If Earth does find itself in the line of fire of any large event, we are reasonably well protected from solar flares. The Earth’s magnetic field and atmosphere shield us from much of the high-energy particles and radiation they produce. But X-rays blasted out in solar flares can penetrate deep into a part of the upper atmosphere called the ionosphere, which can disrupt radio waves, causing radio blackouts.
Coronal mass ejections pose a much greater problem. These charged plasma clouds move through space at speeds approaching 75,000 miles per minute (120,000km per min), continuously producing energetic particles as they go. The Sun emits one coronal mass ejection per week during a solar minimum, but as many as three per day during a solar maximum. They can damage or destroy satellites outside the protection of Earth’s magnetic field and force commercial airlines to re-route flights away from the polar regions. And it can happen very quickly, says Daniel Baker, “just a few hours after the disturbance on the Sun, or sometimes even less.”
When a coronal mass ejection collides with the Earth, it rattles the Earth’s magnetic field and can cause a geomagnetic storm. As the storm moves over long electrical transmission lines it produces current down the line, generating heat that can overload and blowout high-voltage transformers, and bring cities – or entire regions – to a standstill. Just like what happened in Quebec in 1989.
Today’s space-weather forecasters know whether a solar storm will hit the Earth about one or two days in advance. “That’s state of the art,” says Yihua Zheng, a chief forecaster at NASA’s Goddard Space Flight Center Space Weather Center in Greenbelt, Maryland. Zheng and other forecasters use data gathered by a trio of orbiting spacecraft to identify the presence, direction and speed of coronal mass ejections. “Slower [coronal mass ejections] can take between three and five days to reach the Earth, but we care more about the faster ones,” she says, because they tend to be more dangerous.
Zheng is part of a group developing a computer technique, known as ensemble forecasting, to improve NASA’s ability to predict the path and impact of severe solar storms. Ensemble forecasting is already used by meteorologists to track severe weather, like hurricanes, on Earth, but this is the first time is will be used to make sense of space weather.
The technique will allow Zheng and her colleagues to produce as many as 100 computerized forecasts by varying the speed, direction and other parameters of coronal mass ejections in their models, instead of being restricted to one set of solar-storm conditions, as they are now. It would decrease the margin of error, which currently stands at seven or eight hours, and lead to more reliable forecasts. Zheng expects the ensemble forecasting method will be up and running no later than the end of this year – just in time to catch the next solar maximum.
Still, forecasters would be at a major advantage if they could predict a coronal mass ejection in advance, rather than chasing it down after it had erupted. Stathis Ilonidis, a graduate student in physics at Stanford University in California, is tracking the sound waves that travel through the Sun’s interior to predict where a sunspot will appear one or two days in advance. And as coronal mass ejections often come from sunspots, this information could be of enormous help to space weather forecasters.
By some quirk that physicists do not yet entirely understand, sound waves travel faster through sunspots than through the rest of the sun’s interior. Ilonidis measures how long it takes a sound wave to travel from one point on the Sun’s surface to a depth of 37,000 miles and back up to another point. On average, it takes a sound wave an hour to bounce from one point to another 93,000 miles away, but a sound wave passing through a sunspot may shave 12 to 16 seconds off its time. By repeating the process with millions of pairs of points, Ilonidis can chart out a map that identifies regions where travel time is significantly shorter than expected – locations where sunspots are likely to appear. "This is the first time that we can actually detect sunspots inside the sun," says Ilonidis.