skip to main content

gaia data release 3 documentation

9.7 Validation

9.7.1 Galaxies

Sérsic profile

While analysing the space-based studies of galaxy morphology using HST data, only very few galaxies were in common with our list of processed sources: one in common with Trujillo and Aguerri (2004), one with van der Wel et al. (2012), two with Dimauro et al. (2018) and one with dos Reis et al. (2020). The comparison of the Sérsic index, ellipticities and position angles of Gaia DR3 with these studies is given in Figure 9.8. This comparison, although very limited because of the small number of comparison objects shows a good agreement between these studies and the Gaia parameters.

Figure 9.8: Comparison of the Gaia parameters of the Sérsic profile of galaxies with HST: Sérsic index, ellipticities (defined as 1-b/a) where (a,b) are the semi-major and semi-minor axis, position angles [] (from North to East).

There are a few ground-based surveys that adjust a free Sérsic profile (as GAMMA, NASA-Sloan-Atlas (Maller et al. 2009) (hereafter NASATLAS) or the work from Simard et al. (2011)). The Sérsic indexes given in Gaia DR3 and in these surveys are not comparable. The main reason is the atmospheric seeing that modifies the inner light profile of the sources (see Balcells et al. (2003) for similar analysis). Nonetheless, the shape parameters (ellipticity and position angle) of the galaxies derived from space and from ground should globally compare. We present the comparison of the Gaia shape parameters of Sérsic profile with the SDSS DR16 de Vaucouleurs profile in Figure 9.9. The comparison of position angle is excellent, the sources that depart from the diagonal usually exhibit a small ellipticity, for which the position angle parameter is meaningless. The comparison of ellipticities presents a systematic trend: Gaia tends to see galaxies rounder than SDSS. This can be explained by the fact that Gaia is observing the inner part of most galaxies. The fitted ellipticities are influenced by the bulge shape. Another effect concerns the projected ellipticity that is measured and which is varying along the radius, see for instance Ferrari et al. (2004). A more global comparison of the ellipticities given in SDSS DR16 with several surveys exhibits a systematic drift with DES or SPLUS (Mendes de Oliveira et al. 2019) surveys and a good agreement is observed in the comparison with GAMMA and NASA-Sloan-Atlas (NASATLAS) surveys. It is then very difficult to conclude which survey is at the origin of the observed drift.

Figure 9.9: Comparison of the Gaia position angle () (left) and ellipticity (middle) of the Sérsic profile of galaxies with SDSS DR16 (de Vaucouleurs profile), coloured along density. Position angles are defined from North to East. Ellipticity = 1-b/a where (a,b) are the semi-major and semi-minor axis. Comparison of SDSS DR16 ellipticities (de Vaucouleurs profile) with several surveys (right). Coloured lines indicate the median behaviour of the distributions.

de Vaucouleurs profile

To our knowledge, there is no space-based survey that adjust a de Vaucouleurs profile on galaxy. The SDSS DR16 ground-based survey is providing a de Vaucouleurs profile for 390 615 galaxies in common with the list of sources used in this study. The problem of the seeing modifying the shape of the light profile is still present and limit the comparison that can be done. Nonetheless, the de Vaucouleurs profile is a Sérsic profile with index fixed at 4, therefore the comparison may still be relevant. We present the comparison of the parameters derived with Gaia data and by the SDSS in Figure 9.10. The accordance between both radii is reasonable, with no systematic except for the very large objects for which Gaia is in the extrapolation regime. The position angles present a very good accordance as for the Sérsic model. The same systematic of ellipticities is observed as for Sérsic profile.

Figure 9.10: Comparison of the Gaia effective radius (top), position angle [] (middle) and ellipticity (bottom) of the de Vaucouleurs profile of galaxies with SDSS DR16 (de Vaucouleurs profile). The blue line is the median of the distribution. Position angles are defined from North to East. Ellipticity = 1-b/a where (a,b) are the semi-major and semi-minor axis.

Internal coherence

To some extent the comparison of the two profiles fitted for the galaxies allow some internal assessment of the validity of the results. We have compared the position angles and ellipticities derived by the Sérsic and the de Vaucouleurs models (Figure 9.11). To compare the effective radii given by the two models, we selected in the Sérsic model the sources with a Sérsic index close to 4 which corresponds to a de Vaucouleurs profile. The radii should then be comparable (Figure 9.11).

There is an excellent agreement between the parameters fitted by the two models which means that the models are coherent.

Figure 9.11: Comparison of the ellipticities (top) and the position angles (middle) of the Sérsic and the de Vaucouleurs profiles. Comparison of the effective radii derived with the Sérsic and the de Vaucouleurs profiles for a selection of sources that have a Sérsic index close to 4 (bottom). These sources correspond to a typical de Vaucouleurs profile (index=4).

Exponential profile

Theoretical analysis (de Bruijne et al. (2015) and de Souza et al. (2014)) indicate that the on-board video processing unit filters out all disc galaxies. These galaxies are usually modelled by an exponential profile. There are only few hundreds objects for which the free Sérsic profile fitted converged towards a Sérsic index compatible with a pure disc (index <1.5).

Despite this, an exponential profile has also been adjusted on the galaxies as:


where IC is the central light intensity of the exponential profile and rd is the scale length (radius where the intensity drops by a factor e from the central intensity).

The results showed that very few objects were well described by an exponential profile. The bulges of the galaxies usually follow a de Vaucouleurs profile. The fitting of an exponential profile extrapolates, for most objects, the scale length. This tends to lower this quantity as shown in Figure 9.12 presenting the comparison of the effective radius of the profile fitted with the SDSS DR16 exponential profile. The radii fitted are systematically smaller than those given by SDSS. The size of the difference is correlated to the size of the object. These differences mostly arise from the fact that the shape of most Gaia galaxies does not follow an exponential profile, as predicted. Then the fitting process tends to extrapolate the radius with systematic, while SDSS used images of the whole objects with larger influence of the disc. In simulations of bulge plus disc galaxies, the fitting of an exponential profile led to radii following mostly the bulge characteristics.

Figure 9.12: Comparison of effective radius of the exponential profile obtained with Gaia data with the SDSS DR16 values: (left) direct comparison, (right) presented as difference for radii expressed as a function of the number of effective radius included in the AF window.