The Butterfly Effect: For Real?

[Homebrew]

In Jurassic Park, when Jeff Goldblum blabbered on about Chaos Theory and butterflies causing hurricanes, I thought he was full of it. Turns out that butterflies do affect the weather, at least according to the EPA’s EnergyStar PSAs.

I heard this little gem this morning on the radio:

It’s sad that the wind created by one little butterfly flapping its wings can begin a chain reaction that affects weather patterns thousands of miles away.

We have that power, too. When we use products that have earned the Energystar label for efficiency, we reduce energy demand.

For more information call 1-888-STAR YES or visit energystar.gov

Think I’m making this up? Download the mp3 of the PSA here. According to the EPA’s website, “451 stations are airing the ‘Butterfly’ and ‘Breathe’ radio spots in English and Spanish in 263 cities and 49 states.”

So is it true? Are butterflies really a meanace to us all? Or is the EPA actively spreading stupidity?

[Kalhoun]

This sounds like a question for Cecil.

[SentientMeat]

I think Chaos Theory (in this instance we’re talking about a subset called Complexity) is being a little misunderstood here. There is no causal relationship: If butterflies disappeared, no more or fewer hurricanes would result.

The point is that if you worked backwards from the hurricane, trying to quantify the position and velocity of every water molecule in order to “explain” the hurricane, then you would find that a tiny change in some of those initial variables would have macroscopic effects in how the weather developed.

Using this idea to suggest that saving energy might produce fewer hurricanes is a little misleading, however since we do not know how much effect adding CO2 to a complex system balanced at an apparently delicate equilibrium will have on weather patterns in future, perhaps the parallel is justified.

[MC_Master_of_Ceremonies]

Homebrew, this is an often misunderstood concept. It’s not like if we had no butterflies there would be no hurricanes, or everytime a butterfly flaps it’s wings it has any real effect on weather patterns.
Weather is a chaotic system, which means that small, even minute change in the intial state of the system can, over time, lead to huge changes.

So yes it is possible that a hurricane could happen that wouldn’t of happened if every other factor played out exactly the same because a butterfly flapped it’s wings or didn’t flap it’s wings.

[beagledave]

The butterfly effect is discussed at length in this book on chaos theory.

Also see here

The butterfly effect, first described by Lorenz at the December 1972 meeting of the American Association for the Advancement of Science in Washington, D.C., vividly illustrates the essential idea of chaos theory. In a 1963 paper for the New York Academy of Sciences, Lorenz had quoted an unnamed meteorologist’s assertion that, if chaos theory were true, a single flap of a single seagull’s wings would be enough to change the course of all future weather systems on the earth. By the time of the 1972 meeting, he had examined and refined that idea for his talk, “Predictability: Does the Flap of a Butterfly’s Wings in Brazil set off a Tornado in Texas?” The example of such a small system as a butterfly being responsible for creating such a large and distant system as a tornado in Texas illustrates the impossibility of making predictions for complex systems; despite the fact that these are determined by underlying conditions, precisely what those conditions are can never be sufficiently articulated to allow long-range predictions.

[beagledave]

If I remember correctly from the book…even if you were able to place weather instruments (temperature, pressure, humidity etc) 1 meter apart in every direction of the atmosphere…one would not be able to make really long range forecasts because of the microchanges that occur within that 1 meter space.

[Revtim]

I think this discovery describes the concept well.

http://www.ems.psu.edu/info/explore/Chaos.html

Edward Lorenz was a meteorologist at the Massachusetts Institute of Technology in 1961 when he made a serendipitous discovery. He was using a computer model he had created to run a weather simulation. The program took some time to run, and he didn’t have all night. To save time on the second run, he decided to start it from the middle. He knew that the simulation had produced the value 0.506127 at the halfway point, so he started the program in the middle with the value 0.506. He was astonished when the simulation finished up with a value radically different from the first run. The only difference Lorenz introduced in the second run was changing the number from 0.506127 to 0.506. This small change in the starting conditions led to a large disparity at the end of the simulation.

[ultrafilter]

Like others have said, the butterfly effect is not meant to be taken literally. It’s more an illustration of one of the basic ideas of chaos theory.

[scm1001]

It is true in a truely chaotically system. However, is the weather actually truely chaotic?. Eg. perhaps the energy of a butterfly gets damped out?

[ultrafilter]

For anyone who’s interested, I gave an explanation of what the butterfly effect represents in this thread.

[jshore]

*Originally posted by scm1001 *
It is true in a truely chaotically system. However, is the weather actually truely chaotic?. Eg. perhaps the energy of a butterfly gets damped out?

Well, this is a good question. I believe (although I can’t say I know for sure) that even in a chaotic system, there may be some small perturbations that can get damped out.

So, do we really need to know where all the butterflies are to predict the weather (at some fairly long range in time)? I haven’t heard a good answer to this.

By the way, for moderately long range weather prediction (on the 6-10 and 8-14 day scales), the Climate Prediction Center (CPC), which is part of NOAA, is actually running “ensembles” of simulations using their forecase models. The ensemble members differ by perturbations in the initial conditions. This allows them to see how much they can trust the predictions. For example, sometimes the different ensemble members still are fairly tightly clustered around the same solution after 14 days but other times they are very spread out or cluster around a few different solutions. Thus, they will not only issue predictions on a 6-10 day and 8-14 day scale but will also use this and other information to discuss the confidence level for their predictions.
See http://www.cpc.ncep.noaa.gov/products/predictions/610day/fxus06.html and look at the definition for “GFS ensemble”.