Global supercomputer leader Cray Inc. (NASDAQ: CRAY) today announced that
researchers running simulations on the Cray supercomputer at Sandia National
Laboratories have re-created what could have happened 29 million years ago
when an asteroid explosion turned Saharan sand into glass. The greenish
natural glass, which can still be found scattered across remote stretches of
the desert, was used by an artisan in ancient Egypt to carve a scarab that
decorates one of the bejeweled breastplates buried in King Tutankhamen's
tomb.
"Supercomputers now allow us to approach these problems as if we were
conducting actual experiments," said Mark Boslough, the physicist at Sandia
whose theory about the origins of Libyan Desert Glass sparked the research.
"With this class of computer, we can run multiple simulations at such high
resolution and fidelity that we can see phenomena that we wouldn't be able
to predict from first principles. That means we can explore alternate
possibilities as we go. It's more like doing iterative experimental science
than theoretical science."
The Cray supercomputer at Sandia, nicknamed Red Storm, was developed jointly
by Cray and Sandia, a part of the Department of Energy's National Nuclear
Security Administration. Sandia upgraded Red Storm late last year to three
times its original performance level, boosting its performance to more than
100 teraflops, or 100 trillion floating point operations per second. Red
Storm is one of only three supercomputers in the world to exceed the 100
teraflops mark, according to the TOP500 results released last month.
"The Libyan Desert Glass study at Sandia is truly exciting research that
crosses a number of scientific disciplines -- ranging from impact physics
and geology to Egyptology," said Jan Silverman, senior vice president,
corporate strategy and business development at Cray. "We are delighted to
hear about how our highly scalable Cray XT(TM) supercomputer architecture
allows iterative modeling techniques to find the most probable explanation.
Using the computational power of our supercomputers we also see similar
iterative techniques being used to optimize designs from automobiles to
airplanes."
Clues To a Mystery
Until recently Earth scientists believed that natural glass can form by only
two high-temperature processes. Volcanic glass, such as obsidian, can be
produced when lava cools rapidly. Or, in rare cases, a glass known as
tektite can form from the high pressures generated when an asteroid or comet
directly impacts the earth. But compositional studies indicate that Libyan
Desert Glass does not fit either of these two categories. Adding to the
puzzle, scientists generally agree the Libyan glass was somehow formed by a
collision with an object from space, but no one has ever been able to
confirm an impact crater in the region.
Boslough found one clue to the glass mystery in the 1994 collision between
the Comet Shoemaker-Levy 9 and Jupiter. That comet broke up into several
pieces before it made contact with Jupiter's atmosphere, where the
collisions caused fireballs that shot hundreds of miles above the planet.
Boslough conjectured that if such an air burst were to occur above Earth, it
might generate enough heat to fuse surface materials into glass.
Another clue was the Tunguska explosion that flattened a thousand square
miles of forest across Siberia in 1908. Because there is no crater of
sufficient size to have caused this event, it is generally believed that the
Tunguska blast was the result of a meteoroid or comet fragment that exploded
at an altitude of five to 10 kilometers (three to six miles) above the
Earth's surface.
Boslough argues that a similar atmospheric explosion could have created
fireballs large enough and hot enough to produce the Libyan Desert Glass.
Such glass would have been forged in seconds, much like the glass that
formed from super-heated sand at the Trinity site in New Mexico during the
first atom bomb test in 1945. If the asteroid blast occurred above the
Earth, there would be no evidence of a collision in the composition of the
glass and no significant crater in the ground.
Re-creating the Blast
"What I focused on in the simulations was the explosion of the asteroid,"
said Boslough. "As the object entered the atmosphere it had tremendous
kinetic energy. Much of that energy was converted to heat, creating a blast
as hot as the surface of the sun over a large area. The fireball remained in
contact with the Earth's surface for more than 20 seconds. At the same time,
winds behind the blast reached a speed of several hundred meters per second.
The glass formed from the rapid melting and quenching of the sandstone and
alluvium on the ground."
Boslough and his colleagues at Sandia performed high-resolution hydrocode
simulations on Red Storm using the CTH shock-physics code. They postulated a
120-meter diameter stony asteroid hitting the atmosphere at 20 kilometers
per second and breaking up, touching off a blast equivalent to a 110 megaton
bomb and producing intense heat and high-velocity winds.
According to the simulations, this explosion would have been more than
sufficient to melt rocky material on the surface and then cool it quickly,
the conditions necessary to form natural glass. The high winds would have
accelerated the melting process by blowing away the boundary or "melt" layer
that would otherwise insulate the stone from the heat.
Boslough and his group conducted a number of simulations to come up with
their results.
"Multiple iterations are really important for gaining new insights," he
said. "You can't plan out your whole experimental matrix and lock yourself
in. When we vary the parameters, we can see new things. For example, we
observed a large ring vortex during the explosion that acts as a 'lubricant'
for the downward flow of mass and energy. No one had suggested that was
possible before."
For more information about the Libyan Desert Glass study, go to
http://www.sandia.gov/news/publications/technology/2006/0804/glass.html.