A Qatar Airways logo sits on the seat belt clasp of business class seat inside a Boeing Co. 787 Dreamliner aircraft. (Photographer: Chris Ratcliffe/Bloomberg)

Fewer ‘Fasten Seat Belt’ Warnings, Courtesy of IBM

(Bloomberg Businessweek) -- You probably didn’t know you had a barometer in your phone, but International Business Machines Corp. does. IBM is planning to use barometric data from millions of phones (with their owners’ permission) to improve global weather forecasting. One big improvement will be in hour-ahead forecasting of upper-air turbulence, which will allow airlines to avoid choppy air.

The idea that a farmer with a cell phone in Brazil or India could make flying safer for people miles above is one of those weird and wonderful things about information technology in the 21st century. It’s also nice that the people on the ground will benefit: Because IBM’s system covers the whole planet, it will vastly improve day-ahead weather forecasting for farmers and others in parts of the world that have lagged behind the rich countries.

For airlines, finding smoother air has benefits beyond passengers’ comfort. Without good forecasting of turbulence, pilots waste fuel ascending and descending in search of quieter layers of air. And turbulence can rattle planes so hard that they require lengthy inspections and even repairs, delaying flights. No wonder that Delta Air Lines Inc. Chief Executive Ed Bastian appeared with IBM CEO Ginni Rometty at the CES consumer electronics show in January as she announced the new forecasting system. Even before using IBM’s new system, Delta has begun equipping its pilots with a turbulence app on their tablet computers using data that’s already available. “No longer do Delta pilots gauge turbulence by how much their coffee shakes,” the company said in December.

Japan’s Panasonic Corp. used to do weather forecasts for the whole planet but has stopped, so IBM will apparently be the world’s only private company doing global weather forecasts, says Clifford Mass, an atmospheric scientist at the University of Washington who’s working with IBM on how to incorporate crowdsourced barometric data into its forecasts. “I think it is exciting that a big private company like IBM with huge computing power is taking on this challenge,” says Brian Tang, an atmospheric scientist at the University of Albany who isn’t connected to the project.

The IBM system is called GRAF, for Global High-Resolution Atmospheric Forecasting System. Like a camera with more pixels, it can see smaller features, down to the size of individual thunderstorms. And the model is recalculated every hour. That degree of resolution and recency is something that only the developed world has had until now.

For now, IBM isn’t planning to use GRAF for long-range forecasting. Even using the same kind of advanced supercomputers for GRAF that it uses for its work with the U.S. Department of Energy, it can’t look ahead a week or two while maintaining the degree of accuracy that it achieves with its day-ahead forecasts, says Mary Glackin, who is vice president for business solutions and government affairs at the Weather Co., a unit of IBM. For longer-range forecasts, IBM has an artificial intelligence-based system that blends forecasts from government forecasting agencies in the U.S., Europe, Japan, and elsewhere.

It’s an exciting era for meteorologists thanks to more sensors, more computing power, and better models. For Cameron Clayton, the general manager and CEO of the Weather Co., better forecasting is personal. He grew up on a sheep farm in New Zealand. “Literally as a kid some of my earliest memories were around that dining room table talking about ‘All of our lambs died in a snowstorm last week,’ so we weren’t going to make any money that year.”

Oh, in case you were wondering, the purpose of the barometer in your smartphone is to help the phone lock onto GPS. The barometric pressure can be used to calculate altitude, which makes it faster for the phone to find its position with respect to GPS satellites. 

To contact the editor responsible for this story: Eric Gelman at egelman3@bloomberg.net

©2019 Bloomberg L.P.