The following document lists the file deficent.txt
from catalogue V/102.
A plain copy of the file
(without headers/trailers) may be downloaded.
The following is a brief explanation of current problems in and shortcomings
of SKY2000, version 2.0. The user should recognize that these problems exist
and should use the data accordingly. We are working on these things and plan
to eventually have all the major problems solved, although we realize that the
catalog will never be "perfect". We group the following material by topic;
i.e., by type of information, which is essentially the order of the format.
1. Identifications
a. HD numbers are not complete for stars of the second extension
(HD272151 - HD359083)
b. Identification discrepancies involving BD supplemental stars
(stars with lower case "a" and "b" designations) may still be
present. A thorough review of all supplemental stars is planned.
c. Component identifications are not complete for multiple systems. Since
the WDS does not assign components for two-star systems, we need to do
this independently. (The WDS contains pairs of stars in a record, so it
is not necessary to assign component identifiers unless there are more
than two stars in a system.) We must also check for consistency in the
way that we have assigned components.
d. A further complication for the assignment of components stems from the
fact that we have only one components field when it would be preferable
to have separate fields for astrometric and photometric data. Thus, in
cases where positions and proper motions represent, say, only the
primary, while the photometry represents a blend, we must choose to put
either A or AB into the components identifier field following the WDS
number. This is usually chosen according to the astrometry, but we have
not been entirely consistent.
e. There are stars having WDS data for double and multiple stars, but with
no WDS identifiers. These are being isolated, analyzed, and corrected at
the present time.
f. Variable-star names are not complete. We do not know why, but some
problem in a past version of the catalog, probably in the software,
caused some of the information in the GCVS to be missed. We have
discovered this to be mainly the names of variables.
2. Astrometric Data
a. The Tycho Reference Catalogue (TRC) has not yet been incorporated into
SKY2000. Since the Tycho subsets of ACT and the TRC are somewhat
different, there are quite a few thousand stars for which we now
include Tycho data that can be replaced with the considerably better
TRC positions and (especially) proper motions. Users who need the
highest accuracy proper motions and find Tycho data in SKY2000 should
check the TRC and use TRC data when they exist. A few TRC data (source
80) have been used in Version 2.0 during problem solving, but full
processing is planned.
b. Parallaxes from HIP are as good as or better than the best ground-based
data (Gatewood 1998, AJ, 116, 1501), but the short baseline and lower
accuracy methods of Tycho make its results much less reliable. Where
Tycho parallaxes exist in the Yale Parallax Catalogue (van Altena, Lee,
and Hoffleit 1995), we suggest using the latter.
3. Photometric Data
a. Magnitude and color data in the photoelectric fields from sources 94 and
95 are problematical and should not be trusted. They were expected to be
replaced with source 20 data and were not, which indicates some kind of
problem such as misidentification, discrepant data, etc. We intend to
look at these cases and replace or confirm the quality of the data, then
change the source(s).
b. The Bright Star Catalogue (BSC) was a high-priority source for certain
data in SKY2000, but was not intended to be a higher priority source
than 20 (Mermilliod's UBV catalog). Unfortunately, the BSC data were put
into SKY2000 with assigned errors (there are no errors in the BSC), which
is not considered acceptable for data evaluation. We intend to extract
UBV data from source 20 for the BSC stars, but as an interim step, we
have replaced the BSC data with Tycho BV photometry. Since the latter
have a rather short time scale, the errors are often artificially low.
We also lost U-B data with Tycho, so we intend to process source 20
again to retrieve high-quality ground-based data for the bright stars.
c. Another problem with the Tycho photometry is that blended flags ("b")
have not been assigned because that information is not available in TYC.
We are in the process of adding flags to SKY2000 data when it can be
determined that Tycho observed more than one component, but it is not
always easy to identify those cases because we can only compare TYC V
data with visual magnitudes for the components given in the WDS.
d. Photometric data from source 27 (the Russian WBVR catalog) were assigned
errors because there were no errors given in the Northern Hemisphere
Catalog of Red Magnitudes (Warren 1994). However, the original WBVR
catalog (Kornilov et al. 1991) does have error estimates based on
numbers of observations that can be used in SKY2000. For source 27 BVR
data, users are encouraged to look at the original catalog for error
estimates.
e. We intend to replace all RI data on the Johnson system with Cousins
system data, since the latter is a superior system. Users should
substitute Cousins data for the KY2000 Johnson data when they exist.
f. Large numbers of ptv and ptg data are missing from SKY2000. These are
useful for consistency checks for photoelectric data, even though they
are often blends for close companion stars. We intend to reprocess some
of the original source catalogs (e.g., SAO) to put these data back into
SKY2000 (they were removed earlier when a few errors in the original SAO
were found, even though those had been corrected in source 29).
4. Spectroscopic Data
a. Many stars that currently have MK types do not presently have their types
in the catalog. This is because many of the types that were in the old
SKYMAP catalog have been removed because of unreliability. The uncertain
information came from not knowing for sure if the MK types had been
"derived" from one-dimensional spectral types and color data. The vulgar
practice of deriving MK types has been discontinued and we want no such
types in the SKY2000 catalog. Although most of the types from SKYMAP
were legitimate, the original sources had been lost anyway, so we plan to
replace all of those data with newly ingested MK data.
b. MK types with sources in the nineties should be used with great caution
because the original sources are not identified in SKY2000. This is not
what was intended when SKYMAP spectral types were transferred to the new
catalog. Unfortunately, the SKYMAP source replaced the originals when
the MK types were transferred, so we must return to the original sources
to re-retrieve the types. Perhaps this is all for the best because many
more types are now available and errors in the original sources have
sometimes been corrected. It also allows us to consider all the available
types for each star, although choosing among those available is often too
daunting a task to perform on large numbers of objects. A helpful source
is the Selected MK types catalog of M. Jaschek (1978), but the age and
frequent uncertainties of the types chosen are often a problem.
c. The MK types for carbon stars are not all on a consistent classification
system and must be checked for accuracy. It has even been found that
some MK types for C stars are completely missing from SKY2000, even
though they sometimes appear in the MK catalogs that were used as the
original sources for SKYMAP.
5. Radial Velocities
The only radial-velocity data in SKY2000 had originally come from Wilson
(1953), with a few data from the BSC. The Hipparcos mission added a lot of
new data not in the catalog. When we created the equinox and epoch
J2000.0 version of the HIP for SKY2000 usage, we transferred RV data
from the Hipparcos Input Catalogue (HIC) to our version to make it
easier to ingest new RV data into SKY2000. When the HIP J2000.0 was
subsequently processed, the radial velocities went into SKY2000 as coming
from source 66 (HIP) instead of 33 (HIC). This has already been fixed in
our latest proprietary version, but is still present in version 2.0.
New RV catalogs have been published by Barbier-Brossat and Petit. These
will be processed and ingested into SKY2000. In the meantime, users
should consult other catalogs if there is no RV in SKY2000.
6. Double and Multiple Stars
a. As mentioned under item 1, component identifications are far from
complete in Version 2.0 of SKY2000, and there are double stars in the WDS
whose identifiers are missing from the WDS field.
b. Orbital data for visual binaries having published orbits are not yet
included in SKY2000. Actually, we have room only for periods of orbital
pairs at the present time. If we wish to predict what a spacecraft star
tracker will see at any epoch, ephemerides must be computed. This
requires orbital elements, which we may include in a future version of
SKY2000. We intend, however, to include period data before that so that
users can at least have a handle on the motions of orbital pairs.
c. A thorough analysis of bright, orbiting pairs should probably be
undertaken because these can cause significant problems for star
trackers when they are incorrect. Although we have found and corrected
many of the problems, without a complete review (mostly by machine), we
can be quite sure that some problems remain. We do know of
inconsistencies for close pairs that are both in the main HIP and in the
components catalogs. These are subarcsecond pairs that have blended
photometry and the position of only one component from HIPc (because
that was the highest priority source). Thus, the photometric data are
inconsistent with the astrometric information. For those cases, the
blended position from HIP (coded * in duplicity column) should have been
used. An example of this situation is HR 266 (SKY2000
J005646.97+602146.0), which we have treated manually to use the blended
position and motions.
7. Variable Stars
a. As discussed earlier, variable-star names are missing for many GCVS
stars. This situation will be corrected by reprocessing the latest
edition of the GCVS.
b. Some data from the NSV have been put into SKY2000 for the wrong stars,
probably as a result of some software bug in the distant past. This is
not the case for a large percentage of stars, but some have been
detected.
c. As a result of the GCVS and NSV catalogs being higher priority sources
than the UBV catalog (source 20) and because the former were processed
later, V magnitudes from the GCVS and NSV replaced those given in
source 20. For stars that are variable, this procedure is invalid
because the magnitudes and colors come from different sources and were
observed at different times (epochs). To fix this problem, V magnitudes
from the same sources as the colors should be put back into SKY2000 and
then the U and B magnitudes recomputed.
© UDS/CNRS