Artificial intelligence helps physicists predict dangerous solar flares

January 14, 2015 by Leslie Willoughby
This solar flare was captured Jan. 12 by NASA's Solar Dynamics Observatory. Stanford physicists are bringing artificial intelligence techniques to the challenge of predicting such flares. Credit: NASA/SDO and the AIA; EVE; and HMI science teams

Solar flares can release the energy equivalent of many atomic bombs, enough to cut out satellite communications and damage power grids on Earth, 93 million miles away. The flares arise from twisted magnetic fields that occur all over the sun's surface, and they increase in frequency every 11 years, a cycle that is now at its maximum.

Using artificial intelligence techniques, Stanford solar physicists Monica Bobra and Sebastien Couvidat have automated the analysis of the largest ever set of solar observations to forecast solar flares using from the Solar Dynamics Observatory (SDO), which takes more data than any other satellite in NASA history. Their study identifies which features are most useful for predicting solar flares.

Specifically, their study required analyzing vector data. Historically, instruments measured the line-of-sight component of the , an approach that showed only the amplitude of the field. Later, instruments showed the strength and direction of the fields, called vector magnetic fields, but for only a small part of the sun, or part of the time. Now an instrument on a satellite-based system, the Helioseismic Magnetic Imager (HMI) aboard SDO, collects vector magnetic fields and other observations of the entire sun almost continuously.

Adding machine learning

The Stanford Solar Observatories Group, headed by physics Professor Phil Scherrer, processes and stores the SDO data, which takes 1.5 terabytes of data a day. During a recent afternoon tea break, the group members chatted about what they might do with all that data and talked about trying something different.

They recognized the difficulty of forming predictions from many data points when using pure theory and they had heard of the popularity of the online class on machine learning taught by Andrew Ng, a Stanford professor of computer science.

"Machine learning is a sophisticated way to analyze a ton of data and classify it into different groups," Bobra said.

Machine learning software ascribes information to a set of established categories. The software looks for patterns and tries to see which information is relevant for predicting a particular category.

For example, one could use machine-learning software to predict whether or not people are fast swimmers. First, the software looks at features of swimmers – their heights, weights, dietary habits, sleeping habits, their dogs' names and their dates of birth.

And then, through a guess and check strategy, the software would try to identify which information is useful in predicting whether or not a swimmer is particularly speedy. It could look at a swimmer's height and guess whether that particular height lies within the height range of speedy swimmers, yes or no. If it guessed correctly, it would "learn" that the height might be a good predictor of speed.

The software might find that a swimmer's sleeping habits are good predictors of speed, whereas the name of the swimmer's dog is not.

The predictions would not be very accurate after analysis of just the first few swimmers. The more information provided, the better machine learning gets at predicting.

Similarly, the researchers wanted to know how successfully machine learning would predict the strength of solar flares from information about sunspots.

"We had never worked with the machine learning algorithm before, but after we took the course we thought it would be a good idea to apply it to solar flare forecasting," Couvidat said. He applied the algorithms and Bobra characterized the features of the two strongest classes of solar flares, M and X. Though others have used machine learning algorithms to predict solar flares, nobody has done it with such a large set of data and or with vector magnetic field observations.

M-class flares can cause minor radiation storms that might endanger astronauts and cause brief radio blackouts at Earth's poles. X-class flares are the most powerful.

Better flare prediction

The researchers catalogued flaring and non-flaring regions from a database of more than 2,000 active regions and then characterized those regions by 25 features such as energy, current and field gradient. They then fed the learning machine 70 percent of the data, to train it to identify relevant features. And then they used the machine to analyze the remaining 30 percent of the data to test its accuracy in predicting solar flares.

Machine learning confirmed that the topology of the magnetic field and the energy stored in the magnetic field are very relevant to predicting solar flares. Using just a few of the 25 features, machine learning discriminated between active regions that would flare and those that would not flare. Although others have used different methods to come up with similar results, provides a significant improvement because automated analysis is faster and could provide earlier warnings of solar flares.

However, this study only used information from the solar surface. That would be like trying to predict Earth's weather from only surface measurements like temperature, without considering the wind and cloud cover. The next step in prediction would be to incorporate data from the sun's atmosphere, Bobra said.

Doing so would allow Bobra to pursue her passion for solar physics. "It's exciting because we not only have a ton of data, but the images are just so beautiful," she said. "And it's truly universal. Creatures from a different galaxy could be learning these same principles."

Explore further: SDO sees a mid-level solar flare: Nov. 3

More information: Study paper: iopscience.iop.org/0004-637X/798/2/135

Related Stories

SDO sees a mid-level solar flare: Nov. 3

November 4, 2014

The sun emitted a mid-level solar flare, peaking at 5:40 p.m. EST on Nov. 3, 2014. NASA's Solar Dynamics Observatory, which watches the sun constantly, captured an image of the event. Solar flares are powerful bursts of radiation. ...

SDO captures images of two mid-level flares

December 19, 2014

The sun emitted a mid-level flare on Dec. 18, 2014, at 4:58 p.m. EST. NASA's Solar Dynamics Observatory, which watches the sun constantly, captured an image of the event. Solar flares are powerful bursts of radiation. Harmful ...

NASA releases images of first notable solar flare of 2015

January 13, 2015

The sun emitted a mid-level solar flare, peaking at 11:24 p.m. EST on Jan. 12, 2015. NASA's Solar Dynamics Observatory, which watches the sun constantly, captured an image of the event. Solar flares are powerful bursts of ...

Recommended for you

Rocky planet found orbiting habitable zone of nearest star

August 24, 2016

An international team of astronomers including Carnegie's Paul Butler has found clear evidence of a planet orbiting Proxima Centauri, the closest star to our Solar System. The new world, designated Proxima b, orbits its cool ...

Test for damp ground at Mars streaks finds none

August 24, 2016

Seasonal dark streaks on Mars that have become one of the hottest topics in interplanetary research don't hold much water, according to the latest findings from a NASA spacecraft orbiting Mars.

China unveils 2020 Mars rover concept: report

August 24, 2016

China has unveiled illustrations of a Mars probe and rover it aims to send to the Red Planet at the end of the decade in a mission that faces "unprecedented" challenges, state media said on Wednesday.

What do aliens look like? The clue is in evolution

August 19, 2016

Speculating about what aliens look like has kept children, film producers and scientists amused for decades. If they exist, will extra terrestrials turn out to look similar to us, or might they take a form beyond our wildest ...

8 comments

Adjust slider to filter visible comments by rank

Display comments: newest first

cantdrive85
1 / 5 (4) Jan 14, 2015
Well artificial intelligence can't be any worse than astrophysicist intelligence on this topic.
Mike_Massen
4 / 5 (4) Jan 15, 2015
cantdrive85 with the very BEST he could do
Well artificial intelligence can't be any worse than astrophysicist intelligence on this topic
This is a science site, your pre-occupation with AWT predates you significantly, so as you are on a science site Eg Show some actual evidence that AWT has a factor to offer here ?

Any AWT evidence eg affecting Total Solar Insolation reaching Earth, or AWT affecting dynamics of greenhouse gases ?

Eg this:- http://www.chem.a.../sim/gh/

Any Evidence ?

Anything at all (sigh) ?

Anything ?

?
alfie_null
5 / 5 (6) Jan 15, 2015
Well artificial intelligence can't be any worse than astrophysicist intelligence on this topic.

All much preferred to crank intelligence, though.

I'd expect that trying to apply AI to crackpot beliefs, like those you espouse, would yield less than spectacular results. For that reason, I'd guess you hold a dim view of the prospect of employing AI.
vidyunmaya
1 / 5 (6) Jan 15, 2015
Sub: Necessity-Demand :identify drive to Sun
Nature povides Inputs. science in Philosophy helps indexing
Artificial Intelligence groups need best of brains Trust -Nature and philosophy lead to Cosmology studies. This cross-over puts Consciousness in right perspetive. East West interaction needed for science to progress.
Vidyardhicosmology [dot] blogspot [dot]com
PhotonX
5 / 5 (5) Jan 18, 2015
Sub: Necessity-Demand :identify drive to Sun
Nature povides Inputs. science in Philosophy helps indexing
Artificial Intelligence groups need best of brains Trust -Nature and philosophy lead to Cosmology studies. This cross-over puts Consciousness in right perspetive. East West interaction needed for science to progress.
Vidyardhicosmology [dot] blogspot [dot]com
Maybe it's an English-as-a-second-language issue and you just don't express yourself very well in English, or maybe it's just my lack of understanding, but pretty much every post I've seen from you here reads like so much gobbledygook. If you want to attract readers, or answers to what you post, perhaps you shouldn't make them so painful to decipher that few people will bother.
.
I agree that AI needs the best of brains {at least that sentence I can understand), but so do a lot of other fields. Sorry, but the rest of your post I just can't understand, and suspect no one else can either.
russell_russell
not rated yet Jan 18, 2015
Automated analysis is machine learning.
Somehow the label 'A.I.' slip into the article.

Leslie? You want to take responsibility for that?
cantdrive85
1 / 5 (3) Jan 18, 2015
This is a science site, your pre-occupation with AWT predates you significantly, so as you are on a science site Eg Show some actual evidence that AWT has a factor to offer here ?

Any AWT evidence eg affecting TSI reaching Earth, or AWT affecting dynamics of greenhouse gases ?

I'm not a supporter of AWT, obviously you aren't paying any attention whatsoever.

However, here is an article which describes a study trying to determine the previously unaccounted for mechanism (not incorporated in climate models) of joule heating from the solar wind.
http://phys.org/n...cts.html

These unaccounted for phenomena can explain anomalies such as this rapid ice melt in Greenland;
http://phys.org/n...ing.html

Once again, science blames a side effect while missing the drivers. My skepticism of AGW will fade once all the variables are accounted for and similar results are forthcoming.
Mike_Massen
4 / 5 (4) Jan 18, 2015
cantdrive85 muttered
I'm not a supporter of AWT, obviously you aren't paying any attention whatsoever
Apologies but, hey take it easy, I'm very busy 'cant??' somebody - who ?

cantdrive85 claimed
However, here is an article
http://phys.org/n...cts.html
But, how many Watts ?

cantdrive85 asked
Once again, science blames a side effect while missing the drivers
Rubbish, the ONLY sizable DRIVER is Sol via Total Solar Insolation (TSI) & reducing..
http://www.skepti...asic.gif

cantdrive85
My skepticism of AGW will fade once all the variables are accounted for and similar results are forthcoming
This is why FORMAL education in PHYSICS is so VERY important as not only r the physics courses uni structured a particular way to cover the necessary items including the maths but YOU learn issues of comparative magnitudes !

Where is Evidence of ANY study even close to TSI Eg 0.1% ?

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.