On this page we provide advice and instructions for using the GALAH DR3 catalogues. We would recommend taking the time to read Buder et al. (2021), which supplies even more information and quality assessment of GALAH DR3. Sven Buder has also created iPython notebooks tutorials for GALAH DR3.


We highly recommend using the GALAH_DR3_main_allstar_v2.fits catalogue if you want to work with the GALAH DR3 stellar parameters and elemental abundances.

For more information, see our Catalogue documentation.

If your science case contains stellar parameters and elemental abundances, then we recommend the GALAH_DR3_main_allstar_v2.fits catalogue. It contains one entry per star.

There is an extended version called GALAH_DR3_main_allspec_v2.fits that contains one entry per observation, and has abundance information for every spectral line used in the GALAH analysis. This table is for expert use only.

Consult the catalogue documentation for details on the catalogues and acquiring them. If you just want the entire GALAH_DR3_main_allstar_v2 cataloge, it can be directly directly downloaded, or with wget (removing the --spider flag):

wget --spider https://cloud.datacentral.org.au/teamdata/GALAH/public/GALAH_DR3/GALAH_DR3_main_allstar_v2.fits

Overall, we make three recommendations: (1) snr_c3_iraf > 30; (2) only stellar parameters for stars with flag_sp == 0, flag_fe_h == 0; (3) only elemental abundances with flag_X_fe == 0.

For more information, see our Flag bitmask documentation.

The main recommended GALAH_DR3_main_allstar_v2 catalogue contains a number of flags. For most science cases, the two most important to consider are the stellar parameter quality flag (flag_sp) and the iron abundance quality flag (flag_fe_h). flag_sp folds in many potential sources of errors in the input values (e.g., unreliable astrometry, very low signal spectra, reduction problems, possibly binarity). flag_fe_h indicates problems with the stellar parameter determination. By default, we recommend that users only consider stellar parameters (Teff, log g, [Fe/H], broadening velocity, radial velocity) for stars with flag_sp == 0 and flag_fe_h == 0.

For the abundance information there is similarly flags. For the overall α abundance (alpha_fe) and each indivudual elemental abundance X_fe there is a flag_X_fe that indicates problems with the elemental abundance determination. We strongly recommend only considering the abundance of element X when flag_X_fe == 0. As for all of our previous GALAH releases, we want to stress that we discourage the use of element abundances with flag_X_fe > 0 without consideration of the possible systematics that these flagged measurements can introduce.

We would recommend considering only stars with a signal to noise snr_c3_iraf > 30.


For stellar parameters and elemental abundances, use the values found in GALAH_DR3_main_allstar_v2 catalogue. We do not recommend values found only in the GALAH_DR3_main_allspec_v2 catalogue as these are only for expert use and are typically for diagnostic purposes.

For more information, see our Table scheme documentation.

In the catalogues that constitute GALAH DR3, for many parameters we provide only one value, e.g., for the overall α-element abundance there is just alpha_fe. However, some parameters have multiple values calculated by different methods. In this section, we discuss these parameters and in most cases make a recommendation on the value to use.

  • For radial velocity, we strongly recommend the rv_galah.
    • The source of this value differs for each star. For 83 per cent of stars rv_galah == rv_obst (found in the GALAH_DR3_VAC_rv_v2 table), the radial velocity measured by Zwitter et al. (2020) from our observed spectra. The remaining stars use rv_galah == rv_sme_v2, the value measured by SME during the analysis process. The rv_obst value has errors typically five times smaller than the SME-measured value. Consult the use_rv_flag for each star to identify the source of the radial velocity for rv_galah.
    • The other radial velocity values available across the GALAH DR3 catalogues are either from Gaia DR2, or are diagnostic values measured during the analysis (e.g., rv_6708 is the RV measured from just the Li6708 line). See the section of the catalogues documentation on the collated radial velocity measurements (GALAH_DR3_VAC_rv_v2) for more details.
    • Due to bookkeeping error, we include the Gaia radial velocity (and their errors) twice: as dr2_radial_velocity (in GALAH_DR3_VAC_rv_v2 and GALAH_DR3_VAC_GaiaEDR3_v2) and rv_gaia_dr2 (in main_allstar and GALAH_DR3_main_allstar_v2 and GALAH_DR3_main_allspec_v2).
  • For effective temperature, surface gravity, iron abundance ([Fe/H]), broadening velocity, we strongly recommend teff, logg, fe_h, vbroad respectively from the GALAH_DR3_main_allstar_v2 catalogue.
    • These are inferred as part of the spectrum fitting and abundance determination.
    • The other provided values of these are either calculated during an intermediary analysis step (e.g., teff_guess), or as part of a value-added catalogues (e.g., teff_bstep).
  • For stellar luminosity, mass and age, we recommend log_lum_bstep, mass_bstep and age_bstep respectively from the GALAH_DR3_VAC_ages_v2 table.
    • Do not use the lbol, mass and age found in the GALAH_DR3_main_allspec_v2 catalogue.
  • For heliocentric distance, we recommend distance_bstep value from the GALAH_DR3_VAC_ages_v2 table.
    • distance_bstep was calculated as part of the estimation of age, mass, radius etc found in the GALAH_DR3_VAC_ages_v2 catalogue.
    • distance_bstep was used for the vast majority of stars (96 per cent) in the calculation of Galactic kinematic and dynamic parameters in GALAH_DR3_VAC_dynamics_v2. For the other 4 per cent of the stars we mostly use the photogeometric distances r_med_photogeo in the GALAH_DR3_VAC_GaiaEDR3_v2 table as calculated by Bailer-Jones et al. (2020). The distance used for a given spectrum is provided by the use_dist_flag in the GALAH_DR3_VAC_dynamics_v2 table.
  • For reddening, we recommend the ebv value from the GALAH_DR3_main_allstar_v2 table.
  • For parallax, the GALAH DR3 catalogues has the parallax values from both the Gaia DR2 and eDR3 values (the former being called parallax_dr2 in our catalogues). We have also calculated the eDR3 parallax corrected for the zeropoint offset as prescribed in Lindegren et al. (2020), called parallax_corr in our tables.
  • Distinct from the iron abundance ([Fe/H]) calculated from the spectra, the GALAH_DR3_VAC_ages_v2 table includes the initial and current metallicity of the star (meh_ini_bstep and meh_act_bstep) as calculated by BSTEP. These values were not used as part of the abundance analysis.

This table summarizies the parameters in GALAH DR3 for which there are multiple values for a given star. We have bolded our recommended column.

Parameter and
recommended value
main_allstar main_allspec VAC_ages VAC_rv VAC_GaiaEDR3
Radial velocity:
rv_galah
rv_galah,
rv_gaia_dr2
rv_galah,
rv_gaia_dr2,
rv_guess,
rv_5854,
rv_6708,
rv_6722
  rv_galah,
rv_sme_v2,
rv_sme_v1,
rv_obst,
rv_nogr_obst,
dr2_radial_velocity
dr2_radial_velocity
Effective Temperature:
teff
teff,
irfm_teff
teff,
irfm_teff,
init_teff,
teff_guess
teff_bstep    
Surface gravity:
logg
logg logg,
init_logg,
logg_guess
logg_bstep    
[Fe/H]:
fe_h
fe_h,
fe_h_atmo
fe_h,
fe_h_atmo,
init_fe_h_atmo,
feh_guess
     
Broadening velocity:
vbroad
vbroad vbroad,
init_vbroad
     
Stellar luminosity:
log_lum_bstep
  lbol log_lum_bstep    
Stellar mass:
m_act_bstep
  mass m_act_bstep,
m_ini_bstep
   
Stellar age:
age_bstep
  age age_bstep    
Distance:
distance_bstep
r_est_dr2 r_est_dr2 distance_bstep   r_med_geo,
r_med_photogeo
E(B-V):
ebv
ebv,
irfm_ebv
ebv,
irfm_ebv
ebv_bstep    
Parallax parallax_dr2 parallax_dr2     parallax,
parallax_corr
Metallicity     meh_act_bstep,
meh_ini_bstep
   

Joining GALAH DR3 catalogues

Join GALAH DR3 tables using the sobject_id.

The catalogues of GALAH DR3 must be joined or cross-matched using the sobject_id. Do not use the star_id or Gaia source_id to join the catalogues. The value-added catalogues are based upon the extended catalogue which contains measurements per observed spectrum. About 50000 stars were observed multiple times, and therefore have multiple observed spectra. The sobject_id column is our internal ID for each observation and using this column for joining will ensure that you are matching information derived from the same spectrum.

For instance, to join the GALAH_DR3_main_allstar_v2 and GALAH_DR3_VAC_ages_v2 catalogues, using the Data Central, the following ADQL query could be used:

-- Join GALAH_DR3_main_allstar_v2 and GALAH_DR3_VAC_ages_v2
-- and get the effective temperature, surface gravity,
-- iron abundance, and age of each star.
SELECT
  TOP 100
  g_ms.source_id, g_ms.teff, g_ms.logg, g_ms.fe_h,
  g_ages.sobject_id, g_ages.age_bstep, g_ages.e_age_bstep
  -- GALAH_DR3_main_allstar_v2 is galah_dr3.main_star
  FROM galah_dr3.main_star AS g_ms
  -- GALAH_DR3_VAC_ages_v2 is galah_dr3.vac_ages
  JOIN galah_dr3.vac_ages AS g_ages
  	ON g_ms.sobject_id = g_ages.sobject_id