News: Weather's Crystal Ball: Hit Or Hooey?

Weather is hot. The Weather Channel broadcasts to 75 million meteorological voyeurs, and www.weather.com receives 10 million page views per day. Look around your office: The world can be neatly divided into people addicted to watching weather reports and people who believe a coin toss delivers equally accurate information.

Skiers inhabit the former camp. And when it comes to planning ski vacations, they are moths drawn to the light of long-range forecasts. The folks at Weatherplanner.com are banking on it. The free website isn’t satisfied with predicting weather a few days or even weeks in advance. Weatherplanner.com offers predictions a full 365 days ahead.

Plug in a date and a location and two weeks of daily forecasts materialize on screen. The intrigue of looking into Mother Nature’s crystal ball has generated strong cyber traffic, with about 1.6 million visitors per month. Planning a trip to Stowe, Vt., for Valentine’s Day 2001? Expect snow, with high temperatures in the mid 30s to low 40s.

But is this science or nonsense?

“Science,” claims the website. “Most short-range forecasts are developed from analyzing many small details, then running these details sequentially through a series of physical equations.” The Weatherplanner forecasts, however, “focus on the reactions and interactions of many large-scale features, which are run through a different set of physical equations for each daily output and analyzed to determine how various trends will influence large areas and then more specific locations.”

Both the American Meteorological Society and Cliff Mass, a professor of atmospheric sciences at the University of Washington, take the side of nonsense. “The scientific literature consistently shows it’s not possible to provide accurate day-to-day forecasts much more than a week out,” Mass says. He also states that the analog methodology used by Dr. Irving Krick, the meteorologist the site credits for its science, has been discredited by research.

Rather than taking his word, Mass recommends performing your own test. So I did. Using Weatherplanner.com, I planned the dates of a backcountry ski trip in Washington’s Cascade Mountains around the website’s predicted window of fair, clear weather. Weatherplanner got it right two out of five days, which-depending on whether you’re a person who watches weather reports or one who doesn’t-you’ll recognize as the coin-flip accuracy of most weather prognostications.