Tom Spears , The Ottawa Citizen 13 Feb 08;
An eight-year study of ocean life shows a "chaotic" balance of nature, and Dutch scientists say this chaos makes it impossible to predict the rise and fall of wild species - anywhere, ever.
Ecologists and politicians often want computers to show how nature will react if we bulldoze a forest or change the global temperature, but the Baltic Sea study now argues this kind of modeling may not be worth much.
If so, it raises doubts about how we can ever preserve a healthy environment, except through good luck.
The small Baltic creatures such as plankton were isolated from the rest of the ocean and studied for eight years. Each member of the "food web," or network of who eats whom, took turns multiplying and becoming scarce, even though the scientists kept the outside conditions constant.
And they could never figure out a pattern that allowed real predictions of how any species would fare.
"Advanced mathematical techniques proved the indisputable presence of chaos in this food web," they conclude in the journal Nature," adding that "short-term prediction is possible, but long-term prediction is not."
Lead scientist Jef Huisman said it's like trying to forecast the long-term weather: all their forecasts broke down beyond a few weeks.
This is already causing a stir in some biology circles. At Duke University in North Carolina, Stuart Pimm called the work "an outstanding job" and said it raises the possibility that nature just can't be crammed into computer models.
But he adds: "In an important way it's even more interesting than that.
"One of the classic chaotic systems is the weather," he said. "We have computer models that do a good job of predicting the weather in the short term."
But the best supercomputers can't stretch that forecast beyond a week or 10 days, echoing the Baltic Sea models that failed.
"It's the way in which they fail that's interesting," Mr. Pimm says. "They fail spectacularly badly over longer and longer periods of time."
Beyond that short-term forecast, "you simply can't do better than treating it as a game of chance," he says. Past that short-term run of accuracy, the errors shoot up exponentially.
This, he says, is what the Baltic Sea experiment found - but in populations of sea creatures, not in the weather.
"The mark of chaotic systems is not that they're unpredictable (i.e. all the time). It's that in the short term you can predict them really well, but the system deviates from your prediction and deviates ... exponentially. So you start out with a tiny error, and the next day that error doubles, and then doubles again and doubles again so eventually you've got no prediction whatsoever. And that fingerprint of chaos is at the heart of this (Baltic Sea) paper.
"They're not saying, 'Aw, it's really hard to predict plankton systems.' They're finding that after the short-term intervals ... the differences will blow up exponentially."
The question now is whether this chaos will show up in populations of all wildlife from giraffes to bald eagles to crocodiles.
"That's the big question," Mr. Pimm says. "We just don't know," but this research is step one in finding out. "It alerts us to the notion that there may be a lot more stuff like this out there."
