The AMBRE project: A new synthetic grid of high-resolution FGKM stellar spectra

Context. Large grids of synthetic spectra covering a wide range of stellar parameters are essential tools for different stellar and (extra-) Galactic physical applications. They can be used for the automatic parametrisation of stellar spectra such as that performed by the AMBRE project whose main goal is to determine the stellar atmospheric parameters of a few hundred thousands of archived spectra of four ESO spectrographs.
Aims: To fulfil the needs of AMBRE and similar future projects, we computed a grid of synthetic spectra over the whole optical domain for cool to very cool stars of any luminosity (from dwarfs to supergiants) with metallicities varying from 10^-5 to 10 times the solar metallicity, and considering large variations in the chemical composition of the α-elements.
Methods: In these spectrum computations, we used new generation MARCS model atmospheres and the Turbospectrum code for radiative transfer. We also took into account atomic and molecular linelists that were as complete as possible and adopted, in the spectral synthesis, the same physical assumptions and input data as in the MARCS models. This allowed us to present a grid in which there is a high consistency between the atmosphere models and the synthetic spectra.
Results: We present a new grid of 16 783 high resolution spectra over the wavelength range 3000 Å to 12 000 Å computed at a spectral resolution higher than 150 000. Normalised and absolute flux versions are available over a wide range of stellar atmospheric parameters for stars of FGKM spectral types. The parameters covered are 2500 K ≤ T_eff ≤ 8000 K, -0.5 ≤ log(g) ≤ 5.5 dex, and -5.0 ≤ [M/H] ≤ + 1.0 dex. Moreover, for each combination of these stellar parameters, five different values of the enrichment in α-elements were considered (0.0, ± 0.2 dex and ± 0.4 dex around the standard values). This library is thus relevant to any stellar type and luminosity class, present in old and intermediate-age stellar populations with chemical composition varying from extremely metal-poor to metal-rich. This grid is made publicly available through the POLLUX database (about 50% of the spectra have been included in this database) and in FITS format upon request to the authors.