As tropical storm Dorian bore down on Puerto Rico last month, its whirling winds threatening to reach hurricane force, a team of Walmart employees watched closely from an array of screens at the retailer’s Arkansas headquarters 2,000 miles away.
Had the severity grown to crisis levels, Lucas McDonald, a former TV weatherman who leads the chain’s emergency operations, might have called in dozens of workers to support the handful who are posted at the division’s command center in 24/7 shifts. The full-house team—typically assembled only a few times a year—would help coordinate relief efforts, adjust supply routes and disseminate information to affected stores, a playbook the company has perfected through two exceptionally hectic hurricane seasons.
“Right now, we’re having conversations with some of our merchants on when the right time to ship more supplies into places like Florida and the Southeast would be ahead of any possible redevelopment from Dorian after it makes its way through Hispaniola,” McDonald says.
Meanwhile, in Dallas, meteorologists at Southwest Airlines mapped out contingency plans for rerouting and canceling flights given various possible hurricane scenarios. And in the Atlanta nerve center of IBM-owned Weather Company, forecasters relayed storm data and analysis to corporate clients like State Farm, which in turn used it to inform IBM Watson conversational ad units that spread safety information.
“It is all hands on deck,” says Kevin Petty, the Weather Company’s director of science, forecast operations and public-private partnerships. “We have meteorologists on staff talking to our clients to keep them up-to-date informed on the situation and how it might be changing, and how it might impact their operations.”
As data analytics technology and low-cost meteorological equipment have made reams of data more tractable, and climate change has upped the intensity, volatility and, thus, economic toll of extreme or erratic weather events, businesses are increasingly investing in operations like these to better monitor and shape operations around weather conditions.
While exact figures on this trend are hard to come by, a 2017 National Weather Service report estimated that the private weather industry could grow fivefold as “attractive new product offerings” allow businesses to wring more cost savings and revenue growth out of weather information. At the same time, Morgan Stanley estimates that climate-related disasters cost the global economy $650 billion in the last three years, a number that’s only expected to grow.
“What brands and businesses are really looking for now is how they get ahead of that, how they get the data and the insights to help mitigate that impact in advance,” says Randi Stipes, CMO of IBM Watson Media and the Weather Company. “More and more brands now are becoming savvy. They may even have weather operations and meteorologists on staff.”
IBM has been one of the companies on the forefront of this trend since its acquisition of the Weather Company in 2015 for a reported $2 billion-plus. The deal was met with some confusion from industry watchers, who were at a loss to explain what the artificial-intelligence giant might want with the three-decades-old forecaster.
The company revealed an initial answer to that question the next year with the rollout of Deep Thunder, a hyperlocal forecasting platform that integrated machine learning to help retailers understand how granular weather changes could affect consumer buying behavior. Big Blue has also been working to integrate its Watson artificial intelligence platform with the forecasting giant’s data to power everything from weather-targeted ads to logistics coordination.
And now, IBM is poised to add another big piece to that puzzle next month with the release of its Global High-Resolution Atmospheric Forecasting (GRAF) system, an even more advanced hyperlocal forecasting tool that will pull data from millions of sensors in aircraft, barometric pressure meters built into smartphones (if the owner opts in) and even the equipment of amateur weather hobbyists. A back-end powered by the same hardware as the world’s most powerful supercomputers—the U.S. Department of Energy’s Summit and Sierra—will process that massive torrent of inputs, forming it into an hourly updating global forecast that IBM claims will provide double the clarity of existing models in most of the world.