From: LARRY KLAES (ljk4_at_msn.com)
Date: Tue May 20 2003 - 06:53:49 PDT
Telescopes of the World, Unite! A Cosmic Database Emerges
May 20, 2003
By BRUCE SCHECHTER
About a year ago a large group of astronomers began to
assemble what some of them were calling "the world's best
telescope." Their ambitious instrument is still far from
complete, but they recently took it for a test run. Within
minutes, to their joy and astonishment, they had discovered
three or four brown dwarfs, objects that occupy the niche
between planet and star.
"It gave me shivers when I heard about it," said Dr. Alex
Szalay, a Johns Hopkins astronomer who is one of the
telescope's chief architects.
It wasn't the brown dwarfs themselves that excited Dr.
Szalay; hundreds of them have been discovered in the past
decade. But he and many other astronomers believe that the
means used to discover these objects heralds the beginning
of a new era of astronomy, and even a new era of science.
The telescope that Dr. Szalay and his colleagues have
constructed is not built of glass and metal. It is a
virtual observatory, consisting of terabytes of data
collected by dozens of telescopes on Earth and in space,
and the software necessary to mine these data for
scientific gems.
Like much of the rest of science, astronomy has been the
beneficiary - and victim - of Moore's Law, which states
that the capacity of computers and other silicon-based
devices like charge-coupled devices, or C.C.D.'s, doubles
every 18 months. (The C.C.D has largely replaced
photographic film in astronomical cameras.)
Projects like the National Virtual Observatory, which was
created in response to the tsunami of data that is
threatening to drown astronomers, is creating a new branch
of science, Dr. Szalay believes.
Science, he points out, was "originally empirical, like
Leonardo making wonderful drawings of nature." He
continued: "Next came the theorists who tried to write down
the equations that explained the observed behaviors, like
Kepler or Einstein. Then, when we got to complex enough
systems like the clustering of a million galaxies, there
came computer simulations, the computational branch of
science. Now we are getting into the data exploration part
of science, which is kind of a little bit of them all."
Because its primary tools are computers rather than giant,
multimillion telescopes, this new form of astronomy has the
potential to democratize science.
"If at the same time most of the telescopes in the world
are actually putting all of their data online with proper
explanations," Dr. Szalay said, "then it doesn't matter
where somebody is sitting, they can access all the data -
either somebody in Baltimore, or somebody from Africa who
got a Ph.D. in the U.S. and returned there and doesn't have
access to a telescope but suddenly has a bunch of students.
They can actually get to first-class data."
In the past 25 years the number of C.C.D. pixels in all the
world's telescopes has increased by a factor of 3,000, with
each pixel acting as a miniature astronomical instrument.
The result, Dr. Szalay says, is that the total amount of
astronomical data collected every year is doubling even
while the amount spent on astronomy remains constant.
"We are getting overwhelmed," Dr. Szalay says. "With this
explosion it's not just that individual telescopes are
getting more and more data, but also the threshold gets
lower, so that more and more groups are putting big cameras
on their instruments. Even amateur astronomers today can
generate gigabytes of data per night by attaching a digital
camera to their telescope."
The problem is how to mine this vast store of data for the
riches it almost certainly contains. Astronomers have been
busy over the past couple of decades compiling complete
surveys of the sky, encyclopedic catalogs of millions of
astronomical objects viewed at many different wavelengths.
These surveys exist in about 10 different spectral bands,
from X-rays to the infrared, with each survey giving a
different view of the universe.
The surveys contain about 100 terabytes (one terabyte is
1,000 gigabytes) of data, roughly five times as much as the
Library of Congress holds. Unlike the Library of Congress,
however, this information does not reside in a single
place.
"There is no Library of Congress for astronomy," Dr. Szalay
says, and as long as the data continue to accumulate at an
exponential pace, there will never be one. Instead, "there
will always be 8 or 10 large projects that contain 90
percent of the world's data at any one time."
The goal of the National Virtual Observatory is to make
sure that "the current generation of professional and
amateur astronomers are not overwhelmed by the chores of
getting the actual data," Dr. Szalay says. "So we have to
make it simple and easy for them to use the data in a
friendly way."
In the first stage of the project this has meant creating
tools that can search through different databases without
requiring the searchers to be experts in their individual
details. As a kind of shakedown cruise, the researchers at
the National Virtual Observatory decided to focus on the
data contained in two large sky surveys known as the Sloan
Digital Sky Survey, which looks at the sky in the visible
band of the spectrum and the Two Micron All Sky Survey, or
2MASS, which looks at the sky in the infrared.
"The reason we did those two is that they're very deep,
they dig out objects that are very faint, much fainter than
other surveys have been able to generate," said Dr. Bruce
Berriman, a California Institute of Technology astronomer
involved in the demonstration. "Because it goes to very
faint objects you're able to dig out sources that are
unusual or important in ways other projects can't do."
In particular, by combining the surveys they hoped to spot
brown dwarfs. Brown dwarfs are essentially failed stars,
lumps of matter bigger than a planet but not large enough
to kindle the thermonuclear fire of a star. As a result,
they are relatively cool, emit very little light and are
therefore difficult to spot.
The temperature of a star, like that of a glowing piece of
metal, determines the color of light that it emits: the
cooler the star the redder the light. The light from the
brown dwarfs that the astronomers were searching for
straddles the border between the infrared and the visible.
This means that a brown dwarf should appear in the very
shortest wavelength band of the infrared 2MASS survey and
also in the longest wavelength band of the visible Sloan
survey.
An astronomer looking at just the data from, say, the Sloan
survey and seeing an object in a single band would probably
dismiss it.
"Chances are pretty good that that single band detection is
a piece of junk, some sort of artifact in the detectors in
the telescope, a glint off a bright star, any number of
things," said Dr. Davy Kirkpatrick, a member of the Caltech
team. But if that same object also appears in the 2MASS
data then the chances shoot up that it is something worth
looking at more closely.
The astronomers developed a program that could access these
different databases and search them for matches. Within a
few minutes the computer spit out a half a dozen or so
candidates for possible brown dwarfs. Most of these had
been previously noticed in the data, which others had
sifted through manually.
Finding these brown dwarfs was supposed to be the goal of
the demonstration, a debugging run to prove that the
software worked. But the computer also found several
candidates for new brown dwarfs.
"Astronomers' first reaction when you find a new result is
that there's something wrong," Dr. Berriman said. But after
looking at the data more closely "it slowly dawned on us
that this was something real, that this was a brown dwarf
we found."
"Then our eyes started to widen up a little bit at the
prospect of what might be coming in the future," he
continued.
Dr. Szalay says, "This shows how many new things will come
out in this process once hundreds of astronomers are using
it all over the place."
What makes this result even more impressive is that the
overlap of the two surveys covered something less than one
two-hundredth of the sky, and yet they almost instantly
found objects that astronomers, poring over data for weeks,
had previously missed.
Another aspect of the National Virtual Observatory is the
creation of an astronomical search engine, a kind of Google
for astronomy that will allow amateurs and professionals to
find astronomical resources.
"Were you to go to Google right now and type in the word
galaxy, you wouldn't just get a whole bunch of astronomy
sites," Dr. Berriman says. "You'd also find out about the
L.A. Galaxy soccer team. There's even a town in Texas
called Galaxy, with its own Web page. That's no good to
astronomers because there's so much clutter in the sites."
To help solve this problem the National Virtual
Observatory is creating an online registry of astronomical
resources that should be available to the public early next
year.
The registry, and indeed the entire virtual observatory
project, is intended as a tool for anyone interested in
astronomy.
"One of the major components of the registry is collecting
information about the suitability of the resource for
educational purposes, for amateurs, for students," said Dr.
Robert Hanisch of the Space Telescope Science Institute in
Baltimore. "We have a subgroup of our project concentrating
on what kind of information does that clientele need to
know."
The success of the National Virtual Observatory and similar
projects means that exploring the heavens will no longer be
limited to those few hearty individuals willing to sit
freezing on mountaintops, waiting for the clouds to clear.
Adding myriad seeking eyes and pondering brains to those
already contemplating our place in the universe will be the
greatest achievement of this new technology.
http://www.nytimes.com/2003/05/20/science/space/20DWAR.html?ex=1054438550&ei=1&en=1787555f428fb9b4
This archive was generated by hypermail 2.1.6 : Tue May 20 2003 - 07:23:26 PDT