The coming winter in the Washington area will be much snowier and colder than normal - even snowier and colder than last winter - according to the 181-year-old Hagerstown Almanac and television weatherman Gordon Barnes.

On the other hand, the Old Farmers's Almanac, established 187 years ago, predicts that it will be colder than normal, but with less snow than last year.

But then, one of the "grand old men" of meteorology. Hurd C. Willett, professor emeritus of meteorology at the Massachusetts Institute of Technology, forecasts a warmer and wetter winter than normal in the mid-Atlantic states.

And finally to come full circle, the National Weather Service last month predicted warmer and probably drier than normal weather in the mid-Atlantic states, at least through January. Its full winter forecast, through February, will be issued next week.

Long-range weather forecasting is still an iffy thing, the odds at best being only 3 to 2 on predictions of 30 days or more, according to Don Gilman, chief of the National Weather Service long-range prediction group. The Weather Service has been making cautious seasonal predictions for 19 years but has only made them public for the past several years at the request of Congress and press.

The farmers' almanacs, however, have never been hesitant about foretelling the weather, making not only general weather predictions a year in advance but printing 12 months of detailed, almost day-by-day weather forecast - along with tips on the stars, tides, gardening, animanl husbandry, recipes and how to grow old gracefully.

Even Gordon Barnes, who runs a private weather forecasting company, recently gave viewers of WDVM-TV the specific dates of what he predicts will be this winter's major cold spells, warm spells and storms - forecasts laughed at by many meteorologists who consider such specific, long-range predictions as pure guess work.

Barnes has predicted that the first major winter storms will be Dec. 3 through 5 and Christmas Eve and Christmas; the first major cold spells, Dec. 10 through 15 and Dec. 29 through New Year's Eve.

How the predictions are determined is almost as varied as the weather itself. Barnes and Willett, whose winter forecasts appeared in a recent issue of U.S. News and World Report, both subscribe to the sunspot theory - that regularly recurring disturbances on the sun's surface affect weather on earth. Despite that theoretical agreement, the two forecasters' predictions for this winter are almost diametrically opposed; Willett forseeing a warm, wet winter, and Barnes, a cold, snowy one.

"The National Weather Service approaches forecasting in an entirely different way then we do," says Barnes. "They use a complex computer model, while we rely on data going back to 1899 and sunspot activity.

"Every 72 days there is an increase in sunspot activity. This is very important if you have a weather system capable of producing a deviation from normal on or about that date. The deviation becomes more extreme.

"Thus if the normal high is 65 degrees and I see a weather system coming that would bring 70- to 75-degree weather, then I'll predict 75 to 80. I don't know why it works, but it works. December 20 is the next date of sunspot activity, but I'm not yet making any specific prediction for that date," Barnes said.

The Hagerstown Almanac believes lunar cycles affect the weather to some extent, its forecaster William E. O'Toole told The Washington Post earlier this year. O'Toole is a mathematics professor at Mount St. Mary's College in Emmittsburg, Md.

The Old Farmer's Almanac believes that weather is affected by the sun, the moon and cosmic rays, as well as other natural phenomena, and makes its prognostications, it states, with "modern scientific calculations . . . and a formula devised by the founder of this almanac in 1792." Its forecasts are compiled by Richard Head, a former National Aeronautics and Space Administration (NASA) solar physicist.

Among weather proverbs used in formulating the Farmer's Almanac forecasts, according to the almanac, is "The nearer to midday or noon phases of the moon happen, the more foul or wet weather may be expected during the next seven days."

"Garbage," says Gilman, of the National Weather Service, which places little store by the sun, moon or the length of the wooly bear caterpillar's hair in making forecasts.

"It's very problematical whether there are solar effects in weather, certainly nothing you can predict from. Even if it's true that sunspot activity may be connected to major droughts out West, and even that's questionable, sunspots provide little predition capability. Lunar effects are even more subtle and less practical" in forecasting.

The Weather Service has used ocean temperatures and currents in past forecasts, which some meteorologists think are major weather factors. Gilman, however, has said previously that "their significance also is questionable" and they are no longer used by the Weather Service.

The success of the various forecasters has been mixed. Gilman is skeptical about all long-range forecasts and quicky to point out when he's goofed. "We had a bad forecast this summer. There have been good months, but so far it's not been one of our better years." He noted, however, that his long-range weather group correctly predicted last winter's colder than normal weather.

Barnes, who also came close on last winter, is less reticent. "I've had a great year so far."

The almanacs were wrong on most of their detailed predictions last winter, missing the majority of snow storms, although their general winter forecasts were at least partly correct: Hagerstown predicted a colder than normal winter, and Farmer's predicted a snowy but not so cold winter.

For this Thanskgiving holiday, Hagerstown pedicts "fair and cool" weather for the mid-Atlantic states, and Farmer's forecasts rain or freezing rain ending today with fair weather the rest of the weekend.

Barnes' forecast of sunspot-connected storms and unusual weather spells did not include this holiday weekend.

The National Weather Service 30-day forecast last month predicted November would be warmer and drier than normal. Its three-to-five-day forecast, issued Sunday, predicted warm weather for Thanksgiving Day and Friday.