Preparing for Mira spectra in Gaia DR4: Improving the GSP-Spec pipeline in the cold giants domain

The Gaia mission provides powerful astrometric precision, spectro-photometric observations in the visible, and medium resolution (R~11500) spectroscopy (by the Radial Velocity Spectrograph, RVS) around the infrared calcium triplet. The RVS spectra allows for the derivation of stellar radial velocities and, through the analysis of the GSP-Spec module, the estimation of effective temperatures (Teff), surface gravities (log g), and chemical abundances ([M/H], [α/Fe], and [X/Fe] for up to 15 chemical elements). The Gaia DR3 GSP-Spec all-sky catalogue provided these parameters for ~5.6 millions stars, using their mean RVS spectra. Furthermore, the Gaia DR4 will include all epoch RVS spectra for these stars (and more).Access to these parameters for a large sample of Mira pulsating stars, as bright intermediate-age standard candles, would allow for a detailed chemical and dynamical study of the Galaxy's intermediate-age stellar populations far beyond the solar neighborhood, in addition to studies of the variability of Mira stars through their epoch spectra. However, due to the cold temperatures (Teff ≲ 3500 K) and weak surface gravities (log g ≲ 0.5 dex) in these stars' atmospheres, their spectra show no continuum and are difficult to model. Thus, reliably parametrizing these stars remains a challenge.We are working on improving GSP-Spec's performance for the cold giants domain, with the goal of reliably characterizing RVS epoch spectra of Mira variables, opening new paths for a large scale chemical cartography of the Milky Way and its satellites. In this talk, the first results of this analysis with public RVS mean spectra of M-type giants (and reference Teff, log g, and [M/H] values from the literature) will be presented.