6.1.1 Data products in Gaia DR3

Table 6.1 lists the products of the RVS pipeline published in Gaia DR3. The links on the archive name redirect to the archive brief description of each product.

In addition to the radial velocity, the measurement of three other physical characteristics of the stars are provided: the median broadening velocity (in most cases dominated by the projected rotational velocity), the mean spectrum and the median magnitude $G_{\mathrm{RVS}}$. Several other parameters are provided containing information on the measurements of these physical characteristics.

The main data products of the RVS data processing are the radial velocities (radial_velocity) combined over the transits (epochs), which provide the third dimension, the line-of-sight component of the velocity vector that completes the two-dimensional proper motion information (pmra, pmdec).

The stellar radial velocities are obtained by combining the transit data information. For each valid RVS spectrum of each star that enters the pipeline (only the stars brighter than about $G_{\mathrm{RVS}}=14$ have the radial velocity estimated in Gaia DR3, see Section 6.2.1) the transit radial velocity is computed through a fit of the RVS spectrum relative to an appropriate synthetic template spectrum (Section 6.4.8). The atmospheric parameters of the template used are stored in the columns rv_template_teff, rv_template_logg and rv_template_fe_h, and their origin in rv_atm_param_origin.

The combined radial velocity is obtained from the transit data in two different ways depending on the star brightness. For the bright stars (grvs_mag $\le 12$ mag), the method used is the one already used in Gaia DR2: the combined radial velocity is the median of the single-transit radial velocities, corrected by the Gaia velocity with respect to the solar-system barycentre. For the faint stars, for which the single-transit radial velocities are less precise, the single-transit cross-correlation functions are shifted to the solar system barycentric frame and averaged. The mean cross-correlation function obtained in this way is used to measure the star radial velocity. The information on the method used is stored in rv_method_used.

The number of single transit (epoch) observations used to obtain the combined radial velocity is stored in rv_nb_transits, the time covered by these observations in rv_time_duration. For the bright stars, for which the measurement of the single-transit radial velocity is assumed to be precise enough, there is also some information on their transit radial velocities: the total amplitude (max-min), rv_amplitude_robust and the $p$-value for constancy rv_chisq_pvalue are measured in the time series after outliers removal. For the bright stars, the rv_renormalised_gof is also provided; it contains information on the scatter of the source epoch radial velocities compared to the typical epoch uncertainty for similar stars. The combination of the last two parameters may be used to identify potential variable stars.

Another information relative to radial_velocity is the signal to noise ratio (SNR) characterising the combination of the spectra used to obtain the radial velocity, rv_expected_sig_to_noise.

Among the transits used to compute the combined radial velocity, there are also some (rv_nb_deblended_transits) that have undergone deblending. The deblending of the spectra overlapping with spectra from other windows is a new functionality of the DR3 RVS pipeline and is described in detail in Seabroke et al. (2022). While some of the deblended spectra are used to obtain the combined radial velocity and the mean spectra, only the clean non-blended transits are used for the grvs_mag and the vbroad measurements.

vbroad is the line broadening parameter. It is obtained for about 3.5 million bright stars. It is called broadening velocity because, in addition to the projected rotational velocity ($v\sin i$), it includes also other physical effects like macro-turbulence, or residual instrumental effects (LSF model uncertainty) or template mismatch. The source’s broadening velocity is the median of the single-transit broadening velocities, obtained after having excluded the deblended spectra and the spectra observed during the initial time of the mission, when the contamination from water ice was important and resulted in a broadening of the LSF. The associated uncertainty, vbroad_error, is the standard deviation of the epoch measurements, and the number of epoch measurements is stored in vbroad_nb_transits.

grvs_mag is the median of the single-transit $G_{\mathrm{RVS}}$ measurements, and the associated uncertainty, grvs_mag_error, is the error on the median to which a constant of 0.004 mag has been added, in quadrature, to take into account a calibration floor contribution. grvs_mag_nb_transits is the number of transits used in the median, they are all non-blended.

There are about 1 million stars that have their mean RVS spectrum published; they can be identified using the column has_rvs. Their spectrum is available in the table rvs_mean_spectrum. The CCD RVS spectra are extracted, cleaned, deblended (if needed), wavelength calibrated and normalised either to their pseudo-continuum or by scaling with a constant (the latter for cool stars or noisy spectra). The spectra are then shifted to the rest frame, using the epoch radial velocities for the bright stars (in general grvs_mag $\le 12$ mag), or using the combined radial velocity for the faint stars. They are interpolated into a common wavelength array spanning 846–870 nm with a step of 0.01 nm and averaged, taking into account the bad samples (see Section 6.4.9). The information on the SNR estimated in the mean spectra is stored in the column rvs_spec_sig_to_noise.

The transit radial velocities for the Cepheid and RR Lyrae stars are published in the vari_epoch_radial_velocity table. The stars having their epoch radial velocity published can be identified using the column has_epoch_rv. The analysis of the Cepheid and RR Lyrae data is described in Clementini et al. (2023).