terça-feira, 28 de fevereiro de 2017
How to predict and prepare for space weather
SOMETIMES the sun burps. It flings off mighty arcs of hot plasma known as coronal mass ejections (CMEs). If one of these hits Earth it plays havoc with the planet’s magnetic field. Such storms are among the most spectacular examples of what astronomers call space weather, a subject to which a session at this year’s meeting of the American Association for the Advancement of Science (AAAS), in Boston, was devoted. A big CME can have profound effects. In 1859, for instance, a CME subsequently dubbed the Carrington event, after a British astronomer who realised its connection with a powerful solar flare he had observed a few days earlier, generated auroras that could be seen in the tropics. Normally, as the names “northern” and “southern” lights suggest, such auroras (pictured above) are visible only from high latitude. More significant, the Carrington event played havoc with Earth’s new telecommunications system, the electric telegraph. Lines and networks failed, and some operators received severe shocks.
Today, the damage would be worse. A study published in 2013 by Lloyd’s, a London insurance market, estimated that a Carrington-like event now would cause damage costing between $600bn and $2.6trn in America alone. A year before this report came out the sun had indeed thrown off such an ejection—though not in the direction of Earth. A much smaller storm did, however, do serious damage in 1989, by inducing powerful currents in Quebec’s grid, blacking out millions of people. It would therefore be useful, Jonathan Pellish of the Goddard Space Flight Centre, a NASA laboratory, told the meeting, to be able to forecast space weather in much the same way as weather is forecast on Earth. This would permit the most vulnerable equipment to be disconnected, in advance of a CME’s arrival, to prevent damaging power surges.
Sturm und drang
It sounds straightforward enough, but is harder than it sounds. Though CMEs are common, they cause problems on Earth only if they score a direct hit. The so-called “empty” interplanetary space of the solar system is, in fact, suffused by a thin soup of charged particles. These particles interact with moving CMEs in ways that are hard to predict. That makes forecasting a storm’s track difficult. On top of this, CMEs themselves have magnetic fields, with north and south poles, just as Earth does. The way the poles of a CME line up with those of Earth can affect the intensity of the resulting electrical activity.
To try to understand all this better a number of satellites already monitor the sun, looking for, among other things, CMEs. These include a fleet of American environment-modelling craft and also the Solar and Heliospheric Observatory, which is a joint European-American venture launched in 1995. Several new sun-watching instruments are planned for the next couple of years. One is the European Space Agency’s Solar Orbiter. Another is NASA’s Solar Probe Plus. A third is a special telescope, called DKIST, to be built in Hawaii. The eventual goal, said Dr Pellish, is to make space-weather forecasts as easy and routine as terrestrial ones.
Preparing for the extraterrestrial equivalent of hurricanes in this way is surely wise. But space drizzle can cause problems too. Even when the sun is quiet, Earth is bombarded by a steady stream of high-energy subatomic particles. Some come from the sun, which is always shedding matter in small quantities even when it is not throwing off CMEs. Others are cosmic rays, which originate from outside the solar system. Both types, when they smash through the atmosphere, create showers of secondary particles in their wake. And, as Bharat Bhuva, an engineer at Vanderbilt University in Tennessee, described to the meeting, this shrapnel can cause problems with the electronic devices on which people increasingly depend.
If such a particle hits a computer chip, it can inject an electrical charge into the circuit. Since chips work their magic by manipulating packets of charge, that can create all sorts of problems. Dr Bhuva described how, in 2008, the autopilot of a Qantas airliner had been knocked out by a rogue particle. The resulting sudden plunge of about 200 metres injured many of the passengers, a dozen seriously.
Subtler effects can be just as worrying. During a local election in Belgium in 2003, a single scrambled bit of information, almost certainly caused by an errant particle, added 4,096 votes to one candidate’s tally. Since this gave an impossibly high total, the mistake was easily spotted. But had the particle hit a different part of the circuit it might have added a smaller number of votes—enough to change the outcome without anyone noticing. Moreover, as the components from which computer chips are built continue to shrink, they become more sensitive, making the problem worse. A modern computer might expect somewhere between a hundred and a thousand space-drizzle-induced errors per billion transistors per billion hours of operation. That sounds low. But modern chips have tens of billions of transistors, and modern data centres have millions of chips—so the numbers quickly add up.
The trick is to design circuits to cope. That is where Christopher Frost, who works at the Rutherford Appleton Laboratory, near Oxford, thinks he can help. He and his team have modified some particle accelerators in a way that offers designers of electronic equipment the ability to test their products—and, crucially, to test them quickly. Dr Frost’s particle beams are millions of times more intense than the radiation experienced by real-world devices. They deliver in minutes a dose that would take years to arrive naturally.
This sort of pre-emptive action makes sense. The threats from space drizzle (constant, though low-level) and from CMEs (rare, but potentially catastrophic) are real. Hardening equipment against drizzle, and developing forecasts that tell you when to disconnect it to avoid CME-induced power surges, are merely sensible precautions.