NASA Funding Air Turbulence Prediction System

NASA is funding a project run by the National Center for Atmospheric Research (NCAR) in Boulder, Colorado, to identify areas of rapidly developing turbulence and storms over remote areas of the ocean.

The idea is to guide pilots around these areas and avoid a disaster like the one that occurred in June, when 228 people were killed after an Airbus A330 jet flown by pilots from Air France hit a storm and crashed into the Atlantic Ocean.

Scientists are working to predict areas of turbulence, both in clear skies and within storms, by applying artificial intelligence to satellite data and computer generated models of weather, NCAR scientists said.

NCAR already has a system that alerts pilots and air traffic controllers to turbulence at various altitudes over the continental United States.

But alerting pilots who fly over open oceans is harder. A graphic of the path of the Air France flight based on satellite data analyzed by NCAR after the crash shows that the plane flew right through the highest storm clouds, which are often associated with the most intense storms.

"It seems likely that the information provided by a real-time uplink of weather conditions ahead would have, at a minimum, improved the pilots' situational awareness," said NCAR scientist John Williams.

Pilots flying over oceans do get information from weather satellites, but the satellites don't provide as many images as they do over land and don't measure turbulence, according to NCAR.

They may also be updated on weather every four hours and have radar that detects clouds and precipitation. But turbulence and precipitation are not always located near each other.

The new weather system is expected to be tested by pilots next year. So far, scientists have created global maps of turbulence in clear air -- winds that may buffet planes -- and global views of the tops of storm clouds. Now they're looking for turbulence within and around intense storms.

They're analyzing the data using an artificial intelligence technique called "random forests," which creates a series of yes-or-no votes on how elements of a storm may behave, said NCAR scientist Cathy Kessinger.

Also participating in the project are scientists from MIT's Lincoln Laboratory, the Naval Research Laboratory and the University of Wisconsin at Madison.

InformationWeek has published an in-depth report on leading-edge government IT -- and how the technology involved may end up inside your business. Download the report here (registration required).