Determining Stellar Elemental Abundances from DESI Spectra with the Data-Driven Payne

Stellar abundances for a large number of stars are key information for thestudy of Galactic formation history. Large spectroscopic surveys such as DESIand LAMOST take median-to-low resolution (R≲5000) spectra in the fulloptical wavelength range for millions of stars. However, line blending effectin these spectra causes great challenges for the elemental abundancesdetermination. Here we employ the DD-PAYNE, a data-driven method regularised bydifferential spectra from stellar physical models, to the DESI EDR spectra forstellar abundance determination. Our implementation delivers 15 labels,including effective temperature T_ eff, surface gravity log g,microturbulence velocity v_ mic, and abundances for 12 individualelements, namely C, N, O, Mg, Al, Si, Ca, Ti, Cr, Mn, Fe, Ni. Given a spectralsignal-to-noise ratio of 100 per pixel, internal precision of the labelestimates are about 20 K for T_ eff, 0.05 dex for log g, and 0.05 dexfor most elemental abundances. These results are agree with theoretical limitsfrom the Crámer-Rao bound calculation within a factor of two. TheGaia-Enceladus-Sausage that contributes the majority of the accreted halo starsare discernible from the disk and in-situ halo populations in the resultant[Mg/Fe]-[Fe/H] and [Al/Fe]-[Fe/H] abundance spaces. We also provide distanceand orbital parameters for the sample stars, which spread a distance out to∼100 kpc. The DESI sample has a significant higher fraction of distant (ormetal-poor) stars than other existed spectroscopic surveys, making it apowerful data set to study the Galactic outskirts. The catalog is publiclyavailable.