How is AI Helping NASA with Solar Storms?
AI (Artificial Intelligence) is changing the landscape of just about every industry at a rapidly evolving pace – and it’s no surprise that NASA has already developed ways to use AI to the benefit of our planet, long before it became introduced as a function of our own daily lives. In particular, NASA’s AI technology is allowing us to track solar storms, and potentially minimize their impact, in a way that’s truly unprecedented.
What is the Impact of Solar Storms?
Solar storms are the result of the emission of solar material on the sun, and as a result, there is a consistent stream of solar wind, which is the result of these materials being emitted. Then, there are solar eruptions, in which an increase in pressure produces a larger burst – sort of like a volcano erupting. While most solar storms don’t affect us at all, those that produce larger, eruptive emissions can enter our planet’s magnetosphere – Earth’s magnetic environment – and while this doesn’t cause physical danger to us, it is disruptive to our technology. Essentially, it causes storms to take place within our own magnetic atmosphere, which can vary in duration and intensity, thus producing what are known as “geomagnetic disturbances.”
The actual impact of geomagnetic disturbances is something we’ve only recently begun investigating, and as technology has evolved, the impact becomes greater. The more dependent we are on technology in our daily lives, the higher the effect it has on us when these storms do occur. Ultimately, solar storms have the ability to damage our electronic infrastructure on a large scale – although that hasn’t occurred yet. However, if it were to occur, these infrastructure components could be permanently damaged if the storm were severe enough.
However, these storms have done damage in the past, on smaller scales. In 1989, Quebec lost power for 12 hours. And, in 1859, the Carrington Event occurred, in which fires were sparked at telegram stations, preventing communication. In 2023, the same type of storm could cause debilitating interference to our power grid.
There’s even more talk about solar storms right now, because of the upcoming solar maximum event in 2025. The solar maximum occurs once every 11 years, and marks the peak of the sun’s cycle.
How NASA’s AI is Stepping in
Solar storms can be devastating, and the fact that no significant events have taken place in several years doesn’t mean that a real risk is overblown. Thankfully, NASA has taken it upon themselves to develop innovative technology to predict solar storms, so that we can actually make the necessary preparations for them when they are imminent. Keep in mind that we’re still learning about these events – they haven’t been easily trackable until recently, and only now do we have the technology to really explore their impact and observe them on a detailed scale.
An international team of researchers at the Frontier Development Lab, which is a public development team that has partnered with NASA, has developed artificial intelligence technology to observe solar storms in ways that were never possible before. The most important takeaway from this development is that scientists will now be able to predict a solar storm 30 minutes before it occurs, which is similar to the way communications work when a tornado is imminent. This allows time to prepare our own infrastructure on earth for any potential disruptions, or damage.
This technology revolves around a technique called “deep learning,” which involves a process of training a computer to recognize patterns over a period of time in order to detect when an event is going to occur in the near future. In this case, deep learning is used to study the relationship between solar wind and geomagnetic disruptions, observing the measurement of these wind events from heliophysics missions, while simultaneously recording geomagnetic disruptions that occur on Earth.
The name of this particular deep learning model is known as DAGGER (Deep Learning Geomagnetic Perturbation). It’s fully capable of predicting geomagnetic disturbances throughout the world within 30 minutes before they take place, and the predictions update every 60 seconds. It’s already been proven to work, as it was tested in both 2011 and 2015, showing flawless accuracy.
Because DAGGER is open source, it can be utilized by different agencies around the world, including satellite controllers, power grid operators, and more.