IBM 'Deep Thunder' rumbles onto weather modeling scene

Don't you wish you could be wrong half the time and still hold onto your job? Given how much we can all relate to error-prone weather forecasting, it is not surprising that some of the biggest high companies in high-tech, including IBM, are trying to tackle this problem.

IBM's weather modeling technology, which is called Deep Thunder, is focused on providing much more granular information about weather patterns in a region, so that municipal governments and businesses can prepare for the worse, say Lloyd Treinish, chief scientist for Deep Thunder.

"The primary concept is around the idea that aspects of our human infrastructure, utilities, cities, are part of the environment. They interact with the environment and are dependent on the environment. ... This is more than a science experiment," Treinish said.

This is big data at its biggest; the technology uses environmental sensors and other inputs to create visualizations and models that could aid in everything from cutting energy usage to better disaster recovery and emergency services planning. It works by combining real-time inputs with other data, such as the forecasts from the National Weather Service. The technology is being used by the National Oceanic and Atmospheric Administration (NOAA) and it cities including Atlanta, New York, Rio de Janeiro and Burlington, Vt.

The image below offers an idea of the sorts of visualizations that are made possible by Deep Thunder. If you click on the video link, it will take you to an active visualization:

Rio Visualization

In Rio, the focus for Deep Thunder is on improving public safety and avoiding life-threatening mudslides, like the ones that offer occur during summer in the Brazilian city. The idea of having even a day of lead time to predict such situations could help prevent loss of life, Treinish said. The Rio project is part of a larger IBM-developed command and control center being built for the city's operational infrastructure. It is producing a 48-hour forecast that can be updated every 12 hours through a series of visualizations like the ones provided above.

Related posts:

• One answer to forecasting extreme weather
• Cloud-driven weather insurance service harvests $42 million VC round
• What's the forecast? Study suggests average weather changes impact economy
• Intel, the eco-technology innovator