杏吧原创

WHEN it comes to being part of the action, astronomers certainly drew the
short straw. Meteorologists get to chase hurricanes, biologists can watch
mayflies live and die in the palms of their hands, but astronomers can鈥檛 visit
newborn stars to watch fledgling planets form. Nor can they walk across the
dusty plains of Mars to feel the strength of storms there. Apart from the
occasional spacecraft mission to a lonely outpost of the Solar System, they have
to watch the drama of the Universe unfold on a distant stage.

But many astronomers are now trying to bring the complex worlds beyond Earth
back to the laboratory. They are recreating dust storms on Mars and watching the
seeds of planets grow from replicas of interstellar ice and dust. They are even
building miniature models of Jupiter鈥檚 atmosphere to try to answer questions
that have nagged astronomers for years鈥攚hy, for instance, does the giant
planet have stripes?

In a laboratory at Johns Hopkins University in Baltimore, Peter Olson and his
colleague Jean-Baptiste Manneville have shed some light on that mystery. Over
the past two years they have been trying to recreate the striking banded
structure that appears in the outer layer of Jupiter鈥檚 clouds. Jupiter has
between 10 and 12 light and dark bands that circle the globe in both
hemispheres. Observations from Earth show that these bands are produced by
strong winds circulating in alternating directions, and that they have not
changed their latitudes for at least 100 years.

But how do these winds arise? According to Olson, two models have been in the
running to explain the alternating jet system. One theory is that the winds are
almost entirely driven by heating from the Sun, much like the Earth鈥檚 trade
winds. These occur because heat builds up at the surface in the equatorial
region, warming the air and making it rise. The warm air flows at high altitudes
towards the poles. On its way it cools, sinks and flows back to the equator.
Because the Earth spins towards the east, the trade winds flow westwards.

Big bands

A similar process could create the bands on Jupiter. Olson says that
numerical models show that if the thin outer layer of clouds originally
contained a lot of turbulence, then the rotation of the planet could indeed turn
these patterns into the familiar bands. But the Sun鈥檚 heat doesn鈥檛 penetrate the
outer layer of clouds on Jupiter, so this would be possible only if the winds
were confined to a very thin upper layer of its atmosphere.

However, Olson suspected that an alternative model, first suggested in the
1970s by Fritz Busse of the University of Bayreuth in Germany, might be correct.
Busse鈥檚 idea was that the jets of wind may arise from convection currents
whipped up by the intense heat deep within the planet. Jupiter radiates about
twice as much energy as it receives from the Sun. Some of this heat is left over
from the planet鈥檚 creation and some is generated by drops of helium raining down
into the interior.

鈥淲e wanted to know whether or not the deep convection model could produce a
banded structure,鈥� says Olson. So the researchers decided to construct an
experiment that would replicate at least some of the conditions in Jupiter鈥檚
atmosphere. The average density of Jupiter is only about 1.3 times that of
water, so they chose water as a good approximation. They had to create a
temperature difference across the water to generate convection patterns, and
then simulate the force of Jupiter鈥檚 powerful gravity, which would exert a pull
on the atmosphere towards the planet鈥檚 centre.

To do this Olson and Manneville put together a copper sphere, 25 centimetres
wide, nested inside a 30-centimetre Plexiglas sphere. They filled the gap
between the copper and the Plexiglas with water at room temperature, then
circulated chilled antifreeze within the inner copper sphere to create a
temperature difference between the inner and outer layers of several degrees
centigrade.

By spinning the whole apparatus at a steady 13 revolutions per second to
create a centrifugal force, the researchers simulated gravity. The direction of
the force in the laboratory Jupiter is outward, whereas in the real planet,
obviously, it is inward. But with the direction of the heat flow
reversed鈥攆rom the outer surface to the centre鈥攖he 鈥済ravitational鈥�
force and the heat flow are in opposite directions, just as they are on
Jupiter.

Once everything was in motion, Olson and Manneville injected a fluorescent
dye near the outer boundary of the Plexiglas sphere. When it was bathed in
ultraviolet light, several alternating dark and bright bands emerged in the flow
patterns. Bright zones appeared where the dye was flung to the outer surface of
the Plexiglas by the motion of the sphere. Where convection was strong, however,
the dye was drawn away from the surface toward the copper sphere, creating a
dark belt.

Olson reported in August last year that the experiments show that an inner
heat source could easily be driving the Jovian winds (Icarus, vol 122,
p 242). What鈥檚 more, they also found that the number of bands in the model
Jupiter was proportional to the so-called Rayleigh number, a measure of the
temperature difference across the convecting regions. The Rayleigh number for
Jupiter is far too large to recreate in the lab, but when the researchers
projected their findings to Jupiter鈥檚 value, they worked out that there should
be 10 or 12 bands in each hemisphere, just as observed.

In a spin

Olson says that the results support Busse鈥檚 idea that convection takes place
in narrow columns, hundreds or even thousands of kilometres long, that are
aligned from north to south. Individually, these evolve in a complex, chaotic
way. But on a large scale, each creates a bulk convection pattern that moves at
a unique rate in several giant cylinders, concentric with Jupiter鈥檚 rotation
axis. Where the edges of each of these cylinders emerges at Jupiter鈥檚 surface,
it creates one of the distinctive bands (see
Diagram). In other words, the winds
are not just a surface effect鈥擮lson鈥檚 model suggests that they rage right
through the planet.

Compelling as the results seemed, they were far from conclusive. After all,
just how 鈥渞eal鈥� is this model? Jupiter is over 430 million times larger than the
laboratory replica. Olson and Manneville did not include the effects of the
different amounts of solar heating at different latitudes. Nor did they take
account of Jupiter鈥檚 magnetic field, which is likely to exert a pull on the
metallic hydrogen that makes up much of the inner three-quarters of the planet.
And unlike the real Jupiter, their model had a rigid outer boundary.

鈥淟ab modelling never quite covers all the conditions,鈥� says Olson. 鈥淏ut it
does play a provocative role in that it can either complement or contradict
numerical models.鈥� And whether his model is realistic or not, results from the
probe carried by the Galileo spacecraft that plunged into the atmosphere of
Jupiter in December 1995 are suggesting that it may well be correct. If the
bands were produced by superficial solar heating, they would be expected to run
only a few tens of kilometres deep. But during its descent, the Galileo probe
recorded constant wind speeds of around 650 kilometres per hour from the outer
cloud layer to at least 130 kilometres inside.

Harry Swinney of the University of Texas at Austin agrees that simulation is
an important 鈥渞eality check鈥� on the traditional numerical models. 鈥淭hey have
their place, but sometimes they try to do too much,鈥� he says. 鈥淭hey鈥檒l put
everything into it including the kitchen sink.鈥�

A numerical weather model might try to make predictions for every factor
influencing global circulation, such as precipitation, temperature variations,
wind patterns and cloud reflectivity, then calculate how circulation should
evolve over hundreds of hours. Swinney says that, sure enough, you can see new
patterns evolve, but exactly why a given pattern arises is far from clear
because the model depends on so many things.

Swinney and his colleagues Joel Sommeria and Steven Meyers have turned to
laboratory simulation to try to work out another Jovian mystery鈥攚hy its
Great Red Spot, a titanic cyclone rotating anticlockwise at speeds of about 400
kilometres an hour, has managed to rage for so long. 鈥淭here are other persistent
long-lived vortices on Jupiter and the other planets,鈥� says Swinney. 鈥淏ut
nothing like the Great Red Spot, which has been observed for more than 300
years.鈥� By rights, the Red Spot shouldn鈥檛 exist. Sandwiched between fast-moving
easterly and westerly jets in Jupiter鈥檚 southern hemisphere, it should have been
pulled apart long ago.

Drifting spot

Three decades ago, scientists suggested that the Red Spot could be an
atmospheric disturbance positioned over an isolated mountain. Yet this idea bit
the dust when it became clear that the spot sometimes drifts both east and west.
Later theories drew analogies with terrestrial hurricanes, which are driven by
heat rising from the Earth鈥檚 surface or solar heat absorbed near the equator.
But no numerical model could explain why the Jovian storm is so stable.

So in 1988, Swinney and his colleagues began trying to create vortices like
the Red Spot in the laboratory. They use a rotating doughnut-shaped tank, 86
centimetres in diameter, filled with water pumped through a series of inlets and
outlets in its floor. Again, the water mimics the atmosphere, with dye used to
track currents, and the rotation of the tank simulating that of the planet at
mid-latitudes. The researchers have found that they can create large
vortices that appear spontaneously and, like the Red Spot, are oval shaped and
about twice as long as they are wide. They can survive indefinitely, even if the
rotating tank slows down or speeds up a little.

Swinney鈥檚 simulations cannot clear up all the mysteries of Jupiter鈥檚
atmosphere. But they do confirm that you can simulate the turbulent atmosphere
realistically on a small scale. For instance, the model managed to replicate a
sequence of images from the Voyager spacecraft, which visited Jupiter in 1979.
The images showed the striking patterns that emerge over several days when large
vortices merge with smaller ones. 鈥淚t鈥檚 very similar to ours,鈥� says Swinney.

Jupiter鈥檚 atmosphere is not the only one to have come under the spotlight.
For many years, Ronald Greeley of Arizona State University has been
reconstructing the Martian winds in a wind tunnel at the NASA Ames Research
Center in Moffett Field, California. Once used to test the structures of rockets
at low pressure, it now houses the Martian Surface Wind Tunnel (MARSWIT).

鈥淭here isn鈥檛 any other facility that can simulate the surface of Mars like
ours,鈥� says Greeley. The 14-metre-long wind tunnel looks like a chimney lying on
its side. The walls are made of plywood, except in the middle, where two
Plexiglas panels allow scientists to peek inside the tunnel. The end at which
air flows into the tunnel is conical, like the horn of a trumpet, to focus the
wind. To create the same kind of surface wind turbulence that exists on Mars,
the wind passes over small pebbles fixed to the tunnel floor.

Features on the surface of Mars constantly evolve as sand and dust storms
sculpt long corduroy-like grooves and leave sand dunes on its surface. They also
leave bright and dark streaks up to 200 kilometres long. Sometimes dust storms
can engulf the entire planet, and Greeley and his colleagues have been using the
tunnel to investigate one of the biggest mysteries about Mars鈥攈ow the dust
storms can happen at all. The atmospheric pressure on the planet is only about 6
or 7 millibars, less than a hundredth of the average pressure at the Earth鈥檚
surface.

Visits by the Viking landers in 1976 revealed that the winds kick up fine
dust made up of particles a few micrometres thick. To simulate windblown sand in
the MARSWIT, Greeley and his team use silica microspheres mixed with natural
silt. To mimic the Martian dust, they use finely ground walnut shells, which
have a similar shape. They are less dense than their Martian counterparts, and
this compensates for the fact that Mars has just two-thirds the gravity of the
Earth.

To find out the minimum windspeed needed to whip up the dust storms, Greeley
evacuated the wind tunnel to pressures similar to that on Mars. They then
increased the windspeed until the dust started to rise from the surface. Their
results, which they have submitted to Icarus, suggest that the minimum
windspeed for kicking up dust is 450 kilometres per hour over a smooth flat
surface. When the surface was strewn with boulders, however, the dust could be
stirred up by a wind of only 155 kilometres per hour. 鈥淭he minimum velocity to
move things around was one of our best results,鈥� says Greeley. 鈥淲e were able to
nail these down more.鈥�

No one yet knows whether these windspeeds actually match those on Mars. The
fastest winds measured by the Viking lander had speeds of only about 130
kilometres per hour. But NASA鈥檚 Pathfinder spacecraft, which is now on its way
to Mars and due to land in July, carries a series of windsocks that will give
the best measurements to date of windspeeds. Greeley, who is involved in the
Pathfinder mission, is eagerly anticipating the spacecraft鈥檚 data. 鈥淲e can鈥檛
wait to apply our lab results to [Pathfinder鈥檚] results,鈥� he says. The MARSWIT
was used to check that Pathfinder鈥檚 windsocks would not be aerodynamically
unstable and 鈥渇lutter鈥� in Martian conditions.

Greeley has also used the wind tunnel to show that the light streaks on Mars
must be formed when dust settles out of the atmosphere, while the dark streaks
form in the wake of obstructions to the wind, such as large craters. Like Olson,
he feels that joining forces with computer modellers is crucial for getting to
the bottom of planetary geology. 鈥淚t takes this multi-prong approach to let you
gain on the problems in a systematic fashion,鈥� he says. 鈥淣ature is always more
complicated than any model. So it really takes both approaches to come up with
good answers.鈥�

Researchers in Germany are simulating the conditions in the early Solar
System long before Mars and Jupiter came into existence. Five billion years ago,
the Solar System grew out of a disc of hydrogen and helium gas and countless
grains of silicate dust, typically a micrometre or two across. Astronomers
believe that the planets began to form as the dust particles collided as a
result of their random, Brownian motion. Some stuck together. As the particles
grew bigger than 10 micrometres across, friction and turbulence came into play,
increasing the number of collisions and making the particles bigger still. Once
the clumps were about a kilometre in diameter, they would have exerted a large
enough gravitational pull to suck in more particles.

Model spanners

But numerical models throw a spanner in the works, say J眉rgen Blum and
Torsten Poppe of the University of Jena, Germany. These suggest that dust grains
in the early Solar System would have moved too fast to adhere to each other
after impact. Instead of sticking together, small particles would ricochet off
one another. This rebound effect would prevent clumps from growing larger than
about 10 micrometres across (New 杏吧原创, Science, 27 July, p
15).

So why then are fully grown planets flying round the Sun today? In 1994, Blum
and Poppe began constructing a device that would allow them to fire single
grains of silicon dioxide, between 0.5 and 2 micrometers across, at a flat,
polished silica target. This allowed them to measure the upper and lower
velocity limits for particle sticking.

They found that particles in the middle size range of their test sample stuck
together at velocities of about 1.2 metres per second鈥攆our times higher
than predicted by numerical models. Moreover, they found that these particles
often became electrostatically charged when they collided. This electrostatic
effect could help glue them together, and may explain why planets exist.

Blum and Poppe now hope to take artificial planet-building a step further. To
simulate more accurately the low-gravity conditions that existed in the
primordial Solar System, Poppe and Blum have designed the Cosmic Dust
Aggregation Experiment (CODAG), to fly aboard the space shuttle in 1997. A
stream of gas and 2-micrometre particles will be injected into a 1.5-litre
vacuum chamber on the shuttle, forming a weightless, diffuse cloud. The
experiment should reveal the shapes of the particles and how they clump
together.

The CODAG experiment could also point the way to the birthplaces of planets
far beyond the Sun. Blum and Poppe will measure how light is scattered by the
fluffy aggregates in the canister, and compare this to the way light is
scattered by the gas discs around young stars.

鈥淥ur results should enable astronomers to conclude whether the discs of dust
around young stellar objects contain the seeds of planets,鈥� says Poppe. In a few
years, we may be able to focus on faint planets forming round distant stars,
thanks to nothing more than a little box of dust and gas.