ASTRODEBATA / Pradosh Barun Das: Stellar parameters and abundances from Gaia RVS spectra using GALAH DR4 and The Cannon

Stellar parameters and abundances from Gaia RVS spectra using GALAH DR4 and The Cannon

Pradosh Barun Das
Macquarie University, Sydney, Australia

The unprecedentedly large stellar datasets from the current cohort of major spectroscopic surveys, ranging in size from millions (e.g., LAMOST) to billions (e.g., Gaia XP) of stars, are transforming our understanding of the Milky Way’s structure, formation, and chemical evolution. However, extracting precise stellar parameters from low- to moderate-resolution spectra remains difficult, due to limitations in spectral coverage, resolution, and signal-to-noise ratios, with high-resolution spectroscopy remaining the gold standard. The Gaia Radial Velocity Spectrometer (RVS) dataset, eventually expected to cover over 200 million stars, offers a unique opportunity to study Galactic structure and stellar populations, provided accurate stellar parameters and abundances can be derived. To address this, we use "The Cannon", a data-driven algorithm, to transfer “stellar labels” (collectively referring to stellar parameters and abundances) from high-resolution GALAH DR4 data to lower-resolution Gaia DR3 RVS spectra. Our model determines effective temperature (Teff), surface gravity (log g), metallicity ([Fe/H]), and abundances of [Ca/Fe], [Si/Fe], [Ni/Fe] and [Ti/Fe] for over 800,000 RVS spectra (Das et al. 2025). We recover a bimodal distribution in [Ti/Fe] vs. [Fe/H], corresponding to the high-α and low-α populations in the Milky Way disc, marking one of the first such demonstrations of this α-element dichotomy in RVS data. Validation against RVS observations of open and globular cluster stars shows a typical precision of ∼0.02–0.05 dex in metallicity. Additionally, we measure neutron-capture elements (Nd, Zr, Ce) for a subset of ~350,000 red giant stars. We develop a Membership Probability Function (MPF) using MCMC, through which we identify a significantly higher number of Gaia-Enceladus-Sausage (GES) probable members compared to previous spectroscopic samples using only chemical abundances (Das et al., in prep). As the Gaia DR4 release next year dramatically expands the RVS dataset, our methodology will enable accurate stellar parameter and individual abundance determinations for a much larger sample, demonstrating the power of data-driven techniques in large-scale stellar spectroscopy.

Za prenašanje predavanj v živo in snemanje uporabljamo Youtube. Naš kanal je https://www.youtube.com/@astroFMF, kjer lahko najdete stare posnetke in povezavo za ogled predavanja v živo.