The Farmers Almanac threw rain on football fans' parade as far back as August — snow, actually — boldly looking six months into the future and seeing a big storm at Metlife Stadium in New Jersey's Meadowlands on Super Bowl weekend.
On Friday, 16 days prior to the game at which the Denver Broncos hope to represent the AFC and secure a return to NFL supremacy, AccuWeather had a less dire prediction.
Still, it was not one to warm the hearts of anyone blowing their vacation fund to witness it in person. AccuWeather on Friday was calling for a high of 32, and a low of 22, with "snow showers possible."
AccuWeather, it should be noted, now distinguishes itself by daring to post detailed forecasts 45 days out; if you're planning your March 2 in Boulder, for example, the site currently promises a high of 48 and partly sunny.
Time for a reality check.
Meteorologist Bob Henson, a science writer at the National Center for Atmospheric Research in Boulder, said Friday that a detailed and reliable prediction of weather 16 days in the future is a forecasting feat that remains an elusive goal.
In a blog post at the UCAR/NCAR site AtmosNews titled "A Super Forecasting Challenge," Henson tosses cold water on the notion that anyone can say with confidence at this point what the Super Bowl contestants and those in attendance will be enjoying — or suffering through — in the swamplands of north Jersey on Feb. 2.
"Even the official NWS local forecasts that go out seven days need to be viewed with skepticism at the far end, since their skill is little better than climatology at that point," Henson wrote.
"Climatology," in that context, means the average conditions one would expect on a given date, simply based on past weather records.
On the web
Read Bob Henson's blog post "A Super Forecasting Challenge" at www2.ucar.edu/atmosnews
For the record, the weather observation site at Liberty Newark International Airport, the site most applicable for conditions at MetLife Stadium, shows that the average high and low for Feb. 2 is 40 and 25, respectively. The average Feb. 2 temperature there at the scheduled 4:30 p.m. MDT kickoff time is 34 degrees.
As for precipitation, at least 0.1 inch of snow has fallen there on Feb. 2 about 15 percent of the time since 1931. A far more detailed rundown of the weather history at this year's Super Bowl location is offered at biggame.com, a comprehensive site produced by the office of the New Jersey State Climatologist at Rutgers University.
Henson grew up in football country — Oklahoma — but describes himself as no more than a casual football fan. Still, he is well aware of what a communal touchstone the Super Bowl has become in our culture.
"I know when the Broncos have won, because someone always steals my paper. Seriously," he said.
'Too much variability'
Without citing AccuWeather by name, Henson said he puts little credibility in forecasts that dare to reach too far into the future.
"Pinpoint forecasts for a city for any one day become really problematic after a week or two," he said. "Once you get past a week or two, you simply can't pin down, city by city, day by day, what's going to happen. There's too much variability in how weather patterns will unfold, that far out."
Which is not to say that forecasters aren't giving it their best shot.
The National Weather Service simulates the atmosphere out to 384 hours, with model runs that are updated four times each day. The results feed into the NWS outlooks for six-to-10 day intervals and eight-to-14-day intervals.
The main tool used for extended-range forecasting by the NWS is the Global Forecast System model, and Henson's post links to to one, valid for Feb. 1, which shows heavy precipitation from New York City across New England. However, the GFS cycle includes an ensemble of 20 other runs, which vary in what they show for the region.
"All you need is a slight change in timing," Henson said. "Let's say you knew there was going to be a major winter storm somewhere in the first week of February. If it's 14 days out, and your timing is off by 7 percent, that could be the difference between having a nice day and a snowstorm."
Or, he said, the snow line reaching as far south as Princeton, N.J., or only Hartford, Conn.
'Problem of predictability'
According to NCAR senior scientist Mitchell Moncrieff, it comes down to the inherent chaos built into the Earth's ever-changing atmosphere.
"It is because of the sort of fundamental problem of predictability," Moncreiff said. "Small perturbations in the model can grow exponentially, very rapidly... It's the fundamental properties of the equations of motion."
Moncrieff, who focuses on long-term forecasting challenges, cited disturbances to global weather patterns, such as the Madden-Julian Oscillation, which send atmospheric waves from the Indian Ocean, capable of affecting weather in North America 10 days later.
"In the last few weeks we had the polar vortex (affecting the eastern United States), and that is an example of a larger-scale coherence," Moncrieff said. "With these things we don't understand in models quite how they work yet. When we can, then we may be able to get somewhere" with longer-term forecasting.
With all of that said, Henson still dipped a tentative toe into the hazardous waters of long-term forecasting, with an eye toward Super Bowl Sunday in East Rutherford, N.J.
"It's going to snow, or it's not," he ventured.
"That's often what I say about the football outcome. Either the Broncos will win, or they will not."