Astrometry of Landolt Catalog
photometric catalog has a well-known poor astrometry (there are errors
of tens of arcsec for some stars).
How can we get a better astrometry?
This is the method I used to improve the astrometric accuracy of the Landolt
catalog. The first step was to get a 5" accuracy astrometry, the next one
was to match with the USNO catalog.
I used theAladin
tool provided by the Centre de
Données astronomique de Strasbourg, in order to get a digitized
image of (almost) any part of the sky, with RA and DEC coordinates. For
the Landolt fields, I used the
(DSS) of the STScI. Two generations are available, DSS1
and DSS2 with a pixel size of about 1.7" for DSS1 and about 1" for DSS2.
These images gave me the coordinates of each Landolt star, based on the
charts provided by Landolt.
A first step was to localize roughtly each star after its finding chart.
The second one was to try to get its coordinates as accurately as possible.
I used DSS2 (smallest pixel size) to do that. Afterward I noticed that
astrometry of DSS2 is systematically off for a few fields (as for SA110
At this level the accuracy in Landolt coordinates is less than 5".
Finally I used the USNO-A2.0
astrometric catalog to refine the position of each star.
Both my "new" Landolt catalog and the USNO have been matched:
490 Landolt stars (on 526) had one USNO counterpart.
I checked by eye, superimposing the DSS image and the USNO catalog,
the 18 stars with the greatest association distance, ie
- DECUSNO)2]> 4"
8 Landolt stars had two counterparts. I checked them by eye in order to
remove the wrong one.
28 Landolt stars don't have a counterpart, of which:
14 are high proper motion stars:
All of them have their motion properties listed in table 3 of Landolt
(1992), but two stars:
As the plate survey was spread on a quite long period (1979-1988
for DSS1, 1984-1998 for DSS2) it's not possible to quantify the proper
motion just using this.
SA 101 268 is also a high proper star according to CDS's
SA 105 505 is given to have only a motion of 32 mas/yr. I found a ~ 6"
shift between DSS1 and DSS2! It could be interesting to note that this
star is classified as "Variable
The coordinates of these stars are those of the position they had on
DSS2 images, which is not very helpful, I agree...
14 have no USNO counterpart:
In total 498 (on 526) Landolt stars have now USNO coordinates, which means
an accuracy of 0.25". For the remaining 28, I put in the "new" catalog
the coordinates I got by eye (with an accuracy typically < 5").
The astrometrically improved Landolt catalog
To download this "new" Landolt catalog which has exactly the same
shape of the original one. Coordinates are for J2000.0:
First use: quick study of the USNO catalog photometry
We can examine the photometry of the USNO catalog.
Back on the main page.
Guillaume Blanc - august 2001