NCSA and quasars–those distant, tantalizing, extremely bright objects in the night sky, thought to be powered by supermassive black holes–go way, way back.
“Many of the simulations which have given credence to the standard model of quasars and active galactic nuclei were done on NCSA systems,” says Robert Brunner, assistant professor in the department of astronomy at the University of Illinois at Urbana-Champaign and a research scientist at NCSA.
By looking at what kind of light quasars emit, Brunner and other researchers interested in these objects can determine not only how far away they are, but also the physics that governs their extreme luminosity. Furthermore, because they are so bright, quasars can be seen to great distances, allowing astronomers to use them to probe both the physics of the early universe and the formation and evolution of galaxies.
Now Brunner, who leads NCSA's Laboratory for Cosmological Data Mining, is overseeing the processing and storing of data from the Palomar-Quest Survey (PQS), a sky survey that, by imaging the same large region of sky night after night, might help identify gradual changes in the fabric of the universe.
A Wider View
The Palomar-Quest Survey is a collaboration involving the California Institute of Technology, Yale University, Indiana University, and UIUC/NCSA. Charles Baltay, professor of physics at Yale, co-leads the PQS project and is particularly interested in finding and tracking objects that might change dramatically over a period of weeks or years–a characteristic he identifies as “time variability.” Of particular interest, says Baltay, are very high redshift quasars, “which tell us about the early history of the universe,” and supernovae, which “tell you whether the universe is expanding or contracting, and if it's expanding, whether the expansion is getting slower or speeding up.”
The original Quest survey was a collaboration of several North and South American universities in which a camera mounted on the 1.0-m Schmidt telescope at the Llano del Hato National Astronomical Observatory in Venezuela was used to take snapshots of the equatorial sky. This camera consisted of only 16 CCDs, or charge-coupled devices, the solid-state silicon chips inside all digital cameras that convert light into electrical impulses that can be recorded to disk as data.
For the PQS, a larger camera, designed and constructed by Yale and Indiana University, has been mounted on the 48-inch Oschin Schmidt telescope at the Palomar Observatory, located near San Diego and run by the California Institute of Technology. The new camera at Palomar contains 112 CCDs, sufficient to cover a much larger region of the sky. The actual observations are controlled from a special room at Yale. However, the drift-scan survey data are processed and stored at NCSA, chiefly by Adam Rengstorf, a postdoctoral researcher in Brunner's group at UIUC. Says Rengstorf, “Every two minutes, 20 seconds, we get 112 images off the camera–and that continues for up to nine hours a night.” The resulting data flow, says Rengstorf, is just under 7.4 gigabytes an hour.
The next day, the raw data are bundled and transferred to NCSA's UniTree mass storage system, after which the information is processed in a complex series of steps. Earlier this year, this was done using a frame of 32 processors on NCSA's Platinum cluster. This processing pipeline is being converted to operate on generic grid infrastructure, eventually allowing the processing to utilize the TeraGrid. “You have to unpack all the data, do the bookkeeping and organization, and remove the instrumental signatures from the data,” explains Rengstorf. This includes “flattening” the image to compensate for the possibility that a given CCD does not respond to light uniformly across its surface. “Then you have to detect objects, and once you detect them you have to figure out where they are on the different chips.”
After the positions of these objects are calculated, they are mapped according to the U.S. Naval Observatory astrometric catalogue, a very accurate map of the known sky. The brightness of each star is measured. All of this information is written to an output catalog at NCSA and returned to mass storage for use by members of the collaboration.
Timing Is Everything
The Palomar-Quest Survey's emphasis on time variability is an important advance for observational astronomy. It requires imaging a large region of the sky repeatedly over the course of months or years while scanning for changes. This in turn requires using a Schmidt telescope, which has a much wider field of view than most telescopes in use today–including Hubble. The 48-inch Samuel Oschin telescope at Palomar is the second-largest Schmidt telescope in the world.
“People have studied quasar variability before,” says Brunner. “But it's usually to monitor a small number of sources for 10-20 years, or to match new data against much older data. Actually going out and surveying significant fractions of the entire sky–a tenth, a fifth, a quarter–and doing it over and over again for a number of years–that's really a new thing.”
Brunner says that the process has worked well from the very first data transmission in April of this year. “We're an invisible part of the infrastructure now,” he says. And for PQS, invisibility–running smoothly and uneventfully without a hitch–is crucial. “If you want to catch things which are changing in time,” says Baltay, “you don't have the leisure of waiting months and months before you process your data, because by that time the thing is gone.” He hopes that in the future data transmission can be speeded up even more dramatically through automation, so that, for example, raw data received at 8:10 p.m. can be sent, processed, and archived 20 minutes later.
Using data from the new camera that was processed at Caltech led, for example, to the discovery in November 2003 of the planetoid Sedna, the most distant solar system object ever observed and the largest Kuiper Belt object after Pluto. More recently, Palomar-Quest Survey data processed at NCSA yielded a potentially valuable cosmological discovery: a relatively rare, redshift 4.07 quasar that was discovered in follow-up, time-critical observations by a team led by Caltech professor George Djorgovski. While Brunner points out that more distant quasars have been identified, he believes that the real value of the find is that it demonstrates what PQS can do: “It's proof that we're going to be able to systematically find high redshift quasars using the Palomar-Quest Survey.”
Undiscovered Countries
Brunner and Rengstorf are excited about the scientific potential of studying the time variability of quasars and other astrophysical objects. Brunner, whose interest is in cosmological data mining, expects soon to be able to tackle the problem of characterizing quasars according to their variability using data from PQS in conjunction with spectral analysis from other sources. “Optically, quasars and stars might look the same, just little points of light,” says Rengstorf, “but when you take a full spectra, quasars and stars are very distinct. The spectra show you the underlying physics of what's going on and lets you quantify more precisely how far away the quasar is, what its redshift is.”
Baltay also has his sights set on the big picture. As transmission and processing times decrease, the possibility increases of discovering objects which undergo changes in the sky in ever shorter periods of time. What sorts of objects? Baltay enthusiastically admits that he doesn't know. “Typically you find what you look for,” he says. “No one has looked at large areas of the sky with this kind of sensitivity. In the future we might be able to see faint objects at a very rapid time delay–eventually minutes or fractions of an hour. So who knows what we'll find?
“That's part of the fun of looking–you might find something completely new.”
PQS operations at NCSA have been funded in part by the NSF PACI Cooperative Agreement, and NASA grants NAG 5-12578 and NAG 5-12580.
Access Online: http://access.ncsa.uiuc.edu/CoverStories/quest2/
For further information:
http://astronomy.ncsa.uiuc.edu/quest2/index.php