Just when we were thinking that winter heating bills might not bankrupt us after all—the official seasonal forecast from the federal Climate Prediction Center calls for warmer temperatures everywhere except for a triangle from Seattle to North Dakota to southern California—along comes bad news from Siberia. The region had relatively little snow cover throughout October, but November saw "an epic advance" in the amount of land covered in the white stuff, says Judah Cohen, director of seasonal forecasting at Atmospheric and Environmental Research, Inc., an R&D firm in Massachusetts. As Cohen sees it, extensive snow cover in Siberia sets in motion a train of meteorological events, with energy waves propagating to the stratosphere, where they weaken the vortex of winds over the North Pole and … well, suffice it to say that his just-issued forecast calls for a cold start to December in the East and a mild early January, followed by a possible return to the deep freeze around Martin Luther King's birthday.

The contrast between the official U.S. forecasts and those of the new breed of meteorologists like Cohen illustrates a face-off in the small community of long-range weather seers. Traditionalists have basically stuck to what they've been doing for decades. In the official U.S. forecast, for instance, since recent winters have trended warm, due in part to the greenhouse effect, this one is assumed to tilt that way also, other things being equal. The government also bases its winter forecast on El Niño, the warming of the water in the eastern Pacific Ocean around the equator: depending on whether the sea is tepid or chilly, it will shift the all-important jet stream either to where it blocks frigid Canadian air from reaching well into the United States or to where it makes the jet stream as ineffectual as the Jets' defense against Tom Brady.

These traditional methods aren't bad. But in head-to-head comparisons published this past summer, the resulting forecast has been less accurate than that of scientists who grapple more directly with the physics of the atmosphere. Britain's Meteorological Office has begun basing winter forecasts on sea-surface temperatures in the Atlantic. The traditionalists are especially at a disadvantage when the El Niño signal is wishy-washy, as it was five years ago, giving the traditionalists little to go on. In August, the National Science Foundation announced that Cohen, whose research it funds, and his Siberian snow had hit the mark more often than the official winter forecasts.

The way Cohen sees it, the meteorological ball starts rolling in October, when snow begins to pile up across Siberia. The more snow, the more cold, dense air in the lower atmosphere, because the snow cools the air above it, and colder air is denser air. This mass spreads over northern Eurasia and pushes a wave of energy up into the stratosphere. That warms the stratosphere over the Arctic, pushing the river of air that is the jet stream south. The jet stream spills down the eastern face of the Rocky Mountains into the eastern half of the United States, explains Cohen. On a map, the jet stream then looks like a slide starting in the Rockies and dipping south as it travels east. With nothing to block them, cold air masses from the north easily dip well into the Midwest and East. Presto: a colder-than-normal winter. If the October snow cover in Siberia is less extensive than the historical norm, then reverse all the above.

Cohen was alternately pensive and frustrated as he fretted about Siberia. "October snow cover came in low, which is a warm signal," he said. "But the other atmospheric variables that I watch give me a cold signal. Complicating the forecast was an epic advance in Siberian snow cover in early November, which suggests the cold cycle. So I expect a cold start to December in the East." (This was well before the first weekend of the month saw frigid temperatures and snowfalls from the Midwest to the East.) A full-blown cold cycle usually brings four weeks of cold. But because Siberia didn't make up its mind until late in the game, says Cohen, "I expect the cold to last just two to three weeks," with mild weather kicking in from about Christmas into January. "I'm expecting a return to colder weather in the second half of January," he says. "But there have been seasons when the cold cycle never gets re-established. If that occurs, the mild period could persist through January."

The Climate Prediction Center is again using El Niño and recent trends to anchor its winter forecast. It calls for a warmer season than the 30-year norm, especially in the heartland, with the Northwest and Rockies a toss-up. That reflects La Niña—cold in the eastern Pacific—conditions. La Niña shifts tropical rains west, explains Michael Halpert, deputy director of the center. That transfers heat to the atmosphere farther west, which jerks the jet stream back. The jet stream then dives down into the Northwest or the Rockies (allowing cold Canadian air to penetrate), then snakes north again over the Midwest and East (blocking that frigid air).