LRPayne: Stellar parameters and abundances from low-resolution spectra

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. 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_̊m 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 ̧hi^2 minimisation technique. The method operates on spectra degraded to a resolution of R=5000 covering the 4200-6900 Å wavelength range. 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_ . 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. eff​, 0.19±0.23 dex for log(g), and 0.01±0.17 dex for Fe/H 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.