Finding Planets in the Datastream

by Paul Gilster on November 19, 2004

A new project called PlanetQuest will soon offer a way to get involved personally with the hunt for extrasolar planets. The idea is to use the power of distributed computing, as the hugely influential SETI@home project has already done, letting people run data analysis software as a screensaver that operates whenever their computer is idle. PlanetQuest will be designed to hunt for planets by studying high-density star regions looking for occlusions — in other words, for evidence that an extrasolar planet has moved between us and its star.

Laurance Doyle, an astrophysicist at the SETI Institute, notes that while occlusions may be rare (after all, the stellar system must be lined up with ours so that planetary orbits cross our line of vision), the hunt will also yield dividends in terms of our knowledge of variable stars, as well as broader issues like stellar stability and evolution.

But even as we accumulate new data, we still have the problem of managing what we have. Consider the Sloan Digital Sky Survey, which intends to map one-quarter of the entire sky in detail, covering hundreds of millions of celestial objects. The results of the Survey are offered electronically to the scientific community and the public; by Survey’s end, about 15 terabytes of information will be produced, a figure that compares favorably to the 24 or so terabytes of textual information held by the Library of Congress.

Here’s what astronomer and computer scientist Alexander Szalay has to say about the problem of massive data sets in a Johns Hopkins press release:

“The massive amounts of data emerging from our newest instruments – telescopes, particle detectors, gene sequencers – demand a novel method of analysis that coalesces the skills of astronomers, biologists and others with those of the computer scientist, the computational scientist and the statistician… Most scientific data collected today will never be directly examined as ‘raw data’ by scientists; instead, it will be put online into ‘smart databases’ where it will be analyzed and summarized by computer programs before scientists even see or use it.”

Szalay has been working for a decade creating multidisciplinary ways to consolidate huge data sets as part of the science archive for the Sloan Digital Sky Survey. This work has also played a role in the creation of the National Virtual Observatory, which is, according to its Web site, “…developing tools that make it easy to locate,
retrieve, and analyze astronomical data from archives and catalogs worldwide, and to compare theoretical models and simulations with observations.”

Szalay has just received a $1.2 million grant from the Gordon and Betty Moore Foundation to continue his team’s work on data analysis in cases where the sheer amount of information demands new and creative ways to extract good science from the dataflow. Just as digital techniques like adaptive optics have brought unexpected clarity to observations from Earth-bound telescopes, so data manipulation will play an ever larger role in exploring star systems for evidence of distant planets.

{ 1 comment }

David Angus August 15, 2006 at 13:56

I’m interested, particularly in helping to find Earthlike worlds, but planets of any size will do.

Dave the Planetbuilder

Comments on this entry are closed.