Lightning is one of the most unpredictable phenomena in nature. Approximately 100 lightning bolts strike earth’s surface every second, and each lightning bolt can contain up to one billion volts of electricity. Lightning regularly kills both people and animals and sets homes and forests on fire. It’s also been known to ground airplanes.
Researchers at EPFL – a research institute and university in Switzerland – have developed a novel way to predict where and when lightning will strike. The system relies on a combination of standard data from weather stations and artificial intelligence to predict lightning strikes to the nearest 10 to 30 minutes and within a radius of less than 20 miles. The simple and inexpensive system was outlined in a research paper recently published in Climate and Atmospheric Science, a Nature partner journal.
According to researchers, the current systems for predicting lightning strikes are slow, expensive, and complex, relying on external data acquired by satellite and radar. The new inexpensive system from EPFL uses real time data that can be obtained from any weather station, meaning it can cover remote regions that are out of radar and satellite range and where communication networks are lacking. The quick predictions from the system allow alerts to be issued before a storm has even formed.
The system uses a machine-learning algorithm that’s been trained to recognize conditions that lead to lightning. The researchers took into account atmospheric pressure, air temperature, relative humidity, and wind speed, among other things. After training the algorithm, the system was able to predict lightning strikes accurately nearly 80% of the time.
This system is a simple way to predict a complex phenomenon.
Photo, posted December 14, 2018, courtesy of Flickr.