Punxsutawney Phil, the weather forecasting groundhog, will be rudely rustled from his winter slumber Sunday morning to answer the question of the day: shadow or no shadow? Six more weeks of winter or an early spring?
Why this fascination with Phil? Well, scientifically speaking, long-range forecasting is at best a crapshoot.
Jason Samenow, "The Capital Weather Gang" chief meteorologist writes: "Let me state emphatically that no one – with any degree of accuracy – can predict the specific days when cold snaps or storms will occur months in advance."
Which leads us to a shocking conclusion, summed up in an equally shocking headline from a 2012 story by ABC News: "Punxsutawney Phil Isn't Always Right."
Say it ain't so!
Well, just last year Phil (and/or his human collegues) got it wrong – calling for an early spring. In the true tradition of underlings the world over, Phil's handler, Bill Deely, took the blame for the botched forecast, saying his "groundhog-ese" was just rusty enough to misunderstand the famous prognosticator. In any case, it was the second consecutive year for a failed Phil forecast.
According to Stormfax.com, the groundhog method of forecasting is, well, less than reliable. Phil's gotten it right just 39 percent of the time since 1887. Forecast models, probabilities and all else aside, that's worse than a simple coin toss.
Not to point out the obvious, but the safe forecast for Phil would be always to predict six more weeks of winter, since the first day of spring is on March 20, just about a month and a half after his annual shadow-casting observation.
This year, "The Farmers' Almanac" secret forecasting formula predicted a "bitterly cold winter" for most of the nation. Spot on, you say!
Well, maybe, but the publication, in print since 1818, is calling for a winter superstorm to disrupt Sunday's Super Bowl game. Unless the guys using supercomputers to noodle the weather are entirely wrong — and short-term forecasting is, contrary to popular belief, very accurate these days — that's not going to happen.
As for "The Old Farmer's Almanac," we wrote in 2012, that its reliability was questionable, too. The publication, founded in 1792, boasts an accuracy rate of 80 percent, a similar one as claimed by "The Farmers' Almanac." But Jonathan Martin, chairman of the University of Wisconsin-Madison's Department of Atmospheric and Oceanic Sciences, told NPR that his guess is "their success rate is more like half what they say."
Ouch. That's no better than Phil.
As Penn State meteorologist Paul Knight points out, at least in the case of "The Farmers' Almanac," "[they] say from November 5 thru 10, for that whole period: sunny/cool. If one day is sunny and cool, does that count? Does every day have to be sunny and cool? If you held them to every single word for the entire area and every word for the entire period, then I say they might not even be right one third of the time."
Samenow puts Phil and both of the almanacs in roughly the same league, reliability-wise.