Opinion: Long-range storm forecast a tarot reading

There's a little bit of the control freak in every scientist. And perhaps no profession is immune from the illusion that any problem eventually will yield to diligence and intelligence. But in the case of hurricane forecasting, human folly has given way to humbling hilarity.

Take the 2007 season, a fortune-telling exercise no less futile for erring in our favor.

Scientists predicted 17 named storms, including five intense ones. We got 14 storms, only two with winds over 110 miles per hour.

I'm not suggesting an end to long-range hurricane predictions. But maybe, in the interest of full disclosure, they should come stamped: For Amusement Only.


Forecasting has come a long way since the invention of the umbrella. But there are still some things scientists can't predict with certainty -- and probably never will.

"I can't imagine the time in the future where on Jan. 1, they print the chart that goes on the Web and they hand it out to everyone, " said hurricane specialist Hugh Willoughby, now a research professor at Florida International University. "The reason, basically, is the Butterfly Effect: What the atmosphere does depends on its history, and a slightly different history will produce very different results."

The Butterfly Effect -- the notion that the flutter of a butterfly's wings in Brazil can set off a tornado in Texas -- is an elegant illustration of one of the principles of chaos theory: that small variations in the present can produce large variations in the future.

In his 2000 book, Ubiquity, Mark Buchanan uses similar reasoning to explain why some natural phenomena, such as earthquakes and avalanches, elude prediction: "It appears that, at many levels, our world is at all times tuned to be on the verge of sudden, radical change, and that these and other upheavals may all be strictly unavoidable and unforeseeable, even just moments before they strike. Consequently, our human longing for explanation may be terribly misplaced, and doomed always to go unsatisfied."


When it comes to hurricane forecasting, the problem is not so much the science as the reaction to it. Call it the Mother-in-Law Effect: the principle that as the level of nagging increases, the ability to hear dramatically decreases.

The Miami Herald's Martin Merzer reported this week that some researchers worry that errors in the full-season prediction will undermine faith in forecasts of actual storms, which tend to be more accurate.

Others suggested that forecasters have no business trying to predict an entire season before it even starts. "If [William] Gray were honest, " former NOAA hurricane researcher Jeff Masters said in referring to the forecaster, "he would say they have no skill in making predictions that far in advance."

Willoughby takes a softer stance, assuming that most people can tell the difference between wild six-month forecasts and real warnings about impending storms.

"I've never taken year-to-year predictions seriously other than as an exercise in public relations and as a good publicity move to get people started thinking, " he said.

So perhaps the best way to approach the long-term predictions is with the same joyful skepticism with which one would greet the tarot lady at the county fair.

In the meantime, I'm not going to begrudge nature for failing to live up to human prediction. I'm quite glad all those scientists were wrong in May, not least because the hurricane windows I ordered on their advice have yet to arrive. But I could have predicted that.