Skip to content
Weather
Link copied to clipboard

Groundhog says six more weeks of winter. Don't worry: He's usually wrong

When it comes to forecasting, the groundhog is only human. Actually, not even that.

Groundhog Club co-handler Al Dereume holds Punxsutawney Phil, the weather prognosticating groundhog, during the 132nd celebration of Groundhog Day on Gobbler's Knob in Punxsutawney, Pa. Friday, Feb. 2, 2018.
Groundhog Club co-handler Al Dereume holds Punxsutawney Phil, the weather prognosticating groundhog, during the 132nd celebration of Groundhog Day on Gobbler's Knob in Punxsutawney, Pa. Friday, Feb. 2, 2018.Read moreGene J. Puskar / AP Photo

Folklore's most-famous prognosticating rodent has been weathering a slump, but the human meteorologists who spend their careers mud-wrestling with atmospheric chaos are offering no sympathy. They have their own problems.

In a ritual that has become a national metaphor for the repetitiously tedious, Punxsutawney Phil saw his shadow Friday morning, meaning the nation can expect six more weeks of winter. If he had not seen it, winter would have been toast, according to tradition.

So how's he doing? In the last 30 years, Phil is 7-for-30, based on the government's analysis of U.S. temperature data. In short, mathematically, your guess is at least as good as Phil's.

The cold and wintry precipitation expected during a not-so-super weather weekend might be a foretaste of what's to come, as the government's and commercial forecasts favor a chilly and wet, but not necessarily snowy, February.

But a cautionary word: It's not just Phil. The atmosphere, what scientists call a nonlinear chaotic system, continues to bedevil the meteorologists who are trying to figure out how it is going to behave.

They analyze computer models, unavailable to Phil, that simulate the state of the atmosphere into the future. They look at clues, such as temperature changes over the tropical and northwest Pacific and how they might affect weather downstream in the United States, and various large-scale swings in air pressure.

But right now the computer models so successful in shorter ranges are out of their depths beyond about 10 days. (Yes, we know, sometimes 10 hours.) And those clues, or "teleconnections," aren't always reliable.

This winter has been a case study. Pre-season outlooks mentioned La Niña, the anomalous cooling of surface waters over a vast expanse of the tropical Pacific. Historically, the South is warm and dry during La Niña-influenced winters. So what explains the accumulating snows that have fallen atop Atlanta and even Savannah?

"You're not supposed to be getting four snow and ice events in the South," said Paul Pastelok, the veteran seasonal forecaster at AccuWeather Inc. "Teleconnection ideas are getting outdated."

The over-arching problem: The atmosphere is a three-dimensional mix of gases, water vapor, and particulates. It extends from the Earth's surface to the boundaries of space. It is affixed to a sphere rotating at 1,000 mph and is engaged in an ever-changing relationship with the solar furnace that incites its whims. And that's not accounting for the effects of various manmade gases.

Computer models rely on the worldwide observation networks to capture an "initial state" of the atmosphere, to gauge how it changed during the previous several hours and calculate how it will evolve in the hours and days ahead.

But they are at a serious disadvantage even before those calculations begin. Observations are imperfect, with gaps over the oceans, lesser-developed nations, and through the depths of the atmosphere.

"Do we  really know what the winds are out in the Pacific 500 feet above the ocean?" asked Matthew Rosencrans, head of forecasting operations at the government's Climate Prediction Center. "There's a whole lot of nuance in that structure.

"We're not able to capture all that detail all over the globe at this time."

Various computer models use different methods to compensate, and that's one reason that at times the computers appear to be squabbling.

To supplement computer guidance, examine such things as snow cover, which affects temperature and storm tracks; analogs – how current patterns match those of the past; those aforementioned teleconnections; and old-style forecasting expertise.

For all that, the most reliable signal is what has happened in recent winters. "The trend is your friend," said Rosencrans. The nation's winters have become warmer more than other seasons, according to the government.

Citing that trend and La Niña, the climate center came down on the mild side in its winter outlook for the Philadelphia region. So far temperatures have averaged slightly below normal, and snowfall above.

AccuWeather also called for a mild December. "I would have to say that wasn't forecast very well," said Pastelok. However, he added that the snow forecast, around 25 inches, still could work out. The official total stands at 13.2.

Local forecasts issued by TV meteorologists had mixed results.

NBC10's Glenn Schwartz might claim the trophy. He predicted that December would end a degree below normal, and officially it was minus-1.4 in Philadelphia. However, that 8.6 inches of snow measured at the airport was twice his projection. He had January 2 degrees above normal; it was slightly colder than average.

With the caveat that the winter isn't over, he gets bonus points for this observation: "I'm not expecting that we'll see a single huge snowstorm. We will likely get several smaller snow events."

Cecily Tynan at 6abc saw above-normal precipitation and "rain-to-snow events." Since Dec. 1, precipitation deficits range from 30 percent to 45 percent, but her prediction of 18 to 24 inches for the season remains in play. As for the Fox29 outlook for above-normal winter temperatures, that's not looking so good.

Will meteorologists get better at this? Computers, the keys to the future of reliable long-term outlooks, are getting ever more powerful, said Rosencrans, but progress is likely to be incremental.

He predicts that Phil and his successors will be long gone before forecasters will say with any precision what the weather will be like on a given day two or three months in the future.

"The science just isn't there yet," he said.

Phil was unavailable for comment.