In all fairness, the wildly disparate forecasts made by these four respected research groups are rooted in a known, but little-understood scientific phenomenon: The space dust that causes the Leonid meteor shower spreads out more and more every year.

The dust -- mostly particles no larger than sand grains -- is the exhaust of comet Tempel-Tuttle, which orbits the Sun every 33 years.

On each pass, fresh material is boiled off the comet's core by the Sun. And because the comet's orbital path is constantly changing by a small amount, each dust trail is in a slightly different location. Some of the dust Earth will encounter this weekend has been floating through space since before the United States was a country.

Cooke thinks of each trail as a river. As Earth wades through a dust stream, it is shallow at the edges -- less dust. The river is deepest -- containing the highest concentration of dust -- in the middle.

Exactly why the dust moves, and by how much, is what vexes the scientists.

Larger particles are nudged into new paths by the gravity of the Sun, Earth and the other planets, Cooke explained. Smaller bits are less affected by gravity, but they are bullied around somewhat by the pressure of solar radiation -- charged particles that rush constantly outward from the Sun.

One of the most dramatic Leonid showers in history, which occurred in 1833, involved a pass through a dust trail laid down roughly 100 years earlier. The dust was concentrated into a relatively narrow stream, researchers suspect. Some of the 2001 forecasts are based in part on what happened in 1833.

There is no firmly established rate of spreading to factor in, however, and the planet will plunge through at least three dust streams this weekend -- from 1866, 1767 and 1699.

The North American peak will be related to the 1767 stream. The result will almost certainly be something less than what occurred in 1833, when people witnessed a storm of meteors -- one observer in Boston reportedly saw more than 8,000 meteors in just 15 minutes.

For scientists, an hourly rate of 1,000 shooting stars qualifies a meteor shower to be called a "storm."

With all their adjustments in place, each research group now thinks it's got 2001 figured out.

The differences in their methods are subtle. Jenniskens told SPACE.com that his work is based in part on observations of the 1999 and 2000 Leonids from aircraft. He expects that Earth will pass closer to the middle, densest part of the stream than do the other forecasting teams.

Asher and McNaught have looked at historical accounts of meteor showers going back to 1833. Their recent downward revision was based on the inclusion of an "aging parameter," developed by Esko Lyytinen's team, which accounts for how solar radiation might disperse the dust.

Asher said the large uncertainties still associated with meteor forecasting mean that no one knows exactly what's coming Sunday morning. "The estimates of meteor rates by ourselves, by Lyytinen, by Jenniskens, and even our 'old' model, are all within the bounds of possibility," he said.

Cooke, whose expectations are the most modest, thinks the other groups are too optimistic in their calculations of how much dust will be left in the center of a "river" that has been flowing through space for centuries. He said meteor data from the 1833 showers and other historical sources are likely inaccurate, being drawn from nothing more than old newspaper accounts.

So Cooke's calculations are based only on what has been observed each November since 1966.

Yet even Cooke has hedged his bet. Before he left for Hawaii, he coordinated with researchers from NASA and elsewhere, setting up groups that will fan out to New Mexico, Guam and the Gobi desert in Mongolia. Each will train a couple of video cameras on the sky to record shooting stars for later research, in an effort to improve future predictions.

But for those of us who can't fan out, what's the best advice?

For that answer, we turned to Joe Rao, who has written several scientific papers and popular articles about meteor showers. Like many American scientists who've been monitoring the changing forecasts and making travel plans, Rao told us he'll be watching the Leonids from the Southwest.

And what does Rao expect to see?

Rao is familiar with difficult forecasting challenges. He's also a meteorologist, dealing nightly with Mother Nature's whims for viewers of News 12 in Westchester, New York. And he is clearly more comfortable predicting the weather.

But he can't resist throwing his hat into the Leonids forecasting ring. He does so conservatively, recalling that past Leonid meteor showers have produced fewer shooting stars than expected.

He said the forecasts of Esko Lyytinen and his colleagues have been the most consistent over the past couple of years. But even those have been off by a factor of two.

"If pressed for a prediction of rates for the Leonids over North America, I'd suggest 1,000 to 1,200 per hour," he said, though he added that it could go as low as 800.

Either way, that's more than 13 every minute. One every few seconds. An event of a lifetime for anyone lucky enough to see it.