Research & Science Home ESA Public Web Site Sci-Tech Portal ESA Intranet     
Astrophysics Missions Planetary Exploration Missions Solar Terrestrial Science Missions Fundamental Physics Missions Science Faculty

The Hipparcos Intermediate Data Search Facility

The Hipparcos Intermediate Astrometric Data are the one-dimensional coordinates (abscissae) on reference great circles obtained, by an independent analysis but using the same basic satellite data, by the FAST and NDAC Data Reduction Consortia. The data may be retrieved (for individual Hipparcos stars) via the Hipparcos Intermediate Data online search facility. Click here to begin the Intermediate Data Search. A second window is then generated containing the search interface.

The Hipparcos Catalogue contains the derived astrometric parameters for each of the stars observed in the Hipparcos programme. These are classified into (apparently) single systems, component solutions, orbital systems, systems with accelerating photocentres, etc.

To preserve the elementary astrometric data from which these solutions have been derived, the published catalogue includes, on CD-ROM No. 5, the "intermediate astrometric data" file, and the associated reference great-circle data. From these files, the individual observations can be inspected and their fit to the published astrometric solutions can be re-examined - possibly in the light of existing ground-based data or future observations. These intermediate data also preserve access to the individual solutions provided by the two data analysis consortia (FAST and NDAC) before merging to form the final published solution. Thus the intermediate astrometric data allow the quality of the model fitting to be assessed, and possibly refined.

A complete description of the associated data files are given in Section 2.8 of Volume 1 of the Hipparcos and Tycho Catalogues, ESA SP-1200.

The on-line routine provided here allows the intermediate astrometric data from either consortium to be displayed, along with the fitted solution which appears in the published catalogue. The user requests the HIP number of the object to be interrogated, and which of the solutions (FAST or NDAC, or both, or neither) is to be displayed. For the more casual user, either solution will generally give the same qualitative information about the fit. Observations which were not retained in the published solution are coded separately (as open circles). Three plots are generated: the main plot displayed shows the star path and the observed and fitted data points; two others (Statistics 1/2) show statistical information about the residuals for this solution.

Description of the Star Path Plot:

The plot shows the path on the sky of the requested Hipparcos Catalogue object, covering the actual span of observations for this object, typically amounting to a period of about three years centred around the adopted mean catalogue epoch of J1991.25. Equal scales are adopted in both RA and dec; this scale allows the observations to fill the window, irrespective of the magnitude of the astrometric parameters: thus the relative size of the error bars will appear larger for objects with small values of parallax and proper motion, and vice versa. In practice, the error bars are mainly dependent on the magnitude of the object.

The information at the top of the screen shows the data extracted from the header record of the intermediate data file. It includes the reference astrometric parameters: RA and dec at J1991.25 in degrees; parallax (par) in milliarcsec, and proper motion in RA and dec (pma and pmd) in miliarcsec/yr. The Hp magnitude is the value adopted for the merging (not necessarily the same as that given in the main catalogue). For associated errors, consult the main catalogue. The total number of data points in the file, nobs, giving the total number of NDAC+FAST observations, and therefore representing typically (or less than) twice the number of independent data points, is also given. The "code" used in the header for the published astrometric solution has the following meaning (see Section 2.8 for further details of the relevance of this parameter for the interpretation of the plots):

5 single star solution (5 astrometric parameters)
7 acceleration solution (DMSA/G), 7 parameters
9 acceleration solution (DMSA/G), 9 parameters
C component solution (DMSA/C)
O orbital solution (DMSA/O)
V `VIM' solution (DMSA/V)
X stochastic solution (DMSA/X)
- no astrometric solution

The individual observations are shown as blue (FAST) or red (NDAC) points (either, both, or neither can be activated: if neither are displayed, the plot shows simply the adopted solution; this facility may be useful in the case of fainter stars with larger residuals, or in the case of objects with small numerical values of the parallax and proper motion, e.g. HIP 1, 5, etc). Observations rejected from the fit (usually none) are shown as open circles. Further details of these data points are given below.

The straight line shows the barycentric motion of the object, from which only the proper motion of the object is apparent, i.e., it shows the (solved) motion of the object as viewed from the solar system barycentre, without the effect of parallax.

The continuous curve is the modelled stellar path fitted to all of the (accepted) measurements, based on the astrometric parameters given in Fields H8-9 and H11-13 (the curve is generated using an approximate Earth ephemeris, resulting in an rms positional error of about 0.0024 AU over the interval 1988.0 to 1993.0). The amplitude of the oscillatory motion gives the star's parallax, with the linear component representing the star's proper motion.

The FAST/NDAC data points are shown as follows. Solid circles indicate the inferred position of the object at each observation epoch. By definition, these fitted points lie on the continuous modelled path. A straight line (the residual of each observation) joins the fitted point to the observed position line of the star at the relevant epoch. Because the measurement is one-dimensional, the precise location along this position line is undetermined by the observation. The position line (always perpendicular to the residual line) is shown with a length of +/-5 milliarcsec, although the true uncertainty in this dimension is considerably larger (in effect it is unconstrained by the observations). Thus for each "T", the vertical stroke represents the direction of the scanning motion at each epoch. The length of this stroke represents the residual of the measurement. The horizontal part of the "T" represents the (truncated) 1-d position measurement line.

Note that the significance of a given residual can only be assessed in terms of the standard error of that measurement. All of these quantities are contained in the intermediate astrometric data file, and form the basis of the statistical plots which can also be accessed. Similarly, it might not be immediately obvious why a particular data point has been rejected, since rejected observations have not been connected to the fitted curve - thus while it may appear to lie very close to the fitted line, its associated observation epoch may be very different from the fitted star path at that epoch.

The moving circle on the straight line indicates the barycentric position of the objects in steps of 0.1 years. It therefore also shows the direction of motion of the star. The moving circle on the fitted path shows the apparent motion of the object due to the additional effect of parallax, i.e. the motion that the star actually has on the sky as seen from the Earth, also in steps of 0.1 years. The two time steps are synchronised: note the discrepancies (sometimes retrograde) according to the interaction between the position of the object on the sky, the parallax, and the proper motion. Interesting (and surprising!) examples of their relative motions are seen throughout the catalogue.

Care should be exercised in interpreting the intermediate astrometric data for component solutions: the information content for these systems being incompletely characterised by the intermediate astrometric data. For resolved systems (roughly those with separation greater than 0.3 arcsec and magnitude difference less than 3 mag) the observed point is a complicated (and non-linear) function of the geometric and photometric characteristics of the system. Moreover, this function is different (but known) for the FAST and NDAC data. For close, unresolved systems the observed point is generally the photocentre, and the intermediate astrometric data are very useful for investigations of such binary or orbital systems.

Get ASCII Data:

This button calls up the intermediate data file from which the observations have been constructed. See Section 2.8 of Volume 1 for further details on interpreting these data, and for access to the associated reference great circle data from which the corresponding observation epochs have been extracted.

As examples of the possible use of such data: (1) consider that a much better proper motion becomes available (for example from VLBI): the parallax solution may be recomputed using such additional constraints. (2) Suppose that there is independent evidence for an orbital companion to a (presumed) single star (either stellar or - at least in principle - planetary): the residuals of the solution can be examined for evidence of such motion, overlooked in the routine analysis of the Hipparcos data. (3) In cases where the main catalogue solution is at variance with independent data, the two solutions (NDAC and FAST) may be investigated to examined whether they are discrepant.

Show Statistics:

This provides statistical indications of the solution quality, in the form of:

  • histogram of the abscissae residuals for the plotted solution (NDAC and FAST, in milliarcsec), from Field IA8
  • histogram of the corresponding standard errors, from Field IA9
  • the distribution of normalised errors (IA8/IA9)
  • the normalised errors as a function of time, from which errors in the adopted proper motion, or effects of undetected companions may be apparent
  • the normalised errors as a function of time, modulo a user specified period (default: 1 year). In principle, it might be imagined that such a plot might reveal an erroneous parallax or orbital motion; however, such plots do not take into account the direction of scanning, which is generally different in each point. The only simple way of incorporating such angular information would appear to be by calculating periodograms similar to that used to place upper limits on the masses of planetary companions by Perryman et al (Astron. Astrophys. 310, L21, 1996).

The first three plots are shown in Statistics 1, the latter two are shown in Statistics 2. In the first three cases, note that points lying beyond the range of the histogram are placed into the outermost bins.

Acknowledgements: The concept underlying this type of presentation of the Hipparcos astrometric data was formulated by Lennart Lindegren (Lund), who also supplied the algorithms (including the coarse Earth ephemeris) for generating the plots. The www access has been written in java by William O'Mullane (ESTEC/SA).

     Copyright © 2015 European Space Agency. All rights reserved.
This page was first created on 16 April, 2007 and was last updated on 17 July, 2007.
Password Synchronization

The RSSD Web Portal provides a single login facility to simplify access to several other applications within this site.

The login facility has detected that you have passwords to one or more of these other systems (Livelink, Mantis, CVS, etc.) which do not match your current LDAP password.

By clicking on [SYNCHRONIZE] you will be able to reset all these passwords and bring them into line with your main portal password with no further effort on your part.

Synchronizing Passwords - Please Wait

Logging In - Please Wait