Not since the tax man cometh has a public servant been so frequently cursed as the weatherman in winter. A meteorologist who correctly predicts a powder day is credited merely with doing his or her job, but if a called-for dump fails to arrive, the forecaster's name is embellished with vivid adjectives as thousands of ski boots kick the nearest furniture.
Ironically, weather prediction has made huge strides over the past several years. The reasons are as far-flung as a jetliner overhead, a Web cam atop Vail's Two Elk Lodge and an IBM SP supercomputer crunching numbers for the National Weather Service. "Ten years ago, (predicting the weather) was like building a house with a circular saw, a hammer and maybe a pry bar," says John Dee, a private meteorologist who also runs a popular Midwest snow-prediction website. "Now we've got cordless drills and power nail guns and laser levels. It just makes the job so much easier."
Skiers are particularly demanding weather-forecast consumers, looking not just for "where" but also for "what." Though the amount of precipitation that falls in different places may be identical, there's a world of difference between storms that dump feather-light powder or sidewise sleet. "The same half-inch of liquid precipitation can turn into six inches of snow or 12 inches of snow, depending on the temperature, winds and crystal type," says Larry Dunn, meteorologist in charge of the National Weather Service's Salt Lake City office. The weather service in Salt Lake City has been evaluating Alta snow forecasts for several years. They've found that, when wrong, forecasters tend to overestimate the number of small storms and underestimate the size of the larger ones. The Alta analysis suggests that skiers (not just in Utah) are probably more likely to get lucky chasing the occasional big, over-the-knee blizzard than the more frequently predicted smaller, boot-high squalls.
As computer power and memory have grown exponentially, scientists have been able to incorporate vastly expanded amounts of baseline data into increasingly sophisticated computer modeling. Twice daily for years, many of the world's countries have sent aloft weather balloons, 300 miles apart, laden with sensors and transmitters that relay information to the ground. These days, that's just the start. Some 3,000 weather stations now scattered on mountaintops and elsewhere throughout the American West feed weather information into a central database hourly. Even newer are sensors attached to airplanes worldwide that automatically transmit weather conditions during takeoffs and descents.
These technological advancements, along with an evolving understanding of atmospheric behavior, add up to more reliable forecasts. "The computer-model accuracy looking at Day Five now is better than for Day Three in the 1970s," Dunn says. Which means that nowadays, at minimum, forecasters see storms coming. If a storm is predicted for Wednesday night, skiers can have more confidence in a decision to take a day off from the office on Thursday. It's the intensity of the storm that remains slippery.
Ski resorts benefit, too, from increasingly accurate weather forecasts, which allow them to make better decisions on how to spend money or deploy personnel. Stowe Mountain Resort, for example, uses two private forecasters. One provides very localized, very immediate forecasts, while the other peers further into the future, says Rod Kessler, Stowe's vice president of mountain operations. Faced with a prediction of high summit winds, for example, snowmaking crews will focus efforts at lower elevations where the new snow won't blow away. Or, armed with a prediction that a big storm is coming in a few days or even a few weeks, the resort might buy last-minute advertising, he says.
For all these advances, predicting winter storms in the mountains is still "half science and half art-form" says Dee, whose winter-snowfall website (johndee.comm) has received as many as one million hits a day. The problem is that even the most advanced supercomputer can't anticipate small-level changes, such as a simple shift in the Pacific winds. "A butterfly flaps its wings and throws the air currents off 10,000 miles away," says Jim Roemer, a private Vermont meteorologist who operates bestskiweather.com, invoking Edward Lorenz's famous example of chaos theory. (It's no coincidence that Lorenz was an MIT meteorologist who was trying to predict weather patterns.) Shortcomings of computer models are the big reason why the human forecaster hasn't been entirely replaced by a hard drive. Often, meteorologists simply call on their years of experience to sharpen a prediction. The more diligent forecasters can recognize when the computer model is going sour and tweak a forecast accordingly, says Peter Manousos, who works at the National Weather Service's Hydrometeorological Prediction Center, which includes the Winter Weather Desk.
Experience is particularly crucial when trying to predict winter storms down to the local, ski-resort level. "There's something very magical, very special about Little Cottonwood Canyon, and the computer doesn't know that yet," says Dunn, who skis about 40 days each winter in the Salt Lake area.
Indeed, the timing and exact trajectory of storms still give forecasters fits. "Computer models did a pretty good job of showing that there was going to be a large storm off the East Coast (in January of 2000)," Manousos remembers. "The problem was, how close to the coast was it going to be?" One day before the storm hit, models suddenly started to bring the storm closer to the coast, he says. The forecast lurched from "partly cloudy, with a chance of snow showers" to "one foot of snow," with more than two feet eventually falling in parts of North Carolina. "Very humiliating for winter weather forecasters," Manousos adds ruefully.
"We're still in the infancy of weather prediction, though we may be getting into our difficult teen years," says Mark Moore, director of the Seattle-based Northwest Weather and Avalanche Center. "We still have to do a lot of interpolation" about what's happening in the vast spaces between all of those weather balloons, airplanes, weather stations and ships at sea that are gathering atmospheric data. That's a huge amount of real estate in which a whole lot of butterflies are flapping their wings.
Dunn views criticism as part of the job of being a meteorologist. Yet he remains a popular guy with his friends each winter. "No matter how much bashing of the weatherman that I hear," Dunn says, "my ski buddies are ringing my phone off the hook trying to get the inside scoop on which day to go skiing. So there must be some skill in the weather forecast."