Newest constellation: data

At one time, to be an astronomer was to freeze for science. Long, cold nights peering through a telescope gave way to long, frosty nights sitting in a cramped cage at the telescope's focus point, exposing large photographic plates to capture images of stars, nebulae, and galaxies.

Then, the digital revolution allowed astronomers to retreat to the warmth of the telescope's control room, complete with couches, microwave ovens, and the soft strains of Bach on the stereo. High-tech instruments could do the peering - filling the cages where humans used to sit.

Now, researchers are planning a "virtual" observatory that could take full advantage of the digital revolution, open the way to a fresh burst of discovery - and would leave Sir William Herschel and Edwin Hubble scratching their heads in wonder.

Instead of scanning the heavens, astronomers piecing together the history of the universe and its menagerie of objects would scan databases bursting with images and spectra squirreled away from a growing list of astronomical survey projects.

The observatory's planners call it the National Virtual Observatory (NVO), an idea that received high priority in a recent National Research Council report outlining space-science priorities for the next decade.

One indication of the value astronomers are beginning to place on mining databases, such as NVO will build, comes from the Hubble Space Telescope's experience, according to David De Young, a staff scientist at the National Optical Astronomy Observatories headquarters here. "There's more data flying from the space-telescope archives out to the astronomical community than is flying from the telescope," he says. "It's an interesting litmus test of how valuable people feel these archives can be."

Several factors are driving the push for the NVO, Dr. De Young says. First, technology is allowing astronomers to pull in ever larger amounts of data and send it at ever faster speeds.

This, in turn, has allowed scientists to launch projects such as the Sloan Digital Sky Survey, the 2MASS infrared survey, and other efforts. These are filling digital databases with information taken at a range of wavelengths - from radio through the optical segment of the electromagnetic spectrum, and beyond. In addition, this decade astronomers hope to build a 6.5-meter telescope that will image the entire visible sky once a week. Each day, such a telescope alone could pull in enough data to fill the harddrives of more than 160 home computers.

Faced with enormous data-handling needs, the NVO is forcing planners to devise new ways of storing, processing, and opening speedy portals to the wealth of information the observatory will hold. Indeed, the project's architects hope to tap the National Science Foundation for $15 million over the next five years to develop the digital framework for the observatory.

Once in place, the observatory is expected to lead to discoveries, ranging from the large-scale structure of the universe to asteroid hazards in our own solar system. With billions of objects in the data base, astronomers anticipate the discovery of currently unknown types of objects - ones that may appear only in 1 out of 100 million images.

The observatory also may have a profound effect on the sociology of astronomy, De Young says, forcing some researchers to give up the notion that no one can take data better than they can, and prompting funding agencies such as the NSF to earmark more grant money to projects that involve mining archives instead of manning telescopes.

Beyond the scientific gains planners envision for the NVO lie benefits for the public, which will have access to the databases as well.

Projects such as the NVO and GriPhyN, a high-energy physics counterpart, "will go a long way toward improving science education," as students or teachers will be able to craft projects that use real data - and perhaps lead to original discoveries, says Charles Koelbel, program director for advanced computing research at the NSF.

(c) Copyright 2001. The Christian Science Monitor