Forecast: campus weather turns… competitive

When most people talk about the weather, it’s usually to break the ice. But for a group of faculty and students, conversations about the temperature can turn competitive.

For 10 years, UW–Madison has been participating, along with about three dozen other schools, in the National Collegiate Weather Forecasting Contest, which started in the early 1980s. Organized by Penn State, the NCWFC enables undergraduate and graduate students, as well as faculty, to earn bragging rights as the best weather forecasters in their specific categories.

Without using a crystal ball, participants submit forecasted high and low temperatures and anticipated levels of precipitation during a 24-hour period at 13 sites nationwide. Predictions are made every day, Monday through Thursday, for two weeks.

One site is usually a “test” site, where forecasters predict some other weather condition, such as snowfall or visibility near a mountain pass. This year, the test site is Berlin, Germany, the first international city and a location for which there are no modeling data.

“We keep score like golfers do,” explains Jonathan Martin, an atmospheric sciences professor who participates yearly. For every degree a forecaster is off on the temperature, he earns a point. For every level of precipitation he’s off, he gains four points. As in golf, the goal is to accumulate as few points as possible.

For every city, the individual from each of the four categories — freshmen/sophomore, junior/senior, graduate or faculty — who makes the best forecasts during that two-week period wins and will receive a plaque. At the end of the contest, a single forecaster from each category will be honored for making the best, overall predictions.

“Historically, we’ve done quite well,” says Martin. Last year, Daryl Kleist, a graduate student who studies the predictability of weather, took home the plaque for best overall forecasting. He competed with about 150 other graduate students. Martin notes that the graduate category is the most competitive.

“I wasn’t in the lead going into the last city,” recalls Kleist.

Steve Decker, a NCWFC participant at the graduate level who has coordinated the contest at UW–Madison for two years, describes the general tone of the contest as “playful competition” and “not cutthroat.” Yet, according to an undisclosed source, some forecasters occasionally call their competitors at midnight — when official forecasts are posted — to tout their scores.

Kleist often encounters this behavior firsthand, he says, “I’m the person they want to beat!” But he doesn’t seem to mind. After all, Kleist describes himself as competitive. “The contest is just like any other competition. You’re in it to win, you don’t want to see your name at the bottom of the list,” he says.

The contest provides an opportunity to learn the science of forecasting weather. “It allows them to assess whether computer models are valid and to synthesize data assessment skills,” explains Martin.

Kleist agrees. He’s competed in the contest for three years, beginning when he was a senior at UW–Madison. As an undergraduate, he relied mostly on computer models to forecast weather conditions. He admits his predictions generally were off. Now he realizes the importance of considering many factors, including the land’s topography, the surrounding area, cloud coverage, weather in neighboring regions and climate trends. “I’ve learned just how local weather can be,” he says.

But even now, he admits there are days when he’s only 25 percent confident in his predictions. “The weather will always be different, and it’s a matter of figuring out why,” says Kleist, who spends between 30 minutes and an hour each day researching tomorrow’s forecast.

On a regular basis, students, along with faculty, discuss their weather predictions after the official forecast has been posted. Kleist, for instance, often talks about his predictions with his adviser, Michael Morgan, who has collected a few of his own NCWFC plaques over the years.

“It’s nice to talk about Madison weather, but that gets old after a while,” jokes Kleist, who admits that he knows more about the weather in the forecast cities than he does in the city where he lives.

By evaluating their predictions in relation to the actual forecast, Martin says, students begin to recognize factors contributing to weather outcomes.

As part of this learning experience, Kleist, who teaches an upper-level atmospheric science class in which students must participate in the contest, requires the students to write short essays explaining why their forecasts were inaccurate. The students are exempt from the assignment only if their forecasts are better than the teacher’s predictions. So far, Kleist reigns.

Four cities have already been forecasted during this year’s contest. While no one from UW–Madison has made the best forecasts for those cities, two participants, graduate student Decker and undergraduate, Justin Sieglaff, are in the top ten for best, overall forecasting among all categories.

The last contest forecasts will be made on April 17, 2003. To keep track of how UW–Madison, along with its forecasters, stacks up against other NCWFC participants, visit the university’s contest site at http://www.aos.wisc.edu/~nfc/.