What's in the Forecast: Private Weather Predictions


Long-haul truckers can now get hourly forecasts of road surface conditions and temperatures for every mile of major highway in the US, Europe, and parts of China. Autonomous vehicles—once they get rolling—will likely use the same kind of data to help them avoid slippery spots and storms on their way to deliver people and goods.

This valuable information may be coming not from the US National Oceanic and Atmospheric Administration and its vaunted weather agency, or even its European counterpart. Instead, companies may increasingly be paying for access to private data, trend analyses and forecasts, as part of the expansion of a new weather industry that is exploiting modern tools such as machine learning, automated sensing, internet-connected devices, and cloud computing.

For Global Weather Corp., that means developing a system for drivers that can pinpoint the exact weather conditions they are likely to encounter on their route, then steer them away from potential problems. “The system will know not only the weather in the next five miles, but within your vehicle navigation system reroutes you to be quicker,” says Bill Gail, CEO of the Boulder-based company, which is testing the system with several automakers. It’s like Waze meets Weather Underground—at least, that’s the idea. And automated taxis will be able to tell human passengers that certain parts of the city are impassable due to weather conditions, says Gail.

Because road temperature can be as much as 20 degrees different from air temperature, the company’s software program uses a forecast model that takes into account the physics at the surface as well as up in the atmosphere. To do this, Gail’s firm combines observations from the National Weather Service with forecasts developed by his own climate modelers. This blend of public and private might make our lives safer and more efficient, if it results in better or more personalized forecasts. Either way, it’s likely to become increasingly common.

“We are at a tipping point where the technology of weather forecasting, which was dominated by government and is still is, is going to change,” says Shimon Elkabetz, CEO of ClimaCell, a Boston-based startup that collects observations from Internet-connected sensors, airplanes and buildings to create forecasts for clients like JetBlue and the New England Patriots.

The US commercial weather industry—which markets everything from drought forecasts targeted at individual farms to data for Chinese truckers—was already worth an estimated $7 billion in 2017, according to a NOAA study. The demand from the public and from big insurance companies for more accurate forecasts of potentially destructive storms and floods will only grow as climate change juices up extreme weather events.

But commercial interests aren’t the only potential purchasers of this new wellspring of weather data. NOAA itself is also dipping its toe into this world of private data. The agency already buys some commercial information from weather balloons as well as data on lightning strikes.

“There’s a lot of potential that this works out well for the taxpayers,” says Karen St. Germain, deputy director of satellite programs at NOAA. “If we go about these partnerships smartly, we could deliver more actionable information at the same or better price.”

After several years of testing, NOAA is now preparing to purchase readings from commercial satellites that use GPS signals to measure the density of the atmosphere. That data, gathered using a technique called radio occultation, can gauge atmospheric temperature, moisture and pressure.

NOAA’s 2020 budget request includes up to $6 million to purchase radio occultation data from private firms as well as money to continue testing and evaluating the quality of the information. Agency officials say they plan to buy other kinds of commercial weather observations in the next few years, including measurements taken from weather balloons, ground stations, ships and airplanes, and perhaps one day personal cellphones.