There are enough problems on this planet already that we don’t need any new problems coming here from the sun. Unfortunately, we can’t destroy this relentless star just yet, so we’re at its mercy. But NASA anyway can let us know soon when one of its murderous flares will throw our terrestrial systems into confusion.
Understanding and forecasting space weather is a big part of NASA’s job. There’s no air up there, so no one can hear you scream, “Wow, what about this radiation!” That’s why we rely on a series of satellites to detect and communicate this important data to us.
One such measurement is that of the solar wind, “a relentless stream of material from the sun.” Even NASA can’t find anything nice about it! Normally this current is absorbed or dissipated by our magnetosphere, but when there is a solar storm it can be so intense that it overwhelms local defenses.
When this happens, it can put electronics on the fritz, as these charged particles can flip bits or disrupt volatile memory like RAM and solid-state storage. NASA says even telegraph stations were not safe and exploded during the largest recorded solar storm. Carrington event in 1859.
While we can’t prevent these great events, we might be better prepared for them if we knew they were coming. But by the time we know, they’re usually already there. But how can we predict such rare and chaotic events?
A joint project between NASA, the US Geological Survey and the Department of Energy’s Frontier Development Lab has looked into this question, and the answer is exactly what you’d expect: machine learning.
The team collected data on solar flares from multiple satellites that monitor the sun, as well as from ground stations that watch for geomagnetic disturbances (called disturbances), such as those that affect technology. The deep learning model they designed identified patterns in how the former leads to the latter, and they call the resulting system DAGGER: Deep learninGGgeomagnetic pErtuRbatie.
Yes, it’s one piece. But it seems to work.
Using geomagnetic storms that hit Earth in 2011 and 2015 as test data, the team found that DAGGER was able to quickly and accurately predict their effects around the world. This combines the strengths of previous approaches while avoiding their drawbacks. As NASA put it:
Previous prediction models have used AI to make local geomagnetic predictions for specific locations on Earth. Other models that didn’t use AI made global predictions that weren’t very up to date. DAGGER is the first to combine AI’s rapid analysis with real-world measurements from space and around the Earth to generate regularly updated forecasts that are both fast and accurate for locations around the world.
It may take you a while to get a solar alert on your phone telling you to stop or your car to stop working (this won’t really happen…probably), but it can make a big difference if we know there are is a fragile infrastructure that could stop suddenly. A few minutes’ warning is better than no warning!
You can read the article describing the DAGGER model, which is open source by the way, in this issue of the magazine Space weather.