Data Central uses cookies to make your browsing experience better. By using Data Central you agree to its use of cookies. Learn more I agree


Find information, examples, FAQs and extensive descriptions of the data, curated by the survey teams.

GALAH logo


The Galactic Archaeology with HERMES survey is collecting stellar parameters and abundances for one million stars in the Milky Way. GALAH yields a comprehensive view of the formation & evolution of the Galaxy.
Nov. 6, 2020 by J. Simpson
Dec. 14, 2020, 1:12 a.m. J. Simpson

GALAH+ Data Release 3 Overview


The GALAH survey released our Third Data Release (DR3) on 6 November 2020. It provides stellar atmospheric parameters and individual elemental abundances for almost 600,000 stars. It will also soon provide reduced one-dimensional spectra for all the stars.

This page contains an overview of GALAH+ DR3.

These pages listed below contain more information about the data release and the GALAH survey:

What's in GALAH+ DR3?

The Third Data Release of the Galactic Archaeology with HERMES (GALAH) survey provides stellar parameters and elemental abundances for 678,423 spectra for 588,571 mostly nearby stars, that have been observed with the HERMES spectrograph at the Anglo-Australian Telescope between November 2013 and February 2019. The release is fully described in Buder et al. (2020).

With this data release, for the first time, we provide the reduced one-dimensional spectra for the stars.

We catalogue stellar parameters as well as elemental abundances for up to 30 elements per star: Li, C, O, Na, Mg, Al, Si, K, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Rb, Sr, Y, Zr, Mo, Ru, Ba, La, Ce, Nd, Sm, and Eu.

Our catalogue comprises 383,088 (65%) dwarfs and 200927 (34%) giants (and 4556 [1%] unclassified stars).

Based on unflagged chemical composition and age, we find 62.5% young low-alpha stars, 8.8% young high-alpha stars, 26.9% old stars, and 1.8% stars with [Fe/H] < -1. Based on kinematics, we find 4% halo stars.

For ease of use, this release referred to as GALAH+ DR3 presents data from multiple programs as a single catalog. It includes observations from GALAH Phase 1 (bright, main, and faint survey, 476,863 spectra), the K2-HERMES (112,943 spectra) and TESS-HERMES (34,263 spectra) surveys, as well as additional GALAH-related projects including the bulge and observations of more than 75 stellar clusters (54,354 spectra).

The figure below is an overview of distances and photometric information for the spectra (including repeats for some stars) observed as part of GALAH DR3. Panel a) shows the distances of stars in GALAH DR3. Due to the magnitude limited selection of stars, the majority of stars are not only dwarfs but also nearby; that is, within 1 kpc. Only 5.8% of stars are beyond 4 kpc. Panel b) shows a reddened color-absolute magnitude diagram in the optical Gaia passbands. Panel c) shows an analogous diagram made with the infrared 2MASS passbands.


Compared to GALAH DR2, we improve our spectrum analysis with external astro- and photometric information from Gaia DR2 and 2MASS to estimate more accurate stellar surface gravities than before, thus breaking spectroscopic degeneracies. We also use Spectroscopy Made Easy (SME) to analyse the entire spectral data set.

For all targets in DR3, we provide:

  • Reduced one-dimesional spectra across the four wavelength regions of the HERMES spectrograph
  • Heliocentric radial velocities
  • Stellar parameters (effective temperature, surface gravity, iron abundance, microturbulence, broadening)
  • The overall alpha-element abundance, and up to 30 elemental abundances per star
    • light elements: Li, C, O
    • odd-Z elements: Na, Al, K
    • α-elements: Mg, Si, Ca, Ti (and TiII)
    • iron-peak elements: Sc, V, Cr, Mn, Co, Ni, Cu, Zn
    • light and heavy slow neutron capture elements: Y, Ba, La, Rb, Mo, Ru, Nd, Sm
    • rapid neutron capture element: Eu

GALAH Abundances

Sky Coverage

Below is an overview of the distribution of stars observed as part of this data release in Galactic coordinates with the centre of the Galaxy at the origin. Shown are the GALAH main (blue) and faint (orange) targets, which avoid the Galactic plane. The targets of the K2-HERMES follow-up (green) fall within with the K2 campaigns along the ecliptic and show the characteristic tile-pattern of the Kepler telescope. The TESS-HERMES observations (red) are focused on the TESS Southern Continuous Viewing Zone. Other HERMES targets (purple) are distributed across the sky and were observed during several independent programs.

GALAH DR3 Sky Coverage

Data Access

The two main data products of GALAH DR3 are the catalogue of stellar parameters and abundances, and the spectral library. We also provide several value-added catalogues.


The catalogues can be accessed either by:

  • Direct download of the FITS files;
  • The query services on the AAO Data Central webpage.
    • Currently the galah_dr3.vac_rv table is not available through this service.

There are two main catalogues:

Table name File name Description
galah_dr3.main_star GALAH_DR3_main_allstar_v1.fits (853 MB) Strongly recommend the use of this table. One entry per star observed. Radial velocity, stellar parameters and abundance data. Important data from Gaia, 2MASS, and WISE.
galah_dr3.main_spec GALAH_DR3_main_allspec_v1.fits (2.2 GB) One entry per observation. Radial velocity, stellar parameters for each observation. Also contains abundances derived for each individual line. Important data from Gaia, 2MASS, and WISE.

For science cases involving stellar parameters, it is highly recommended that you only consider values where flag_sp = 0 and flag_fe_h = 0. For science cases involving the abundance of element x, it is highly recommended that you only consider x_fe where flag_x_fe = 0. See the Table Schema page for detailed descriptions of these flags.

There are several value-added catalogues:

Table name File name Description
galah_dr3.vac_ages GALAH_DR3_VAC_ages_v1.fits (243.7 MB) Ages and masses of the stars estimated using isochrones
galah_dr3.vac_dynamics GALAH_DR3_VAC_dynamics_v1.fits (502.1 MB) Kinematic and dynamic information for each star/spectrum
galah_dr3.vac_rv GALAH_DR3_VAC_rv_v1.fits (21.6 MB) Higher precision radial velocities measured essentially noiseless median spectra


GALAH DR3 provides the reduced one-dimensional spectra for each star in the main catalogue. For each star there are four files, one for each of the four HERMES cameras, and each spectrum file contains:

  • The reduced spectrum
  • The variance of the reduced spectrum
  • The pseudo-continuum normalized spectrum
  • The variance of the pseudo-continuum normalized spectrum
  • The sky spectrum used for sky subtraction

These files can be accessed via the Single Object Viewer or as a Bulk Download. Instructions on how to retrieve GALAH DR3 spectra can be found on the Data Access page.

Changes from the Second Data Release

Data Reduction

There have been two major improvements to the data reduction:

  1. The main improvement is the wavelength solution, which is now more stable at the edges of green and red CCDs, where we lack arc lines. This has been achieved by monitoring the solution and fixing the polynomial describing the pixel-to-wavelength transformation, if deviations from a typical or average solution are detected.
  2. Cross-talk is now parametrized differently. It can only be measured in larger gaps between every 10th spectrum.


There are two main changes to the analysis methods for DR3 compared to DR2.

  1. We are using astrometric information from the Gaia mission to break spectroscopic degeneracies. This requirement has resulted in about 5000 stars that were in GALAH DR2 not being in GALAH DR3. See the data reduction and analysis documentation for more details.
  2. GALAH DR3 uses Spectroscopy Made Easy for all the stellar parameter and elemental abundance determination. For the second data release of the GALAH survey we made use of data-driven approaches to improve both speed and precision of the spectroscopic analysis. Although the data-driven approaches were successful for the majority of GALAH DR2 stars, we know that these approaches can suffer from signal aliasing, they can learn unphysical correlations between the input data and the output stellar labels, and the results are not necessarily valid outside the parameter space of the training set. We found in DR2 that the data-driven approaches meant that stars at the periphery in stellar label space, e.g. high temperature or low metallicity did not receive optimal labels from the data driven process.

Below is a comparison of GALAH DR2 (upper panels) and GALAH DR3 (lower panels, this release). The smooth light blue background indicates all measurements, whereas the colormap shows the number of unflagged measurements at each point. The stellar parameters and abundances from GALAH DR2 appear more tightly constrained, but we note that this is an artefact of the data-driven approach, which tends to find solutions closer to the mean parameter/abundance patterns.

Improvements of GALAH DR3 over DR2

Need help?

Questions about GALAH DR3 that are not answered in the documentation?

Contact us:

Nov. 6, 2020 by J. Simpson
Dec. 14, 2020, 1:12 a.m. J. Simpson