. It's important to note that there are countless other tiny events that have just as much of an impact on the way the system develops. When you realise that, it becomes a statement about unpredictability, not one about the significance of butterflies.

@PlasmaHH: right. There's going to be hurricanes. However, assuming the atmosphere to be a chaotic system, everything that happened "long enough ago" is significant in determining precisely when and where they occur and what path they take. The point is just that when you run the model with and without the flap, eventually you get hurricanes in different places, or one has one at a particular time and the other doesn't. The only misleading part is to focus on the notion of whether the flap (or non-flap) therefore "caused" the hurricane. It is one of many causes.

@HotLicks, Re quantum noise. Lorenz's "butterfly" talk was a hypothetical argument. He was saying that even if the laws of physics were completely deterministic, we still would have no hope of long-term weather prediction because an input as small as a butterfly's wing-flap has consequences that never die out and, as chaos theory proves, never average out either. Not sure what you mean by "quantum noise," but it sounds like you are claiming that the laws of physics are not deterministic. Well, OK, but how does denying Lorenz's hypothesis does diminish his argument?

@jameslarge - His argument is still valid as a hypothetical, it's just that it's significance in the real world is swamped by true randomness.

What's the timescale? It's easy to imagine that after a large amount of time even the most minuscule disturbance would produce a drastic change.

From what I've read, the original detraction to Lorenz's thoughts was with regard to the flap of a sea gull's wing off the coast of Brazil causing a tornado in Texas. There are two problems with this. One problem is that sea gulls are the ugly vultures of the seas. They aren't that pretty. Another problem is that they aren't small enough to make the metaphor stick. Butterflies are pretty, they are small, and they are innocuous. How can a butterfly in Brazil possibly cause a tornado in Texas?

A young Austrian man failed to get into art school; one thing led to another, and the United States dropped two nuclear warheads on the sovereign nation of Japan.

@indextwo: I was about to construct something like this, but your real life example fits too well.

The example of that young Austrian who caused the dropping of two nucleair bombs is somewhat (at least) misleading. The energy released by different flappings of a butterfly´s wings will dissipate quickly in the surounding turbulences. The higher level turbulences won´t be infuenced anymore. Only if you change the whole weather system (not only the boundary on Earth)a bit will there be a change in outcome. The best way to know is to compare two identical Earths with two slightly different butterflies. A bit difficult. So we must rely on our thinking and doing so conclude there is no effect.