How IoT is improving weather forecasts

In the Pacific Northwest, I’ve always been amazed at how it could be sunny as I’m traveling on a freeway and turn into a storm one mile later, then return to sunshine the next mile. Fluctuations like this, common in mountainous areas, aren’t captured in weather forecasts. 

The world of Internet of Things (IoT) offers a new opportunity for improving the accuracy of weather forecasts. Forecasting with IoT uses an army of devices that potentially could reach into the billions, said Gartner, which predicts that IoT will grow to 5.8 billion enterprise and automotive endpoints in 2020, a 21% increase from 2019.

SEE: Securing IoT in your organization: 10 best practices (free PDF) (TechRepublic)

It does this by transforming a type of  IoT into the “Weather of Things” (WoT) that can collect weather data from wireless signals, connected cars, drones, and other IoT devices.

The reporting now has the potential to be reported at a very “micro” level.
“Companies want more precision in their weather forecasts and analytics,” said Shimon Elkabetz, CEO of ClimaCell, a weather technology company. “Until recently, most weather forecasters relied on government weather models because it was too costly for them to develop their own models.”

Elkabetz said the data-collection capabilities of IoT sensors and devices can be leveraged for more detailed, and accurate, weather forecasts.

“The IoT around us can be used as weather sensors and can be incorporated into weather forecasts,” Elkabetz said. “These more localized devices can superimpose more granular weather data on top of the government weather models that weather reporters already use. The end result is a more accurate weather forecast.” In other words, it is possible to construct a weather forecast that can take into account “micro” systems of weather, like rain that lasts for one mile, and that would never show up in a broader weather report.

Different industries have different weather sensitivities, but there is no doubt that a closer understanding of weather conditions that are reported by IoT devices and sensors that are in a particular geographic location, can help.

SEE: How IBM uses Fortran and POWER9 in GRAF, the first “high-resolution, hourly forecast model” (TechRepublic)

Here are some use cases for IoT weather sensors:

• Major airlines can plan routes, schedule events like deicing at airports, and reduce risks if they obtain a more detailed look at weather forecasts for areas where they fly. 

• Firefighters and energy companies can plan their operations for the paths that wildfires are likely to take, based on weather reports that also incorporate micro conditions, such as wind.

• Construction companies can know when and where to pour concrete, preventing washouts and cost overruns.

• Logistics providers will have more detailed weather information at their disposal for the planning of routes.

SEE: IBM, The Weather Company launch GRAF weather forecasting system globally (ZDNet)

“To get information to decision makers in different industries, we use a dashboard that acts as a front end to the analytics, artificial intelligence, and machine learning that we use to operate on weather information and weather models,” Elkabetz said.

Tools like this don’t solve every weather forecast issue, but they do leverage IoT to produce and report more weather information at a micro level than current weather models do not support. This makes weather reporting more accurate, and that’s good for business.

Also see