Solar flare prediction?

Maybe I lost some friends with the Spice Girls blog. Let's turn to more familar topics.

Solar flares, the subjects of my own research, were in the news recently. A group of scientists in the USA believe they have a new solar flare prediction technique. I thought I'd discuss this new idea here, mostly because it was in the news but also because it's pretty wild! If correct - and that's probably a big "if" - its implications go a long way beyond satellite engineering to genuinely new physics.

A solar flare is an explosion on the Sun, a sudden release of energy from the Sun's magnetic field. Patches of the Sun's surface brighten up briefly, typically for some minutes, and radio waves, X-rays and ultraviolet light are emitted. A big flare is usually accompanied by a cloud of material expelled from the Sun into interplanetary space: a Coronal Mass Ejection (CME). When a CME arrives at Earth it can trigger the sort of magnetic storms that give us beautiful displays of the Northern Lights, but may also shut down power lines and electronic communications. CME's can be a threat to satellites and technologies that rely on them: telecommunications, GPS, etc.

A group of American researchers in Purdue University have announced a possible new method for predicting flares, maybe more than a day before they actually occur. Our technologically sophisticated society is more and more reliant on satellites and telecommunications so there is a lot of interest in flare prediction. So far such forecasts are of the "20% chance of an X-class flare in the next 24 hours" type. You can find examples at e.g. the Solar Influences Data Center in Brussels. It would be extremely useful if there was a technique that could say, "there will be an X-class flare 24 - 28 hours from now".

Professors Jenkins and Fischbach, the Purdue researchers, have been studying radioactive decay. Suppose we have a lump of a radioactive substance, say 1 kg of silicon 32 (which we write ³²Si - not the usual sort of silicon found in bathroom sealants and breast implants, but a radioactive isotope). We switch on a Geiger counter, a radiation detector (for younger readers: here's a great video demonstrating a Geiger counter in action. They were often seen in Cold War era science fiction, e.g. when the scientists in The Thing from Another World find the crashed alien spaceship). Normally the counter goes 'click' roughly 10 times per minute (on average; some minutes it will be 8, or 9, or 11, or 12; more rarely 3, say, or 20; much more rarely 0 or 30; etc.). This represents the normal, 'background' level of radiation, due mostly to tiny quantities of radioactive substance found everywhere. If we bring the counter near to the ³²Si it starts to click more rapidly. This tells us that this substance is radioactive, possibly dangerous if we get too close. We can go further and count the clicks. Then we can calculate how rapidly nuclei of ³²Si change into something else, i.e. decay. Different substances decay more or less rapidly. Some are gone in microseconds, some last for billions of years. ³²Si decays rapidly enough to give reasonable numbers of clicks in the counter, but not so rapidly that it's all gone before the experiment is over.

The rate of decay is a property of the nucleus of the atom of the particular isotope. We believe it doesn't depend on anything else: how hot the stuff is, presence of other substances, magnetic fields.... We don't expect the decay rate ever to vary. We believe it will be the same, for the same substance, everywhere in the Universe. Professors Fischbach and Jenkins find that the rate of decay of certain substances seems to vary during the year. The number of clicks per minute in the Geiger counter (for example) varies very slightly, at about the tenth of a percent level; slightly greater than average in January and slightly less in July. More than one group of scientists has made this discovery so it looks like it's real - but it's definitely unexpected!

Earth's distance from the Sun is not quite constant, varying by about 3% over the course of a year. Fischbach and Jenkins note that the decay rate of ³²Si is greatest when Earth is closest to the Sun: could radioactive decay on Earth be influenced by something to do with the Sun? They go on to speculate that almost massless, subatomic particles called neutrinos might be the means for the Sun to influence radioactivity here on Earth. Vast numbers of neutrinos continually escape from the nuclear furnace of the Sun's deep core. They are detected - with difficulty - here on Earth, allowing us to confirm our ideas of what happens deep inside the Sun. Neutrinos would have to behave in ways we don't presently know about for slight variations in their numbers to have such effects on laboratory radioactivity. So this is either wrong, or extremely interesting!

Maybe the rate of decays is actually constant - as almost all nuclear scientists would expect - and the detectors, rather than the ³²Si itself, behave slightly different in winter and in summer. Some scientists have proposed detailed explanations along these lines. It seems to me that this might be checked by making measurements north and south of the equator. We might expect them to be six months out of phase. I don't know if anybody has done this.

Fischbach and Jenkins go further, speculating that it is solar magnetism - sunspots and flares - that influences radioactive decay rates here on Earth. On a couple of occasions they claim to have seen a decrease in the rate of decays starting 39 hours before a major flare and this is what they think could give an early warning of flares. Suppose the neutrino mechanism is correct. Some neutrinos would indeed be produced in a major flare, at the same time as the gamma-rays that can be detected by spacecraft near Earth. Their numbers would be absolutely tiny, however, compared to the number continually being produced in the Sun's core. A solar flare is a huge event by earthly standards, but it involves a miniscule portion of the Sun's enormous bulk.

(Reuven Ramaty was the guru of nuclear processes in solar flares. I was once at a meeting where Reuven was very scornful of suggestions that neutrino rates on Earth were being influenced by flares, for exactly this reason - not nearly enough neutrinos ).

To make neutrinos we would need energetic ions. In flares, ions do indeed gain high energies and some neutrinos would result. But the same ions would definitely make gamma-rays as well (very energetic, penetrating radiation like X-rays). We would detect this gamma-radiation at the same time as we think the neutrinos are being produced, at the moment with the NASA Fermi Gamma-ray Space Telescope. Anybody can look for gamma-rays before big solar flares with the RHESSI data broswer. They aren't there, at least not easy to see in these measurements. This is even interesting: it lets us rule out some ideas for solar flares that involve accumulating energy in the form of energetic ions before the flare happens.

There are some other comments that might be made about predicting flares a day and a half before they happen, by any means at all. Flares result because the turbulent flow of gas in the Sun's outer layers twists and stresses the magnetic field. This is a noisy process with random elements; I'm not convinced that even big flares are inevitable this far in advance. But there are other people who could comment more expertly then me.

I think the measurements these gentlemen start from are really interesting, because diferent people have obtained similar results at different times. They may be pointing to a matter of detail in how radioactivity is measured, technically interesting but little more. They might be pointing to some aspect of radioactivity, neutrinos etc. that hasn't previously been properly understood. This would be very exciting! I'm very sceptical about the suggested connection to flares and magnetic activity, however. I'll be keeping an eye on this topic to see how it plays out.