This material originally posted to sci.astro.amateur, and MPML.
This past January I posted a long (but I hope informative) message about current and upcoming comprehensive star catalogues. Since several things in that post have now gone from future-tense to present-tense, here is a revised version. In particular, the Hipparcos and Tycho data are now publicly avialable, and Web links are provided below.
Under this heading are the two venerable equinox B1950.0 catalogues: the AGK3 and the SAO. The AGK3 has about 200,000 stars north of -2 Dec, containing positions, proper motions, and photo-blue magnitudes complete to about blue magnitude 8.5, but with many stars down to mag. 12 or so. This used to be the catalogue of choice for the northern hemisphere, since it had twice as many stars as the SAO catalogue, and the positions were more accuracte. At present, the positions in the AGK3 are good to about 0".75.
The SAO catalogue was compiled in the 1960s from a large number of sources, reduced as well as possible (at that time) to a common system. The catalogue contains about 260,000 stars for the whole sky (less dense than the AGK3 in the north). It was intended for satellite tracking (from the ground) and also for attitude-control on orbiting spacecraft. Despite having these origins as an engineering tool, the compilers tried to make it useful for astronomy as well. At present the positions in the SAO have rather large errors, typically around 1".2 in the north, and progressively worse south of about -20 Dec, with many cases of errors up to 10" in the far-southern sky. The reason for this is the sheer age of the positions that went into the catalogue, which have a mean ending epoch around 1940. As a result, when the proper motions derived from the older data are extrapolated to the present, substantial errors creep in, even if the original positions are good.
For most purposes there are again two "catalogues of precision" in use for doing astrometry, making charts, etc. One is the ACRS ("Astrographic Catalog Reference Stars"), which was produced about 1990 by the U. S. Naval Observatory. It is an all-sky catalogue of about 380,000 stars, compiled as was the SAO from a large number of older sources. In this case, the analysis of the input data was much better than before, and more stars could be included as well from modern observations. The current accuracy of the positions is about 0".3, a factor of four better than the SAO; the motions are factors of six to ten more accurate.
The other catalogue is the PPM ("Positions and Proper Motions"), which was compiled in a similar way by the Astronomisches Rechen-Institut in Heidelberg. This catalogue contains altogether about about 470,000 stars with mean errors similar to the ACRS. Since many people get confused: in the northern sky, the PPM lists the photo-blue magnitudes directly from the AGK3. These magnitudes will thus be about 0.5 to 1.0 (typically) fainter than visual magnitudes. In the south, the magnitudes are a mix of photo-visual and photo- blue magnitudes, and even some real visual estimates surviving from hundred- year-old catalogues such as the Cordoba Durchmusterung. There is a code in the PPM-South that tells the source of the magnitude for each star. In all cases, the magnitudes scatter from "truth" by around +/- 0.3-0.5 mag. For most of the naked-eye stars, however, no matter where in the sky they are, accurate photoelectric V magnitudes are given (to 0.1 mag. precision).
Far more comprehensive, but of lower accuracy, is the Space Telescope Guide Star Catalogue (GSC). The GSC version 1.1, which now appears in many software packages intended for the amateur market, including those produced by contributors to this newsgroup. It contains about 16,000,000 'things' from digitally-scanned Schmidt plates. Most of these are stars and galaxies, but also emulsion flaws, asteroids, parts of galaxies, pieces of diffraction spikes emanating from bright stars; there are also bits of dust, dandruff flakes, and whatnot included in the GSC. It is also an engineering product (to point and track Space Telescope), not a true astronomical catalogue, no matter that both professionals and amateurs are using as if it were the latter. For the star positions, the GSC v1.1 used as the reference net the AGK3 in the north and the SAO in the south, there being no other reference catalogues available when the plates were scanned. On top of this problem, the equations used to convert the x,y pixel coordinates of the scans into RA and Dec did not adequately model the distortions of the Schmidt plates. As a result, in the corners of the plates, the star positions are frequently off by 2"-3", plus having the general errors caused by the old star catalogues. Mean errors, however, are about 0".8 at present.
It is worth re-emphasizing that the GSC is _not_ an inventory of the sky, only a list of stars for controlling Space Telescope. Thus there are plenty of fairly bright stars missing from the catalogue that will screw you up if you compare charts made from them with views at the eyepiece. Along the Milky Way, the catalogue starts to become incomplete around mag. 11.5, even though there are a lot of mag. 13 and 14 stars given in the same area. And there are certainly plenty of mag. 10 stars missing from the GSC.
This is a good place to mention the magnitudes in the GSC: they suck! In the north (specifically north of about +3 Dec), the GSC derives from short- exposure yellow-light plates taken in a special survey at Palomar. South of +3 Dec, the deep blue-light survey plates from the UK Schmidt in Australia were used. Along the southern Milky Way, in order to avoid severe crowding problems, a series of V-band plates was taken with the UK Schmidt. Special plates were taken for such areas as the Magellanic Clouds, M31, etc. The magnitudes were calibrated with sequences of about six stars near the centers of each field. In the immediate region of the those sequences, say within a circle 1 degree across, the GSC magnitudes are often pretty good. Outside that area, all bets are off, and the GSC magnitudes scatter over a range of two to three whole magnitudes. Sometimes they're okay, but usually they're off by half a magnitude or more. And since they're based on only a single plate, the internal errors are no better than +/- 0.3-0.4 mag. at best. In the north, of course, the magnitudes are in the yellow, but in the south they're blue-light magnitudes (i.e. fainter than visual for most stars). However, except for the southern near-Milky Way regions, _none_ of the magnitudes are on the standard V ("visual") system. In the north, they're redward of visual, so red stars come out too bright; in the south, they're blue, so red stars come out too faint. The bottom line is not to use the GSC magnitudes for anything. No, not anything.
What the Naval Observatory has done is to prepare a new star catalogue combining the old AC x,y positions (reduced to RA/Dec) and new positions measured on plates taken with a 20cm Twin Astrograph in Washington DC. This catalogue obviously doesn't cover the whole sky (yet), but reaches to -18 Dec. In this area are some 750,000 stars, nominally complete to photo-blue mag. 10.5, with a lot of stars included that are up to one magnitude fainter. This is triple the areal density of the ACRS or the PPM, and further, the accuracy of the positions is three times better, approaching 0".1! This is limited now mainly by the fundamental reference frame of positions on the sky. You can download the "Twin Astrograph Catalogue" (TAC) in one-degree strips directly from the USNO at:
http://aries.usno.navy.mil/ad/tac.html
It's three times larger than the PPM, so have plenty of disc space ready! Be sure to read the introductory text at the Web site, since there are some caveats on the quality and completeness of the data for certain zones. This Web page also has a link to the original AC re-reductions, too. There's some interesting historical reading provided about the origins of the AC.
The TAC doesn't include all the stars contained in the AC re-reductions, but new positions at similar precision for these fainter stars can be obtained from another up-to-the-minute catalogue: GSC, version 1.2. This is a re-reduction of the GSC using this time the PPM star catalogue as the reference net, and also taking into account very carefully the distortions of the Schmidt plates. The GSC v1.2 is not available yet 'in toto', but can be explored on a star-by-star basis at the Web site for the "Catalogues and Surveys Branch" (CASB) at Space Telescope:
http://www-gsss.stsci.edu/gsc/gsc12/gsc12_form.html
The claim is that the GSC v1.2 is reliable at the 0".3 level, and this appears to be the case for several examples tested by asteroid occultation predictions. Thus the data are five to ten times better than v1.1! The main CASB Web area homepage is at:
...which contains other interesting stuff.
Even the GSC is small change nowadays, thanks to complete deep scans of the original sky survey plates in both hemispheres in two colors. How deep? Magnitude 20. How big? How does 500,000,000 stars sound?! That's something like one-half percent of _all_ the stars in the Milky Way galaxy! Compared to the GSC, this is like a hippopotamus squatting on a pocket Bible. The ten degree square region centered on the Large Magellanic Cloud contains about 15,000,000 stars---as many stars as are in the entire GSC. The sky survey that's readily available now has been produced by the U. S. Naval Observatory's Flagstaff station. The products from the "Proper Motion Machine" (PMM, not PPM!) are described at another USNO Web site:
This page includes links to three on-line search engines (CDS, ESO, Lowell) as well as third-party software for reading the data.
There are two main catalogues. "SA1.0" is a selected list of a mere 54 million stars between mag. 16 and 19 uniformly distributed around the sky. It is intended to be used as an astrometric reference net for large telescopes and narrow-field instruments, such as for amateurs doing asteroid and comet astrometry with commercial CCDs. Often there are only a few GSC stars in a field, not really enough to get a good asteroid position. This catalogue will allow even small chips to get enough reference stars to do such measurements. Obviously, this is not a catalogue to use to make star charts, since the stars are chosen by their distribution, not for an inventory of the sky. If you do astrometry, head to the Web page above to read about getting a copy of the CD this catalogue is on, although it is now being included in astrometric reduction software packages.
The catalogue everyone should be excited about is "A1.0". This is a real inventory of the Schmidt sky survey plates, done from the original plates (not copies), and reaching even fainter than you can see on the POSS prints, comprising 488,000,000 'detections', with some star/nonstar discrimination. Using the catalogue we have on-line at Lowell, I have extracted positions for mag. 19 variable stars in the thickest parts of the Aquila and Cygnus Milky Way. This comes on 10 CDs (6 gigabytes), and will not be generally available (at least for now) except to professionals, partly because USNO-Flagstaff is not in the business of spending their time cutting CD-ROMs to sell, but mainly because of the complicated legal entanglements involved.
Because a lot of people have their fingers in the pie, it is not clear how these can be used in commercial products. Besides the Navy, there's also the National Geographic (who's paid for both surveys in the north), the European Southern Observatory and the Anglo-Australian Observatory (for the south), SERC/PPARC, Cal Tech, and Space Telescope Science Institute, all of whom have claims to various intellectual property rights and copyrights on the source material. As the project maestro Dave Monet has said, "don't make me wake up the lawyers". However, because the material can be freely searched on-line, I don't see that this is much of a hindrance for ordinary amateur visual/CCD observing.
If you talk to anybody making star catalogues now, they'll all tell you that everything is going to be swept aside by Hipparcos. The Hipparcos spacecraft operated in the early 1990s to obtain parallaxes, positions, plus B and V magnitudes for stars. The results have been under tight wraps, but have now been released. The parallax part of the mission (Hipparcos) produced parallaxes good to about 1 milliarcsecond (one thousandth of an arcsecond) and high-precision proper motions for about 100,000 bright stars. Another instrument on the same spacecraft, called Tycho, has produced positions of lesser (but still high) precision plus B and V magnitudes for one million stars---complete to mag. 10.5, and lots of stars to 11.5 (not quite as complete as the TAC).
You can do star-by-star searches using the CDS-Strasbourg "VizieR" look-up facility:
http://vizier.u-strasbg.fr/cgi-bin/VizieR
This page produces a catalogue "order form". In item 1 here, type in the catalogue number, which is I/239. Submit this, then on the succeeding page select either the Hipparcos or Tycho portion of the catalogue, and continue. This brings you to the search page, which allows searches by star name or by position, and selection of what data you would like to see.
The entire dataset is also available via ftp from the CDS:
http://cdsweb.u-strasbg.fr/cgi-bin/Cat?I/239
The files come to 427Mb, so the VizieR service is likely to be more useful for most folks.
As a result of using the Hipparcos/Tycho data in the last two weeks, I can offer the following advice. First, note especially that the positions supplied are for epoch 1991.25 and equinox 2000. Because the time baseline of the positions was only a few years, the proper motions derived from the data are rather poor compared to up-to-date astrometric catalogues mentioned above. Thus I would avoid using the proper motions given particularly by Tycho. For most purposes it is also better to use positions for both epoch and equinox 2000, such a those in the PPM.
The Hipparcos parallaxes are the best available, but bear in mind that most stars are far enough away that the errors of these observations are a substantial fraction of the parallax itself. In other words, for most stars the parallax is not statistically different from zero. For some the data show only that the star is not in the room with you, and sometimes not even that when the errors are larger than the parallax itself. (There are also a fair number of spurious negative parallaxes.) When using the data, be sure to check the value given for the uncertainty.
The VizieR search results also show various data resulting from the photometry done by the instruments. The corrected V magnitudes shown as one of the earlier columns of data seem to be adjusted very closely to the standard V system. Also given in later columns are the raw data, uncorrected for what appears to be a strong color term in the B and V results. The individual magnitudes here seem to be "wrong", but the B-V colors listed seem to be correct when compared to Landolt equatorial standard stars. Again, be sure to look at the errors listed, as these can be quite large for fainter stars in the Tycho catalogue. At mag. 10, the per-observation error of the magnitudes was about 0.4 magnitudes. Only by having made one hundred or more observations was the error of the mean beat down to an acceptable level. The large per-observation errors however, mean that small-amplitude variables among the fainter stars went unnoticed, "lost in the noise".
Many variable-star hunters have despaired that the Hipparcos results will mop up all the bright variable stars that are not already known. I can confirm, however, that this is simply not so. Yes, there are some 8000 new variables detected, but because the data were taken in a temporally very irregular fashion, not every variable was caught, particularly eclipsing binaries of the Algol type for bright stars, and just about any type fainter than mag. 9.0 or so. Plenty of work still to be done!
More interesting for most amateur observers is that a new large-scale star atlas, the "Millenium Star Atlas", is being produced from the Tycho data by Roger Sinnott and colleagues at Sky Publishing. When it comes out later this year, it will likely be a significant advance on the Uranometria and Herald-Bobroff atlases. (Is it going to be perfect? No.) Advertising for this has already appeared, where further details can be sought.
You can find out more generally about the Hipparcos/Tycho mission and its products at the ESA Web site:
http://astro.estec.esa.nl/SA-general/Projects/Hipparcos/hipparcos.html
The high-precision positions from Hipparcos/Tycho mean that all previous ground-based fundamental reference frames are obsolete. Now that the data are available to mere mortals, _everyone_ will re-reduce their position catalogues using the Hipparcos stars as the reference frame. The GSC will be redone again, A1.0 will become A2.0 (or something), the Twin Astrograph Catalogue will get re-reduced; new star catalogues will be started to extend the high-precision to fainter limits. Already the USNO-Washington has planned a digital sky survey to 15th magnitude that will have the same accuracy as Hipparcos. This is scheduled to start in Chile in 1998.
You read/hear about a lot of amazing stuff coming out of astronomy, but it's mostly about specific objects, or an obscure new discovery. Here are some things that will really change the way we do even amateur astronomy. Watch this space!
Abstract. The USNO CCD Astrograph (UCA) started an astrometric survey in February 1998 at Cerro Tololo, Chile. This first, preliminary catalog (UCAC1) includes data taken up to November 1999 with about 80% of the Southern Hemisphere covered. Observing continues and full sky coverage is expected by mid 2003 after moving the instrument to a Northern Hemisphere site in early 2001. The survey is performed in a single bandpass (579-642 nm), a 2-fold overlap pattern of fields, and with a long and a short exposure on each field. Stars in the magnitude range of 9 to 14 have positional precisions of <= 20 mas. At the limiting magnitude of about R=16 the positional precision is 70 mas. The UCAC aims at a density (stars per square degree) larger than that of the Guide Star Catalog (GSC) with a positional accuracy similar to Tycho. The UCAC program is a major step towards a high precision densification of the optical reference frame in the post-Hipparcos era, and the first stage, the UCAC1 contains over 27 million stars. Preliminary proper motions are included, which were derived from Tycho-2, Hipparcos and ground-based transit circle and photographic surveys for the bright stars (V <= 12.5) and the USNO A2.0 for the fainter stars. The accuracy of the proper motions varies widely, from 1 to over 15 mas/yr. The UCAC1 is available on CD-ROM from the U.S. Naval Observatory.
You can find more information at http://ad.usno.navy.mil/ucac/
A note on the UCAC page, updated 2002 Aug 8, states
UCAC2: will become available not earlier than the end of 2002. Production has been further delayed , we are just short in manpower and want to do a good job. The new catalog will supersede the UCAC1 and will cover declinations from -90 to about +35..+50 degrees. Proper motions will be provided, based on SPM, NPM, AGK2 and other catalogs.
added Sep 14, 2002
Some folks may have an interest in the upcoming USNO-B1.0 star catalogue. This was described this afternoon (13 Sep) in a presentation by Dave Monet at USNO-Flagstaff. The catalogue contains just over a billion (1e10^9) stars, about twice as many as its predecessor, USNO-A2.0. (This is roughly 1 percent of all the stars in the Galaxy.) Most detections are from the well-known series of five northern and four southern Schmidt photographic surveys in three colors. Some additional astrograph plate series were also scanned. At the bright end, the catalogue is salted with Tycho-2.
Any object detected on any two or more plates was accepted, which will remove various problems that appeared in A2.0 for faint red/blue stars (and quasars etc). There will be a crude star/galaxy separator (actually only star/non-star) that is claimed to be about 90 percent correct, but which is not fully analyzed. Monet suggested that things larger than 30"-40" will be broken up into various pieces by the blob-finder, so large galaxies will be represented by large numbers of entries for various bits and pieces. (This is already the case for USNO-A2.0, so no real change here.)
Astrometric errors of 0".15 are claimed, presumably at some mean epoch of the plates used for any particular entry. It should be possible to verify this using UCAC, the Sloan calibration fields, and the ICRF galaxies directly. Proper motions will have errors of something like 4mas/year. "Blobs" which moved less than 2".5 in the 40-year span of the longest-epoch baseline (i.e. < 0".06 per year) are assumed to have zero motion and will have 0 in the proper-motion data columns. The public product will show positions at epoch and equinox J2000 having a close tie to the Hipparcos-based ICRS.
Photometry is based largely on the Guide Star Photometric Catalogue version 2, which will probably mean a modest improvement over A2.0, but large systematics may be present plate-by-plate (this remains to be tested). The magnitudes will be in the native systems of the plate-series, not on standard systems like Johnson-Cousins BVRI.
Monet sees this catalogue as transitional to B1.1, which he suggested could be completed in another year if things go smoothly. This will cross- match the final 2MASS catalogue (which also must be delivered in the coming month or two), for which there is a 99 percent correlation in USNO-B, but with better astrometry and of course fairly reliable photometry in the near-IR. Delivery depends on funding and close cooperation of USNO and the 2MASS groups, with the latter corporate marriage spoken of in terms of "if".
Delivery. 100Gb discs will be sent to a few selected parties such as the data centers. USNO-Flagstaff will serve images and plotted charts, but will not provide the catalogue in bulk. (The scanned images have in fact been available for some time at: http://www.nofs.navy.mil/data/FchPix/cfra.html) Access to star-lists will be through data-center facilties like the Strasbourg VizieR service. There are no plans for general bulk distribution, so no personal copies. If they really FedEx a hard drive copy to Strasbourg on Monday, as Monet said was the plan, it could be that mere mortals can access the data via VizieR sometime in October.
...
Which catalog is the best one for me?
If you are doing wide-field astrometry a la the surveys, and can deal with the relatively bright stars (and modest star-density) of UCAC, then use it. The precision and accuracy of the _astrometry_ is greatly superior to B1.0. Without being able to look at specific fields in detail, I cannot say anything about the B1.0 _photometry_. UCAC will not cover the whole sky for a few years yet.
If your fields are narrower, and/or you want to try using a big star catalogue to subtract out known objects to find asteroids/comets as part of an image-detection pipeline, then you'll want to explore how useful B1.0 is for your task. You may simply need the greater star-density and fainter limits because your frames are saturated at brighter than 17th or something and the field is only 10'x10' and there simply aren't enough stars---and you've noticed that per-star position errors in A2.0 are getting pretty mushy (in the north) due to the long interval since the plates were taken (now 50 years on average). B1.0 will solve those problems.
As to the plethora of catalogues, I sure don't know about any others besides UCAC and B1.0 that will cover the whole sky at visible wavelengthsm, and are generally publicly available. SDSS will cover only 1/4 of the sky and only at high galactic latitude, and saturates at mag. 15. There are catalogues from Cambridge (APM) and Edinburgh (SuperCOSMOS), but they are difficult to access and don't cover the whole sky, and the APM is no different or better than USNO-A2.0. FAME was cancelled, GAIA and other spacecraft are a decade downstream.
So just two catalogues to worry about: UCAC for brighter than 15th, B1.0 for fainter than 15th.