Joanne Canen

As anyone who's been caught in a downpour without an umbrella knows, predicting the weather is no simple task. This is especially true of long-term forecasting. Yet since 1818, the Farmers' Almanac—published annually from its Lewiston, Maine offices—has been trying to do just that for the continental United States. Can you really predict the weather a year from now?

Penn State meteorologist Paul Knight is more than a little skeptical.

"The ability to predict events that far in advance is zero," says Knight. "There's no proven skill, there's no technique that's agreed upon in science to be able to do that."

The Farmers' Almanac's Web site explains that its forecaster (referred to only by his pseudonym, Caleb Weatherbee) uses a "top secret mathematical and astronomical formula, that relies on sunspot activity, tidal action, planetary position and many other factors" to predict weather sixteen months in advance for seven different U.S. climate zones.

According to Knight, the Almanac's secrecy is part of the problem.

"If you have something that's really innovative and shows skill, then bring it before your peers," he says. "You don't have to show us everything in case you want to make a business out of it, but give us some idea."

Until the modern era of meteorology, many people relied on publications like the Farmers' Almanac for long-range weather forecasts. Today, radar, satellites, and computer simulations help meteorologists more accurately predict everything from an afternoon rain shower to a tropical cyclone.

Says Knight, even with the advanced technology, it is still difficult for meteorologists to forecast things like rainfall very far in advance. "For precipitation, I don't know anyone short of the prophet Elijah who has any skill in foretelling precipitation more than a couple weeks, perhaps a month in advance," he says.

"If we get it right every once in a while people say, 'Well, if they can do that, then why can't the Farmers' Almanac be right?' Well, because they're using very different techniques. One is scientific; the other is behind the curtain."

Knight points out that the Farmers' Almanac words its predictions imprecisely, making it difficult to assess their accuracy.

"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. In fact, they might be right 10 percent of the time." Acknowledges Knight, "I don't think they're holding themselves to that degree of accuracy, and I don't think other people are either."

Perhaps the strongest long-term predictor when it comes to weather is climate history, the way the weather "normally" behaves in a given region. But awareness of the past record, Knight points out, requires little to no scientific skill.

"I could say things like October 8 to 15 in this area: generally dry, very cool weather expected; first frost and freezes in the valleys. And I would be right probably eight out of 10 years," says Knight. "I could say February 12 to 19: heavy snow along parts of the eastern seaboard. I'm going to be right seven out of 10 years. There is some relative frequency to these things, but to say that this is of great scientific accuracy would be a real misnomer."

While Knight feels that there's nothing wrong with reading the Farmers' Almanac (and with 4 million copies sold each year, it has its loyal readers), just don't expect it to be very accurate, he warns.

"If you want to use that for some kind of planting purpose or guidance more power to you," says Knight. "And if it works for you, great. But will it work regularly? No."

So feel free to read the Farmers' Almanac, and watch the Weather Channel, too, but you might also want to carry an umbrella with you ... just in case.

Paul Knight, M.S., CCM, is an instructor of meteorology in the College of Earth and Mineral Sciences and is co-host and producer of Weather World. He is also Pennsylvania State Climatologist. Knight can be reached at pgk2@psu.edu.