Radial velocities for the HIPPARCOS-Gaia Hundred-Thousand-Proper-Motion project

Context. The Hundred-Thousand-Proper-Motion (HTPM) project will determine the proper motions of ~113 500 stars using a ~23-year baseline. The proper motions will be based on space-based measurements exclusively, with the Hipparcos data, with epoch 1991.25, as first epoch and with the first intermediate-release Gaia astrometry, with epoch ~2014.5, as second epoch. The expected HTPM proper-motion standard errors are 30-190 μas yr^-1, depending on stellar magnitude.
Aims: Depending on the astrometric characteristics of an object, in particular its distance and velocity, its radial velocity can have a significant impact on the determination of its proper motion. The impact of this perspective acceleration is largest for fast-moving, nearby stars. Our goal is to determine, for each star in the Hipparcos catalogue, the radial-velocity standard error that is required to guarantee a negligible contribution of perspective acceleration to the HTPM proper-motion precision.
Methods: We employ two evaluation criteria, both based on Monte-Carlo simulations, with which we determine which stars need to be spectroscopically (re-)measured. Both criteria take the Hipparcos measurement errors into account. The first criterion, the Gaussian criterion, is applicable to nearby stars. For distant stars, this criterion works but returns overly pessimistic results. We therefore use a second criterion, the robust criterion, which is equivalent to the Gaussian criterion for nearby stars but avoids biases for distant stars and/or objects without literature radial velocity. The robust criterion is hence our prefered choice for all stars, regardless of distance.
Results: For each star in the Hipparcos catalogue, we determine the confidence level with which the available radial velocity and its standard error, taken from the XHIP compilation catalogue, are acceptable. We find that for 97 stars, the radial velocities available in the literature are insufficiently precise for a 68.27% confidence level. If requiring this level to be 95.45%, or even 99.73%, the number of stars increases to 247 or 382, respectively. We also identify 109 stars for which radial velocities are currently unknown yet need to be acquired to meet the 68.27% confidence level. For higher confidence levels (95.45% or 99.73%), the number of such stars increases to 1071 or 6180, respectively.
Conclusions: To satisfy the radial-velocity requirements coming from our study will be a daunting task consuming a significant amount of spectroscopic telescope time. The required radial-velocity measurement precisions vary from source to source. Typically, they are modest, below 25 km s^-1, but they can be as stringent as 0.04 km s^-1 for individual objects like Barnard's star. Fortunately, the follow-up spectroscopy is not time-critical since the HTPM proper motions can be corrected a posteriori once (improved) radial velocities become available.

The results data file is only available in electronic form at the CDS via anonymous ftp to cdsarc.u-strasbg.fr (130.79.128.5) or via http://cdsarc.u-strasbg.fr/viz-bin/qcat?J/A+A/546/A61