In This Story
Now in its fifth year, ɫƵ’s contest again received entries from around the world, with teams attempting to predict peak bloom in five different locations. But what local folks want to know is: how will all that “snowcrete” and our wacky winter impact the trees at the Tidal Basin and around Washington, D.C.?
“Typically, trees need a certain amount of cold weather and then a certain amount of warm weather to bloom,” said Jonathan Auerbach, an assistant professor in the Department of Statistics and the competition co-organizer. “If they don't get enough of either, then the bloom is delayed. For this reason, many contestants are predicting a late bloom this year.”
In fact, the average contestant is predicting a peak bloom in the D.C. area on April 2, about a week later than in recent years.
“One challenge many contestants chose to tackle this year was the effect of prolonged snow ground cover,” said David Kepplinger, also an assistant professor in the Department of Statistics and a competition co-organizer. He noted that some of the entries use snow coverage in their models.
A key challenge is while cherry trees tend to bloom earlier as the climate warms, complex weather patterns make annual predictions extremely difficult. “It wasn’t a record amount of snow or a record amount of cold, but it was a unique combination of the two,” Kepplinger said.
The contest received more than 30 entries, which is typical.
Last year, artificial intelligence (AI) made its appearance as a new tool for predictions. Auerbach said many contestants actually did better than AI. “This year we did the opposite of AI. We held a hackathon in person and got the students to network and meet,” he said, referring to an event on February 21 at Georgetown University.
The contest asks contestants to submit predictions for trees in Washington, D.C., New York City, Kyoto, Liestal-Weideli in Switzerland, and Vancouver, B.C. Entries must include a compelling narrative and reproducible analysis with any data and code they used.