APOGEE DR17 Synopsis

Data Release 17 provides spectra, radial velocities, and stellar atmospheric parameters and individual elemental abundances for more than 657,000 stars, sampling all major components of the Milky Way. This is the final Data Release from the APOGEE-1 and APOGEE-2 programs.

Summary of Major Updates in DR17:

For those familiar with prior Data Releases from APOGEE, here is a brief summary of especially noteworthy changes to the the dataset and data products:

  • New observations taken from August 2018 through the end of the survey, November 2020 from APO, and January 2021 from LCO, are included.
  • All previously released data taken as part of SDSS-III/APOGEE-1 and SDSS-IV/APOGEE-2 have been homogeneously processed using the DR17 pipeline.
  • The Radial Velocity and Visit Combination code have been revised to provide better radial velocities, especially for the faintest targets.
  • Stellar Parameters and Chemical Abundances have been derived using a new set of spectral synthesis grids that use the Synspec spectral synthesis code and include NLTE effects for four elements: Na, Mg, K, Ca.
  • Gaia EDR3 data products are included in the summary data files.
  • MEMBER and MEMBERFLAG columns have been added to provide likely membership information for some star clusters and for dwarf galaxies.

A more complete list of modifications to the data processing and analysis is given below.

DR17 Scope

DR17 contains 657,000 unique APOGEE targets in 2,261 fields accumulated through 2,660,000 individual visits. For all targets, DR17 provides visit spectra, heliocentric radial velocities, combined spectra, stellar atmospheric parameters (Teff, log g, vmicro, [M/H], [C/M], [N/M], [α/M], vsini or vmacro), and abundance determination for up to 20 species are presented (C, C I, N, O, Na, Mg, Al, Si, S, K, Ca, Ti, Ti II, V, Cr, Mn, Fe, Co, Ni, Ce). In DR17, Ge, Rb, and Yb were not attempted whereas P, Cu, Nd, and 13C were attempted but judged to be have been unsuccessful.

Instructions on how to access the data can be found on Data Access. Eight Value Added Catalogs have been provided with DR17 to enhance these products that are listed in the DR17 Value-Added Catalogs section below. Six more Value Added Catalogs will be released in 2022.

DR17 Observations

DR17 Sky Coverage

DR17 is the final release of data from APOGEE-2 and includes all data from the SDSS-III/APOGEE and SDSS-IV/APOGEE-2 surveys. The observation cutoff date for APOGEE-2N is November 07, 2020 (MJD 59160) and for APOGEE-2S is January 20, 2021 (MJD 59234). The DR16 cutoff date was August 18, 2018 and, thus, DR17 includes, roughly, an additional two years of observations in both hemispheres.

The figure below displays the DR17 APOGEE spectroscopic coverage in Galactic coordinates where each field is color-coded by the total number of targets to demonstrate the target density by field.


Distribution of APOGEE fields in DR17 overlaid on an all-sky image from 2MASS. Each field is color-coded by the number of targets in that field. <em>Image credit: C. Hayes</em>
Distribution of APOGEE fields in DR17 overlaid on an all-sky image from 2MASS. Each field is color-coded by the number of targets in that field. Image credit: C. Hayes

DR17 APOGEE Targets

DR17 contains data for approximately 657,000 unique APOGEE targets in 2,261 fields, of which roughly 372,000 are from the Main Red Star Sample. A total of 2,660,000 individual visits are included in DR17. A general breakdown of targets in DR17 follows below:

43,200 targets toward the Bulge
68,000 targets toward the Halo

  • 10,500 targets toward Halo streams
  • 56,500 targets toward the smooth Halo

163,000 targets co-observed on MaNGA plates
152,000 targets toward the Disk
22,600 targets toward dwarf satellite galaxies

  • 14,000 targets toward the Large and Small Magellanic Clouds
  • 6,700 targets toward the the Sgr Core and Stream
  • 1,900 targets toward dwarf spheroidal satellites

43,500 targets toward Star Clusters

  • 14,00 targets toward Young Clusters
  • 28,500 targets toward Open or Globular Clusters

132,000 targets for Photometric Objects of Interest (POIs) in the Kepler, K2, CoRoT, and TESS footprints
19,500 targets in Ancillary Programs
92,000 targets in Contributed Programs
1,200 targets in NMSU 1-meter Programs

The Using Targets/Samples page provides guidance for identifying these target categories, while the target selection algorithms are described in the Targeting Information page. Two papers, Beaton et al. (2021), and Santana et al. (2021), summarize the final targeting algorithms for APOGEE-2S and APOGEE-2N, respectively.

Data Analysis Changes

The subsections below summarizes major and specific changes for the DR17 data analysis pipelines relative to DR16. This is meant to be a quick list for experienced users and the full details of these changes are described in Holtzman et al. (in prep.). The Using APOGEE data pages (parameters, abundances, spectra, radial velocities, and targets/samples) discuss the implications of these changes for typical use. Examples and Tutorials provide practical guidance for using the data products.

Changes to Targeting Information

  • A new bitmask, APOGEE2_TARGET4, was created for DR17, but is currently unpopulated.
  • The following problems from prior DRs have been resolved:
    • Stars in K2 fields are now identifiable by their sub-selection (asteroseismology, planet hosts, etc.).
    • Stars selected to be candidate members of Globular Clusters are now identified.
  • The meaning of several APOGEE2 Targeting Flags were altered and homogeneously applied to the APOGEE-2 targets:
    • APOGEE2_TARGET1=29: “Faint Target” is set for any targets whose magnitudes preclude them from reaching S/N=100 in the number of visits assigned to a field in its design. Stated differently, these targets were never intended to read S/N=100.
    • APOGEE2_TARGET2=16:Gaia Overlap” is set for any targets that were selected using Gaia data data products. Typically, only proper motions were used.
    • APOGEE2_TARGET1=10: “Globular Cluster Candidate” is used to identify all targets selected as candidate members of Globular Clusters using proper motions and/or color-magnitude criteria.
    • APOGEE2_TARGET2=20: “Halo Members” is set for halo stars targeted in both the Northern and Southern hemispheres that were identified from prior surveys. In the APOGEE-2N, we used SEGUE and in APOGEE-2S we used SkyMapper photometric metallicities.
    • APOGEE2_TARGET2=0: “K2 GAP Program” is set for stars in K2 fields selected from the Galactic Archaeology Program (GAP).
  • The full listing of photometry and proper motion sources used in the targeting input catalogs are now identified. The bulk of the non-standard photometry or astrometry sources are from ancillary or contributed programs that often required special datasets.

The full details of the targeting flag changes are described in Santana et al. (2021).

Changes to Data Processing

  • The Radial Velocity (RV) code has been overhauled and uses a package called Doppler.
  • The visit combination process has also been overhauled.

The full details of these changes are described in Holtzman et al. (in prep.).

Changes to Analysis Procedures

Some major changes to Analysis Procedures from the previous data release include:

  • ASPCAP was re-written.
  • The DR17 ASPCAP release uses on a new synthetic grid constructed with Synspec that includes NLTE treatment for Na, Ca, Mg, and K
  • The following changes were made to the abundances measured:
    (1) Ce improved by using multiple lines,
    (2) Though measured, no abundances for P and Cu were deemed reliable,
    (3) Measurements for Ge, Rb, and Yb were not attempted.
  • Supplemental ASPCAP analyses were performed using alternative libraries:
    (1) Synspec LTE (only),
    (2) Turbospectrum 20 LTE with spherical radiative transfer calculations for giants, and
    (3) Turbospectrum 20 LTE with plane parallel radiative transfer calculations for giants
  • STAR_BAD has been revised to be less restrictive

The full details of these changes are described in Holtzman et al. (in prep.).

Changes to Parameter Calibrations

Some major changes to Parameter and Abundance Calibrations from the previous data release include:

  • log g calibrations based on a neural network trained on a combination of true parameters from asteroseismology (giants, warm dwarfs) or isochrones (cool dwarfs)
  • Parameter dependent calibration applied to Teff and log g
  • A zeropoint calibration is applied to elemental abundances with different zero-points determined for dwarfs and giants from solar metallicity stars in the solar neighborhood
  • Uncalibrated values are now populated to _SPEC tags for all parameters and abundances
  • Named tags are only populated for stars meeting quality restrictions.

The full details of these changes are described in Holtzman et al. (in prep.).

Changes to the Data Model

The following changes have occurred to the data model:

  • Bad, missing, or invalid floating point entries are now NaN instead of “-9999” in summary files from the SAS, but the CAS will still use -9999
  • apStar files no longer contain parametric fits to the combined LSF
  • ap1dvisit files now contain catalog information
  • The VISIT_PK tag in the allStar file now contains the entry number for all spectra of a given target, not just those for the FIELD+APOGEE_ID combination.
  • New Tags have been included in the allStar file:
    (1) A new bitmask, MEMBER, and a corresponding string, MEMBERFLAG, that indicate if a star is considered a member of a dwarf galaxy or star cluster using simple criteria,
    (2) A new bitmask, RV_FLAG, has been added that conveys details regarding the RV-related flags in STARFLAG
  • The Gaia related tags are now structured as GAIAEDR3 instead of GAIA to indicate the source of the Gaia data. Additional Gaia eDR3 columns have been added for the convenience of the user (Gaia Collaboration, Brown et al. 2021). We note that the APOGEE team performs the match between the APOGEE target and the Gaia catalogs. We encourage those using these data in their analyses to cite the original source of the data; to encourage this, the references are provided below.
    The full list of tags is as follows:
    GAIAEDR3_SOURCE_ID (Gaia Collaboration, Brown et al. 2021, Gaia Collaboration, Prusti et al. 2016),
    GAIAEDR3_PARALLAX, GAIAEDR3_PARALLAX_ERROR, GAIAEDR3_PMRA, GAIAEDR3_PMDEC, GAIAEDR3_PMRA_ERROR, GAIAEDR3_PMDEC_ERROR (Lindegren et al. 2021),
    GAIAEDR3_G_MEAN_MAG, GAIAEDR3_BP_MEAN_MAG, GAIAEDR3_RP_MEAN_MAG (Riello et al. 2021),
    GAIAEDR3_RADIAL_VELOCITY, GAIAEDR3_RADIAL_VELOCITY_ERROR (Seabroke et al. 2021),
    GAIAEDR3_R_MED_GEO, GAIAEDR3_R_LO_GEO, GAIAEDR3_R_HI_GEO,GAIAEDR3_R_MED_PHOTOGEO, GAIAEDR3_R_LO_PHOTOGEO, GAIAEDR3_R_HI_PHOTOGEO (Bailer-Jones et al. 2021).
  • Because no abundances were attempted for Yb, Rb, and Ge, there are no associated named tags (e.g., no XX_FE, XX_FE_SPEC, XX_FE_ERR, XX_FE_FLAG)
  • No raw FERRE files are included

DR17 Value Added Catalogs

Note: All VACs were produced with the the synspec version of the AllStar Summary File. (See this Caveat).

The following Value Added Catalogs (VACs) are available for the DR17 datasets:

  • Open Clusters -- Provides membership analysis for stars in the vicinity of known open clusters
  • The Joker -- Provides stellar orbital parameters from modeling multi-epoch RV measurements using The Joker
  • SB2 -- Provides spectroscopic binary identifications
  • APOGEE NET -- Provides an independent analysis of APOGEE spectra using neural nets
  • Red Clump -- Provides a catalog of red clump stars using spectro-photometric selections.
  • AstroNN -- Provides stellar parameters, elemental abundances, distances, and ages from a neural network-based code trained on the spectra; orbital parameters are provided using these distances and GalPy.
  • StarHorse -- Provides distances and extinctions using spectro-photometry for all stars in DR17
  • APOGEE-Centric Fire Simulations -- Provides mock catalogs with APOGEE-error model using the Ananke framework to sample the Latte Simulations

Additional Value Added Catalogs will be made public in 2022.