Astronomy & Astrophysics manuscript no. aa202039657 ©ESO 2020
December 4, 2020
Gaia Early Data Release 3
Summary of the contents and survey properties
Gaia Collaboration, A.G.A. Brown 1, A. Vallenari 2, T. Prusti 3, J.H.J. de Bruijne 3, C. Babusiaux 4, 5, M.
Biermann6, O.L. Creevey 7, D.W. Evans 8, L. Eyer 9, A. Hutton10, F. Jansen3, C. Jordi 11, S.A. Klioner 12,
U. Lammers 13, L. Lindegren 14, X. Luri 11, F. Mignard7, C. Panem15, D. Pourbaix 16, 17, S. Randich 18, P.
Sartoretti5, C. Soubiran 19, N.A. Walton 8, F. Arenou 5, C.A.L. Bailer-Jones20, U. Bastian 6, M. Cropper 21,
R. Drimmel 22, D. Katz 5, M.G. Lattanzi 22, 23, F. van Leeuwen8, J. Bakker13, C. Cacciari 24, J. Castañeda
25, F. De Angeli8, C. Ducourant 19, C. Fabricius 11, M. Fouesneau 20, Y. Frémat 26, R. Guerra 13, A.
Guerrier15, J. Guiraud15, A. Jean-Antoine Piccolo15, E. Masana 11, R. Messineo27, N. Mowlavi9, C. Nicolas15, K.
Nienartowicz 28, 29, F. Pailler15, P. Panuzzo 5, F. Riclet15, W. Roux15, G.M. Seabroke21, R. Sordo 2, P. Tanga
7, F. Thévenin7, G. Gracia-Abril30, 6, J. Portell 11, D. Teyssier 31, M. Altmann 6, 32, R. Andrae20, I.
Bellas-Velidis33, K. Benson21, J. Berthier 34, R. Blomme 26, E. Brugaletta 35, P.W. Burgess8, G. Busso 8, B.
Carry 7, A. Cellino 22, N. Cheek36, G. Clementini 24, Y. Damerdji37, 38, M. Davidson39, L. Delchambre37, A.
Dell’Oro 18, J. Fernández-Hernández40, L. Galluccio 7, P. García-Lario13, M. Garcia-Reinaldos13, J.
González-Núñez 36, 41, E. Gosset37, 17, R. Haigron5, J.-L. Halbwachs 42, N.C. Hambly 39, D.L. Harrison 8, 43,
D. Hatzidimitriou 44, U. Heiter 45, J. Hernández13, D. Hestroffer 34, S.T. Hodgkin8, B. Holl 9, 28, K.
Janßen46, G. Jevardat de Fombelle9, S. Jordan 6, A. Krone-Martins 47, 48, A.C. Lanzafame 35, 49, W. Löffler6, A.
Lorca10, M. Manteiga 50, O. Marchal42, P.M. Marrese51, 52, A. Moitinho 47, A. Mora10, K. Muinonen 53, 54, P.
Osborne8, E. Pancino 18, 52, T. Pauwels26, J.-M. Petit 55, A. Recio-Blanco7, P.J. Richards56, M. Riello 8, L.
Rimoldini 28, A.C. Robin 55, T. Roegiers57, J. Rybizki 20, L.M. Sarro 58, C. Siopis16, M. Smith21, A.
Sozzetti 22, A. Ulla59, E. Utrilla10, M. van Leeuwen8, W. van Reeven10, U. Abbas 22, A. Abreu Aramburu40, S.
Accart60, C. Aerts 61, 62, 20, J.J. Aguado58, M. Ajaj5, G. Altavilla 51, 52, M.A. Álvarez 63, J. Álvarez Cid-Fuentes
64, J. Alves 65, R.I. Anderson 66, E. Anglada Varela 40, T. Antoja 11, M. Audard 28, D. Baines 31, S.G.
Baker 21, L. Balaguer-Núñez 11, E. Balbinot 67, Z. Balog 6, 20, C. Barache32, D. Barbato9, 22, M. Barros 47,
M.A. Barstow 68, S. Bartolomé 11, J.-L. Bassilana60, N. Bauchet34, A. Baudesson-Stella60, U. Becciani 35, M.
Bellazzini 24, M. Bernet11, S. Bertone 69, 70, 22, L. Bianchi71, S. Blanco-Cuaresma 72, T. Boch 42, A.
Bombrun73, D. Bossini 74, S. Bouquillon32, A. Bragaglia 24, L. Bramante27, E. Breedt 8, A. Bressan 75, N.
Brouillet19, B. Bucciarelli 22, A. Burlacu76, D. Busonero 22, A.G. Butkevich22, R. Buzzi 22, E. Caffau 5, R.
Cancelliere 77, H. Cánovas 10, T. Cantat-Gaudin 11, R. Carballo78, T. Carlucci32, M.I Carnerero 22, J.M.
Carrasco 11, L. Casamiquela 19, M. Castellani 51, A. Castro-Ginard 11, P. Castro Sampol11, L. Chaoul15, P.
Charlot19, L. Chemin 79, A. Chiavassa 7, M.-R. L. Cioni 46, G. Comoretto80, W.J. Cooper 81, 22, T. Cornez60,
S. Cowell8, F. Crifo5, M. Crosta 22, C. Crowley73, C. Dafonte 63, A. Dapergolas33, M. David 82, P. David34, P.
de Laverny7, F. De Luise 83, R. De March 27, J. De Ridder 61, R. de Souza84, P. de Teodoro13, A. de Torres73,
E.F. del Peloso6, E. del Pozo10, M. Delbo 7, A. Delgado8, H.E. Delgado 58, J.-B. Delisle 9, P. Di Matteo5, S.
Diakite85, C. Diener8, E. Distefano 35, C. Dolding21, D. Eappachen86, 62, B. Edvardsson87, H. Enke 46, P. Esquej
88, C. Fabre89, M. Fabrizio 51, 52, S. Faigler90, G. Fedorets53, 91, P. Fernique 42, 92, A. Fienga 93, 34, F. Figueras
11, C. Fouron76, F. Fragkoudi94, E. Fraile88, F. Franke95, M. Gai 22, D. Garabato 63, A. Garcia-Gutierrez11, M.
García-Torres 96, A. Garofalo 24, P. Gavras 88, E. Gerlach 12, R. Geyer 12, P. Giacobbe22, G. Gilmore 8, S.
Girona 64, G. Giuffrida51, R. Gomel90, A. Gomez 63, I. Gonzalez-Santamaria 63, J.J. González-Vidal11, M.
Granvik 53, 97, R. Gutiérrez-Sánchez31, L.P. Guy 28, 80, M. Hauser20, 98, M. Haywood 5, A. Helmi 67, S.L.
Hidalgo 99, 100, T. Hilger 12, N. Hładczuk13, D. Hobbs 14, G. Holland8, H.E. Huckle21, G. Jasniewicz101, P.G.
Jonker 62, 86, J. Juaristi Campillo6, F. Julbe11, L. Karbevska9, P. Kervella 102, S. Khanna 67, A. Kochoska 103,
M. Kontizas 44, G. Kordopatis 7, A.J. Korn 45, Z. Kostrzewa-Rutkowska1, 86, K. Kruszy´nska 104, S. Lambert
32, A.F. Lanza 35, Y. Lasne60, J.-F. Le Campion105, Y. Le Fustec76, Y. Lebreton 102, 106, T. Lebzelter 65, S.
Leccia 107, N. Leclerc5, I. Lecoeur-Taibi 28, S. Liao22, E. Licata 22, H.E.P. Lindstrøm22, 108, T.A. Lister 109,
E. Livanou44, A. Lobel26, P. Madrero Pardo11, S. Managau60, R.G. Mann 39, J.M. Marchant110, M. Marconi 107,
M.M.S. Marcos Santos36, S. Marinoni 51, 52, F. Marocco 111, 112, D.J. Marshall113, L. Martin Polo36, J.M.
Martín-Fleitas 10, A. Masip11, D. Massari 24, A. Mastrobuono-Battisti 14, T. Mazeh 90, P.J. McMillan 14, S.
Article number, page 1 of 21
arXiv:2012.01533v1[astro-ph.GA]2Dec2020
A&A proofs: manuscript no. aa202039657
Messina 35, D. Michalik 3, N.R. Millar8, A. Mints 46, D. Molina 11, R. Molinaro 107, L. Molnár
114, 115, 116, P. Montegriffo24, R. Mor 11, R. Morbidelli 22, T. Morel37, D. Morris39, A.F. Mulone27, D.
Munoz60, T. Muraveva 24, C.P. Murphy13, I. Musella 107, L. Noval60, C. Ordénovic7, G. Orrù27, J. Osinde88, C.
Pagani68, I. Pagano 35, L. Palaversa117, 8, P.A. Palicio 7, A. Panahi 90, M. Pawlak 118, 104, X. Peñalosa
Esteller11, A. Penttilä 53, A.M. Piersimoni 83, F.-X. Pineau 42, E. Plachy 114, 115, 116, G. Plum5, E. Poggio 22,
E. Poretti 119, E. Poujoulet120, A. Prša 103, L. Pulone 51, E. Racero36, 121, S. Ragaini24, M. Rainer 18, C.M.
Raiteri 22, N. Rambaux34, P. Ramos 11, M. Ramos-Lerate122, P. Re Fiorentin 22, S. Regibo61, C. Reylé55, V.
Ripepi 107, A. Riva 22, G. Rixon8, N. Robichon 5, C. Robin60, M. Roelens 9, L. Rohrbasser28, M.
Romero-Gómez 11, N. Rowell39, F. Royer 5, K.A. Rybicki 104, G. Sadowski16, A. Sagristà Sellés 6, J.
Sahlmann 88, J. Salgado 31, E. Salguero40, N. Samaras 26, V. Sanchez Gimenez11, N. Sanna18, R. Santoveña
63, M. Sarasso 22, M. Schultheis 7, E. Sciacca 35, M. Segol95, J.C. Segovia36, D. Ségransan 9, D.
Semeux89, S. Shahaf 90, H.I. Siddiqui 123, A. Siebert 42, 92, L. Siltala 53, E. Slezak7, R.L. Smart 22, E.
Solano124, F. Solitro27, D. Souami 102, 125, J. Souchay32, A. Spagna 22, F. Spoto 72, I.A. Steele 110, H.
Steidelmüller12, C.A. Stephenson31, M. Süveges28, 126, 20, L. Szabados 114, E. Szegedi-Elek 114, F. Taris32, G.
Tauran60, M.B. Taylor 127, R. Teixeira 84, W. Thuillot34, N. Tonello 64, F. Torra 25, J. Torra†11, C. Turon 5,
N. Unger 9, M. Vaillant60, E. van Dillen95, O. Vanel5, A. Vecchiato 22, Y. Viala5, D. Vicente64, S. Voutsinas39,
M. Weiler11, T. Wevers 8, Ł. Wyrzykowski 104, A. Yoldas8, P. Yvard95, H. Zhao 7, J. Zorec128, S. Zucker
129, C. Zurbach130, and T. Zwitter 131
(Affiliations can be found after the references)
Received ; accepted
ABSTRACT
Context. We present the early installment of the third Gaia data release, Gaia EDR3, consisting of astrometry and photometry for 1.8 billion
sources brighter than magnitude 21, complemented with the list of radial velocities from Gaia DR2.
Aims. A summary of the contents of Gaia EDR3 is presented, accompanied by a discussion on the differences with respect to Gaia DR2 and an
overview of the main limitations which are present in the survey. Recommendations are made on the responsible use of Gaia EDR3 results.
Methods. The raw data collected with the Gaia instruments during the first 34 months of the mission have been processed by the Gaia Data
Processing and Analysis Consortium (DPAC) and turned into this early third data release, which represents a major advance with respect to
Gaia DR2 in terms of astrometric and photometric precision, accuracy, and homogeneity.
Results. Gaia EDR3 contains celestial positions and the apparent brightness in G for approximately 1.8 billion sources. For 1.5 billion of those
sources, parallaxes, proper motions, and the (GBP − GRP) colour are also available. The passbands for G, GBP, and GRP are provided as part of the
release. For ease of use, the 7 million radial velocities from Gaia DR2 are included in this release, after the removal of a small number of spurious
values. New radial velocities will appear as part of Gaia DR3. Finally, Gaia EDR3 represents an updated materialisation of the celestial reference
frame (CRF) in the optical, the Gaia-CRF3, which is based solely on extragalactic sources. The creation of the source list for Gaia EDR3 includes
enhancements that make it more robust with respect to high proper motion stars, and the disturbing effects of spurious and partially resolved
sources. The source list is largely the same as that for Gaia DR2, but it does feature new sources and there are some notable changes. The source
list will not change for Gaia DR3.
Conclusions. Gaia EDR3 represents a significant advance over Gaia DR2, with parallax precisions increased by 30 per cent, proper motion
precisions increased by a factor of 2, and the systematic errors in the astrometry suppressed by 30–40% for the parallaxes and by a factor ∼ 2.5
for the proper motions. The photometry also features increased precision, but above all much better homogeneity across colour, magnitude, and
celestial position. A single passband for G, GBP, and GRP is valid over the entire magnitude and colour range, with no systematics above the 1%
level.
Key words. catalogs - astrometry - parallaxes - proper motions - techniques: photometric - techniques: radial velocities
1. Introduction
We present the first installment of the third intermediate Gaia
data release, Gaia Early Data Release 3 (Gaia EDR3), which
is based on the data collected during the first 34 months of the
mission. This early part of Gaia DR3 consists of an updated
source list, astrometry, and broad band photometry in the G,
GBP, and GRP bands. In addition, an updated list of radial velocities
from Gaia DR2, cleaned from spurious values, is included.
Gaia EDR3 represents a significant improvement in both the
precision and accuracy of the astrometry and broad-band photometry.
The factor two improvement in proper motion precision
provides new views of the fine structure of Galactic phase
space. The suppression of systematic errors enables, for the first
time, a measurement in the optical of the acceleration of the solar
system barycentre with respect to the rest frame of distant
extragalactic sources (Gaia Collaboration et al. 2020c), which
is a beautiful confirmation of the superb astrometric quality of
Gaia EDR3. Likewise, the photometry is significantly improved
over Gaia DR2; it is much more homogeneous over the sky, as
well as over source brightness and colour, where a single passband
for G, GBP, and GRP can now be used over the entire magnitude
and colour range, with no systematics above the 1% level.
The full Gaia DR3 release is expected in 2022 and will enrich
the current release with the following: updated and new radial
velocities; astrophysical parameters for sources based on the
blue and red prism photometer (BP and RP) spectra, as well as
spectra from the radial velocity spectrograph (RVS); the mean
BP and RP prism- and RVS-spectra for a subset of sources; an
Article number, page 2 of 21
Gaia Collaboration et al.: Gaia Early Data Release 3
extended catalogue of variable stars; the first catalogue of binary
stars, including eclipsing, spectroscopic, and astrometric binaries;
astrometry for a larger sample of solar system objects, and
reflectance spectra for a small subset of asteroids; quasi-stellar
object (QSO) host and galaxy morphological characterisation;
and the light curves for all sources in a field centred on M31. We
stress here that the source list, astrometry, and broad-band photometry
will not be updated from Gaia EDR3 to Gaia DR3, both
releases being based on the same number of input observations.
This paper is structured as follows. In Sect. 2 we provide a
short overview of the improvements and additions to the data
processing that led to the production of Gaia EDR3. We summarise
the contents of the early installment of the third data
release in Sect. 3 and comment on the quality of this release
in Sect. 4. In Sect. 5 we discuss the major differences between
Gaia EDR3 and Gaia DR2. In Sect. 6 we comment on the completeness
of Gaia EDR3 and some of the known limitations
which the user of the data should keep in mind. Additional guidance
on the use of Gaia EDR3 is provided in Sect. 7. In Sect. 8
we provide updates to the Gaia data access facilities and documentation
available to the astronomical community. Gaia started
collecting scientific data in July 2014 (Gaia Collaboration et al.
2016b) and is currently in its extended mission phase, the nominal
60 month mission having been concluded on July, 16 2019.
We conclude with a look ahead at the extended Gaia mission and
the next data releases in Sect. 9. Throughout the paper we make
reference to other Gaia Collaboration and Gaia Data Processing
and Analysis Consortium (DPAC) papers that provide more details
on the data processing and validation for Gaia EDR3. All
these papers (together with the present article) can be found in
the Astronomy & Astrophysics Special issue on Gaia EDR3.
2. Data processing for Gaia EDR3
As described in detail in Gaia Collaboration et al. (2016b), Gaia
measurements are collected with three instruments. The astrometric
instrument collects images in Gaia’s white-light G-band
(330–1050 nm); the blue (BP) and red (RP) prism photometers
collect low resolution spectrophotometric measurements of
source spectral energy distributions over the wavelength ranges
330–680 nm and 640–1050 nm, respectively; and the radial
velocity spectrometer (RVS) collects medium resolution (R ∼
11 700) spectra over the wavelength range 845–872 nm centred
on the Calcium triplet region (Cropper et al. 2018).
We repeat here for convenience the way events on board
Gaia, including the data collection, are timed. The times are
given in terms of the on board mission timeline (OBMT) which
is generated by the Gaia on board clock. By convention OBMT
is expressed in units of six hour (21 600 s) spacecraft revolutions
(Gaia Collaboration et al. 2016b). The approximate relation between
OBMT (in revolutions) and barycentric coordinate time
(TCB, in Julian years) at Gaia is
TCB J2015.0+(OBMT−1717.6256 rev)/(1461 rev yr−1
) . (1)
The 34 month time interval covered by the observations
used for Gaia EDR3 starts at OBMT 1078.3795 rev
= J2014.5624599 TCB (approximately July 25, 2014,
10:30:00 UTC), and ends at OBMT 5230.0880 rev =
J2017.4041495 TCB (approximately May 28, 2017,
08:45:00 UTC). This time interval contains gaps caused
by both spacecraft events and by on-ground data processing
problems. This leads to gaps in the data collection or stretches
of time over which the input data cannot be used. Which data
are considered unusable varies across the Gaia data processing
systems (here astrometry and photometry), and as a consequence
the effective amount of input data used differs from one system
to the other. We refer to the specific Gaia EDR3 data processing
papers (listed below) for the details.
The pre-processing for all Gaia instruments is described in
Hambly et al. (2018) and includes the removal of the effects of
non-uniformity of the charge-coupled device (CCD) bias levels.
A summary of the entire data processing system for Gaia
is given in Gaia Collaboration et al. (2016b). The sub-sections
below summarise the major improvements in the data processing
for Gaia EDR3 with respect to Gaia DR2.
2.1. Source list
A given data processing cycle for Gaia starts with the creation
of the list of sources that will be treated. The series of CCD measurements
recorded as a source travels across the focal plane (referred
to collectively as a ‘transit’, Torra et al. 2020), are grouped
and assigned to known sources on the sky or to newly ‘created’
sources, corresponding to groups of transits at a celestial position
where previously no source was catalogued. The starting
point for creating the source list is the previous Gaia data release,
or the Initial Gaia Source List (Smart & Nicastro 2014)
in the case of Gaia DR1. As pointed out in Gaia Collaboration
et al. (2018) the source list may evolve from one release to the
next due to the merging of groups of transits previously assigned
to two or more sources, the splitting of a group of transits into
two or more sources, or changing the list of transits assigned
to a source. The changes in the source list from Gaia DR1 to
Gaia DR2 were significant but from Gaia DR2 to Gaia EDR3
the source list has largely stabilised, the changes being at the 2–3
per cent level overall.
A full account of how the source list is created for
Gaia EDR3 (and Gaia DR3) can be found in Torra et al. (2020),
who note the following significant improvements with respect
to Gaia DR2. The identification of spurious on-board detections
(caused among others by bright star diffraction spikes,
bright cosmic rays, or major planets in the solar system transiting
across or near a telescope field of view) has been improved
which leads to a much cleaner list of sources and associated tran-
sits.
The algorithm that groups together transits and assigns them
to sources has been improved with respect to the treatment of
high-proper motion stars and variable stars. High proper motion
stars are seen as groups of observations stretched out over
the sky which were mistaken for multiple sources in previous
releases. They are now reliably recognised as belonging to the
same source. The grouping of transits and their association to
sources contains a stage where the magnitude of the source observed
during a transit is taken into account. For highly variable
sources this can lead to a splitting of the transits over multiple
sources. This type of error is now prevented through a postprocessing
step which can recognise clusters of detections very
close together on the sky, but disjoint in time, as belonging to the
same (variable) source.
A more comprehensive analysis and cleaning of the
observation-to-source matching results led to less sources with
highly significant negative parallaxes or too large parallaxes (see
appendix C in Lindegren et al. 2018), which also removes spuriously
high proper motions. The treatment of close source pairs
was improved to deal with the pairs with separations below
400 mas which were erroneously considered duplicate sources
in Gaia DR2. These now appear as two sources in Gaia EDR3.
Article number, page 3 of 21
A&A proofs: manuscript no. aa202039657
The separation limit below which two sources are considered
duplicates was lowered to 180 mas.
For quantitative information on the above please refer to
Torra et al. (2020). We stress here that the source list for
Gaia EDR3 and Gaia DR3 will be the same and thus the process
described in Torra et al. (2020) applies to both releases.
2.2. Astrometric data processing improvements
The astrometric processing for Gaia EDR3 is described in Lindegren
et al. (2020b), who note the following major improvements
with respect to the processing for Gaia DR2. The basic
inputs for the Astrometric Global Iterative Solution (AGIS; Lindegren
et al. 2012) are the source image locations in the Gaia
CCD pixel stream, translated to observation times and acrossscan
locations. The image locations are determined in a process
called ‘image parameter determination’ (IPD; Rowell et al.
2020), by fitting a point spread function (PSF) or line spread
function (LSF) to the 2D or 1D observation windows containing
the image samples. The modelling of the PSF and LSF has been
very much improved with the introduction of time and source
colour dependencies, among many other enhancements. The details
are described in Rowell et al. (2020). It is shown in Lindegren
et al. (2020b) that for sources fainter than G = 13 the chromatic
effects on image locations are almost completely removed
during the image parameter determination stage, thus mostly
eliminating an important source of systematic errors. Eventually
for future Gaia data releases this should remove the need
for colour terms in the astrometric calibrations. The colour of a
source is included in the form of the effective wave number νeff,
which is derived from the flux as a function of wavelength in the
BP and RP prism spectra.
For the first time the image parameter determination and astrometric
solution were iterated. Specifically, after the first IPD
run an AGIS solution was produced, which was then used to improve
the PSF and LSF calibrations for a second round of IPD
(which benefits from improved source positions from the first
AGIS run), followed by the Gaia EDR3 production run of AGIS.
This leads to a more self-consistent set of image locations and
source astrometric parameters. In addition, the image fluxes estimated
as part of the IPD are improved, which in turn improves
the G-band photometric processing. A diagram illustrating this
data processing flow can be found in Lindegren et al. (2020b).
The astrometric calibration model was improved and extended.
Several new dependencies were introduced to better handle
the locations of saturated images, the effects of charge transfer
inefficiency, and imperfections in the PSF and LSF models
that leave residual effects at the sub-pixel level, and as a function
of the rate at which sources move across the telescope field
of view, perpendicular to the scan direction (caused by the spin
axis precession of Gaia).
The Gaia EDR3 astrometric processing includes a model
(Velocity error and effective Basic Angle Calibration, VBAC)
that can calibrate out the effects of spin-related instrument distortions,
in particular distortions over time scales comparable to the
six hour spin period of Gaia. An important component of these
distortions are the basic angle variations, of which the term proportional
to the cosine of the spacecraft spin phase Ω can lead to
a global parallax bias if left uncorrected (Butkevich et al. 2017).
The VBAC model introduces additive corrections to the basic
angle variation corrections calculated on the basis of the basic
angle monitor measurements (cf. Lindegren et al. 2018). Lindegren
et al. (2020b) describe the successful attempt to fit the
coefficient of the cos Ω term, which leads to a reduction in the
size of the overall parallax zero point, which for Gaia EDR3 is
−17 µas (compared to −29 µas for Gaia DR2), as estimated from
quasar parallaxes.
An additional calibration model, which handles telescope focal
length and optical distortion variations over smaller scales
than handled by VBAC, was introduced to further reduce instrument
distortion related systematics. Lastly, an ad-hoc correction
was introduced to ensure that the bright star reference frame
has no net spin with respect to the reference frame defined by
quasars, an issue that was described in detail in Lindegren et al.
(2018) and Lindegren (2020a,b).
Overall the Gaia EDR3 astrometry shows a reduction in the
median uncertainties at G = 15 by a factor 0.71 for the positions
and parallaxes and 0.44 for the proper motions. At the
bright end (G < 12) the gain is larger, a factor 0.43 for positions
and parallaxes and 0.27 for the proper motions, thanks to much
improved calibrations. The overall parallax zero point has improved
as mentioned above, and the variance over the sky in the
systematic errors (as estimated from quasars) has been reduced
by 30–40% for the parallaxes and by a factor of ∼ 2.5 for proper
motions. Lindegren et al. (2020b,a) provide many more details
(see also Sect. 7.1).
2.3. Photometric data processing improvements
The photometric data processing for Gaia EDR3 is described in
Riello et al. (2020), where only the processing for the broad band
photometry is considered. The processing and calibration of the
spectra will be described in forthcoming papers (De Angeli et al.
2021; Carrasco et al. 2020; Montegriffo et al. 2021). The BP and
RP spectra are still being validated internally to DPAC at the time
of the Gaia EDR3 release, by employing them in the astrophysical
characterisation of sources. The spectra will be published
as part of Gaia DR3 (for a subset of sources only). Riello et al.
(2020) describe the following improvements to the broad-band
photometric processing.
The G-band photometry benefits from the improvements implemented
for the astrometric instrument image parameter determination
stage. As described in Rowell et al. (2020), this includes
a better PSF and LSF modelling, better treatment of saturated
images, the masking out of suspected disturbing sources
and a more precise determination of the background flux for each
observation window. This leads to more accurate and robust Gband
flux estimates.
The broad band photometry benefits from a detailed evaluation
of the observations potentially affected by neighbouring
sources in crowded fields. Although the crowding effects were
not corrected, the crowding evaluation led to a cleaner list of internal
calibration sources. The background flux in BP and RP
due to stray light and astronomical sources is better determined,
with higher spatial resolution to follow smaller scale variations.
The range of time over which observations free from telescope
throughput losses (due to contamination; Gaia Collaboration
et al. 2016b) are available was much extended. This allowed
for better sky coverage of internal calibrator sources and more
flexibility to select an optimal set of calibrators, well distributed
in colour and magnitude. The external calibration used to determine
the passbands is based on a much larger set of calibrators,
covering a wider spectral range, where in Gaia DR2 only
the spectrophotometric standard stars (Pancino et al. 2012) were
used which limited the aspects of the passbands that could be
determined reliably.
These improvements, and the larger set of input observations,
have led to broad-band photometry which is significantly
Article number, page 4 of 21
Gaia Collaboration et al.: Gaia Early Data Release 3
Table 1. Number of sources of a certain type, or the number of sources
for which a given data product is available in Gaia EDR3.
Data product or source type Number of sources
Total 1 811 709 771
5-parameter astrometry 585 416 709
6-parameter astrometry 882 328 109
2-parameter astrometry 343 964 953
Gaia-CRF3 sources 1 614 173
ICRF3 sources for frame orientation 2269
Gaia-CRF3 sources for frame spin 429 249
G-band 1 806 254 432
GBP-band 1 542 033 472
GRP-band 1 554 997 939
Radial velocity 7 209 831
Table 2. Distribution of the Gaia EDR3 sources in G-band magnitude.
The distribution percentiles are shown for all sources and for those with
a 5-p, 6-p, and 2-p astrometric solution, respectively, as well as the
sources for which a radial velocity is available in Gaia EDR3.
Magnitude distribution percentiles (G)
Percentile All 5-p 6-p 2-p vrad
0.135% 11.7 10.6 15.1 15.7 6.7
2.275% 15.1 13.7 17.4 18.8 9.2
15.866% 17.9 16.3 18.9 20.1 11.2
50% 19.7 18.1 19.9 20.7 12.4
84.134% 20.6 19.4 20.5 21.0 13.2
97.725% 21.0 20.4 20.8 21.2 14.1
99.865% 21.5 20.8 20.9 21.7 15.1
better than Gaia DR2 photometry in both precision and accuracy.
Especially at the bright end (G < 13) large gains were made, and
many of the systematic effects reported for Gaia DR2, such as
imprints from the zodiacal light and the scan law features (Evans
et al. 2018; Arenou et al. 2018), have been removed or greatly
suppressed (cf. Fabricius et al. 2020). In addition the problem
that two passbands are needed to describe the Gaia DR2 G-band
photometry (Maíz Apellániz & Weiler 2018) has been resolved,
with only one passband needed for Gaia EDR3 for each of G,
GBP, and GRP.
Despite the improvements in the astrometry and photometry,
several limitations remain in Gaia EDR3 which require taking
care when using the data. In Sect. 6 we summarise the known
limitations of the present Gaia data release and point out, where
relevant, the causes. In Sect. 7, and in Lindegren et al. (2020b)
(astrometry), Riello et al. (2020) (photometry), and Fabricius
et al. (2020) (catalogue validation) we provide additional guidance
on the use of Gaia EDR3 results. The reader is strongly
encouraged to read these papers and the online documentation1
to understand the limitations in detail.
3. Overview of the contents of Gaia EDR3
Gaia EDR3 contains new astrometry and broad-band photometry,
as well as radial velocities from Gaia DR2. Basic statistics
on the source numbers and the overall distribution in G can
1
http://gea.esac.esa.int/archive/documentation/GEDR3/
index.html
be found in Tables 1 and 2. The overall quality of Gaia EDR3
results in terms of the typically achieved uncertainties is summarised
in Table 3. The contents of the main components of
the release, of which the magnitude distributions are shown in
Figs. 1 and 2, are summarised in the following.
3.1. Astrometric data set
The astrometric data in Gaia EDR3 comprises three subsets:
5-parameter solutions (5-p) For these sources the colour information
(νeff derived during the processing for Gaia DR2) was
of high enough quality to assume that any chromatic effects
were removed during the IPD stage, thus allowing for a standard
5-parameter (α, δ, , µα∗, µδ) astrometric solution.
6-parameter solutions (6-p) For these sources no colour information
of sufficient quality was available, thus forcing an
estimate of the image locations with a PSF or LSF model for
a default source colour. This means that chromatic effects
have to be accounted for in the astrometric solution by estimating
νeff for the source along with the astrometric parameters.
Thus for these sources the 5 astrometric parameters and
the so-called pseudo-colour are published along with a 6 × 6
covariance matrix listing, in addition to the astrometric covariance
matrix entries, the uncertainty on the pseudo-colour
and the correlations between the estimated colour and the
astrometric parameters.
2-parameter solutions (2-p) As for previous releases there are
sources for which insufficient data of the required quality is
available to justify publishing a full 5-p or 6-p solution. For
these sources a 5- or 6-parameter solution is in fact made, but
with so-called galactic priors on the parallaxes and proper
motions (Michalik et al. 2015; Lindegren et al. 2018, the fallback
solution). Only the positions and their covariance matrix
are published for these sources. In principle all sources
at G > 21 have a 2-p solution. However this boundary is
not strict because the limit in G was decided on the basis of
the Gaia DR2 value for the magnitude or that based on the
on-board estimate. Hence there are 5-p and 6-p solutions at
G > 21, and sources with 2-p solutions at G ≤ 21 for which
in principle a 5-p or 6-p solution was possible. See Lindegren
et al. (2020b) for details on how the decision was taken
to use the fall-back solution.
The three solution types can be identified through the
astrometric_params_solved field in the Gaia EDR3 archive
which equals 3, 31, and 95, respectively for 2-p, 5-p, and 6-p
astrometric solutions. Fig. 2 shows the distribution of the three
solution types over magnitude. We note the prevalence of 6-p
solutions at G < 4 and the relative increase in 6-p solutions
around G = 11. In both cases this reflects the source list evolution
at these magnitudes, where for a large fraction of sources the
change in source identifier (ID) prevented looking up the colour
calculated for the corresponding source in Gaia DR2. The 100%
fraction of sources with 2-p solutions at G > 21 is by construction
(Lindegren et al. 2020b).
Along with the astrometry, new data quality indicators are
published as part of Gaia EDR3. The renormalised unit weight
error, introduced after Gaia DR2 was published (RUWE; Lindegren
2018)2
, is now part of the release. New quality indicators,
related to the image parameter determination process, provide
2
https://www.cosmos.esa.int/web/gaia/
dr2-known-issues#AstrometryConsiderations
Article number, page 5 of 21
A&A proofs: manuscript no. aa202039657
5 10 15 20
Mean G [mag]
100
102
104
106
108
Numberper0.1magbin
DR2
EDR3
DR2 vrad
EDR3 vrad
GCRF3
GCRF3 alignment
GCRF3 spin
Fig. 1. Distribution of the mean values of G
for all Gaia EDR3 sources shown as histograms
with 0.1 mag wide bins. The distribution of the
Gaia DR2 sources (dashed lines, for the full
catalogue and for the radial velocity sample) is
included for comparison. The other histograms
are for the main Gaia EDR3 components as indicated
in the legend.
5 10 15 20
Mean G [mag]
0.0
0.2
0.4
0.6
0.8
1.0
Fractionper0.1magbin
5-parameter astrometry
6-parameter astrometry
2-parameter astrometry
Fig. 2. Distribution of the fraction of astrometric solution types as a
function of G-band magnitude.
indications whether a source is one of a close pair (possibly a binary)
or whether the data suffers from nearby disturbing sources.
The indicators are as follows:
ipd_gof_harmonic_amplitude This statistic measures the
amplitude of the variation of the image parameter determination
goodness of fit (reduced χ2
) as function of the position
angle of the scan direction. A large amplitude indicates
that the source is double, in which case the phase (next item)
indicates the position angle of the pair, modulo 180 degrees.
ipd_gof_harmonic_phase This statistic measures the phase
of the variation of the IPD goodness of fit as function of the
position angle of the scan direction
ipd_frac_multi_peak This field provides information on the
observation windows used for the astrometric processing of
this source. It provides the fraction of windows for which
the IPD algorithm has identified a double peak, meaning that
the detection may be a resolved double star (either an optical
pair or a physical binary).
ipd_frac_odd_win Percentage of transits with truncated windows
or multiple gates applied to a window. A high percentage
indicates that a source is disturbed due to nearby sources
in a crowded field or due to a nearby bright (G < 13) source.
The reference epoch for all sources is J2016.0 (TCB). This
epoch near the middle of the observation interval included in
Gaia EDR3 was chosen to minimise correlations between the
position and proper motion parameters. This epoch is 0.5 year
later than the reference epoch for Gaia DR2, which should be
accounted for when comparing source positions between the two
releases.
As in previous releases all sources were treated as single stars
when solving for the astrometric parameters. For a binary the
parameters may thus refer to either component, or to the photocentre
of the system, and the proper motion represents the mean
motion of the component, or photocentre, over the 2.8 years of
data included in the solution. Depending on the orbital motion,
this could be significantly different from the proper motion of
the same object in Gaia DR2, and significantly different from
the proper motion of the centre of mass of the binary.
The positions and proper motions are provided with respect
to the third realisation of the Gaia celestial reference frame
(Gaia-CRF3) which is aligned with the International Celestial
Reference Frame (ICRF) to about 0.01 mas root-mean-square
(RMS) at epoch J2016.0 (TCB), and non-rotating with respect to
the ICRF to within 0.01 mas yr−1
RMS. The Gaia-CRF3 is materialised
by 1 614 173 QSOs and aligned to the ICRF through
a subset of 2269 QSOs. The construction and properties of the
Gaia-CRF3 and the comparison to the ICRF are described in
Gaia Collaboration et al. (2020b).
3.2. Photometric data set
The photometric data set contains the broad band photometry in
the G, GBP, and GRP bands. The mean value of the G-band fluxes
is reported for practically all sources while for about 85 per cent
of the sources the mean values of the GBP and GRP fluxes are provided.
For a small fraction of the sources any of the three bands
may be missing (see Sect. 6.3.2). As for Gaia DR2, the photometric
data processing consisted of three categories, ‘Gold’,
‘Silver’, and ‘Bronze’, which represent decreasing quality levels
of the photometric calibration achieved, where in the case of the
Bronze sources no colour information is available. The photometric
processing category of each source is indicated in the released
catalogue by a numeric field (phot_proc_mode) assuming
values 0, 1, and 2 for gold, silver, and bronze sources respectively.
At the bright end the photometric uncertainties are dominated
by calibration effects which are estimated to contribute
2.0, 3.1, and 1.8 mmag RMS per CCD observation, respectively
for G, GBP, and GRP (Riello et al. 2020).
Two new data quality indicators are included with the
photometry which allow filtering of sources according to
the probability that their photometry is affected by crowding
effects. The fields phot_bp_n_blended_transits and
Article number, page 6 of 21
Gaia Collaboration et al.: Gaia Early Data Release 3
Table 3. Basic performance statistics for Gaia EDR3. The astrometric uncertainties and the Gaia-CRF3 alignment and rotation limits refer to
epoch J2016.0 TCB.
Data product or source type Typical uncertainty
Five-parameter astrometry (position) 0.01–0.02 mas at G < 15
0.05 mas at G = 17
0.4 mas at G = 20
1.0 mas at G = 21
Five-parameter astrometry (parallax) 0.02–0.03 mas at G < 15
0.07 mas at G = 17
0.5 mas at G = 20
1.3 mas at G = 21
Five-parameter astrometry (proper motion) 0.02–0.03 mas yr−1
at G < 15
0.07 mas yr−1
at G = 17
0.5 mas yr−1
at G = 20
1.4 mas yr−1
at G = 21
Six-parameter astrometry (position) 0.02–0.03 mas at G < 15
0.08 mas at G = 17
0.4 mas at G = 20
1.0 mas at G = 21
Six-parameter astrometry (parallax) 0.02–0.04 mas at G < 15
0.1 mas at G = 17
0.5 mas at G = 20
1.4 mas at G = 21
Six-parameter astrometry (proper motion) 0.02–0.04 mas yr−1
at G < 15
0.1 mas yr−1
at G = 17
0.6 mas yr−1
at G = 20
1.5 mas yr−1
at G = 21
Two-parameter astrometry (position only) 1–3 mas
Systematic astrometric errors (averaged over the sky) < 0.05 mas
Gaia-CRF3 alignment with ICRF 0.01 mas at G = 19
Gaia-CRF3 rotation with respect to ICRF < 0.01 mas yr−1
at G = 19
Mean G-band photometry 0.3 mmag at G < 13
1 mmag at G = 17
6 mmag at G = 20
Mean GBP-band photometry 0.9 mmag at G < 13
12 mmag at G = 17
108 mmag at G = 20
Mean GRP-band photometry 0.6 mmag at G < 13
6 mmag at G = 17
52 mmag at G = 20
phot_bp_n_contaminated_transits (and similar for RP)
indicate the number of transits for a given source believed to be
affected by ‘blending’ or ‘contamination’. The former refers to
the presence of another source in the same observation window
or very nearby, and the latter refers to the presence of sources in
the wider neighbourhood of the target source, which are bright
enough to contribute flux to the observation window. The ratio
between the numbers in these fields and the total number of transits
(phot_bp_n_obs or phot_rp_n_obs) can be used to identify
sources for which the photometry is possibly less reliable
(see Riello et al. 2020, for a more detailed discussion).
3.3. Radial velocity data set
The radial velocity spectrograph data processing relies on the
preliminary astrometric solution (AGIS-3.1 in Lindegren et al.
2020b) in order to have sufficiently accurate source positions to
fix the correct wavelength scale of the RVS spectra. This means
that the RVS processing can only start later during a processing
cycle. In the planning of the Gaia EDR3 and Gaia DR3 releases
it was not possible to accommodate the full sequence of
RVS processing, radial velocity estimation, and validation, to allow
new radial velocities to be published as part of Gaia EDR3.
These will instead appear with Gaia DR3 in 2022. At the time of
writing, the Gaia DR3 RVS processing is finished and the results
are undergoing internal validation through the use of epoch radial
velocities in the non-single star pipeline, and of RVS spectra
in the astrophysical characterisation of Gaia sources.
In order to keep Gaia EDR3 maximally useful it was decided
to copy Gaia DR2 radial velocities to this release. The evolution
of the source list (Torra et al. 2020) necessitated a careful tracing
of Gaia DR2 sources to their Gaia EDR3 counterparts, before
assigning the DR2 radial velocities to the latter. The opportunity
was used to also clean up the list of radial velocities from the
Article number, page 7 of 21
A&A proofs: manuscript no. aa202039657
potentially spurious values identified by Boubert et al. (2019).
The process is described in detail in Seabroke et al. (2020).
The result is that 7 209 831 out of 7 224 631 Gaia DR2 radial
velocities were transferred to Gaia EDR3, where 97% of the
Gaia EDR3 sources with a radial velocity have the same source
ID as in Gaia DR2. The radial velocities that were discarded
correspond to sources that could not be traced to Gaia EDR3
or were shown, from a comparison with Gaia DR3 preliminary
radial velocities and investigations of the raw observations, to
have unreliable or erroneous radial velocities. The histograms
presented in Seabroke et al. (2020) show that in particular in the
tails of the distribution (at |vrad| > 600 km s−1
) radial velocities
have been removed. It is also noteworthy that of the ∼ 70 000
potentially spurious radial velocities identified by Boubert et al.
(2019), 96% were retained as reliable in Gaia EDR3. We note
that all radial velocity related fields in the Gaia EDR3 archive
are prefixed with ‘dr2_’, leading to dr2_radial_velocity,
dr2_radial_velocity_error, etc. For the detailed characteristics
of the radial velocity data set (precision, accuracy, limitations)
please refer to the relevant Gaia DR2 papers (Sartoretti
et al. 2018; Katz et al. 2019; Gaia Collaboration et al. 2018).
The distribution of the astrometric, photometric, and radial
velocity data sets in G is shown in Fig. 1, where for comparison
the distribution for Gaia DR2 is also shown. We note the
improved completeness at the faint end, at magnitudes close to
G = 21. With respect to Gaia DR2 there are noticeably fewer
radial velocities at G > 15 and at G < 4, the latter due to the
source list evolution at the bright end. The distribution of the
Gaia-CRF3 sources shows a sharp drop at G = 21 which reflects
that only QSOs at G < 21 were used for the construction
of the reference frame. The magnitude distribution is also shown
for the Gaia-CRF3 sources used for ensuring that the spin of the
reference frame is zero, and for the Gaia-CRF3 sources used for
the alignment of the reference frame (2269 ICRF3 sources with
counterparts in Gaia EDR3).
4. Scientific performance and potential of
Gaia EDR3
Gaia EDR3 is accompanied by four papers that provide basic
demonstrations of the scientific quality of the results included in
this release. The following topics are treated. The Gaia EDR3
proper motions of quasars reveal a systematic pattern that can
be ascribed to the acceleration of the solar system barycentre
with respect to the rest frame of distant extragalactic sources.
The value and direction of the acceleration can be determined
from these data. That this measurement is now possible is testament
to the much improved quality of the astrometry, in particular
thanks to the suppression of systematic errors (Gaia Collaboration
et al. 2020c). The Gaia catalogue of nearby stars presents
and characterises a carefully selected sample of sources located
within 100 pc from the sun (Gaia Collaboration et al. 2020e). A
study of the Galactic anti-centre region illustrates the increased
richness in phase space unveiled by the more precise Gaia EDR3
proper motions (Gaia Collaboration et al. 2020a). The structure
and properties of the Magellanic Clouds are studied in Gaia Collaboration
et al. (2020d).
We note in the following a couple of specific areas of improvement
in the Gaia EDR3 astrometry and photometry compared
to Gaia DR2. For further insights into the increased data
quality see Fabricius et al. (2020) and the papers cited above.
The overall completeness of the catalogue at the faint end
has increased somewhat as can be appreciated from Fig. 1. At
the bright end Gaia EDR3 has a similar level of incompleteness
as Gaia DR2. The spatial resolution of Gaia EDR3 has slightly
improved with respect to Gaia DR2. This can be seen from the
number counts of source pairs as a function of their separation
presented in Fabricius et al. (2020). The counts drop below the
expected line for a random source distribution at ∼ 1.5 arcsec
(2.2 arcsec for Gaia DR2). The improved resolution can
also be seen in the increased completeness of visual pairs from
the Washington Double Star catalogue (Fabricius et al. 2020).
The completeness in close source pairs decreases rapidly below
about 0.7 arcsec which is to be expected as the treatment
of sources in crowded fields has not fundamentally changed between
Gaia DR2 and Gaia EDR3, even if the crowding effects
were better characterised during the data processing.
The better treatment of high proper motion stars at the source
list creation stage and the generally cleaner source list has led
to a much more reliable sample of high proper motion stars
in Gaia EDR3. This is demonstrated in Fabricius et al. (2020)
where the diagram of proper motion vs. parallax shows a striking
improvement from Gaia DR2 to Gaia EDR3, with most of
the high proper motion stars now located on the positive parallax
side of the 500 km s−1
tangential velocity locus. The suppression
of spurious parallaxes has also removed a lot of unrealistically
high proper motions for stars with negative parallaxes.
5. Changes from Gaia DR2 to Gaia EDR3
In Gaia Collaboration et al. (2018) it was emphasised that
Gaia DR2 should be treated as independent from Gaia DR1 due
to the evolution of the source list and the photometric system.
This still holds when using Gaia EDR3 instead of Gaia DR2,
the two releases should be treated as independent and in particular
the user of the data is warned against making source
by source comparisons between the two releases. Comparison
should only be done at the statistical level for well defined samples
of sources.
We repeat here the change in reference epoch from J2015.5
for Gaia DR2 to J2016.0 for Gaia EDR3, which should be taken
into account when propagating Gaia EDR3 source positions into
the future or past. The photometric system has changed in terms
of the better internal calibration leading to much smaller magnitude
and colour terms, and a new set of passbands is presented
in Riello et al. (2020). Synthetic photometry for the prediction
of Gaia observations should be updated with the new passbands.
5.1. Source list evolution
As stressed in Torra et al. (2020) the source lists of the Gaia data
releases should be treated as completely independent. Although
extensive efforts are made to ensure that a physical source retains
its identifier across releases, changes in the identifier associated
to a source will occur in a small fraction of cases. Ideally,
a given Gaia DR2 source with its associated transits appears in
Gaia EDR3 (and Gaia DR3) with the same source ID and the
same transits, plus the new transits added for Gaia EDR3. This
would allow a simple matching of the sources across the two
releases through the source ID. However, for the reasons mentioned
in Sect. 2.1 (and elaborated in Torra et al. 2020), this
should not be done, instead the DR2 to DR3 match table in
the Gaia archive (dr2_neighbourhood) should be used to trace
sources from Gaia DR2 to Gaia EDR3. This prevents mistakes
in cross-identifications of sources due to the source list evolu-
tion.
Article number, page 8 of 21
Gaia Collaboration et al.: Gaia Early Data Release 3
Another form of source list evolution is in the transits assigned
to a given source ID. For a small fraction of sources a significant
fraction of the transits that were assigned to the source
for Gaia DR2 may have been reassigned to other sources or
discarded altogether. This means that the source character may
change from Gaia DR2 to Gaia EDR3. The Gaia EDR3 archive
contains the following fields to help in understanding why a
source with the same source ID in Gaia DR2 and Gaia EDR3
may have significantly different source parameters (astrometry,
photometry, etc):
matched_transits The total number of field of view transits
m matched to a given source ID.
new_matched_transits The number of transits n newly included
in the transit list of an existing source ID.
matched_transits_removed The number of transits r removed
from the transit list of an existing source ID.
The fraction of transits retained for a source ID from Gaia DR2
to Gaia EDR3 is given by (m−n)/(m−n+r). Fig. 15 in Torra et al.
(2020) shows this fraction as a function of G-band magnitude,
which is lower than 100% for a significant number of source IDs
at G 12 and G 19. For source IDs with large changes in
the transit list one should be careful when making comparisons
between Gaia DR2 and Gaia EDR3.
As an example we look a bit more closely at the 2179 sources
for which the G-band magnitude in Gaia DR2 is less than 5.
These can be traced to Gaia EDR3 with the following Astronomical
Data Query Language (ADQL) query in the Gaia archive:
select dr2.source_id , dr2.phot_g_mean_mag ,
edr3.*
from gaiadr2.gaia_source as dr2
join gaiaedr3.dr2_neighbourhood as edr3
on dr2.source_id=edr3.dr2_source_id
where dr2.phot_g_mean_mag <5
order by dr2.source_id asc
This results in a list of 2208 matches from a Gaia DR2 to a
Gaia EDR3 source. In 34 cases there are two possible matches
in Gaia EDR3 for the Gaia DR2 source, where in all such cases
one of the matches is rather far (> 0.9 arcsec) from the target
position. This leaves 2174 ‘best’ matches for very bright
Gaia DR2 sources in Gaia EDR3, hence 5 sources in Gaia DR2
at GDR2 < 5 have no counterpart in Gaia EDR3. Of the matched
sources about 40% changed source ID, among which a prominent
example is β Pictoris (Gaia DR2 4792774797545105664
→ Gaia EDR3 4792774797545800832).
These drastic changes occur mostly at the bright end as
shown in Fig. 1 in Fabricius et al. (2020). For all other magnitudes
the changes in source ID occur for a few per cent of the
sources, except over the range 9 < G < 12 where up to almost
10 per cent of the sources changed ID.
6. Using Gaia EDR3 data: Completeness and
limitations
Gaia EDR3 represents a significant improvement in Gaia astrometry
and broad-band photometry, but as pointed out in Rowell
et al. (2020), Torra et al. (2020), Lindegren et al. (2020b),
and Riello et al. (2020), there are still many improvements to
be made to the data processing. This implies that there are limitations
which should be kept in mind when using Gaia EDR3
data. Next, we describe how the data were filtered between the
data processing and release publication stages and what the main
limitations are that the user should be aware of.
6.1. Gaia EDR3: Validation and source filtering
For details on how the quality of the Gaia EDR3 data were assessed
we refer to the astrometric and photometric processing
papers (Lindegren et al. 2020b; Riello et al. 2020) for validation
at the pipeline level, while a more global view of the data
quality is provided in Fabricius et al. (2020). Here we describe
the main filtering that was applied before accepting a source for
publication.
6.1.1. Astrometry
The filtering of the astrometric data set was very similar to the
procedure used for Gaia DR2. The results were filtered by requiring
that a source was observed by Gaia at least five times
(five focal plane transits), and that the semi-major axis of the
position uncertainty ellipse is less than 100 mas. In contrast to
Gaia DR2, no filtering on astrometric excess noise was done.
The parallax and proper motions are determined only for sources
satisfying the requirement that they are brighter than G = 21,
the number of ‘visibility periods’ used is at least 9 (a visibility
period represents a group of observations separated from other
such groups by at least four days), and the semi-major axis of
the 5-dimensional uncertainty ellipse is below a magnitude dependent
threshold. We refer to Lindegren et al. (2020b) for the
details. For sources that do not meet these requirements only the
positions are reported in Gaia EDR3. We remind the reader that
the sources with parallaxes and proper motions fall into the two
categories of 5-p and 6-p astrometry solutions (see Sect. 3.1).
For source pairs closer together than 0.18 arcsec only one source
was retained (detailed criteria in Lindegren et al. 2020b), which
is then marked as a duplicate_source in the Gaia EDR3
archive.
6.1.2. Photometry
Unlike the previous releases, sources were not discarded from
Gaia EDR3 if no G-band photometry was available. There are
in fact some 5.5 million sources in Gaia EDR3 without a value
for G in the published catalogue. For these sources the values
of G could only be estimated after the processing and validation
stages were finished and they will be provided through a separate
channel (see Sect. 7.2). The criteria to publish photometry
for a source are: the G-band is only provided for sources with
phot_g_n_obs ≥ 10; the GBP-band is only provided for sources
with phot_bp_n_obs ≥ 2; the GRP-band is only provided for
sources with phot_rp_n_obs ≥ 2. We note that any of the photometric
bands can be absent for a given source. No filtering on
the flux excess factor was done (in contrast to Gaia DR2).
6.2. Survey completeness
Fig. 1 shows that the completeness of the Gaia survey has improved
slightly with respect to Gaia DR2 at the faint end, between
G ≈ 19 and G ≈ 21. The fraction of bright stars missing
at G < 7 is unchanged with respect to Gaia DR2. The brightest
star in Gaia EDR3 is at G = 1.73.
The large-scale completeness limit, as estimated by the 99th
percentile of the G magnitude distribution (Fabricius et al. 2020),
varies between G ∼ 20 at low galactic latitudes (b 30) and
around the Magellanic Clouds, to G ∼ 22 at higher latitudes.
The survey limit variations over the sky show clear imprints of
the Gaia scanning pattern.
Article number, page 9 of 21
A&A proofs: manuscript no. aa202039657
In crowded regions the capability to observe all stars is reduced
(Gaia Collaboration et al. 2016b). In combination with
the still limited data treatment in crowded areas this means that
the survey limit in regions with densities above a few hundred
thousand stars per square degree can be substantially brighter
than G = 20. Fabricius et al. (2020) show that the completeness
as measured on OGLE fields is 100% up to source densities of
2 × 105
deg−2
, while at higher densities the completeness has
improved with respect to Gaia DR2, staying close to 100% up
to 6 × 105
stars deg−2
and dropping to 50% at densities over
8 × 105
deg−2
. Fabricius et al. (2020) also studied the completeness
in a sample of 26 globular clusters which were observed
previously with the Hubble Space Telescope. On average they
find a completeness at G = 17 of ∼ 80% for densities below
5 × 105
stars deg−2
, ∼ 50% at 5 × 105
–2 × 106
stars deg−2
, and
∼ 15% at 2 × 106
–2 × 107
stars deg−2
, with strong variations
across the clusters and between the cores and the outer regions.
In the very densest regions the incompleteness can be so severe
as to give the appearance of holes in the source distribution.
As described in Sect. 4, the effective angular resolution of
the Gaia EDR3 source list has slightly improved with respect
to Gaia DR2, with incompleteness in close pairs of stars starting
below about 1.5 arcsec. Refer to Fabricius et al. (2020) for
details. Fabricius et al. (2020) note that among the source pairs
with separations between 0.18 and 0.4 arcseconds there may be
many spurious solutions.
6.3. Limitations
6.3.1. Astrometry
The major gain in the precision of parallaxes and proper motions
from Gaia DR2 to Gaia EDR3 is complemented by a significant
reduction in the systematic errors, which is evident from the reduced
variance in the parallax and proper motion bias variations
over the sky, and confirmed by the beautiful measurement of the
acceleration of the solar system barycentre with respect to the
distant universe. Nevertheless the following characteristics of the
astrometry should be kept in mind.
The two solution types, 5-parameter and 6-parameter, behave
differently in terms of uncertainties and systematics, with the 6-p
astrometry in general being less precise. For the 5-p solutions the
astrometric uncertainties are underestimated by ∼ 5% at the faint
end (G > 16) and by up to ∼ 30% at the bright end (G < 14).
For the 6-p solutions the numbers are ∼ 20% and up to ∼ 40%,
respectively. The underestimation of the uncertainties increases
in crowded areas such as the Large Magellanic Cloud, and for
sources with indications that they may have companions or be
part of a partially resolved double. The details can be found in
Fabricius et al. (2020).
By examining the distribution of negative parallaxes, Fabricius
et al. (2020) estimate that among the sources with formally
high quality parallaxes ( /σ > 5) some 1.6% of the 1.5 billion
5-p and 6-p astrometric solutions are spurious, meaning that
the listed parallax may be significantly in error despite the formally
high precision. Fabricius et al. (2020) show that the fraction
of spurious solutions is strongly dependent on magnitude
and source density on the sky. For faint sources (G 17 for
6-p astrometric solutions and G 19 for 5-p solutions) and in
crowded regions the fractions of spurious solutions can reach 10
per cent or more. It should be stressed that the spurious astrometric
solutions in Gaia EDR3 produce smaller errors on the
astrometric parameters than was the case for Gaia DR2.
The global parallax zero point for Gaia EDR3, as measured
from quasars, is −17 µas. The RMS angular (i.e. source
to source) covariances of the parallaxes and proper motions on
small scales are ∼ 26 µas and ∼ 33 µas yr−1
, respectively. Details
on the angular covariances can be found in Lindegren et al.
(2020b). The parallax zero point (and the proper motion systematics)
varies as a function of magnitude, colour, and celestial
position. This is described in detail in Lindegren et al. (2020a).
6.3.2. Photometry
The increased precision and homogeneity of the Gaia EDR3
broad band photometry make it harder to assess the external accuracy
of the photometry. Nevertheless Riello et al. (2020) show
that the discontinuities that appeared at G = 13 and G = 16,
when comparing the Gaia DR2 photometry to APASS (Henden
et al. 2015, 2016) and SDSS DR15 (Aguado et al. 2019), have
disappeared in Gaia EDR3. As shown in Fabricius et al. (2020),
the same is true of the strong saturation effects in G at the bright
end, and the significant variation of the G-band zero point with
magnitude present in Gaia DR2. The effects are now below the
0.01 magnitude level for most sources. We stress again that a single
passband now suffices for all three bands, G, GBP, and GRP.
The following issues with the photometry were revealed following
internal investigations, all described in detail in Riello et al.
(2020) and Fabricius et al. (2020).
For faint red sources the flux in the BP band is overestimated
which leads to these sources appearing much bluer in (GBP −
GRP) than they should be. This can be recognised for example
in open cluster colour magnitude diagrams as a blue-ward turn
of the lower main sequence in G vs. (GBP − GRP). This issue is
caused by the rejection of observations with G-band fluxes below
1 e−
s−1
, where the rejection was also applied to the BP and RP
observations. This does not cause problems for G and GRP, but
at the faint end leads to overestimated BP fluxes.
During the internal validation of the Gaia EDR3 photometry
a small tail of sources going as faint as G ≈ 25.5 was noticed.
Such faint sources will never be observed by Gaia, even when
taking into account the fuzziness of the nominal G = 20.7 survey
limit. The problem was traced to sources with unreliable colours
for which the application of the internal photometric calibration
failed (Riello et al. 2020; Fabricius et al. 2020). As a result it
was decided to remove from the Gaia EDR3 catalogue the unreliable
fluxes and magnitudes, which means there are sources
for which any of the three bands could be missing. All in all
there are 5 455 339 sources for which no G-band flux is available
in the main Gaia EDR3 catalogue. For these sources the
G-band flux was estimated ad-hoc by calibrating the sources assuming
default colours. The values are available as a separate
table through the Gaia EDR3 ‘known issues’ pages3
. For 54 125
of the sources without G-band fluxes this ad-hoc calibration was
not possible. An indication of their brightness will be provided
based on the on-board magnitude estimate.
At the image parameter determination stage (which precedes
the astrometric and photometric data processing) the G-band
fluxes (and locations) of sources for which no reliable colour information
was available (from Gaia DR2) were estimated with
a PSF or LSF model for a default source colour. This concerns
the sources for which in the astrometry 6-p solutions were derived,
which mitigated the remaining chromatic effect in the
source image locations. Such a colour effect is also present in
3
https://www.cosmos.esa.int/web/gaia/
edr3-known-issues
Article number, page 10 of 21
Gaia Collaboration et al.: Gaia Early Data Release 3
the fluxes but this was unfortunately not accounted for in the
photometric processing. However, it is possible to correct the
published G-band photometry for sources with 6-p solutions
(astrometric_params_solved = 95) to bring them onto the
photometric system of the 5-p sources. The correction formula
is presented in Riello et al. (2020). We stress that this issue is not
related to the photometric pass-bands.
For bright and extremely blue sources (G < 13, GBP −GRP <
−0.1) there is a residual trend of about 5 mmag/mag for sources
in the range 8 < G < 13, when comparing the Gaia EDR3 magnitudes
to synthetic magnitudes derived from BP and RP spectra.
This is only seen in G and is probably related to deficiencies in
the PSF modelling for bright stars. At G < 8 the residuals are
dominated by saturation effects.
7. Using Gaia EDR3 data: Additional guidance
Here we provide some further advice on the use of the
Gaia EDR3 data. This concerns issues specific to this release.
The papers listed in Sect. 4 provide extensive examples of how
to use Gaia EDR3 data responsibly, and we remind the reader
of the need to be careful when estimating distances from parallaxes
with relatively large uncertainties (Luri et al. 2018). We
note again, as stressed in Sect. 5.1, that tracing sources from
Gaia DR2 to Gaia EDR3 should not be done by blindly matching
source IDs. The Gaia DR2 to Gaia EDR3 match table
(dr2_neighbourhood) should be used for this purpose.
7.1. Astrometry
Fabricius et al. (2020) make the following recommendation for
dealing with spurious astrometric solutions. Even when selecting
only a sample of high-quality parallaxes, for example /σ >
5, one should select also the corresponding sample with negative
parallaxes ( /σ < −5) in order to ascertain what fraction
of the positive parallaxes may in fact be spurious. The parameter
ipd_gof_harmonic_amplitude is useful in identifying spurious
solutions as shown in Fabricius et al. (2020), where values
above 0.1 in combination with a ruwe larger than 1.4 are indicative
of resolved doubles, which are still not correctly handled in
the astrometric processing, and may cause spurious solutions.
For the sources with 6-p solutions there will in many cases
be independent colour information available from photometric
or spectroscopic observations, which may provide a superior estimate
for νeff than the value estimated during the Gaia EDR3
astrometric data processing. In these cases it is possible to incorporate
the better colour information to update the astrometry
for the 6-p solution to more precise values. The formulae
for calculating the updated astrometry are given in the appendix
of Lindegren et al. (2020b), where it is demonstrated that for
significant correlations (coefficients larger than 0.3) between the
pseudo-colour and the astrometric parameters, real gains in precision
and accuracy can be expected.
Lindegren et al. (2020a) present an extensive investigation of
the parallax zero point variations as a function of source brightness,
colour, and celestial position. The samples used in this investigation
are quasars, the Large Magellanic Cloud, red clump
stars – sources for which the parallax is precisely known from
independent estimates – and wide pairs of co-moving stars for
which the parallaxes should be the same. As a result of the detailed
characterisation of the zero point variations, a correction
recipe is presented, separately for sources with 5-p and 6-p astrometry,
which allows removing the parallax bias as a function
of source magnitude, colour, and ecliptic latitude. It should be
stressed that this is a tentative recipe, primarily intended as an illustration
of how corrections could be derived for other samples
of sources for which precise independent distance information
is available. The recipe is not intended to be applied blindly and
has not been applied to the published Gaia EDR3 parallaxes.
Python code to apply the recipe will be made available as part of
the Gaia EDR3 access facilities4
.
7.2. Photometry
Riello et al. (2020) provide guidance on the use of the
Gaia EDR3 photometry which we summarise here. Further insights
into the photometric data are presented by Fabricius et al.
(2020).
The GBP flux of faint sources is likely to be biased
(Sect. 6.3.2) and one can elect to filter on the value of GBP, retaining
only the brighter bias-free sources. This introduces undesirable
selection effects and a better alternative may be to use
the (G −GRP) colour, for example when studying the lower main
sequence.
The G-band photometry for sources with 6-p astrometric solutions
should be corrected to account for the use of a default
colour at the flux estimation stage, the correction formula is presented
in Riello et al. (2020). In Appendix A we show how to calculate
the corrections on the fly as part of a Gaia EDR3 archive
query, and also present Python code that can be used for the same
purpose.
For sources where the G magnitude is missing, the value
can be looked up in separate tables to be provided through the
Gaia EDR3 ‘known issues’ web pages. While this may seem
unnecessarily cumbersome, this choice was deliberately made to
ensure that it is very clear to the user that the G-band magnitudes
for these sources are from a very different origin (from an
ad-hoc calibration, and for a small number of sources from the
on-board magnitude estimate) and not directly comparable to the
main catalogue photometry.
As was the case for Gaia DR2, at the bright end (G < 8 for G
and G 4 for GBP and GRP) the magnitudes should be corrected
for saturation effects. The correction formulae can be found in
the appendix of Riello et al. (2020).
An important point made in Riello et al. (2020) is that the
flux excess factor in Gaia EDR3 is much more representative of
astrophysical inconsistencies between the fluxes in BP and RP
with respect to the flux in G, for example due to the extended nature
of a source or its non-standard (non-stellar) spectral energy
distribution (although Fabricius et al. (2020) show that some
artefacts in the photometry can still be traced in the flux excess
factor). It is thus not possible to easily identify problematic photometry
through the flux excess factor, and using this quantity
in the construction of samples should be done with care. Refer
to Riello et al. (2020) for detailed guidance. They present a corrected
version of the flux excess factor which is recommended
for use instead of the raw phot_bp_rp_excess_factor value
listed in Gaia EDR3. The corrected version can be calculated
from the formula presented in section 6 of Riello et al. (2020).
Appendix B shows how to include the correction of the flux excess
factor within an ADQL query, and presents Python code to
achieve the same.
As described in Sect. 3.2, Gaia EDR3 contains fields from
which a metric can be constructed that indicates how likely it is
that the photometry of a given source is affected by crowding.
4
https://gitlab.com/icc-ub/public/gaiadr3_zeropoint
Article number, page 11 of 21
A&A proofs: manuscript no. aa202039657
This metric should be used with some care, which is explained
further in Riello et al. (2020).
Finally, due to the evolution of the source list and the improvements
in the photometry we strongly discourage comparisons
between Gaia EDR3 and Gaia DR2 photometry, in particular
on a source by source basis. Comparisons at the sample
level are likely to reveal mostly differences due to changes in the
photometric system and errors in Gaia DR2.
8. Gaia EDR3 access facilities
The Gaia EDR3 data will be available through the archive hosted
by ESA5
, with the facilities as described in Gaia Collaboration
et al. (2016a) and Gaia Collaboration et al. (2018). The data
is also accessible at the partner and affiliated data centres in
Europe, the United States, Japan, Australia, and South Africa.
These data centres provide their own access facilities, but do not
necessarily host all data contained in the ESA Gaia archive. We
note the following enhancements and changes.
The pointing of the Gaia telescopes as a function of time is
available as the table commanded_scan_law. The pointing for
the entire 34 month period covering Gaia EDR3 is available at
10 second intervals, and allows one to reconstruct how often and
at what scan angles a given position on the sky was observed by
Gaia. We note that the commanded pointing is provided which
may deviate from the actual attitude of Gaia by up to 30 arcsec.
In addition gaps in the data collection due to spacecraft or onground
problems are not accounted for.
The Gaia Universe Model Snapshot (GUMS, Robin
et al. 2012) and the corresponding simulated Gaia catalogue
(GOG, Luri et al. 2014) are now available as part of the
Gaia EDR3 archive in the tables gaia_universe_model and
gaia_source_simulation, respectively. They correspond to
version 20 of GUMS and GOG and are described in detail in
the on-line documentation 6
. We note that a few issues in the
simulations could not be corrected on time. Notably, young star
kinematics were wrongly set, such that their astrometry should
be corrected by the user before using the simulation. Essentially,
the stars with ages less than 0.15 Gyr should follow the Milky
Way rotation curve, while they do not (their mean rotation velocity
was erroneously set to 0). In addition, RR Lyrae stars are
missing from GUMS, and the number of outliers in GOG, both
in astrometry and photometry, is larger than expected from the
simulated errors.
The astrometric performance predictions for Gaia DR4 and
beyond have been updated, based on an extrapolation of the
Gaia EDR3 performance. The new predictions will appear on the
Gaia science performance pages7
. The tables agn_cross_id
and frame_rotator_source provide the source IDs of the
Gaia-CRF3 sources.
Pre-computed cross matches to other large surveys are provided.
We recommend using these cross-matches because they
have been carefully validated and their use facilitates reproducing
analyses of Gaia EDR3 data combined with other survey
data. The details of the cross-match procedure are provided
in Marrese et al. (2020) (see also Marrese et al. 2017, 2019).
Pre-computed cross-matches are provided for the following surveys:
Hipparcos (new reduction, van Leeuwen 2007); Tycho-2
5
https://archives.esac.esa.int/gaia
6
http://gea.esac.esa.int/archive/documentation/GEDR3/
Data_processing/chap_simulated/
7
https://www.cosmos.esa.int/web/gaia/
science-performance
(Høg et al. 2000), merged with the Tycho Double Star Catalogue
(Fabricius et al. 2002); 2MASS (point source and extended
catalogue merged, Skrutskie et al. 2006); SDSS DR13 (Albareti
et al. 2017); Pan-STARRS1 (Chambers et al. 2016); SkyMapper
DR2 (Onken et al. 2019); AllWise (Wright et al. 2010); URAT1
(Zacharias et al. 2015); GSC2.3 (Lasker et al. 2008); APASS
DR9 (Henden et al. 2016, 2015); and RAVE DR5 (Kunder et al.
2017).
9. Conclusions
Gaia EDR3 represents another significant advance in the series
of data releases resulting from the Gaia mission. Based on 34
months of input data, this release features major improvements
in the astrometry and broad-band photometry, where for the first
time the astrometry and photometry benefit from iterative processing
between the determination of image locations and fluxes,
and the astrometric solution. Other significant improvements are
the increased robustness and stability of the source list and a
much more sophisticated modelling of the PSF and LSF for the
astrometric instrument. Next to the significant increase in the
precision of the astrometry and photometry, the suppression of
systematic errors is a major component of the improvements.
Gaia EDR3 represents the first installment of the Gaia DR3
release planned for publication in 2022. Gaia DR3 will feature
new data products of which the BP, RP, and RVS spectra
(to be released for a subset of sources) and the non-single
star catalogue represent qualitative changes in character with respect
to Gaia DR2. The planned contents are: astrometry and
broad-band photometry from Gaia EDR3 will remain unchanged
for Gaia DR3, and the same is true for the source list; an expanded
radial velocity survey (some 30 million stars brighter
than GRVS ∼ 14); astrophysical parameter estimates based on
the parallaxes, broad-band photometry, and BP, RP, and RVS
spectra, where the latter are a new element enabling a richer astrophysical
characterisation of sources; an order of magnitude
larger sample of variable stars, with their light curves, classifications,
and astrophysical properties; a non-single star (mostly binary
stars) catalogue based on the analysis of epoch astrometry,
epoch radial velocities and the light curves of eclipsing binaries;
analyses of extended objects (galaxies, and QSO hosts); epoch
astrometry and photometry, as well as orbits, for an expanded
list of over 100 000 solar system objects; mean BP, RP, and
RVS spectra, for a subset of astrophysically well-characterised
sources; reflectance spectra for ∼ 5000 asteroids derived from
BP and RP spectra; the Gaia Andromeda Photometric Survey
(GAPS), which consists of the broad-band photometric timeseries
for all sources in a 5.5◦
radius field centred on M31.
There is thus much to look forward to in Gaia DR3, with the
GAPS data set providing a taste of what is to come in Gaia DR4.
We stress that epoch astrometry and epoch radial velocities will
not appear in Gaia DR3 (except for solar system object epoch
astrometry).
Looking ahead, the Gaia spacecraft is currently in good overall
health. The spacecraft operations since the appearance of
Gaia DR2 have largely been smooth, with little to no degradation
of the detectors in Gaia’s focal plane, except for the steadily increasing
radiation damage. The latter is however still well below
the anticipated levels before the launch of Gaia. The last decontamination
of the telescopes and focal plane took place in August
2016 and no further decontamination was needed since. The current
evolution of the throughput of the telescopes suggests that
also in the future no further decontamination is needed. The only
limiting factor to the lifetime of Gaia as a high precision astromArticle
number, page 12 of 21
Gaia Collaboration et al.: Gaia Early Data Release 3
etry mission is the amount of propellant for the micro-propulsion
system. This is predicted to be exhausted in early 2025, after
which time the attitude and spin rate of Gaia can no longer be
maintained at the levels of precision needed for the astrometry.
With this end-of-life date for Gaia in mind, and the end of the
nominal mission planned for mid-2019, the process of applying
for an extended mission was started already in 2016. The nominal
Gaia mission ended on July 16, 2019 and Gaia has been
in extended mission operations since that date. The mission extension
is formally approved to the end of 2022 at the time of
writing, with good hopes of the mission continuing to its estimated
end-of-life. This would bring the total mission lifetime to
10 years, implying a 40 per cent improvement on the precision
of all data products with respect to a five year mission, and a
factor of almost three improvement for the proper motions.
In this context the community can look forward to two major
data releases, Gaia DR4 and Gaia DR5, both incorporating
data from the extended Gaia mission. Gaia DR4 will be based
on 66 months of input data (which is already in hand), while
Gaia DR5 will include all data collected over the entire (nominal
+ extended) Gaia mission. The extra half year of data from
the extended mission included in the Gaia DR4 data processing
is motivated by the wish to include part of the one year period
between July 16, 2019 and July 29, 2020 when Gaia was operated
with a reversed direction of the precession of the spin axis
around the direction to the sun. This was introduced to break
the degeneracy between the across-scan rate at which sources
move across the focal plane and their parallax factor (see appendix
B in Lindegren et al. 2020b, for more details). Including
the first 6 months of the reverse precession period in the inputs
for Gaia DR4 is expected to already significantly mitigate the
effects of this degeneracy.
The main new features of Gaia DR4 are the publication of
a list of exoplanets discovered with Gaia and the publication of
all the time series data, meaning epoch astrometry, broad-band
photometry, radial velocities, as well as BP, RP, and RVS spectra
for all sources. This will be a significant expansion in the volume
of data released. We leave to the imagination of the reader the
expanded scientific possibilities.
Acknowledgements. This work presents results from the European Space
Agency (ESA) space mission Gaia. Gaia data are being processed by the
Gaia Data Processing and Analysis Consortium (DPAC). Funding for the
DPAC is provided by national institutions, in particular the institutions participating
in the Gaia MultiLateral Agreement (MLA). The Gaia mission website
is https://www.cosmos.esa.int/gaia. The Gaia archive website is
https://archives.esac.esa.int/gaia. The Gaia mission and data processing
have financially been supported by, in alphabetical order by country:
the Algerian Centre de Recherche en Astronomie, Astrophysique et Géophysique
of Bouzareah Observatory; the Austrian Fonds zur Förderung der wissenschaftlichen
Forschung (FWF) Hertha Firnberg Programme through grants
T359, P20046, and P23737; the BELgian federal Science Policy Office (BELSPO)
through various PROgramme de Développement d’Expériences scientifiques
(PRODEX) grants and the Polish Academy of Sciences - Fonds
Wetenschappelijk Onderzoek through grant VS.091.16N, and the Fonds de la
Recherche Scientifique (FNRS); the Brazil-France exchange programmes Fundação
de Amparo à Pesquisa do Estado de São Paulo (FAPESP) and Coordenação
de Aperfeicoamento de Pessoal de Nível Superior (CAPES) - Comité
Français d’Evaluation de la Coopération Universitaire et Scientifique avec le
Brésil (COFECUB); the National Science Foundation of China (NSFC) through
grants 11573054 and 11703065 and the China Scholarship Council through
grant 201806040200; the Tenure Track Pilot Programme of the Croatian Science
Foundation and the École Polytechnique Fédérale de Lausanne and the
project TTP-2018-07-1171 ’Mining the Variable Sky’, with the funds of the
Croatian-Swiss Research Programme; the Czech-Republic Ministry of Education,
Youth, and Sports through grant LG 15010 and INTER-EXCELLENCE
grant LTAUSA18093, and the Czech Space Office through ESA PECS contract
98058; the Danish Ministry of Science; the Estonian Ministry of Education
and Research through grant IUT40-1; the European Commission’s Sixth Framework
Programme through the European Leadership in Space Astrometry (ELSA)
Marie Curie Research Training Network (MRTN-CT-2006-033481), through
Marie Curie project PIOF-GA-2009-255267 (Space AsteroSeismology & RR
Lyrae stars, SAS-RRL), and through a Marie Curie Transfer-of-Knowledge
(ToK) fellowship (MTKD-CT-2004-014188); the European Commission’s Seventh
Framework Programme through grant FP7-606740 (FP7-SPACE-2013-
1) for the Gaia European Network for Improved data User Services (GENIUS)
and through grant 264895 for the Gaia Research for European Astronomy
Training (GREAT-ITN) network; the European Research Council (ERC)
through grants 320360 and 647208 and through the European Union’s Horizon
2020 research and innovation and excellent science programmes through
Marie Skłodowska-Curie grant 745617 as well as grants 670519 (Mixing and
Angular Momentum tranSport of massIvE stars – MAMSIE), 687378 (Small
Bodies: Near and Far), 682115 (Using the Magellanic Clouds to Understand
the Interaction of Galaxies), and 695099 (A sub-percent distance scale from
binaries and Cepheids – CepBin); the European Science Foundation (ESF), in
the framework of the Gaia Research for European Astronomy Training Research
Network Programme (GREAT-ESF); the European Space Agency (ESA)
in the framework of the Gaia project, through the Plan for European Cooperating
States (PECS) programme through grants for Slovenia, through contracts
C98090 and 4000106398/12/NL/KML for Hungary, and through contract
4000115263/15/NL/IB for Germany; the Academy of Finland and the Magnus
Ehrnrooth Foundation; the French Centre National d’Etudes Spatiales (CNES),
the Agence Nationale de la Recherche (ANR) through grant ANR-10-IDEX-
0001-02 for the ’Investissements d’avenir’ programme, through grant ANR-15CE31-0007
for project ’Modelling the Milky Way in the Gaia era’ (MOD4Gaia),
through grant ANR-14-CE33-0014-01 for project ’The Milky Way disc formation
in the Gaia era’ (ARCHEOGAL), and through grant ANR-15-CE31-0012-
01 for project ’Unlocking the potential of Cepheids as primary distance calibrators’
(UnlockCepheids), the Centre National de la Recherche Scientifique
(CNRS) and its SNO Gaia of the Institut des Sciences de l’Univers (INSU),
the ’Action Fédératrice Gaia’ of the Observatoire de Paris, the Région de
Franche-Comté, and the Programme National de Gravitation, Références, Astronomie,
et Métrologie (GRAM) of CNRS/INSU with the Institut National
Polytechnique (INP) and the Institut National de Physique nucléaire et de
Physique des Particules (IN2P3) co-funded by CNES; the German Aerospace
Agency (Deutsches Zentrum für Luft- und Raumfahrt e.V., DLR) through
grants 50QG0501, 50QG0601, 50QG0602, 50QG0701, 50QG0901, 50QG1001,
50QG1101, 50QG1401, 50QG1402, 50QG1403, 50QG1404, and 50QG1904
and the Centre for Information Services and High Performance Computing
(ZIH) at the Technische Universität (TU) Dresden for generous allocations
of computer time; the Hungarian Academy of Sciences through the Lendület
Programme grants LP2014-17 and LP2018-7 and through the Premium Postdoctoral
Research Programme (L. Molnár), and the Hungarian National Research,
Development, and Innovation Office (NKFIH) through grant KH_18-
130405; the Science Foundation Ireland (SFI) through a Royal Society - SFI
University Research Fellowship (M. Fraser); the Israel Science Foundation (ISF)
through grant 848/16; the Agenzia Spaziale Italiana (ASI) through contracts
I/037/08/0, I/058/10/0, 2014-025-R.0, 2014-025-R.1.2015, and 2018-24-HH.0 to
the Italian Istituto Nazionale di Astrofisica (INAF), contract 2014-049-R.0/1/2
to INAF for the Space Science Data Centre (SSDC, formerly known as the
ASI Science Data Center, ASDC), contracts I/008/10/0, 2013/030/I.0, 2013-
030-I.0.1-2015, and 2016-17-I.0 to the Aerospace Logistics Technology Engineering
Company (ALTEC S.p.A.), INAF, and the Italian Ministry of Education,
University, and Research (Ministero dell’Istruzione, dell’Università e
della Ricerca) through the Premiale project ’MIning The Cosmos Big Data
and Innovative Italian Technology for Frontier Astrophysics and Cosmology’
(MITiC); the Netherlands Organisation for Scientific Research (NWO) through
grant NWO-M-614.061.414, through a VICI grant (A. Helmi), and through a
Spinoza prize (A. Helmi), and the Netherlands Research School for Astronomy
(NOVA); the Polish National Science Centre through HARMONIA grant
2018/06/M/ST9/00311, DAINA grant 2017/27/L/ST9/03221, and PRELUDIUM
grant 2017/25/N/ST9/01253, and the Ministry of Science and Higher Education
(MNiSW) through grant DIR/WK/2018/12; the Portugese Fundação para
a Ciência e a Tecnologia (FCT) through grants SFRH/BPD/74697/2010 and
SFRH/BD/128840/2017 and the Strategic Programme UID/FIS/00099/2019 for
CENTRA; the Slovenian Research Agency through grant P1-0188; the Spanish
Ministry of Economy (MINECO/FEDER, UE) through grants ESP2016-80079C2-1-R,
ESP2016-80079-C2-2-R, RTI2018-095076-B-C21, RTI2018-095076B-C22,
BES-2016-078499, and BES-2017-083126 and the Juan de la Cierva
formación 2015 grant FJCI-2015-2671, the Spanish Ministry of Education, Culture,
and Sports through grant FPU16/03827, the Spanish Ministry of Science
and Innovation (MICINN) through grant AYA2017-89841P for project ’Estudio
de las propiedades de los fósiles estelares en el entorno del Grupo Local’
and through grant TIN2015-65316-P for project ’Computación de Altas Prestaciones
VII’, the Severo Ochoa Centre of Excellence Programme of the Spanish
Government through grant SEV2015-0493, the Institute of Cosmos Sciences
University of Barcelona (ICCUB, Unidad de Excelencia ’María de Maeztu’)
through grants MDM-2014-0369 and CEX2019-000918-M, the University of
Barcelona’s official doctoral programme for the development of an R+D+i
Article number, page 13 of 21
A&A proofs: manuscript no. aa202039657
project through an Ajuts de Personal Investigador en Formació (APIF) grant,
the Spanish Virtual Observatory through project AyA2017-84089, the Galician
Regional Government, Xunta de Galicia, through grants ED431B-2018/42 and
ED481A-2019/155, support received from the Centro de Investigación en Tecnologías
de la Información y las Comunicaciones (CITIC) funded by the Xunta
de Galicia, the Xunta de Galicia and the Centros Singulares de Investigación de
Galicia for the period 2016-2019 through CITIC, the European Union through
the European Regional Development Fund (ERDF) / Fondo Europeo de Desenvolvemento
Rexional (FEDER) for the Galicia 2014-2020 Programme through
grant ED431G-2019/01, the Red Española de Supercomputación (RES) computer
resources at MareNostrum, the Barcelona Supercomputing Centre - Centro
Nacional de Supercomputación (BSC-CNS) through activities AECT-2016-
1-0006, AECT-2016-2-0013, AECT-2016-3-0011, and AECT-2017-1-0020, the
Departament d’Innovació, Universitats i Empresa de la Generalitat de Catalunya
through grant 2014-SGR-1051 for project ’Models de Programació i Entorns
d’Execució Parallels’ (MPEXPAR), and Ramon y Cajal Fellowship RYC2018-
025968-I; the Swedish National Space Agency (SNSA/Rymdstyrelsen); the
Swiss State Secretariat for Education, Research, and Innovation through the
Mesures d’Accompagnement, the Swiss Activités Nationales Complémentaires,
and the Swiss National Science Foundation; the United Kingdom Particle
Physics and Astronomy Research Council (PPARC), the United Kingdom Science
and Technology Facilities Council (STFC), and the United Kingdom
Space Agency (UKSA) through the following grants to the University of Bristol,
the University of Cambridge, the University of Edinburgh, the University
of Leicester, the Mullard Space Sciences Laboratory of University College
London, and the United Kingdom Rutherford Appleton Laboratory (RAL):
PP/D006511/1, PP/D006546/1, PP/D006570/1, ST/I000852/1, ST/J005045/1,
ST/K00056X/1, ST/K000209/1, ST/K000756/1, ST/L006561/1, ST/N000595/1,
ST/N000641/1, ST/N000978/1, ST/N001117/1, ST/S000089/1, ST/S000976/1,
ST/S001123/1, ST/S001948/1, ST/S002103/1, and ST/V000969/1. This work
made use of the following software: Astropy, a community-developed core
Python package for Astronomy (Astropy Collaboration et al. 2013, 2018, http:
//www.astropy.org), IPython (Pérez & Granger 2007, https://ipython.
org/), Jupyter (https://jupyter.org/), Matplotlib (Hunter 2007, https:
//matplotlib.org), SciPy (Virtanen et al. 2020, https://www.scipy.org),
NumPy (Harris et al. 2020, https://numpy.org), and TOPCAT (Taylor 2005,
http://www.starlink.ac.uk/topcat/). This work has made use of NASA’s
Astrophysics Data System. We thank the referee, Andy Casey, for a careful reading
of the manuscript.
References
Aguado, D. S., Ahumada, R., Almeida, A., et al. 2019, ApJS, 240, 23
Albareti, F. D., Allende Prieto, C., Almeida, A., et al. 2017, ApJS, 233, 25
Arenou, F., Luri, X., Babusiaux, C., et al. 2018, A&A, 616, A17
Astropy Collaboration, Price-Whelan, A. M., Sip˝ocz, B. M., et al. 2018, AJ, 156,
123
Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., et al. 2013, A&A, 558,
A33
Boubert, D., Strader, J., Aguado, D., et al. 2019, MNRAS, 486, 2618
Butkevich, A. G., Klioner, S. A., Lindegren, L., Hobbs, D., & van Leeuwen, F.
2017, A&A, 603, A45
Carrasco et al. 2020, A&A in prep.
Chambers, K. C., Magnier, E. A., Metcalfe, N., et al. 2016, ArXiv e-prints
[arXiv:1612.05560]
Cropper, M., Katz, D., Sartoretti, P., et al. 2018, A&A, 616, A5
De Angeli et al. 2021, A&A in prep.
Evans, D. W., Riello, M., De Angeli, F., et al. 2018, A&A, 616, A4
Fabricius, C., Høg, E., Makarov, V. V., et al. 2002, A&A, 384, 180
Fabricius, C., Luri, X., Arenou, F., et al. 2020, A&A in prep.
Gaia Collaboration, Antoja, T., McMillan, P. J., et al. 2020a, A&A in prep.
Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018, A&A, 616, A1
Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2016a, A&A, 595, A2
Gaia Collaboration, Klioner, S., & et al. 2020b, A&A in prep.
Gaia Collaboration, Klioner, S., Mignard, F., et al. 2020c, submitted to A&A
Gaia Collaboration, Luri, X., Chemin, L., et al. 2020d, submitted to A&A
Gaia Collaboration, Prusti, T., de Bruijne, J. H. J., et al. 2016b, A&A, 595, A1
Gaia Collaboration, Smart, R. L., Sarro, L. M., et al. 2020e, submitted to A&A
Hambly, N. C., Cropper, M., Boudreault, S., et al. 2018, A&A, 616, A15
Harris, C. R., Millman, K. J., van der Walt, S. J., et al. 2020, Nature, 585,
357–362
Henden, A. A., Levine, S., Terrell, D., & Welch, D. L. 2015, in American Astronomical
Society Meeting Abstracts, Vol. 225, American Astronomical Society
Meeting Abstracts #225, 336.16
Henden, A. A., Templeton, M., Terrell, D., et al. 2016, VizieR Online Data Catalogue,
2336
Høg, E., Fabricius, C., Makarov, V. V., et al. 2000, A&A, 355, L27
Hunter, J. D. 2007, Computing in Science and Engineering, 9, 90
Katz, D., Sartoretti, P., Cropper, M., et al. 2019, A&A, 622, A205
Kunder, A., Kordopatis, G., Steinmetz, M., et al. 2017, AJ, 153, 75
Lasker, B. M., Lattanzi, M. G., McLean, B. J., et al. 2008, AJ, 136, 735
Lindegren, L. 2018, technical note GAIA-C3-TN-LU-LL-124
Lindegren, L. 2020a, A&A, 633, A1
Lindegren, L. 2020b, A&A, 637, C5
Lindegren, L., Bastian, U., Biermann, M., et al. 2020a, A&A in prep.
Lindegren, L., Hernández, J., Bombrun, A., et al. 2018, A&A, 616, A2
Lindegren, L., Klioner, S., Hernández, J., et al. 2020b, A&A in prep.
Lindegren, L., Lammers, U., Hobbs, D., et al. 2012, A&A, 538, A78
Luri, X., Brown, A. G. A., Sarro, L. M., et al. 2018, A&A, 616, A9
Luri, X., Palmer, M., Arenou, F., et al. 2014, A&A, 566, A119
Maíz Apellániz, J. & Weiler, M. 2018, A&A, 619, A180
Marrese, P. M., Marinoni, S., Fabrizio, M., & Altavilla, G. 2019, A&A, 621,
A144
Marrese, P. M., Marinoni, S., Fabrizio, M., & Altavilla, G. 2020, A&A in prep.
Marrese, P. M., Marinoni, S., Fabrizio, M., & Giuffrida, G. 2017, A&A, 607,
A105
Michalik, D., Lindegren, L., Hobbs, D., & Butkevich, A. G. 2015, A&A, 583,
A68
Montegriffo et al. 2021, A&A in prep.
Onken, C. A., Wolf, C., Bessell, M. S., et al. 2019, PASA, 36, e033
Pancino, E., Altavilla, G., Marinoni, S., et al. 2012, MNRAS, 426, 1767
Pérez, F. & Granger, B. E. 2007, Computing in Science and Engineering, 9, 21
Riello, M., De Angeli, F., Evans, D. W., et al. 2020, submitted to A&A
Robin, A. C., Luri, X., Reylé, C., et al. 2012, A&A, 543, A100
Rowell, N., Davidson, M., Lindegren, L., et al. 2020, submitted to A&A
Sartoretti, P., Katz, D., Cropper, M., et al. 2018, A&A, 616, A6
Seabroke, G., Fabricius, C., Teyssier, D., et al. 2020, A&A in prep.
Skrutskie, M. F., Cutri, R. M., Stiening, R., et al. 2006, AJ, 131, 1163
Smart, R. L. & Nicastro, L. 2014, A&A, 570, A87
Taylor, M. B. 2005, in Astronomical Society of the Pacific Conference Series,
Vol. 347, Astronomical Data Analysis Software and Systems XIV, ed.
P. Shopbell, M. Britton, & R. Ebert, 29
Torra, F., Castañeda, J., Fabricius, C., et al. 2020, submitted to A&A
van Leeuwen, F. 2007, Hipparcos, the New Reduction of the Raw Data, Astrophysics
and Space Science Library. Vol. 350 edn. (Springer)
Virtanen, P., Gommers, R., Oliphant, T. E., et al. 2020, Nature Methods, 17, 261
Wright, E. L., Eisenhardt, P. R. M., Mainzer, A. K., et al. 2010, AJ, 140, 1868
Zacharias, N., Finch, C., Subasavage, J., et al. 2015, AJ, 150, 101
1
Leiden Observatory, Leiden University, Niels Bohrweg 2, 2333 CA
Leiden, The Netherlands
2
INAF - Osservatorio astronomico di Padova, Vicolo Osservatorio
5, 35122 Padova, Italy
3
European Space Agency (ESA), European Space Research and
Technology Centre (ESTEC), Keplerlaan 1, 2201AZ, Noordwijk,
The Netherlands
4
Univ. Grenoble Alpes, CNRS, IPAG, 38000 Grenoble, France
5
GEPI, Observatoire de Paris, Université PSL, CNRS, 5 Place Jules
Janssen, 92190 Meudon, France
6
Astronomisches Rechen-Institut, Zentrum für Astronomie der Universität
Heidelberg, Mönchhofstr. 12-14, 69120 Heidelberg, Ger-
many
7
Université Côte d’Azur, Observatoire de la Côte d’Azur, CNRS,
Laboratoire Lagrange, Bd de l’Observatoire, CS 34229, 06304
Nice Cedex 4, France
8
Institute of Astronomy, University of Cambridge, Madingley Road,
Cambridge CB3 0HA, United Kingdom
9
Department of Astronomy, University of Geneva, Chemin des
Maillettes 51, 1290 Versoix, Switzerland
10
Aurora Technology for European Space Agency (ESA), Camino
bajo del Castillo, s/n, Urbanizacion Villafranca del Castillo, Villanueva
de la Cañada, 28692 Madrid, Spain
11
Institut de Ciències del Cosmos (ICCUB), Universitat de Barcelona
(IEEC-UB), Martí i Franquès 1, 08028 Barcelona, Spain
12
Lohrmann Observatory, Technische Universität Dresden, Mommsenstraße
13, 01062 Dresden, Germany
13
European Space Agency (ESA), European Space Astronomy Centre
(ESAC), Camino bajo del Castillo, s/n, Urbanizacion VilArticle
number, page 14 of 21
Gaia Collaboration et al.: Gaia Early Data Release 3
lafranca del Castillo, Villanueva de la Cañada, 28692 Madrid,
Spain
14
Lund Observatory, Department of Astronomy and Theoretical
Physics, Lund University, Box 43, 22100 Lund, Sweden
15
CNES Centre Spatial de Toulouse, 18 avenue Edouard Belin, 31401
Toulouse Cedex 9, France
16
Institut d’Astronomie et d’Astrophysique, Université Libre de
Bruxelles CP 226, Boulevard du Triomphe, 1050 Brussels, Belgium
17
F.R.S.-FNRS, Rue d’Egmont 5, 1000 Brussels, Belgium
18
INAF - Osservatorio Astrofisico di Arcetri, Largo Enrico Fermi 5,
50125 Firenze, Italy
19
Laboratoire d’astrophysique de Bordeaux, Univ. Bordeaux, CNRS,
B18N, allée Geoffroy Saint-Hilaire, 33615 Pessac, France
20
Max Planck Institute for Astronomy, Königstuhl 17, 69117 Heidelberg,
Germany
21
Mullard Space Science Laboratory, University College London,
Holmbury St Mary, Dorking, Surrey RH5 6NT, United Kingdom
22
INAF - Osservatorio Astrofisico di Torino, via Osservatorio 20,
10025 Pino Torinese (TO), Italy
23
University of Turin, Department of Physics, Via Pietro Giuria 1,
10125 Torino, Italy
24
INAF - Osservatorio di Astrofisica e Scienza dello Spazio di
Bologna, via Piero Gobetti 93/3, 40129 Bologna, Italy
25
DAPCOM for Institut de Ciències del Cosmos (ICCUB), Universitat
de Barcelona (IEEC-UB), Martí i Franquès 1, 08028 Barcelona,
Spain
26
Royal Observatory of Belgium, Ringlaan 3, 1180 Brussels, Bel-
gium
27
ALTEC S.p.a, Corso Marche, 79,10146 Torino, Italy
28
Department of Astronomy, University of Geneva, Chemin d’Ecogia
16, 1290 Versoix, Switzerland
29
Sednai Sàrl, Geneva, Switzerland
30
Gaia DPAC Project Office, ESAC, Camino bajo del Castillo, s/n,
Urbanizacion Villafranca del Castillo, Villanueva de la Cañada,
28692 Madrid, Spain
31
Telespazio Vega UK Ltd for European Space Agency (ESA),
Camino bajo del Castillo, s/n, Urbanizacion Villafranca del
Castillo, Villanueva de la Cañada, 28692 Madrid, Spain
32
SYRTE, Observatoire de Paris, Université PSL, CNRS, Sorbonne
Université, LNE, 61 avenue de l’Observatoire 75014 Paris, France
33
National Observatory of Athens, I. Metaxa and Vas. Pavlou, Palaia
Penteli, 15236 Athens, Greece
34
IMCCE, Observatoire de Paris, Université PSL, CNRS, Sorbonne
Université, Univ. Lille, 77 av. Denfert-Rochereau, 75014 Paris,
France
35
INAF - Osservatorio Astrofisico di Catania, via S. Sofia 78, 95123
Catania, Italy
36
Serco Gestión de Negocios for European Space Agency (ESA),
Camino bajo del Castillo, s/n, Urbanizacion Villafranca del
Castillo, Villanueva de la Cañada, 28692 Madrid, Spain
37
Institut d’Astrophysique et de Géophysique, Université de Liège,
19c, Allée du 6 Août, B-4000 Liège, Belgium
38
CRAAG - Centre de Recherche en Astronomie, Astrophysique et
Géophysique, Route de l’Observatoire Bp 63 Bouzareah 16340 Algiers,
Algeria
39
Institute for Astronomy, University of Edinburgh, Royal Observatory,
Blackford Hill, Edinburgh EH9 3HJ, United Kingdom
40
ATG Europe for European Space Agency (ESA), Camino bajo del
Castillo, s/n, Urbanizacion Villafranca del Castillo, Villanueva de
la Cañada, 28692 Madrid, Spain
41
ETSE Telecomunicación, Universidade de Vigo, Campus LagoasMarcosende,
36310 Vigo, Galicia, Spain
42
Université de Strasbourg, CNRS, Observatoire astronomique de
Strasbourg, UMR 7550, 11 rue de l’Université, 67000 Strasbourg,
France
43
Kavli Institute for Cosmology Cambridge, Institute of Astronomy,
Madingley Road, Cambridge, CB3 0HA
44
Department of Astrophysics, Astronomy and Mechanics, National
and Kapodistrian University of Athens, Panepistimiopolis, Zografos,
15783 Athens, Greece
45
Observational Astrophysics, Division of Astronomy and Space
Physics, Department of Physics and Astronomy, Uppsala University,
Box 516, 751 20 Uppsala, Sweden
46
Leibniz Institute for Astrophysics Potsdam (AIP), An der Sternwarte
16, 14482 Potsdam, Germany
47
CENTRA, Faculdade de Ciências, Universidade de Lisboa, Edif.
C8, Campo Grande, 1749-016 Lisboa, Portugal
48
Department of Informatics, Donald Bren School of Information and
Article number, page 15 of 21
A&A proofs: manuscript no. aa202039657
63
CITIC - Department of Computer Science and Information Technologies,
University of A Coruña, Campus de Elviña S/N, 15071,
A Coruña, Spain
64
Barcelona Supercomputing Center (BSC) - Centro Nacional de
Supercomputación, c/ Jordi Girona 29, Ed. Nexus II, 08034
Barcelona, Spain
65
University of Vienna, Department of Astrophysics, Türkenschanzstraße
17, A1180 Vienna, Austria
66
European Southern Observatory, Karl-Schwarzschild-Str. 2, 85748
Garching, Germany
67
Kapteyn Astronomical Institute, University of Groningen,
Landleven 12, 9747 AD Groningen, The Netherlands
68
School of Physics and Astronomy, University of Leicester, University
Road, Leicester LE1 7RH, United Kingdom
69
Center for Research and Exploration in Space Science and Technology,
University of Maryland Baltimore County, 1000 Hilltop Circle,
Baltimore MD, USA
70
GSFC - Goddard Space Flight Center, Code 698, 8800 Greenbelt
Rd, 20771 MD Greenbelt, United States
71
EURIX S.r.l., Corso Vittorio Emanuele II 61, 10128, Torino, Italy
72
Harvard-Smithsonian Center for Astrophysics, 60 Garden St., MS
15, Cambridge, MA 02138, USA
73
HE Space Operations BV for European Space Agency (ESA),
Camino bajo del Castillo, s/n, Urbanizacion Villafranca del
Castillo, Villanueva de la Cañada, 28692 Madrid, Spain
74
CAUP - Centro de Astrofisica da Universidade do Porto, Rua das
Estrelas, Porto, Portugal
75
SISSA - Scuola Internazionale Superiore di Studi Avanzati, via
Bonomea 265, 34136 Trieste, Italy
76
Telespazio for CNES Centre Spatial de Toulouse, 18 avenue
Edouard Belin, 31401 Toulouse Cedex 9, France
77
University of Turin, Department of Computer Sciences, Corso
Svizzera 185, 10149 Torino, Italy
78
Dpto. de Matemática Aplicada y Ciencias de la Computación,
Univ. de Cantabria, ETS Ingenieros de Caminos, Canales y Puertos,
Avda. de los Castros s/n, 39005 Santander, Spain
79
Centro de Astronomía - CITEVA, Universidad de Antofagasta,
Avenida Angamos 601, Antofagasta 1270300, Chile
80
Vera C Rubin Observatory, 950 N. Cherry Avenue, Tucson, AZ
85719, USA
81
Centre for Astrophysics Research, University of Hertfordshire,
College Lane, AL10 9AB, Hatfield, United Kingdom
82
University of Antwerp, Onderzoeksgroep Toegepaste Wiskunde,
Middelheimlaan 1, 2020 Antwerp, Belgium
83
INAF - Osservatorio Astronomico d’Abruzzo, Via Mentore Maggini,
64100 Teramo, Italy
84
Instituto de Astronomia, Geofìsica e Ciências Atmosféricas, Universidade
de São Paulo, Rua do Matão, 1226, Cidade Universitaria,
05508-900 São Paulo, SP, Brazil
85
Mésocentre de calcul de Franche-Comté, Université de FrancheComté,
16 route de Gray, 25030 Besançon Cedex, France
86
SRON, Netherlands Institute for Space Research, Sorbonnelaan 2,
3584CA, Utrecht, The Netherlands
87
Theoretical Astrophysics, Division of Astronomy and Space
Physics, Department of Physics and Astronomy, Uppsala University,
Box 516, 751 20 Uppsala, Sweden
88
RHEA for European Space Agency (ESA), Camino bajo del
Castillo, s/n, Urbanizacion Villafranca del Castillo, Villanueva de
la Cañada, 28692 Madrid, Spain
89
ATOS for CNES Centre Spatial de Toulouse, 18 avenue Edouard
Belin, 31401 Toulouse Cedex 9, France
90
School of Physics and Astronomy, Tel Aviv University, Tel Aviv
6997801, Israel
91
Astrophysics Research Centre, School of Mathematics and Physics,
Queen’s University Belfast, Belfast BT7 1NN, UK
92
Centre de Données Astronomique de Strasbourg, Strasbourg,
France
93
Université Côte d’Azur, Observatoire de la Côte d’Azur, CNRS,
Laboratoire Géoazur, Bd de l’Observatoire, CS 34229, 06304 Nice
Cedex 4, France
94
Max-Planck-Institut für Astrophysik, Karl-Schwarzschild-Straße
1, 85748 Garching, Germany
95
APAVE SUDEUROPE SAS for CNES Centre Spatial de Toulouse,
18 avenue Edouard Belin, 31401 Toulouse Cedex 9, France
96
Área de Lenguajes y Sistemas Informáticos, Universidad Pablo de
Olavide, Ctra. de Utrera, km 1. 41013, Sevilla, Spain
97
Onboard Space Systems, Luleå University of Technology, Box 848,
S-981 28 Kiruna, Sweden
98
113
IRAP, Université de Toulouse, CNRS, UPS, CNES, 9 Av. colonel
Roche, BP 44346, 31028 Toulouse Cedex 4, France
114
Konkoly Observatory, Research Centre for Astronomy and Earth
Sciences, MTA Centre of Excellence, Konkoly Thege Miklós út
15-17, 1121 Budapest, Hungary
115
MTA CSFK Lendület Near-Field Cosmology Research Group
116
ELTE Eötvös Loránd University, Institute of Physics, 1117,
Pázmány Péter sétány 1A, Budapest, Hungary
117
Ru ¯der Boškovi´c Institute, Bijeniˇcka cesta 54, 10000 Zagreb, Croa-
tia
118
Institute of Theoretical Physics, Faculty of Mathematics and
Physics, Charles University in Prague, Czech Republic
119
INAF - Osservatorio Astronomico di Brera, via E. Bianchi 46,
23807 Merate (LC), Italy
120
AKKA for CNES Centre Spatial de Toulouse, 18 avenue Edouard
Belin, 31401 Toulouse Cedex 9, France
121
Departmento de Física de la Tierra y Astrofísica, Universidad Complutense
de Madrid, 28040 Madrid, Spain
122
Vitrociset Belgium for European Space Agency (ESA), Camino
bajo del Castillo, s/n, Urbanizacion Villafranca del Castillo, Villanueva
de la Cañada, 28692 Madrid, Spain
123
Department of Astrophysical Sciences, 4 Ivy Lane, Princeton University,
Princeton NJ 08544, USA
124
Departamento de Astrofísica, Centro de Astrobiología (CSICINTA),
ESA-ESAC. Camino Bajo del Castillo s/n. 28692 Villanueva
de la Cañada, Madrid, Spain
125
naXys, University of Namur, Rempart de la Vierge, 5000 Namur,
Belgium
126
EPFL - Ecole Polytechnique fédérale de Lausanne, Institute of
Mathematics, Station 8 EPFL SB MATH SDS, Lausanne, Switzer-
land
127
H H Wills Physics Laboratory, University of Bristol, Tyndall Avenue,
Bristol BS8 1TL, United Kingdom
128
Sorbonne Université, CNRS, UMR7095, Institut d’Astrophysique
de Paris, 98bis bd. Arago, 75014 Paris, France
129
Porter School of the Environment and Earth Sciences, Tel Aviv
University, Tel Aviv 6997801, Israel
130
Laboratoire Univers et Particules de Montpellier, Université Montpellier,
Place Eugène Bataillon, CC72, 34095 Montpellier Cedex
05, France
131
Faculty of Mathematics and Physics, University of Ljubljana, Jadranska
ulica 19, 1000 Ljubljana, Slovenia
Article number, page 16 of 21
Gaia Collaboration et al.: Gaia Early Data Release 3
Appendix A: G-band corrections for sources with
6-parameter astrometric solutions
Table A.1 shows how to formulate an ADQL query, to be executed
in the Gaia EDR3 archive, that contains an on-thefly
calculation of the corrected G-band fluxes or magnitudes.
These queries are somewhat complex and create a performance
overhead. Hence downloading the requisite Gaia EDR3 fields
and calculating the corrections a posteriori may be more efficient.
Example Python code to do this is included in Table
A.2. The Python code is also available as a Jupyter notebook
at the following link: https://github.com/agabrown/
gaiaedr3-6p-gband-correction.
Appendix B: Calculating the corrected flux excess
factor
Table B.1 shows how to formulate an ADQL query, to be
executed in the Gaia EDR3 archive, that contains an onthe-fly
calculation of the corrected flux excess factor. This
query is somewhat complex and incurs a performance overhead.
Hence downloading the requisite Gaia EDR3 fields
and calculating the corrections a posteriori may be more efficient.
Example Python code to do this is included in Table
B.2. The Python code is also available as a Jupyter notebook
at the following link: https://github.com/agabrown/
gaiaedr3-flux-excess-correction.
Article number, page 17 of 21
A&A proofs: manuscript no. aa202039657
Table A.1. Example queries that can be submitted to the Gaia archive in the Astronomical Data Query Language to retrieve corrected G-band
photometry.
Query that includes a calculation of the G-band flux correction. The condition ‘bp_rp > -20’ ensures that no correction is attempted
in case the (GBP − GRP) colour is not available (‘bp_rp is not null’ does not work). The condition on random_index
is included to retrieve example data for a random sample of sources.
select source_id , astrometric_params_solved , bp_rp, phot_g_mean_mag , phot_g_mean_flux ,
if_then_else(
bp_rp > -20,
case_condition(
phot_g_mean_flux * (1.00525 -0.02323*greatest(0.25, least(bp_rp, 3))
+0.01740*power(greatest(0.25, least(bp_rp, 3)),2)
-0.00253*power(greatest(0.25, least(bp_rp, 3)),3)),
astrometric_params_solved != 95,
phot_g_mean_flux ,
phot_g_mean_mag < 13,
phot_g_mean_flux ,
phot_g_mean_mag < 16,
phot_g_mean_flux * (1.00876 -0.02540*greatest(0.25, least(bp_rp, 3))
+0.01747*power(greatest(0.25, least(bp_rp, 3)),2)
-0.00277*power(greatest(0.25, least(bp_rp, 3)),3))
),
phot_g_mean_flux
) as phot_g_mean_flux_corr
from gaiaedr3.gaia_source
where random_index between 1000000 and 1999999
Query that includes a calculation of the G-band magnitude correction. We note the type-cast ‘to_real()’ of the return value of the
conditional part of the query.
select source_id , astrometric_params_solved , bp_rp, phot_g_mean_mag , phot_g_mean_flux ,
if_then_else(
bp_rp > -20,
to_real(case_condition(
phot_g_mean_mag - 2.5*log10( (1.00525 -0.02323*greatest(0.25, least(bp_rp, 3))
+0.01740*power(greatest(0.25, least(bp_rp, 3)),2)
-0.00253*power(greatest(0.25, least(bp_rp, 3)),3)) ),
astrometric_params_solved != 95,
phot_g_mean_mag ,
phot_g_mean_mag < 13,
phot_g_mean_mag ,
phot_g_mean_mag < 16,
phot_g_mean_mag - 2.5*log10( (1.00876 -0.02540*greatest(0.25, least(bp_rp, 3))
+0.01747*power(greatest(0.25, least(bp_rp, 3)),2)
-0.00277*power(greatest(0.25, least(bp_rp, 3)),3)) )
)),
phot_g_mean_mag
) as phot_g_mean_mag_corr
from gaiaedr3.gaia_source
where random_index between 5000000 and 5999999
Article number, page 18 of 21
Gaia Collaboration et al.: Gaia Early Data Release 3
Table A.2. Python code for calculating the corrections to the G-band photometry for sources with 6-parameter astrometric solutions.
import numpy as np
def correct_gband(bp_rp, astrometric_params_solved , phot_g_mean_mag , phot_g_mean_flux):
"""
Correct the G-band fluxes and magnitudes for the input list of Gaia EDR3 data.
Parameters
----------
bp_rp: float, array_like
The (BP-RP) colour listed in the Gaia EDR3 archive.
astrometric_params_solved: int, array_like
The astrometric solution type listed in the Gaia EDR3 archive.
phot_g_mean_mag: float, array_like
The G-band magnitude as listed in the Gaia EDR3 archive.
phot_g_mean_flux: float, array_like
The G-band flux as listed in the Gaia EDR3 archive.
Returns
-------
The corrected G-band magnitudes and fluxes. The corrections are only applied to
sources with a 6-parameter astrometric solution fainter than G=13, for which a
(BP-RP) colour is available.
Example
gmag_corr , gflux_corr = correct_gband(bp_rp, astrometric_params_solved ,
phot_g_mean_mag , phot_g_mean_flux)
"""
if np.isscalar(bp_rp) or np.isscalar(astrometric_params_solved) or \
np.isscalar(phot_g_mean_mag) or np.isscalar(phot_g_mean_flux):
bp_rp = np.float64(bp_rp)
astrometric_params_solved = np.int64(astrometric_params_solved)
phot_g_mean_mag = np.float64(phot_g_mean_mag)
phot_g_mean_flux = np.float64(phot_g_mean_flux)
if not (bp_rp.shape == astrometric_params_solved.shape \
== phot_g_mean_mag.shape == phot_g_mean_flux.shape):
raise ValueError(’Function parameters must be of the same shape!’)
do_not_correct = np.isnan(bp_rp) | (phot_g_mean_mag <=13) | \
(astrometric_params_solved != 95)
bright_correct = np.logical_not(do_not_correct) & (phot_g_mean_mag >13) & \
(phot_g_mean_mag <=16)
faint_correct = np.logical_not(do_not_correct) & (phot_g_mean_mag >16)
bp_rp_c = np.clip(bp_rp, 0.25, 3.0)
correction_factor = np.ones_like(phot_g_mean_mag)
correction_factor[faint_correct] = 1.00525 - 0.02323*bp_rp_c[faint_correct] + \
0.01740*np.power(bp_rp_c[faint_correct],2) - \
0.00253*np.power(bp_rp_c[faint_correct],3)
correction_factor[bright_correct] = 1.00876 - 0.02540*bp_rp_c[bright_correct] + \
0.01747*np.power(bp_rp_c[bright_correct],2) - \
0.00277*np.power(bp_rp_c[bright_correct],3)
gmag_corrected = phot_g_mean_mag - 2.5*np.log10(correction_factor)
gflux_corrected = phot_g_mean_flux * correction_factor
return gmag_corrected , gflux_corrected
Article number, page 19 of 21
A&A proofs: manuscript no. aa202039657
Table B.1. Example query that can be submitted to the Gaia archive in the Astronomical Data Query Language to retrieve the corrected flux excess
factor presented in Riello et al. (2020).
Query that includes a calculation of the correction of the flux excess factor. The condition ‘bp_rp > -20’ ensures that no correction
is attempted in case the (GBP − GRP) colour is not available (‘bp_rp is not null’ does not work). We note the type-cast
‘to_real()’ of the return value of the conditional part of the query. The condition on random_index is included to retrieve
example data for a random sample of sources.
select source_id , bp_rp, phot_bp_rp_excess_factor ,
if_then_else(
bp_rp > -20,
to_real(case_condition(
phot_bp_rp_excess_factor - (1.162004 + 0.011464*bp_rp
+ 0.049255*power(bp_rp ,2)
- 0.005879*power(bp_rp ,3)),
bp_rp < 0.5,
phot_bp_rp_excess_factor - (1.154360 + 0.033772*bp_rp
+ 0.032277*power(bp_rp ,2)),
bp_rp >= 4.0,
phot_bp_rp_excess_factor - (1.057572 + 0.140537*bp_rp)
)),
phot_bp_rp_excess_factor
) as phot_bp_rp_excess_factor_corr
from gaiaedr3.gaia_source
where random_index between 1000000 and 1999999
Article number, page 20 of 21
Gaia Collaboration et al.: Gaia Early Data Release 3
Table B.2. Python code for calculating the corrected flux excess factor presented in Riello et al. (2020).
import numpy as np
def correct_flux_excess_factor(bp_rp, phot_bp_rp_excess_factor):
"""
Calculate the corrected flux excess factor for the input Gaia EDR3 data.
Parameters
----------
bp_rp: float, array_like
The (BP-RP) colour listed in the Gaia EDR3 archive.
phot_bp_rp_flux_excess_factor: float, array_like
The flux excess factor listed in the Gaia EDR3 archive.
Returns
-------
The corrected value for the flux excess factor, which is zero for "normal" stars.
Example
-------
phot_bp_rp_excess_factor_corr = correct_flux_excess_factor(bp_rp,
phot_bp_rp_flux_excess_factor)
"""
if np.isscalar(bp_rp) or np.isscalar(phot_bp_rp_excess_factor):
bp_rp = np.float64(bp_rp)
phot_bp_rp_excess_factor = np.float64(phot_bp_rp_excess_factor)
if bp_rp.shape != phot_bp_rp_excess_factor.shape:
raise ValueError(’Function parameters must be of the same shape!’)
do_not_correct = np.isnan(bp_rp)
bluerange = np.logical_not(do_not_correct) & (bp_rp < 0.5)
greenrange = np.logical_not(do_not_correct) & (bp_rp >= 0.5) & (bp_rp < 4.0)
redrange = np.logical_not(do_not_correct) & (bp_rp > 4.0)
correction = np.zeros_like(bp_rp)
correction[bluerange] = 1.154360 + 0.033772*bp_rp[bluerange] +
0.032277*np.power(bp_rp[bluerange],2)
correction[greenrange] = 1.162004 + 0.011464*bp_rp[greenrange] + \
0.049255*np.power(bp_rp[greenrange],2) \
- 0.005879*np.power(bp_rp[greenrange],3)
correction[redrange] = 1.057572 + 0.140537*bp_rp[redrange]
return phot_bp_rp_excess_factor - correction
Article number, page 21 of 21