Pitt Researchers Tackle Flood of Space Data With $1.6 Million Project

Issue Date: 
July 18, 2011

constellationsIn the past 15 years, the wealth of data available to astronomers through new and larger digital-optimized telescopes has revolutionized our understanding of the Universe’s inner workings. However, these vast stores of images have also generated a cosmic headache as thousands of researchers have produced an unorganized torrent of different measurements.

In response, University of Pittsburgh researchers have undertaken a four-year, $1.6 million project supported by the National Science Foundation to create a single online source that astronomers worldwide can use to organize and quickly share their latest celestial observations. A publicly available tool called AstroShelf would allow researchers to add new measurements about astronomical objects, report their findings in real time, and work with one another’s data, explained principal investigator Alexandros Labrinidis, a professor of computer science in Pitt’s School of Arts and Sciences. Labrinidis leads the project with Pitt computer science professors Panos Chrysanthis and Liz Marai, as well as with Pitt professors of physics and astronomy Jeffrey Newman, Michael Wood-Vasey, and Arthur Kosowsky.

Large “survey” projects have been a key development in astronomy during the past decade, allowing astronomers the world over to pore over uniformly high-quality images of deep space. Studies of distant galaxies were previously limited to a few astronomers with access to the most powerful ground-based telescopes, explained Newman, a nationally recognized expert in astronomical surveys. New satellites and enormous digital cameras on specially designed telescopes have now obtained detailed images of vast swaths of the night sky, capturing cosmic evolution and activity spanning more than 10 billion years.

The popularity of large surveys mushroomed after 2002 with the public release of photos from the Sloan Digital Sky Survey, the most influential survey project and one in which Pitt had a part, Newman said. Now there are dozens of projects that document the active lives of celestial bodies; Pitt cosmologists alone help lead or participate in approximately 12 different undertakings. And the resulting data can be enormous. For instance, Newman and Wood-Vasey lead Pitt’s participation in the Large Synoptic Survey Telescope (LSST), a telescope with a 27-foot diameter and fitted with a 3-billion-pixel camera under construction in Chile. Scheduled to be operational by the end of the decade, the LSST will scan the sky every three days and generate roughly 30 terabytes (more than 5,000 DVDs) of information every night for 10 years. In just a few nights of operation, it will match the amount of data collected by the Sloan survey.

This dizzying flood of information also comes in a variety of forms, Newman said. In the absence of an agreed-upon format, researchers create data sets with a wide variety of parameters and names for celestial bodies. In addition, smaller, potentially important projects lack the communications heft of well-funded efforts and become lost in the din. Newman compared conducting research within the current jungle of information to planning a large dinner party that requires purchasing each ingredient at a different store—with the staff at each store speaking a different language.

The Pitt team’s first step is to create a computational framework that will let astronomers link their observations to specific experiments, models, or other observations. Labrinidis and Chrysanthis will lead the design of an annotation framework to create these links along with an automated work-flow system that will streamline many of the tasks astronomers currently perform manually, such as coordinating follow-up observations. Marai will develop a visual interface allowing astronomers to work directly with images of the sky as they construct and test computational models of the Universe. Together, the three of them will create software to interactively analyze experimental results, as well as to construct and test hypotheses.

AstroShelf also will allow images obtained using different types of light—such as infrared or X-rays—to be correlated with ease to reveal obscured details of distant galaxies and black holes. This feature will be designed using data from two projects for which Newman is a key member: DEEP3 (the Deep Extragalactic Evolutionary Probe 3) and AEGIS (the All-Wavelength Extended Groth Strip International Survey). DEEP3 examines the light of more than 50,000 faraway galaxies to determine the bodies’ composition and distance from Earth. The AEGIS project combines images of the Groth Strip—an area the width of four full moons near the “handle” of the Big Dipper—from the largest ground- and space-based telescopes to document how galaxies have grown and changed over the past 10 billion years.

AstroShelf also will enable researchers to record and share preliminary results of their analyses, Labrinidis said. This capability would be particularly valuable for studying such transient events as supernovae, which are visible from Earth for only a short time and often require multi-institutional coordination to fully observe. Pitt’s Wood-Vasey is centrally involved in discovering transient events for one such project, Pan-STARRS (Panoramic Survey Telescope and Rapid Response System), a large survey that uses the world’s most powerful digital camera to search the sky for explosive and violent deep-space events.