(Bloomberg) -- As Hurricane Matthew rolled toward Florida earlier this week, the meteorologists at Verisk Analytics were tracking its every move.
They weren’t just plotting the usual meteorological stuff, though, and making the usual forecasts: wind speed, rainfall, sea-swell.
They also were armed with new-fangled intelligence — like a model that displays the types, and current density level, of the trees surrounding electricity lines — that allowed them to spit out an estimate of how many households and businesses would lose power. (Days before the storm hit, it was already at least hundreds of thousands.) That’s crucial information for Verisk clients in the Florida insurance industry trying to determine how much cash they’ll need to meet liability claims.
It’s the kind of report that was unthinkable just a few years ago, illustrating a new-found sophistication and skill that’s helped make commercial weather forecasting a multibillion-dollar industry.
Vaisala Oyj, an instrument maker, says this is meteorology’s “golden age,” with advances in science and technology allowing for better data collection that today’s more-powerful computers can then turn into more accurate long- and short-term forecasts. Meteorologists are no longer just prognosticators but advisers, too.
“If your business is built around just telling people what the weather is, you are part of the past,” said Steve Bennett, chief operating officer at Austin, Texas-based Riskpulse. The company projects weather risks for energy producers, retailers, automakers, food-shipping companies and commodity investors.
About one-third of U.S. gross domestic product, or $6 trillion, is generated by industries vulnerable to changes in the weather, according to John Dutton, president of Prescient Weather Ltd. Their need for more detailed, timely information is expanding the range of companies providing data and equipment.
International Business Machines Corp. is using its Watson artificial-intelligence platform to create better weather-forecast models, aiming to improve accuracy by 5 percent this year. It also is applying its computing power to load forecasts for utilities and outage predictions, which are affected by weather, said Mark Gildersleeve, president of business solutions at the Weather Co. IBM acquired the company’s digital assets for an undisclosed amount. The acquisition is part of IBM’s efforts to drive growth with expanded data analytics.
Fine-tuned forecasts for renewable energy, such as solar, could save utilities about $450 million in the next 20 to 25 years, said Jeff Lazo, a project scientist with the National Center for Atmospheric Research, citing a 2016 study by the National Center for Atmospheric Research for the U.S. Department of Energy.
In August, the National Weather Service began using its Cray XC40 supercomputer to simulate 2.7 million stream and river locations in the contiguous 48 states to create hourly reports. The goal is to generate street-by-street flood projections. The service also is rebuilding its Global Forecasting System model. When the new system is operational, in about three years, it will provide more accurate simulations of clouds and storms and have a resolution of about 6 miles or less, compared with about 8 miles currently.
Some companies are creating their own internal systems. Home-improvement retailer Lowe’s Cos. built identical emergency-command centers 50 miles (80 kilometers) apart in North Carolina to track the weather and contracted with AccuWeather Inc., in State College, Pennsylvania, for tailored forecasts.
In January, an eight-day outlook predicted a blizzard would bury the East Coast, allowing the company to stock stores from North Carolina to Maine with storm-related supplies before the first flake fell, said Rick Neudorff, the centers’ operations manager. Manhattan ultimately got 27.5 inches (70 centimeters) of snow, closing New York for a day.
Weather adds $1 to each of the almost 300 million over-the-road shipments in the U.S., according to Riskpulse Chief Executive Officer Matt Wensing. In some cases, it could be much more: A company could save as much as $1,000 if a forecast for mild weather between San Francisco and Chicago meant it didn’t need a refrigerated truck, he said.
High tech drones
Raytheon Co. estimates the market for weather instrument and data-collection products could be $15 billion during the next five years. It builds satellites that gather information from space as well as unmanned aircraft that fly into storms. Raytheon sends the data they amass to the U.S. Air Force and 140 U.S. National Weather Service offices, according to Susan Massihzadeh, the company’s director of earth and space observing solutions in Aurora, Colorado.
Its battery-powered Coyote drone can enter the center of a hurricane and fly as long as an hour, remotely controlled by engineers to collect information that can help predict the storm’s track and intensity. The $22,000 drones have a range of about 30 miles and can skim as low as 100 feet above the Atlantic, “where energy gets extracted from the ocean,” said Joe Cione, a research meteorologist in Boulder, Colorado, with the National Oceanic and Atmospheric Administration’s Hurricane Research Division and Earth System Research Laboratory’s Physical Science Division. “It is how the storm feeds itself.”
Cione was flying with hurricane hunters on Thursday afternoon, trying to determine the best time to deploy Coyote.
Through September, the U.S. experienced 12 weather and climate disasters with losses exceeding $1 billion, according to NOAA. These events killed 68 people and cost $26.9 billion. While advances in meteorology will never prevent the destruction, more detailed, timely forecasts can help everyone prepare and react more effectively.
Businesses have been affected by storms “for time immemorial but never felt that the weather was actionable,” said Jim Block, chief meteorological officer at Schneider Electric SE, an energy-management company. Now, technologies allow them “to make better use of this information.”
Copyright 2017 Bloomberg. All rights reserved. This material may not be published, broadcast, rewritten, or redistributed.