A first-of-its-kind artificial intelligence system built by Nasa in collaboration with IBM can predict violent solar storms threatening satellites and power grids two hours before they erupt from the Sun, scientists say.
The AI system, named Surya Heliophysics Foundational Model, was trained on nearly a decade of observations from Nasa’s Solar Dynamics Observatory.
Powerful solar storms energise the Earth's upper atmosphere and can cause GPS errors or complete signal loss to satellite communications, as well as risks to power grids in the form of induced electric currents on land.
Predicting severe solar storms has proved difficult as they are brewing millions of miles away on the Sun, but take just minutes to hours to reach the Earth.
"Our society is built on technologies that are highly susceptible to space weather," said Joseph Westlake, director of the heliophysics division at Nasa’s Washington headquarters.
Such violent space weather originates in the Sun’s dark spots called active regions, which increasingly speckle its surface in the lead up to the peak of its 11-year activity cycle.
“If we can predict there’s a big active region forming, we can give Nasa an early warning that hey, it will probably generate a lot of storms,” said solar physicist Spiros Kasapis from Princeton University.
“Just as we use meteorology to forecast Earth's weather, space weather forecasts predict the conditions and events in the space environment that can affect Earth and our technologies,” Dr Westlake explained.
Researchers hope the new AI model can be used to provide early warnings to satellite operators and help scientists predict how the Sun’s radiation affects the Earth’s upper atmosphere.
“Applying AI to data from our heliophysics missions is a vital step in increasing our space weather defence to protect astronauts and spacecraft, power grids and GPS, and many other systems that power our modern world,” Dr Westlake said.
“We want to give Earth the longest lead time possible... Our hope is that the model has learned all the critical processes behind our star’s evolution through time so that we can extract actionable insights,” said project lead scientist Andrés Muñoz-Jaramillo from the Southwest Research Institute.
Early results suggest Surya – Sanskrit for the Sun – can generate visual predictions of solar flares two hours into the future, setting a new benchmark in the use of AI for space weather prediction.
The open-access AI system is now available on Hugging Face, GitHub, and via IBM’s TerraTorch library for researchers across the world to test and explore.
Traditional AI systems to predict such space weather disturbances require extensive labelling of data by researchers.
On the contrary, Surya builds directly on the Solar Dynamics Observatory’s already labelled long-term high-resolution record of the Sun going back nearly 15 years.
Over these years, images have been captured of the Sun for every 12 seconds in multiple wavelengths, along with precise magnetic field measurements, scientists say.
“This stable, well-calibrated dataset, spanning an entire solar cycle, is uniquely suited for training AI models like Surya, enabling them to detect subtle patterns in solar behaviour that shorter datasets would miss,” Nasa said in a statement.
“By developing a foundation model trained on Nasa’s heliophysics data, we’re making it easier to analyse the complexities of the Sun’s behaviour with unprecedented speed and precision,” said Kevin Murphy, chief science data officer at the American space agency’s headquarters in Washington.
Space solar panels to provide most of Europe’s green energy by 2050, scientists say
NASA’s nuclear gamble on the moon faces growing skepticism
NASA Curiosity rover discovers coral-like flower-shaped rocks on Mars
There will be a rare black moon this weekend. Here’s what to know and how to watch
Scientists get amazing rare glimpse inside dying star as it explodes
