Federal agency offering $800,000 prize for anyone who can reliably forecast weather beyond 2 weeks


Now that we've achieved a good 7-day forecast (no really, we have), researchers are turning to the next frontier: forecasting the weather weeks in advance.

It's a lofty goal, the code to which hasn't really been cracked. So far, long-range forecasts have been modestly successful, but they haven't proved useful in predicting drought-enhancing conditions or extreme rainfall.

To inject some outside-the-box thinking, one federal agency is offering $800,000 in cash prizes to any person or group that can develop a system that skillfully forecasts general weather patterns - temperature and precipitation - beyond two weeks.

Day-to-day forecasting can be a challenge in itself, let alone trying to figure out what might happen three or four weeks down the line.

Computer models have improved, along with our understanding of the way weather responds to things like ocean temperatures, El Niño and a steadily-warming planet. But then forecasters then need to assemble those pieces of knowledge in a way that generates something useful.

An accurate long-range forecast would be invaluable to water and land management bureaus. Cities could conserve water prudently before a dry spell instead of scrambling to reduce usage during a drought.

It's of particular interest to the U.S. Bureau of Reclamation, whose mission is to "manage, protect and develop" water and hydroelectric power in the West - a challenge after five years of extreme drought.

This is where "sub-seasonal" forecasting comes in - the time between a 10-day forecast and an entire season. The forecasts are a few weeks into the future, and if they're accurate, they could be a drought-management game changer.

"There's been a lot of collective interest in this topic," said Ken Nowak, a research scientist at the Bureau of Reclamation, "Groups across the west have begun to prioritize it more. It seemed like a ripe opportunity for us to make a contribution to the field, being a water management agency."

The traditional grant process can weed out an applicant that doesn't have the "correct" background or credentials, Nowak explained. It can prevent out-of-the-box ideas from breaking onto the scene.

The bureau's contest crowdsources the problem, utilizing more brains and perhaps some that come from different backgrounds and ways of thinking.

They expect graduate students and university research groups to participate, but they hope to cast a wider net into fields that may not be directly related to weather, Nowak said, like signal processing, physics and electrical engineering.

"This [competition] awards on performance and not on potential," Nowak went on. "Whoever gets the best idea will be the one rewarded."

The forecasting challenge "seeks to improve on existing sub-seasonal forecasts and asks Solvers to develop systems that perform demonstratively better than an existing baseline forecast for temperature and precipitation over a 15-42 day time frame," according to the Innocentive website hosting the event.

The Bureau of Reclamation is partnering with NOAA - specifically the Earth Science Research Lab and the Climate Prediction Center - to evaluate the methods and results.

The rules require winners to hand over their application and its code to the government agencies. Furthermore, the prize money may only be given to a U.S. citizen, permanent resident or corporation, though the contest will accept programs from anyone.


Author Information:

Angela Fritz is an atmospheric scientist and The Post's deputy weather editor. She has a B.S. in meteorology and an M.S. in earth and atmospheric science.

(c) 2016, The Washington Post · Angela Fritz



Recommended for you