快猫短视频

DAMNED weather forecasters. I was looking forward to a spectacular storm, but
all I got was a night of shivering on the roof.

No, I wasn鈥檛 hoping to get drenched; I wanted to see my first Northern
Lights. Space weather forecasters had predicted that on the nights of 8 and 9
June, a magnetic storm raging around the Earth would create auroras in the night
sky much further south than usual. But like thousands of others, I went to bed
disappointed鈥攖here were no low-latitude auroras that night or the next.
The Siberians had been the only ones to enjoy the show.

Perhaps I should have taken those predictions with a pinch of salt. In May, a
huge magnetic storm had smacked into Earth at over 3 million kilometres an hour,
spraying auroras as far south as Oklahoma. Large as the storm was, space weather
forecasters failed to give us a warning. And with the Sun now at a peak in its
11 year activity cycle, the next one could be worse.

Yet an end to mistaken forecasts might be just around the corner. 鈥淚 think
we鈥檙e in the situation that terrestrial weather forecasters were 30 years ago
when they first got satellite data,鈥� says Bernard Jackson of the University of
California, San Diego. Jackson鈥檚 new forecasting method picked up a disturbance
a couple of days before the storm hit in May. What鈥檚 more, his results cast
doubt on the accepted origins of space storms.

Most scientists agree that a storm is caused by the Sun spewing out a
million-kilometre-long blob of plasma, a coronal mass ejection. CMEs merge with
the solar wind, the stream of ionised gas that is constantly flowing off the
Sun, and form disturbances in the speed, density and magnetic field of the
normally smooth wind. If a CME hits Earth, we have a storm on our hands.

The CME batters the Earth鈥檚 magnetic field, inducing damaging surges of
current in national power grids. In the worst case, the CME鈥檚 field merges with
that of the Earth, allowing high-energy particles to pour into our atmosphere.
In June teams working on NASA鈥檚 Polar Spacecraft and Japan鈥檚 Geotail satellite
simultaneously announced the first sightings of this process, known as
reconnection. As well as producing pretty auroras, these particles can confuse
or kill satellites and lead aeroplanes astray
(see 快猫短视频, 27 February 1999, p 28).
Given a few days notice, satellite operators could
reschedule important manoeuvres and power companies could redistribute load.

At the moment, forecasts come from NASA鈥檚 Space Environment Center in
Boulder, Colorado, which relies on observations from the Solar and Heliospheric
Observatory (SOHO), launched back in 1996. SOHO monitors the surface and
interior of the Sun from its position 1.5 million kilometres sunward of the
Earth. 鈥淲hen SOHO is down, our forecasters are seriously unhappy,鈥� says Ernie
Hildner, director of the centre.

Most valuable among SOHO鈥檚 instruments is a telescopic camera which snaps the
Sun鈥檚 corona every 20 minutes. Simon Plunkett at NASA Goddard Space Flight
Center monitors the images, looking for Earth-bound CMEs. CMEs usually show up
brightly because they are denser than the rest of the solar wind and hence
scatter more light.

But the ones headed straight for us are harder to see, because Plunkett鈥檚
pictures have a blind spot. To photograph the Sun鈥檚 corona, the camera needs its
own artificial eclipse: a black disc obscuring the bulk of the Sun. CMEs headed
directly towards us appear as a whitish halo around this disc. It鈥檚 hard to work
out their direction, and if they fade quickly they may not show up at all.
Hildner reckons that about a quarter of CMEs that are going to cause storms
evade the plethora of instruments focused on the Sun, including SOHO鈥檚
telescopes.

Long time coming

What鈥檚 more, over 60 per cent of predicted storms never materialise. Much as
we knock terrestrial weather forecasters, it鈥檚 hard to imagine them doing as
badly as that. And of storms that do hit, some take only two days to blow in and
others as many as five. The latest method of predicting arrival time
(快猫短视频, 1 July, p 16)
still has a 12-hour margin of error. It鈥檚 clear
that we need to do more than keep tabs on the Sun鈥攚e need to probe the
secrets of the solar wind.

The best attempt so far is ACE (Advanced Composition Explorer), another NASA
satellite lying near SOHO. ACE probes the direction, intensity, speed and
magnetic field of the solar wind, and beams the information down to Earth at the
speed of light. The trouble is, ACE gives only about an hour鈥檚 notice of likely
trouble. In an increasingly wired world, that鈥檚 not enough.

鈥淔or serious forecasts we need to know what will happen to a CME further out,
how much will it slow down, whether it will rotate to Earth, will it hit full-on
or just a glancing blow,鈥� explains Jackson, 鈥渟o we need to be able to track them
well out into space.鈥� His aim is to create a continually updated map of the
weather between Sun and Earth.

Unfortunately, CMEs fade fast. We can see them because light from the Sun
scatters off electrons in their plasma. But that illumination and the density of
the plasma both fall off swiftly further out from the Sun, so CMEs soon become
almost invisible. SOHO鈥檚 instruments, for example, cannot follow them more than
a quarter of the way from Sun to Earth.

The solution is to tune in to radio waves. Back in the 1960s, Antony Hewish
at the University of Cambridge founded the field known as interplanetary
scintillation. His radio telescopes in Cambridge were set up to look at space
weather. By chance, they also discovered pulsars, when graduate student Jocelyn
Bell noticed blips in the signal that were so regular they couldn鈥檛 be caused by
the messy solar wind. The sources themselves were pulsating. Hewish suggested
that these sources were rotating neutron stars, picking up a Nobel prize.

The search for space weather was a less surprising success. Hewish鈥檚
telescopes were trained on very distant sources, such as radio galaxies, looking
at them through the solar wind. Denser plasma absorbs more radio waves, so if
the plasma in the wind is turbulent these sources twinkle or 鈥渟cintillate鈥� every
fraction of a second or so. It is like watching a patch of hot air rising off a
road in summer.

By looking for regions where the scintillation was strong, Hewish could
detect structures in the wind and follow their progress past the Earth. So in
principle, the radio data provide a way to spot the ominous precursors of
storms.

Sadly, simply watching for scintillation isn鈥檛 enough for a good prediction
of the weather. When you see a source twinkling, all you know is that there is
some turbulence moving across your line of sight鈥攜ou can鈥檛 tell how far
along that line it is. A fair guess is to assume that any scintillation is
caused by plasma at the point nearest to the Sun along each line of sight, but
that鈥檚 not reliable enough for a useful forecast.

So the challenge is turning a two-dimensional pattern of twinkles into a
three-dimensional weather map. Something like this is already done with
computer-aided tomography: when doctors make a CAT scan of the body, they take
many line-of-sight readings and use a computer to build up a three-dimensional
model. But the radio telescope arrays can only see a limited number of sources.
鈥淚n medicine you can take as many slices as you want, but we can only take a
few,鈥� says Jackson.

To get around the lack of data, Jackson鈥檚 model starts off with an educated
guess at the current space weather, based on known activity at the Sun鈥檚
surface. The computer calculates what the radio telescopes would see if the
guess were correct, and compares it with what the telescopes actually saw. If it
doesn鈥檛 match, the computer modifies its guess and tries again. After many
iterations, the result is a 3D model of the weather.

Jackson needed observations of the wind as close to the Sun as possible.
Plasma has a characteristic threshold frequency below which radio waves are
absorbed. Denser plasma has a higher threshold, and the density of plasma gets
so high near the Sun itself that only a very high-frequency radio telescope can
cope. The Cambridge telescopes, for example, could only look about halfway in
towards the Sun because of just this problem.

But hope is at hand in the form of Masayoshi Kojima at Nagoya University,
Japan. Kojima and his collaborators run an array of four telescopes running at
327 megahertz, about three times the frequency of the Cambridge array. They can
penetrate plasma as close as 15 million kilometres from the Sun, just a tenth of
the distance to Earth.

Wind speed

Each telescope is fixed in place and can only look close to the Sun at around
noon, but because the telescopes are at different longitudes they take their
snapshots at different times. In this way Kojima and Jackson can measure the
velocity of any structures in the solar wind, and predict how long they will
take to blow in.

So it was that on 22 May Jackson saw a spurt in the velocity of the wind. A
turbulent region of plasma was headed straight at us, and it was coming fast.
The sighting wasn鈥檛 considered worrying by the Boulder weather centre because
SOHO had seen no corresponding solar activity. But then the storm hit on 24 May.
Maybe the forecasters will take more notice next time.

The Japanese are so impressed by results like this that they鈥檙e planning to
spend $20 million on a larger radio telescope array, which will be able
to see more sources and so give finer resolution in the model. Meanwhile, Hewish
is delighted to hear that his work has been continued. His vision is an array of
radio telescopes at different longitudes and on several radio frequencies
dedicated to producing weather maps 24 hours a day. According to Hewish, four
arrays evenly spaced round the globe would do the job.

And radio isn鈥檛 the end of the story. 鈥淭here鈥檚 something about an image in
visible light that just makes things so much easier,鈥� says Simon Plunkett on
SOHO. Jackson鈥檚 CAT scan is not fussy about the kind of data it takes, so he鈥檚
adapting it for a new visible-light observatory called the solar mass ejection
imager (SMEI). SMEI will be launched by NASA and the US Air Force at the end of
2001. It will hold cameras sensitive enough to see starlight scattering off the
solar wind, so CMEs won鈥檛 fade before they make it beyond the cameras鈥� central
blind spot.

SMEI will make modest inroads into that troubling 60 per cent of false
positives, the storms that are predicted when we see a CME heading towards us,
but somehow fail to strike. Some are CMEs that just drift off course, so by
monitoring their speed and direction Jackson can catch them at it and stifle a
false alarm. Unfortunately, most of the culprits are avoiding us for another
reason鈥攖hey reach the Earth, but slide around our magnetic field. 鈥淚f the
CME鈥檚 magnetic field orientation is like Earth鈥檚, the storm will simply bounce
off. It not, it will smack right into us,鈥� says Hildner.

We still have no way to measure the magnetic field of a CME. Jackson has a
few ideas, however. We can measure the field very close to the Sun by analysing
the spectrum of iron in the corona, which is shifted by a magnetic field. And
Jackson鈥檚 model shows that CMEs spread out and rotate like huge teardrops, with
the narrow end anchored in the Sun鈥檚 corona
(see Diagram). Perhaps the CME
structure remains a cohesive magnetic loop. If that鈥檚 true, we could track the
field from the corona right out to Earth. Jackson reckons that developing such a
model will take 10 to 15 years, so it should be ready in time for the next peak
in solar activity.

Meanwhile, Jackson鈥檚 success has spawned a new mystery. The fast-moving
structures he saw in May weren鈥檛 much denser than the rest of the wind. A CME,
on the other hand, should be a lot denser. And there have been several occasions
on which interplanetary scintillation has revealed storms missed by
observatories looking out for CMEs.

Hewish thinks that some storms are caused instead by gusts in the solar wind.
Wind blowing at more than twice the usual speed can rush out of regions called
coronal holes on the surface of the Sun. Coronal holes are already accepted as a
cause for moderately bad space weather around Earth, but Hewish rates them as
far more dangerous. 鈥淚鈥檓 regarded as a bit of a crank on this.鈥� If Hewish is
right, interplanetary scintillation might be the best way to predict storms
because it can pick up high-speed wind like that seen before May鈥檚 giant storm.
But Jackson thinks that was something different again. 鈥淭he event we saw didn鈥檛
rotate with the Sun, so if it was associated with a coronal hole it was a very
transient one,鈥� he says.

Some storms could be even stealthier. There may be regions in the wind where
the magnetic field is reversed. These 鈥渕agnetic bubbles鈥� could also shake up the
Earth鈥檚 magnetosphere and be just as dangerous as CMEs, but they would be
invisible to any known space-weather equipment. 鈥淚f we get SMEI up there and
we鈥檙e still missing storms,鈥� says Hildner, 鈥淚 would have to seriously consider
the possibility of magnetic bubbles.鈥� In which case, come the next peak in solar
activity, we might be just as much in the dark.