Mon. Not. R. Astron. Soc. 000, 1–?? () Printed 29 October 2018 (MN LATEX style file v2.2) Local Stellar Kinematics from RAVE Data I. Local Standard of Rest B. Co¸skuno˘glu1⋆, S. Ak1, S. Bilir1, S. Karaali2, E. Yaz1, G. Gilmore3, G. M. Seabroke4,5, O. Bienaym´e6, J. Bland-Hawthorn7, R. Campbell8, K. C. Freeman9, B. Gibson10, E. K. Grebel11, U. Munari12, J. F. Navarro13, Q. A. Parker14,16, A. Siebert6, A. Siviero12,15, M. Steinmetz15, F. G. Watson16, R. F. G. Wyse17, T. Zwitter18,19 1Istanbul University, Science Faculty, Department of Astronomy and Space Sciences, 34119, University-Istanbul, Turkey 2Beykent University, Faculty of Science and Letters, Department of Mathematics and Computer, Beykent, 34398, Istanbul, Turkey 3Institute of Astronomy, Madingley Road, Cambridge, CB3 OHA, UK 4e2v Centre for Electronic Imaging, Planetary and Space Sciences Research Institute, The Open University, Milton Keynes, UK 5Mullard Space Science Laboratory, University College London, Hombury St Mary, Dorking, RH5 6NT, UK 6Observatoire de Strasbourg, Strasbourg, France 7Sydney Institute for Astronomy, School of Physics, University of Sydney, NSW 2006, Australia 8Western Kentucky University, Bowling Green, Kentucky, USA 9RSAA, Australian National University, Canberra, Australia 10University of Central Lancashire, Preston, UK 11Astronomisches Rechen-Institut, Zentrum f¨ur Astronomie der Universit¨at Heidelberg, Heidelberg, Germany 12INAF Astronomical Observatory of Padova, 36012 Asiago (VI), Italy 13Department of Physics and Astronomy, University of Victoria, Victoria, BC, Canada V8P 5C2 14Macquarie University, Sydney, Australia 15Astrophysikalisches Institut Potsdam, Potsdam, Germany 16Australian Astronomical Observatory, Sydney, Australia 17John Hopkins University, Baltimore, Maryland, USA 18University of Ljubljana, Faculty of Mathematics and Physics, Ljubljana, Slovenia 19Center of Excellence SPACE-SI, Ljubljana, Slovenia. ABSTRACT We analyze a sample of 82850 stars from the RAVE survey, with well-determined velocities and stellar parameters, to isolate a sample of 18026 high-probability thin-disc dwarfs within 600 pc of the Sun. We derive space motions for these stars, and deduce the solar space velocity with respect to the Local Standard of Rest. The peculiar solar motion we derive is in excellent agreement in radial U⊙ and vertical W⊙ peculiar motions with other recent determinations. Our derived tangential peculiar velocity, −1 V⊙ agrees with very recent determinations, which favour values near 13 km s , in disagreement with earlier studies. The derived values are not significantly dependent on the comparison sample chosen, or on the method of analysis. The local galaxy arXiv:1011.1188v1 [astro-ph.GA] 4 Nov 2010 seems very well dynamically relaxed, in a near symmetric potential. Key words: Galaxy: kinematics and dynamics, solar neighbourhood, solar motion, local standard of rest, stellar content 1 INTRODUCTION which is defined as the rest frame of a star at the location of the Sun that would be on a circular orbit in the gravitational The kinematics of stars near the Sun provide vital informa- potential of the Galaxy. This implies the determination of tion regarding the structure and the evolution of the Galaxy. the solar peculiar velocity components relative to that stan- Since the observations are relative to the Sun, which is not dard, U⊙, V⊙ and W⊙. These values are also of considerable stationary, the kinematical parameters we obtain need to be intrinsic interest, as their determination as a function of the corrected accordingly to the Local Standard of Rest (LSR), spatial distribution and kinematic properties of the com- parison stellar sample is a test of the symmetry of the local ⋆ E-mail: [email protected] Galactic potential as a function of, for example, velocity dis- c RAS 2 Co¸skuno˘glu et al. persion. One can test if thin disk slowly moving stars are, or ues of V⊙ than that of DB98. Piskunov et al. (2006) de- are not, affected by a Galactic bar, among many other con- termined a value different from that of DB98 by about 7 siderations. Uncertainty in the solar kinematic parameters km s−1 in the V component of the solar motion, when de- can have profound implications for Galactic structural anal- termined with respect to open clusters in the solar neigh- yses (cf. e.g. McMillan & Binney 2010) and may be affected bourhood. Binney (2010) fitted distribution function models systematically in turn by presumptions about larger scale to two sets of velocity distributions and obtained a differ- Galactic stellar populations (cf. e.g. Sch¨onrich et al. 2010). ence of about 6 km s−1 compared to the value of DB98. There have been several very recent determinations of the This higher value of V⊙ has been confirmed by Reid et al. solar peculiar velocity, including some using RAVE (RAdial (2009) and Rygl et al. (2010) in their works related to ra- Velocity Experiment, Steinmetz et al., 2006, see also Section dio frequency astrometry of masers in regions of massive 2) data as we use here (e.g. Pasetto et al. 2010). A summa- star formation (though see McMillan & Binney (2010) for rized list of recent determinations of the space velocity com- a different analysis). Sch¨onrich et al. (2010) applied a par- ponents of the Sun appearing in the literature is given in ticular chemodynamical model of the Galaxy to DB98’s data Table 1, while Francis & Anderson (2009) provide an even and determined a value for the V⊙ solar velocity component longer list of references. Given the astrophysical significance consistent with recent high values. The key factor in the of systematic sample- and method-dependent effects in the Sch¨onrich et al. (2010) analysis, and of others, is the rele- 2 determination, we consider here a further determination of vant range of radial velocity dispersions σR which is used the solar space velocity, using a complementary technique. in the Jeans’ analysis. Sch¨onrich et al. (2010) showed that 2 −1 2 The determination of the solar space motion relative for σR>600 (km s ) the Hipparcos data define a straight 2 −1 2 to a Local Standard of Rest (LSR) is a long-standing chal- line, whereas for σR6400 (km s ) they deviate from such lenge. The robust determination of the systematic motion of a simple fit. If one omits stars with low dispersions, and ex- some stellar population relative to some agreed LSR would trapolates the linear fit for stars with high dispersion to zero, itself be of considerable interest. Hence many studies have the linear fit for stars with high dispersions, one confirms the been carried out since that of Homann (1886) to estimate low value of DB98. Actually, this is the procedure used by these parameters. However, there is still uncertainty about DB98, who also ignored stars with low velocity dispersion the numerical values of the solar velocity components, rela- due to their probable lack of dynamical equilibrium. While tive to local thin disc stars, especially for the V⊙ component, it is true that low velocity dispersion stars are likely to be due to the complexity involved in compensating for the ve- most affected by the effects of dissolving star clusters and locity “lag” of sample stars. The difference between the re- the non-axisymmetric gravitational potential of spiral arms, sults of different researchers originates from different data as one does need to consider carefully the offset character of 2 well as different procedures used. The very different results the relation between Va and σR at low velocity dispersions derived by Dehnen & Binney (1998) and Sch¨onrich et al. to avoid bias. (2010), both of whom who used the Hipparcos data, are Here we use a new sample of stars and a different an illustration of the continuing model-dependence of the methodology to re-estimate the U⊙, V⊙ and W⊙ solar veloc- result. ity components. 1) We use RAVE data extending to larger The radial (U⊙) and vertical (W⊙) components of the distances than the Hipparcos sample. 2) We applied the fol- solar motion, as a function of the relevant stellar popula- lowing constraints to obtain a sample of main sequence stars: tion, can be obtained in principle by direct analysis of the i) we selected stars with surface gravity 4<log g<5, ii) we mean heliocentric velocities of solar neighbourhood stars, omitted stars with (J − H)0<0.05 and (J − H)0>0.55 to subject to appropriate sample selection. As one can see in avoid the blue horizontal branch and possible red giant stars, Table 1, the various determinations are in tolerable agree- iii) we excluded stars with space velocity errors larger than −1 ment. However, the determination of the component in the 25 km s , iv) we separated stars into populations (see Sec- direction of Galactic rotation (V⊙) is complicated by sys- tion 5) and we used the thin disc population which is not tematics: the mean lag, i.e. asymmetric drift Va, with re- contaminated by thick disc/halo as a preferred sample in spect to the LSR which depends on the velocity disper- our work. The last constraint is especially important in es- sion (σ) of the stellar sample in question. An extensive timating V⊙, because it excludes thick disc and halo stars discussion of the radial Jeans’ equation on which analysis which have relatively large asymmetric drift. Thus, it limits of the asymmetric drift is based, including its derivation, the range of V velocity component, which in turn minimizes relevant approximations, and applications, can be found in possible astrophysical large-scale dynamical asymmetry ef- Gilmore, Wyse & Kuijken (1989). At its simplest, a single fects. population asymmetric drift/Jeans’ analysis assumes a lin- The RAVE survey is described in Section 2, the data ear relation between (suitable values of) the asymmetric are presented in Section 3, Sections 4 and 5 are devoted to drift Va of any stellar sample and its squared radial veloc- kinematics and population analysis, respectively.