Measuring 14 elemental abundances with R=1,800 LAMOST spectra

The LAMOST survey has acquired low-resolution spectra (R=1,800) for 5 million stars across the Milky Way, far more than any current stellar survey at a corresponding or higher spectral resolution. It is often assumed that only very few elemental abundances can be measured from such low-resolution spectra, limiting their utility for Galactic archaeology studies. However, Ting et al. (2017) used ab initio models to argue that low-resolution spectra should enable precision measurements of many elemental abundances, at least in theory. Here we verify this claim in practice by measuring the relative abundances of 14 elements from LAMOST spectra with a precision of $\lesssim$ 0.1 dex for objects with ${\rm S/N}_{\rm LAMOST}$ > 30 (per pixel). We employ a spectral modeling method in which a data-driven model is combined with priors that the model gradient spectra should resemble ab initio spectral models. This approach assures that the data-driven abundance determinations draw on physically sensible features in the spectrum in their predictions and do not just exploit astrophysical correlations among abundances. Our analysis is constrained to the number of elemental abundances measured in the APOGEE survey, which is the source of the training labels. Obtaining high quality/resolution spectra for a subset of LAMOST stars to measure more elemental abundances as training labels and then applying this method to the full LAMOST catalog will provide a sample with more than 20 elemental abundances that is an order of magnitude larger than current high-resolution surveys, substantially increasing the sample size for Galactic archaeology.