Rodeo Season

SNR professor competes in Sub-Seasonal Climate Forecast Rodeo

Tony Lupo, professor of atmospheric science in the College of Agriculture, Food and Natural Resources, and Joe Renken, a self-taught meteorologist for KOPN, have teamed up for the National Integrated Drought Information System’s (NIDIS) Sub-Seasonal Climate Forecast Rodeo.

The duo began their partnership two years ago when Renken approached Lupo about joining together to research the field of long-range forecasting. Their research method looks at the Bering Sea Rule and the Recurring Rossby Wave Train – part of the Organic Forecasting Method developed by Renken and Josh Herman, a collaborator of the team’s. Their partnership and researching methods were previously highlighted and explained in greater detail last year. 

The rodeo is sponsored by The Bureau of Reclamation, in partnership with the National Oceanic and Atmospheric Administration, the United States Geological Survey and the United States Army Corps of Engineers. A total of 188 research groups from across the globe have entered to help improve the field of long-range forecasting. This time frame encompasses the three–to-six-week range and, according to NIDIS, would allow for better preparation for the onset of droughts or other extreme weather events.

Lupo agrees. “The three-to-six-week time frame has been a long neglected niche,” he said. “Forecasting in this time frame is statistical and would benefit us in agricultural applications, construction, etc.”

The rodeo began in April of 2017 and will continue until April of 2018. Every other week, each team submits a forecast of the precipitation and temperature for the western United States. Forecasts are made for both the three and four weeks and five and six weeks range and scored based on correlation with the outcome. At the end of the rodeo the top performing teams will be awarded a total of $800,000 in prizes. The top three finishers in each of the four categories will be awarded $100,000, $50,000 and $25,000, respectively. However, to be eligible for prize money, teams need to perform better than the computer model “CFSv2” as well as “DampedPersistence,” a model that uses the previous three to four weeks to predict the following three to four weeks.

Lupo and Renken’s team’s performance has been promising thus far. Currently they place fifth in the three-to-four-week temperature category, second in the three–to-four-week precipitation category, fourth in the five–to-six-week temperature category and third in the five–to-six-week precipitation category. Combined scores do not receive prizes, but the group ranks first when aggregating the three-to-four-week scores and third when aggregating the five-to-six-week scores. While a great deal of the competition remains, the current rankings would result in $75,000 for their team.

Those who wish to follow along may check in at https://www.drought.gov/drought/challenge_contest. The team’s name is “lupoa13.”

Lupo is excited and optimistic about the remainder of the event.

“We can see how our research works in real-life applications,” Lupo said. “If we should win, we’ll contribute to bettering the U.S. capability to forecast in this time period.”