Aug. 9, 2012 at 6:17 PM ET
On June 28, if anyone in the greater Washington D.C. area ran a weather-modeling service called Deep Thunder, they would have known a derecho windstorm was about to rip trees from the ground, knock out power and leave millions of people stifling in relentlessly sticky heat.
No one did.
As a result, millions of people suffered for days on end as utilities scrambled to restore electricity. Many businesses were idled through the Fourth of July holiday. Untold millions in wages and revenue were lost. At least 13 people died.
The company recently tested the service with weather data available the day before the twice-a-decade derecho hit. No weather forecasters in the area predicted the event. IBM, in hindsight, says Deep Thunder would have nailed it.
The example is a good sales pitch. It’s also what’s possible with today’s computer technology combined with mountains of data and the algorithms to make sense of it all.
Deep Thunder combines available data from public and private weather, geological and space agencies along with information about a client’s infrastructure and readiness requirements to generate a business-specific forecast. It has a resolution of 1 kilometer and is effective out to three days.
“The concept is to be able to predict the impact and response due to severe weather events,” Treinish said.
The forecasts are customized with visualizations that are easy for decision makers to understand.
An electric utility client on the Gulf Coast, for example, could get a tailored forecast that allows them to position repair crews in areas where the strongest winds are expected from an approaching hurricane.
Big agriculture might use the information to plan planting or fertilizer application to avoid soaking rains.
A trucking company could use it to route drivers around snowy mountain passes or windy stretches of highway.
Deep Thunder has been under development since 1996 and rolled out to clients over the past few years, Treinish noted.
Rio de Janeiro in Brazil, host of the 2016 Summer Olympics, is using the technology to help mitigate flood and landslide risks as well as improve emergency response to severe weather events.
IBM isn’t alone in the race to bring computing power and algorithms to crunch weather data. A team of ex-Googlers recently started The Climate Corporation to issue crop insurance based on forecasts it generates from weather data, climate models, soil samples, and other information, for example.
IBM is marketing the service as a tool that allows businesses to make more efficient and optimal decisions that save property and lives, especially at a time when frequent severe weather events begin to look like the new normal.
“The weather affects much of our daily lives — everything from sports to produce prices – and although we don’t have the technology to change it, at the very least, we can better plan for it,” Trenish noted in a recent IBM blog post about the Deep Thunder.