On the accuracy of close stellar approaches determination

The aim of this paper is to demonstrate the accuracy of our knowledge of close stellar passage distances in the pre-GAIA era.We used the most precise astrometric and kinematic data available at the moment and prepared a list of 40 stars nominally passing (in the past or future) closer than 2 pc from the Sun.We used a full gravitational potential of the Galaxy to calculate the motion of the Sun and a star from their current positions to the proximity epoch. For this calculations we used a numerical integration in rectangular, Galactocentric coordinates. We showed that in many cases the numerical integration of the star motion gives significantly different results than popular rectilinear approximation.We found several new stellar candidates for close visitors in past or in future. We used a covariance matrices of the astrometric data for each star to estimate the accuracy of the obtained proximity distance and epoch. To this aim we used a Monte Carlo method, replaced each star with 10 000 of its clones and studied the distribution of their individual close passages near the Sun. We showed that for contemporary close neighbours the precision is quite good but for more distant stars it strongly depends on the quality of astrometric and kinematic data. Several examples are discussed in detail, among them the case of HIP 14473. For this star we obtained the nominal proximity distance as small as 0.22 pc 3.78 Myr ago. However there exist strong need for more precise astrometry of this star since the proximity point uncertainty is unacceptably large.