DESI Data Release 1: Stellar Catalogue

TL;DR

DESI Data Release 1 (DR1) presents a remarkably large stellar Value-Added Catalogue (VAC) derived from 13 months of DESI observations, including coadded and single-epoch spectra for over 4 million stars and more than 1 million single-epoch measurements. The analysis employs two pipelines, RVS (RVSpecFit with a neural-network spectrum emulator) and SP (FERRE), to derive radial velocities, stellar parameters, and abundances, with a temperature- and gravity-dependent calibration that brings [Fe/H] into agreement with high-resolution surveys to about 0.1 dex. DR1 uncovers extensive Galactic structure through distant and metal-poor stars, clusters, streams, and dwarfs, and provides a detailed validation showing generally good agreement with external surveys while documenting systematic biases and the need for calibrations. The release also introduces valuable single-epoch measurements and crossmatches with Gaia, enabling studies of RV variability and halo substructure, and sets the stage for DR2, which will expand the sample by roughly a factor of three and enhance processing and calibration efforts.

Abstract

In this paper we present the stellar Value-Added Catalogue (VAC) based on the DESI Data Release 1. This VAC contains stellar parameter, abundance and radial velocity measurements for more than 4 million stars. It also contains, for the first time, measurements from individual epochs for more than a million stars with at least two observations. The main contribution to the catalogue comes from the bright program of the main survey, which includes $\sim $2.5 million stars, and the backup program, which includes $\sim $ 1 million stars. The combined magnitude range for the stars in the catalogue extends from ${\it Gaia}$ G $\sim 12$ to G $\sim 21$. For the magnitude range $17.5<G<21$ this catalogue represents a factor of 10 increase in the number of stars with radial velocity and abundance measurements compared to existing surveys. Despite DESI's resolution (R $\sim 2500-5000$), the median radial velocity uncertainty for stars in the catalogue is better than 1 km s$^{-1}$. The stellar parameters and abundances of stars in DESI are measured by two independent pipelines, and after applying a temperature-dependent calibration, [Fe/H] abundances of high signal-to-noise stars are accurate to better than $\sim$ 0.1 dex when compared to high-resolution surveys. The catalogue probes different Galactic components including a particularly large number of distant stars: tens of thousands of stars further than 10 kpc, and thousands further than 50 kpc. The catalogue also contains several thousand extremely metal-poor stars with ${\rm [Fe/H]}<-3$. The released sample of stars includes measurements for thousands of stars that are members of dwarf galaxies, open and globular clusters as well as members of several dozen stellar streams. The next public DESI data release is expected in less than two years and will contain three times as many stars as DR1.

Figures (21)

• Figure 1: Density of objects on the sky that are included in the VAC, were observed in the main survey and are classified as stars. Each panel shows a different program of the main survey. The units for the density are in objects per square degree. The dashed line shows the Galactic plane.
• Figure 2: Left 3 panels: The colour-magnitude distribution in DECaLS photometric bands for objects included in the VAC. Different panels (from left to right) show objects from the bright program of the sv3 survey, dark and bright programs of the main survey respectively. Right panel: The colour-magnitude distribution of targets in the backup program included in the VAC in Gaia photometric bands. All the panels only show objects classified as a star by Redrock. No extinction correction was applied to the magnitudes.
• Figure 3: The colour-magnitude distribution for different target classes in the VAC observed in the bright program of the main survey. The description of the selections is provided in Cooper2023. The most numerous targets are the MWS_MAIN_BLUE and MWS_MAIN_RED, where the former selects blue sources in the $16\lesssim r\lesssim19$ magnitude range, the latter targets red sources but also applies parallax and proper motion cuts to prioritise distant giants. The other large target class is MWS_BROAD, which is a magnitude limited selection of any object not selected by MWS_MAIN_BLUE and MWS_MAIN_RED categories. The target boundaries in the plot appear somewhat blurred as selections are applied to dereddened magnitudes, while we show raw magnitudes. The colour-magnitude distributions for high priority targets, such as white dwarfs, stars within 100 pc and blue horizontal branch stars are shown in the bottom three panels. On each panel we also give the bitmask value that can be used to select these targets using MWS_TARGET column in the catalogue.
• Figure 4: Histogram of the number of repeated observations in DESI DR1 for Gaia sources in this VAC. Curves of different colours (grey, blue, green, orange, brown) show the histograms for subsets of objects where the time separation between the last and first observation in DR1 is less than 1 day, between 1 and 10 days, 10 and 100 and more than 100 days respectively.
• Figure 5: Distribution of the surface gravity and effective temperatures measured by two DESI stellar pipelines (RVS on the left, SP on the right). The figure only shows stars observed in the main survey with the signal-to-noise in the R arm above 10.