1. “Long-term forecast? Your guess is as good as ours.”

WEATHER FORECASTERS have gotten pretty good at nailing the outlook two to three days ahead. The problem is, “everyone wants to know what the weather is going to do next weekend,” says Paul Karpowicz, president of the Meredith Broadcasting Group, and forecasting the weather a week or more down the road isn’t so easy.

Eric Floehr, founder of ForecastAdvisor, which tracks the accuracy of predictions, looked at high-temperature forecasts from the National Weather Service, AccuWeather and other organizations for 2008 and compared their
numbers with actual temperatures. When predicting highs for the following day, they were off by about three degrees; when forecasting nine days out, they missed by nearly seven degrees. Doug Young, performance branch chief at the National Weather Service, says his organization’s precipitation forecast for seven days out is only 55 percent accurate. “You’re almost flipping a coin at that point,” he says. Ray Ban, consultant for The Weather Channel, says the best forecasters can do is try to convey the uncertainty of long-range predictions.

2. “We’re pretty accurate—as long as the sun is shining.”

ONE OF THE MOST important things forecasters can do is tell you when bad weather is on the way. Unfortunately, they’re not very good at predicting rain. That’s especially true in summer, when most rainfall comes from thunderstorms, which are small, unpredictable and hard to track. It’s often difficult to tell where they’re headed or whether they’ll produce any rain. Most models for forecasting weather divide the country into a grid of squares that cover about 55 square miles each, though some have smaller squares. Whatever the square’s size, rain, snow, sun and temperature are forecast as a whole for each one. Since most thunderstorms are smaller than the squares, it’s tough to predict exactly where it will rain. “Forecasters are terrible at telling you if rain is going to fall where you live tonight,” says William Gallus, a meteorology professor at Iowa State University.

Even when forecasters can track a larger storm’s direction and speed, it’s very hard to determine its intensity and whether or not it will weaken or gain traction. That’s why most forecasters err on the side of caution, says
Floehr, issuing predictions like “50 percent chance of rain countywide.”

3. “We’re often more show biz than science.”

THERE ARE NO specific job requirements for TV weather people. Anyone can do it. Some have degrees in meteorology; others have no background in science. In markets where weather matters—such as hurricane-prone parts of Florida—real weather experts often dominate local newscasts. In Miami, for example, the No. 1 local news station, WPLG, features a former director of the National Hurricane Center. But in most fair-weather markets, “if you picked 100 weather forecasters,” says Tom Herwitz, a former president of station operations at Fox, “you’d find a lot of comedians and radio DJs.”

At some stations the evening-news weatherperson, who can earn between $100,000 and $300,000 a year, may not even be involved in putting together forecasts, which are based on data from the National Weather Service and private firms like WSI Corp., which is owned by The Weather Channel Cos. “Being a trained meteorologist where it’s sunny and warm every day is less important,” says Larry Rickel, president of consulting firm Broadcast Image Group. “You can have more personality.”