An inexpensive tool combines machine learning and AI to forewarn of lightning bolts.

A team of researchers from École Polytechnique Fédérale de Lausanne, Switzerland, has found a way to use artificial intelligence (AI) to predict lightning strikes within 10 to 30 minutes.

However, the area covered at a single time is a 30-kilometer radius (approximately 18.6 miles).

“We have used machine learning techniques to successfully hindcast nearby and distant lightning hazards by looking at single-site observations of meteorological parameters,” the authors wrote in a new paper published earlier this month.

AI: Machine learning the four variables

The researchers trained an AI algorithm to recognize patterns from data on four weather variables. These were air pressure at station level (QFE), air temperature, relative humidity, and wind speed. They sourced the data from 12 Swiss weather stations in both urban and hilly locations between 2006 and 2017.

They compared the output from the AI algo with data from lightning location systems.

“Evaluation results show that the model has statistically considerable predictive skill for lead times up to 30 min,” the scientists said.

In other words, the algo correctly forecast strikes from these giant electrical charges 80% of the time.

AI prediction of lightning strikes: A cheap and simple system

Because the system uses data from any weather station, there is no need for radar or satellite data. It can, therefore, be put to use in remote and inaccessible regions out of radar or satellite range and where communication networks are unavailable.

There are many benefits of predicting lightning strikes. It can help save lives by warning people to stay away from risky areas. Firefighting missions can prepare in advance to tackle fires.

Related Story: How AI, Drones, U-Cal and Salesforce Watch Out for Sharks