• Choose a category
• All
• Classic-Fiction

By Root 422 0

a tiny variation in the initial data—say a few millionths of a degree difference in temperature, much too small to actually measure, or a tiny variation in wind speed—something surprising happens. For the first few hours of the simulated run, the "new storm" replicates exactly the course and intensity of the old one. But then, the behaviors of the two virtual storms begin to diverge, and eventually they differ quite radically—one might veer sharply north, the other continue on its westward course; one may die, the other bear down on Florida. Some theories suggest—though still unproven—that something as small as the effect of a flock of birds flying into the originating "warm moist spot" may end up changing a storm's history. Or perhaps the storm simply falls over in a light breeze. Or the low-pressure system may pass over a tiny atoll or island and be thoroughly changed in its nature. The errant butterfly of ecological legend may not be enough to change a storm's course, but if an embryo disturbance passes over a single resort hotel, that might well be enough.21

A scientist at MIT, Edward Lorenz, independently discovered the same phenomenon when he did the same storm-modeling run on two different computers, the old one using data to three decimal places, the new one to six. His model storm, too, produced major deviances after setting off similarly—a relatively tiny change, from 3.461 in the first run to 3.461154 in the second, produced large differences in both the storm's intensity and its predicted path. It was Lorenz who called the phenomenon "sensitive dependence on initial conditions," now more popularly known as chaos theory22

Will we ever be able to achieve absolute accuracy in storm forecasting? Only, says Chris Fogarty of the Canadian Hurricane Centre in Dartmouth, if we have data inputs every few inches across the planet, both vertically and horizontally, something that is clearly impossible. With that many sensors, there'd be no room for people. And even if we were to blanket the earth with a sensor for every molecule, what then? The atmosphere contains more molecules than there would be electrons in any computer, so the calculations would end up being slower than the reality they were forecasting— you'd have a forecast that arrived after the event being predicted.23

The charts of historic hurricane tracks pinned to bulletin boards in hurricane centers everywhere need to be interpreted with great caution. Both the tracks and their numbers are truly unpredictable, classically chaotic. To speak of "increasing numbers of severe storms," or of "the storm of the century," or to predict the numbers that will occur in any one year is, in Zebrowski's phrase, to "entertain a statistical delusion." Because hurricanes, like other natural forces, are chaotic systems, you cannot predict how many will happen by looking at what happened in the past. There may be apparent patterns, but they are illusions. So if you know that from 1951 to 2000 the average annual number of named tropical storms was ten, six of which became hurricanes, or if you know that last year there were fifteen named storms and nine hurricanes, this average and this raw number are entirely useless to predict what will happen this year, or next.

IV

So great is the appetite for data about hurricanes that pilots have been flying into their deadly vortexes for more than sixty years, breaking through the maelstrom of the eyewall into the calm center of the eye itself, a beautiful and terrible place. They did this at first because they were told to (they were, at the time, military pilots in a wartime situation) and partly because the pilot's code of permanent insouciance meant that the challenge was irresistible. It was no surprise, then, that the pilots who did so referred to their flights as penetrations, and to the storms as females—the practice of assigning storms female names had started only a few years earlier.

The first deliberate penetration into an eye was in the summer of 1943. The pilot was Joseph Duckworth, at the time a flight instructor and