Understanding Catalog Capriciousness
Webmaster Note: This article first appeared in the April 1997 issue of Sky & Telescope Magazine. At that time, the Hipparcos Catalog had not been released. Today, TheSky with Gaia star catalog gives astronomers and citizen scientists an unparalleled tool for astronomical research and discovery.
Astronomical Computing
written by Stuart J. Goldman
Reprinted with permission.
Behold the differences in deep-sky catalogs. Thomas Bisque prepared this synopsis for the galaxy cluster Stephan’s Quintet in Pegasus. On a background image from the RealSky CDROM, he overlaid the galaxies’ positions from the Revised New General Catalogue (yellow crosses), NGC 2000.0 (aqua circles), and Principal Galaxy Catalogue (green ovals). The field is 8 arc-minutes wide. Courtesy Software Bisque.
SOFTWARE for amateur astronomers used to be limited to showing the planets against a background of naked-eye stars. Today, high-capacity data storage has made millions of stars and deep-sky objects instantly available, while Pentium and PowerPC processors have eased users’ impatience in redrawing vivid sky representations. But such enormous detail is still accompanied by a liability. Astronomical software simulates the sky, and no simulation is perfect. As comprehensive as the astronomical catalogs used by software are, no database is totally accurate or complete. Each reference has its own peculiarities, which have origins in how the data were selected, compiled, or processed. And while a source may be 99.99 percent correct, the 0.01-percent shortfall still amounts to 100 errors per million entries. Perhaps the best-known historical example of a catalog goof is entry 102 in the Messier Catalog – the object never existed. Instead, M102 was galaxy M101 logged with two positions. And then there is the case of the star 34 Tauri. It doesn’t exist either. John Flamsteed observed the “star” in December 1690 and added it to his catalog. However, the object was in fact the planet Uranus, witnessed 91 years before William Herschel discovered it (S&T- January 1978, page 52).
Today’s software can
reveal the limitations of
astronomical databases.
Errors in catalogs are becoming more prominent with the growing capabilities of astronomical software coupled with the escalating size of databases and instantaneous communication via the Internet. Amateurs need to keep the vagaries of astronomical catalogs in mind when using software. Otherwise, like a motorist who discovers that the street he or she is on isn’t plotted on a map, you could find yourself on a road of confusion.
VERY STARRY SKIES Arguably, the “new age” of astronomical catalogs began with the 1989 release of the Hubble Guide Star Catalog. It picked up where other star tallies (primarily the SAO Catalog) had left off at about 9th magnitude. Desktop planetarium software was transformed as the number of stars available increased 60fold to 15 million. Star plots down to 15th magnitude were instantly appreciated by amateur astronomers who, by using large telescopes or CCD cameras, were probing ever deeper into the cosmos.
As revolutionary as the GSC was, it was not errorless. As explained by Ray Villard in the December 1989 issue, page 583, the GSC was created by scanning Schmidt telescope plates. The data were filtered and classified as stellar or nonstellar (some 3.6 million objects). The GSC had a single purpose – to support the operation of the Hubble Space Telescope. So the catalog does not include stars that would not serve as guide stars.
Bisque used the 101 disc version of the Digitized Sky Survey for this image of globular cluster Omega Centauri (RealSky does not cover the sky south of about declination -15′). The position of this object in the RNGC (red cross) is 10 arc-minutes off. NGC 2000.0 is more on target. Courtesy Software Bisque.
Clearly, the catalog will not be as large as the GSC. The value of this project is not in the number of stars but in its un- unprecedent accuracy. Stellar positions will be known to 0.02 arc-seconds or better. Furthermore, the data will include proper motions for some stars to 1 arc-second per millennium.
The Hipparcos harvest will surely be incorporated into future astronomical software, boosting the level of accuracy.
Its availability will also heighten the “play” factor of the programs, as mouse pad astronomers can take into account the space motion of stars. Will astronomers find errors in the Hipparcos data? Probably.
Stars brighter than about 9th magnitude are typically not listed in the stellar portion of the GSC, since their images were so large on the Schmidt plates that they were classified as nonstellar by the inventory algorithms. Most of the objects flagged as nonstellar are galaxies, but this catchall category also includes plate flaws (reflections and diffraction spikes), asteroid trails, and pairs of unresolved stars. The GSC also does not contain stars that are close to bright stars. As the picture below illustrates, there are “holes” around many naked-eye stars because the fields on the Schmidt plates were overexposed.
There are even more bizarre holes in the data. If you direct your GSC-capable program to right ascension 17′ 20m, declination +33′ 10′ in Hercules (2000.0
coordinates), you will find a 1 degree-long strip devoid of faint stars. This is not a curiously delineated dark nebula but an improperly scanned region of sky There are other, smaller holes too. Use a cloudy night to find them.
The GSC has been revised to correct a variety of subtle problems. As of this writing, the final touches on version 1.2 were being completed, which will, among other amends, fix systematic plate distortions.
The next phase of star-catalog utility is about to dawn. The bounty of the European Space Agency’s Hipparcos satellite will be released publicly on June 1st. The mission’s two long-anticipated databases will embody precise positions and magnitudes for one million stars of 11th magnitude and brighter.
Just as the GSC changed the face of astronomy software regarding stars, another innovation is under way with the introduction of the Palomar Observatory Sky Survey (POSS) for home use. What was once available only as photographic prints (costing as much as $15,000) and then as a digitally scanned version on 101 CD-ROMs (for about $3,000), can now be had on eight CD-ROMs from the Astronomical Society of the Pacific (for $250; see the January issue, page 65).
RealSky is an amazing reference self. However, when combined with Software Bisque’s TheSky version 4.0, so far the only popular-level program that can directly access the POSS images) the software takes on an added life as amateurs begin to use this source, they are noticing some peculiarities. For example, when you call up a region from RealSky – say, Stephan’s Quintet in Pegasus – and then overlay the catalog positions, they don’t match.
Thomas Bisque of Software Bisque has looked at catalog errors with regard to the RealSky CD-ROM for several months, since people began contacting his company with questions. “I am surprised how many NGC and other catalog errors there are when playing around with TheSky and RealSky,” he told Sky & Telescope. “In fact, many customers are now noticing more and more of these errors and are now questioning why TheSky’s positions for so many objects are wrong.”
The positions don’t match TheSky because the input catalogs do not list objects with arc-second accuracy. These include J. L. E. Dreyer’s original New General Catalogue (NGC) in 1888, its subsequent Index Catalogues (IC), and recent incarnations such as the Revised New General Catalogue (RNGC) by Jack W Sulentic and William G. Tifft in 1973, and NGC 2000.0 by Roger W. Sinnott in 1988. Furthermore, an ultraprecise “center” is superfluous for logging the position of a diffuse nebula or open star cluster. As amateur astronomers probe deeper into the night, they need maps showing not only the stars in the right place, but deep-sky objects as well. One accurate catalog for extragalactic objects is the Principal Galaxy Catalogue (PGC), produced by Lyon Observatory in France. This database contains positions for 73,000 galaxies. Other teams of amateur and professional astronomers have worked to correct positions in the NGC and IC Among them are Larry Mitchell and Emil Bonanno, who wanted to improve the accuracy of E.L.B. Software’s MegaStar Star Atlas. Version 3.0 of this program incorporates position and orientation-angle changes to nearly 5,000 objects after Mitchell painstakingly compared MegaStar plots to POSS charts.
Besides the precision problem, the NGC and IC have had others. Throughout the century, the catalogs have been gradually refined as misidentifications and misplots have been fixed. For example, NGC 2000.0 incorporated hundreds of corrections to the previous versions by astronomers Brent A. Archinal and Brian A. Skiff.
Perhaps what will be the definitive augmentation of the NGC and IC is an ongoing venture accessible via the Internet. The goal of the NGC/IC Project is to identify all of the original objects in Dreyer’s compilations and assemble images and characteristics of each into one database. Major contributors to the project include Skiff, Ronald J. Buta, Steve Coe, Harold G. Corwin Jr., Robert Erdmann, and Steve Gottlieb.
The project has no deadline, so it will be continually updated as needed. Professional and amateur astronomers are welcome to provide whatever assistance they can. The project has a site on the World Wide Web (http://www.ngcicproject.org; editor’s note this URL has been updated) that offers details about the endeavor, some “puzzles” that Web surfers are encouraged to solve, and soon the database itself.
Amateur astronomers need to keep the capriciousness of catalogs in mind when preparing or analyzing observations. If you find something unusual, your first response should be to question your reference instead of concluding the sky has changed.
This article first appeared in the April 1997 issue of Sky & Telescope Magazine and was reprinted here with permission from the Editors.
