LRPayne: Stellar parameters and abundances from low-resolution spectra

¹ Dipartimento di Fisica e Astronomia, Universitá di Padova, vicolo dell’Osservatorio 2, 35122 Padova, Italy
² INAF – Osservatorio Astronomico di Padova, vicolo dell’Osservatorio 5, 35122 Padova, Italy

^★ Corresponding author: This email address is being protected from spambots. You need JavaScript enabled to view it. ; This email address is being protected from spambots. You need JavaScript enabled to view it.

Received: 18 July 2025
Accepted: 7 November 2025

Abstract

Aims. This paper introduces LRPayne, a novel algorithm designed for the efficient determination of stellar parameters and chemical abundances from low-resolution optical spectra, with a primary focus on data from large-scale galactic surveys such as WEAVE.

Methods. LRPayne employs a model-driven approach, utilising a fully connected artificial neural network (ANN), trained on a library of 70 000 synthetic stellar spectra generated using iSpec with 1D model atmospheres and the Turbospectrum synthesis code. We trained the network to predict normalized flux given stellar labels (T_eff, log(g), [Fe/H], v_mic, v_max, and v sin i, plus 24 individual elemental abundances). We subsequently derived stellar parameters from observed spectra by finding the best-fit synthetic spectrum from the ANN using a χ² minimisation technique. The method operates on spectra degraded to a resolution of R=5000 covering the 42006900 Å wavelength range.

Results. Internal accuracy tests on synthetic spectra show a median interpolation error of less than 0.13% for 90% of the validation sample. The method accurately recovers most of the input labels from synthetic spectra, even at a signal-to-noise ratio (S/N) of 20, with some expected challenges for elements such as Li, K, and N. Validation on the observed spectra of 25 Gaia FGK benchmark stars and 42 metal-poor stars reveals good agreement with the literature values. For the stellar parameters, the mean differences are 22±87 K for T_eff, 0.19±0.23 dex for log(g), and 0.01±0.17 dex for [Fe/H]. Abundances for elements such as Na, Mg, Si, and most Fe-peak elements (Cr, Ni, V, and Sc) are well-recovered. We note challenges for oxygen, manganese in metal-rich giants, aluminium in metal-poor stars and dwarfs, and deriving log g in hot metal-poor dwarfs, partly due to non-local thermodynamic equilibrium effects and line characteristics.

Conclusions. LRPayne demonstrates the possibility of extracting precise stellar parameters and chemical abundances from a large number of low-resolution spectra. Its strong performance across different kinds of stars makes it well-suited for current and future large surveys. The abundance results from LRPayne will be very useful for studying stellar nucleosynthesis and the chemical evolution of our Galaxy on a large scale.