A World of Data to Help Your Plans to Go Skiing - Hillary Rosner [The New York Times]
March 11, 2008
After a weekend of unusually wet snow, Jim Steenburgh stood atop Alta's Sugarloaf Pass, surveying the mostly blue sky. Clouds were forming to the south, moving in over Sundance. But no, Dr. Steenburgh predicted, the weather would not come this way.
By lunchtime, fat flakes were falling, and soon it was snowing so hard you could barely see the terrain. Dr. Steenburgh, 40, was asked if he wanted to revise his forecast. He did not. "It'll only be a quarter-inch accumulation," he predicted.
Back at his office, Dr. Steenburgh, professor and chairman of the meteorology department at the University of Utah, wasted no time in tracking the day's atmospheric events. At 5:36, he sent an e-mail note: only a trace of snow had accumulated, and the snow water equivalent measured on the mountain the amount of water contained in the snow, related to its weight was just .02 inch, a very dry snowfall.
Dr. Steenburgh knows the Wasatch Range and its particular weather patterns better than just about anyone except, perhaps, his longtime ski buddy, Mike Kok (pronounced coke), the meteorologist for the Alta Weather Program. Dr. Steenburgh and Mr. Kok are among a breed of mountain meteorologists who have mastered the art of applying vast quantities of meteorological data from satellites, models and weather stations to the specific alpine features of their home turf.
Dr. Steenburgh is an expert in locally important forces like the lake effect. The Great Salt Lake heats and moistens the air, helping to give the Cottonwood Canyons here about 20 percent more snow than they would otherwise have. During the 2002 Winter Olympics in Salt Lake City, Dr. Steenburgh's lab provided weather forecasts. But despite the fact that Dr. Steenburgh is a self-described "weather weenie" with a deep scientific interest in the goings-on of the atmosphere, at heart he is after the same information as everyone else. "When I go resort skiing," Dr. Steenburgh said, "I want to know whether my 7-ysar-old is going to complain that it's too cold or too nasty."
For the answers, meteorologists like Dr. Steenburgh and Mr. Kok rely on sophisticated computer models and technology like satellite imaging. But they are still wrong sometimes, precisely because there is so much to take into consideration. Mountains create their own weather systems and present a suite of phenomena like lake effect and orographic lift, which is the forcing of air to rise by rising terrain, like a mountain.
Mountain meteorology, like skiing, requires not just technical skill but intuitive understanding of movement and motion and flow. "The atmosphere is a fluid," said Larry Dunn, meteorologist in charge of the National Weather Service in Salt Lake City, who skis with Dr. Steenburgh and Mr. Kok. "It behaves just like any other fluid, and you can use physics and mathematics to describe the behavior of that fluid in equations. If you knew the exact state of the atmosphere over the whole planet right now, if you knew where the cold air was, the warm air, the moisture, the wind and its directions, everywhere at all levels of the atmosphere, then you could plug that initial condition into those equations and solve them out into the future and get the exact state of the atmosphere tomorrow or the day after."
While weather has intrigued mankind throughout history, the tools of modern forecasting are only about a decade old. In the late 19th century, the telegraph first enabled simultaneous weather observations from different regions to be pieced together into rudimentary maps of weather patterns. In the early 1900s, European mathematicians first sought to predict the weather using equations, but by 1922 the science had advanced only to a point: If you could figure out a way to squeeze 64,900 people together into the same room to perform calculations, you could, theoretically, generate a weather "prediction" although weeks after that weather had come and gone. In the 1950s, computers rose to the task.
Today, weather data comes from satellite imagery and soundings measurements of atmospheric temperature and humidity; from weather monitoring stations on the ground and on buoys in the ocean; and from balloons equipped with weather instruments and transmitters, sent up in 12-hour intervals from sites around the world.
In this country, much of this information is delivered in raw form, viar satellite feed, to the Boulder, Colo., offices of Unidata, a quasi-governmental research corporation that collects the data, renders it readable and sends it out, along with software to analyze it, to meteorologists around the country.
At Unidata, a climatologist, Jeff Weber, watches the data stream in 25. gigabytes of it every day. "There's a lot of number crunching," Dr. Weber said billions of calculations, in fact, involving things like solar radiation, precipitation, wind speed and direction, and vapor pressure, for every series of coordinates on the satellite grid, at each level of the atmosphere not to mention a consideration of the laws of thermodynamics.
"In the old days, we received maps from computer model forecasts at only a couple of levels in the atmosphere," Dr. Steenburgh said. "Today, thanks to high-speed networking, we don't receive the maps. Instead we get output right from the computer models, and we generate the maps ourselves."
Dr. Steenburgh writes software programs to download exactly the information he needs. "We can slice and dice this output much like an M.R.I, of the atmosphere and get a much better picture of what is happening," he said.
Weather models essentially take every shred of available weather information and use them to form a threedimensional picture of the atmosphere. They then solve a series of equations that ask what will happen 12 hours, three days or a week out. There are global models, national models and smaller-scale regional models, all of which are run on supercomputers.
Local forecasting "always starts with the big picture how highs and lows are moving," Dr. Steenburgh said, adding: "Then we make adjustments for what the local terrain does. How does the mountain range affect that front Every place has a different local climatology, so it's about adjusting the characteristics of your mountains."
Partly hidden in the trees at 9,600 feet, just off Alta's Main Street trail, Collins Mid-Mountain weather station looks like something the Brady Bunch children might have built in their backyard with a 1970s kit and a little help from Dad. A metal bucket catches snow, which is melted using antifreeze. The weight is recorded to establish the snow water equivalent. Giant measuring sticks one outfitted with an ultrasound "gizmo," as Dr. Steenburgh called it gauge the snow depth from the ground up. A cylindrical device topped off by a ring of plastic triangles that resemble a truck's mud flaps measures precipitation.
Despite its homemade appearance, this is one of the best automated mountain weather stations in the country. Snowfall, snow-depth and snow-density measurements are recorded hourly and relayed by radio signal to the top of the Collins chairlift, where they are transmitted by phone line to the base. Almost instantaneously, the weather watchers of Little Cottonwood Canyon have another set of inputs for their arsenal.
Hourly automated observations like this are extremely useful to forecasters like Mr. Kok, but they are even better when they are "ground-truthed" by a person. "The computer doesn't show you what kind of snow crystals are forming," said Daniel Howlett, Alta's assistant snow safety director who is known as Howie. "All the computers in the world won't replace being outside."
Mr. Howlett and his boss both have weather monitoring equipment in their yards, which they check at 3:45 a.m. They call Mr. Kok, who issues daily forecasts at 4:30 a.m. and 4:30 p.m.; when it is snowing, he is up at 2 a.m. consulting the maps. As soon as they have received Mr. Kok's forecast, Mr. Howlett's patrollers begin their avalanche planning for the morning, and then hit the slopes to set off charges.
The point of weather forecasting is practical and defensive, after all. And nothing is more practical than avoiding avalanches.
Hillary Rosner [The New York Times]