20.5.3 qso_candidates
This table contains parameters derived from various modules dedicated to the classification and characterisation of sources considered as QSO candidates. Together with those, the QSOs used to define the GaiaCRF3 are also listed in this table. This table has been constructed with the intention to be complete rather than pure and, as such, it will contain a large fraction of nongenuine extragalactic sources. Purer samples can be drawn using dedicated flags or queries. Please refer to Chapter 12 of the online documentation for details about how this table was built, its content, and for recommendations regarding its exploitation.
Columns description:
All Gaia data processed by the Data Processing and Analysis Consortium comes tagged with a solution identifier. This is a numeric field attached to each table row that can be used to unequivocally identify the version of all the subsystems that were used in the generation of the data as well as the input data used. It is mainly for internal DPAC use but is included in the published data releases to enable end users to examine the provenance of processed data products. To decode a given solution ID visit https://gaia.esac.esa.int/decoder/solnDecoder.jsp
A unique single numerical identifier of the source obtained from gaia_source (for a detailed description see gaia_source.source_id).
astrometric_selection_flag : Flag indicating if the source is part of the astrometric selection (boolean)
This flag indicates whether the source is part of the astrometric selection of the qso_candidates table that was constructed based on the principles originally that were used to select the sources defining the GaiaCRF3 and published in agn_cross_id. See also Gaia Collaboration et al. (2022g).
This boolean flag indicates whether the source is part of the GaiaCRF3, which is published through agn_cross_id.
See also Gaia Collaboration et al. (2022g).
vari_best_class_name : Name of best class, see table vari_classifier_class_definition for details of the class (string)
Best class name with corresponding classification score in best_class_score of table vari_classifier_result. For the ‘n_transits:5+’ classifier in DR3, the following classes are published: ‘ACVCPMCPROAMROAPSXARI’, ‘ACYG’, ‘AGN’, ‘BCEP’, ‘BEGCASSDORWR’, ‘CEP’, ‘CV’, ‘DSCTGDORSXPHE’, ‘ECL’, ‘ELL’, ‘EP’, ‘LPV’, ‘MICROLENSING’, ‘RCB’, ‘RR’, ‘RS’, ‘S’, ‘SDB’, ‘SN’, ‘SOLAR_LIKE’, ‘SPB’, ‘SYST’, ‘WD’, and ‘YSO’. The ‘GALAXY’ classifications are largely due to artificial variability (see Section 10.3.3 of the release documentation for details) and their probabilities are published exclusively in column vari_best_class_score of table galaxy_candidates.
See vari_classifier_class_definition for a detailed description of this classifier and its published classes.
It describes a quantity between 0 and 1 which is related to the (median) normalised rank of the confidence of the classifier(s) in the identification of the best class (vari_best_class_name). In the special case of class ‘EP’, all scores are set to 1. See Section 10.3.4 of the release documentation for details.
The fractional variability is calculated as $\sqrt{{\text{MAD}}^{2}(F){\text{RMS}}^{2}({\sigma}_{F})}/\text{median}(F)$, where $\text{MAD}(F)$ is the Median Absolute Deviation (MAD) of the fieldofview transit fluxes $F$ in the G band, ${\text{RMS}}^{2}({\sigma}_{F})$ the mean square of flux uncertainties ${\sigma}_{F}$, and $\text{median}(F)$ is the median flux of the fieldofview transits in the G band.
Index of the firstorder structure function (SF; Simonetti et al. 1985), i.e., slope in the logSF vs logTau space, where Tau is the time lag. The SF is expressed in magnitude squared and computed from fieldofview transit magnitudes in the G band. The index is linked to that of the Fourier power spectrum and indicates the type of noise process at work.
structure_function_index_scatter : Standard deviation of the index of the structure function (double)
Standard deviation of the index $\alpha $ of the structure function (i.e., slope in the logSF vs logTau space). The structure function is expressed in magnitude square. The index is linked to the power of the Fourier power spectrum and indicates the type of noise process at work. AGN should show $\alpha =0.5$, in between flicker noise ($\alpha =0$) and shot (random walk) noise ($\alpha =1$).
Quasar variability metric from fieldofview transit magnitudes in the G band in log format, $\mathrm{log}({\chi}_{\mathrm{QSO}}^{2}/\nu )$, from Butler and Bloom (2011), after adaptation to Gaia data.
Nonquasar variability metric from fieldofview transit magnitudes in the G band in log format, $\mathrm{log}({\chi}_{\mathrm{false}}^{2}/\nu $), from Butler and Bloom (2011), after adaptation to Gaia data.
vari_agn_membership_score : Membership score (0=lowest,1=highest) of source to be of AGN type (double)
Membership score (0=lowest,1=highest) of source to be of AGN type.
Each specific object module has its own method of computing this score so this value cannot be perceived as a probability, nor can it readily be compared between different specific object modules without recalibration (apart from the extremes 0 and 1).
For SOSAGN (Chapter 10.4), it is defined as the Mahalanobis distance from the AGIS3.1 reference set of QSOs, inverted and rescaled by a Gaussian from 0 to 1, which takes into account five variability features (and their covariances) from the Gaia time series: the Butler & Bloom parameters (QSOvar and nonQSOvar), the structure function slope, the fractional variability, and the Abbe (or von Neumann) parameter.
classprob_dsc_combmod_quasar : Probability from DSCCombmod of being a quasar (data used: BP/RP spectrum, photometry, astrometry) (float)
Probability that the object is of the named class.
This is the overall probability for this class, computed by combining the class probabilities from DSCSpecmod (which classifies objects using BP/RP spectra) and DSCAllosmod (which classifies objects using several astrometric and photometric features). It is important to realise that the DSC classes are defined by the training data used, and that this may produce a narrower definition of the class than may be expected given the class name. This is a posterior probability that includes the global class prior, given in the documentation.
classprob_dsc_combmod_galaxy : Probability from DSCCombmod of being a galaxy (data used: BP/RP spectrum, photometry, astrometry) (float)
Probability that the object is of the named class.
This is the overall probability for this class, computed by combining the class probabilities from DSCSpecmod (which classifies objects using BP/RP spectra) and DSCAllosmod (which classifies objects using several astrometric and photometric features). It is important to realise that the DSC classes are defined by the training data used, and that this may produce a narrower definition of the class than may be expected given the class name. This is a posterior probability that includes the global class prior, given in the documentation.
classlabel_dsc : Class assigned by DSC based on the probability from its Combmod classifier (string)
Class assigned by DSC based on the probability from its Combmod classifier.
DSCCombmod provides a normalized posterior probability vector across several classes. Note that this incorporates the global class prior, as explained in the documentation. This class label is set to that class with the largest probability above 0.5. If no probability is above 0.5, this class label is ‘unclassified’. If users want to perform classification using a different threshold, or by adopting a different prior, they should use the DSC probability vectors.
classlabel_dsc_joint : Class assigned by DSC based on the probability from its Specmod and Allosmod classifiers (string)
DSCSpecmod and DSCAllosmod each provide a normalized posterior probability vector across several classes. Both incorporate the global class prior, as explained in the documentation. If the ‘quasar’ class probability from both Specmod and Allosmod are above 0.5, this class label is set to ‘quasar’. If the ‘galaxy’ class probability from both Specmod and Allosmod are above 0.5, this class label is set to ‘galaxy’. (Note that these two cases are mutually exclusive.) Otherwise the class label is set to ‘unclassified’.
Class label of the neuron that represents the source, as assigned by the OA module in Apsis. See Section 11.3.12 for further details.
The redshift of the source estimated from the analysis of the BP/RP spectra by the QSOC Apsis module.
The redshift of the source is inferred from the crossmatch of a quasar template to the observed BP/RP spectra in order to produce the cross correlation function that corresponds to the negative of the ${\chi}^{2}$ values evaluated in each trial redshift, plus a constant. The construction of the cross correlation function and subsequent selection of the redshift is described in Section 11.3.14 of the online documentation.
Lower confidence level of the redshift estimate based on the BP/RP spectra analysis. This is the 16th percentile of the PDF over redshift as computed by QSOC.
Upper confidence level of the redshift estimate based on the BP/RP spectra analysis. This is the 84th percentile of the PDF over redshift as computed by QSOC.
ccfratio_qsoc : Value of the crosscorrelation function used to derive the redshift from QSOC, relative to the maximum value (float)
Ratio of the value of the cross correlation function evaluated at the selected redshift estimate to the maximum of the cross correlation function over all redshifts. Values are in the range [0,1]. When different from one, low values of this parameter indicate an inappropriate fit of the quasar template to the observed spectrum (i.e. a high ${\chi}^{2}$) while large values indicate potential degeneracy in the redshift determination.
See Section 11.3.14 of the online documentation for further details.
The parameter zscore is defined in Section 11.3.14 of the online documentation and takes values between 0 and 1. Low values of zscore indicate that at least one emission line commonly found in the spectra of QSOs is not present.
The processing flags report the potential errors that can occur during the prediction process from QSOC.
The flag value is a binary combination (binary OR) of:

•
0: Z_NOWARNING. The processing of this source raised no warning flag.

•
1: Z_AMBIGUOUS. The cross correlation function has more than one maximum with ccfratio_qsoc $>$ 0.85, meaning that at least two redshifts lead to a similar ${\chi}^{2}$ and the solution is ambiguous.

•
2: Z_LOWCCFRATIO. The selected redshift leads to a small value of the cross correlation function when compared to the maximum value of the cross correlation function. Equivalently, ccfratio_qsoc $$ 0.9.

•
4: Z_LOWZSCORE. The selected redshift leads to a low zscore_qsoc, meaning that at least one emission line is either missing or is strongly damped. Equivalently, zscore_qsoc $$.

•
8: Z_NOTOPTIMAL. We did not choose the redshift having the lowest ${\chi}^{2}$ (i.e. ccfratio_qsoc $$ 1).

•
16: Z_BADSPEC. Raised if one of the following conditions is met:

–
the number of BP or RP spectral transits (${N}_{\mathrm{BP}}$ and ${N}_{\mathrm{RP}}$ hereafter), is lower than 10 transits, or

–
$G>$ 20.5 mag, or

–
$G>19+0.03\times ({N}_{\mathrm{BP}}10)$ mag, or

–
$G>19+0.03\times ({N}_{\mathrm{RP}}10)$ mag.
This allows the user to filter out uncertain predictions in a simple way without having to explicitly deal with the aforementioned formula, nor with ${N}_{\mathrm{BP}}$ and ${N}_{\mathrm{RP}}$.

–
For a definition of ccfratio_qsoc and zscore_qsoc, see Section 11.3.14 of the online documentation.
The number of transits used to reconstruct the image and to analyse the object morphology.
intensity_quasar : Fitted intensity of the quasar at its centre (double, Flux[e${}^{}$ s${}^{1}$])
Light intensity at the centre of the quasar as obtained by fitting an exponential profile for the central quasar and a Sersic profile for the host galaxy.
intensity_quasar_error : Error on the fitted intensity of the quasar at its centre (double, Flux[e${}^{}$ s${}^{1}$])
Uncertainty on the light intensity at the centre of the quasar as obtained by fitting an exponential profile for the central quasar and a Sersic profile for the host galaxy.
intensity_hostgalaxy : Fitted intensity of the host galaxy at the effective radius (double, Flux[e${}^{}$ s${}^{1}$])
Light intensity at the centre of the host galaxy as obtained by fitting an exponential profile for the central quasar and a Sersic profile for the host galaxy.
intensity_hostgalaxy_error : Error on the fitted intensity of the host galaxy at effective radius (double, Flux[e${}^{}$ s${}^{1}$])
Formal uncertainty on the light intensity of the host galaxy at the effective radius as obtained by fitting an exponential profile for the central quasar and a Sersic profile for the host galaxy.
Effective radius containing half of the total luminosity of the Sersic profile used to fit the host galaxy.
radius_hostgalaxy_error : Error on the fitted effective radius of the host galaxy (double, Angle[mas])
Uncertainty on the effective radius containing half of the total luminosity of the Sersic profile used to fit the host galaxy.
Sersic index as obtained by fitting a Sersic profile for the host galaxy.
The Sersic profile is a mathematical function that describes how the intensity of a galaxy varies with the distance from its centre. The Sersic index describes how steep is this variation.
Formal uncertainty on sersic_index.
Ellipticity (defined as 1($b/a$) where ($b/a$) is the axis ratio) of the host galaxy obtained by the morphological fitting of a Sersic profile for the host galaxy.
Formal uncertainty on the ellipticity of the host galaxy as obtained by the morphological fitting of a Sersic profile.
Position angle of the host galaxy obtained by the morphological fitting of the Sersic profile with respect to Celestial North Pole.
posangle_hostgalaxy_error : Error on the fitted position angle of the host galaxy (double, Angle[deg])
Formal uncertainty on the position angle of the host galaxy obtained by the morphological fitting of the Sersic profile with respect to Celestial North Pole.
Flag indicating the presence of a detectable host galaxy. False means that no host galaxy could be detected, while True means that a measurable host galaxy was detected around the quasar.
This value represents the mean squared error between the integrated flux of all observed samples (from the Sky Mapper and Astrometric Field) and the integrated flux of synthetic samples produced with the fitted profile.
morph_params_corr_vec : Vector form of the upper triangle of the correlation matrix for the fitted morphological parameters (double[15] array)
Correlation matrix of the fitted profile parameters in the following order:

0:
intensity of the quasar

1:
intensity of the host galaxy

2:
radius of the host galaxy

3:
Sersic index of the host galaxy

4:
ellipticity of the host galaxy

5:
position angle of the host galaxy
Only nonzero, nonunity, correlation coefficients from the correlation matrix M are provided here. They are served as a linear array of constant size $S=n(n1)/2$ corresponding to the full normal matrix of dimension $n\times n$. The ordering of the elements in the linear array follows a columnmajor storage scheme, i.e.:
$\mathbf{M}=\left[\begin{array}{ccccccc}\hfill 1\hfill & \hfill C[0]\hfill & \hfill C[1]\hfill & \hfill C[3]\hfill & \hfill C[6]\hfill & \hfill \mathrm{\cdots}\hfill & \hfill C[S(n1)]\hfill \\ \hfill \hfill & \hfill 1\hfill & \hfill C[2]\hfill & \hfill C[4]\hfill & \hfill C[7]\hfill & \hfill \mathrm{\cdots}\hfill & \hfill C[S(n2)]\hfill \\ \hfill \hfill & \hfill \hfill & \hfill 1\hfill & \hfill C[5]\hfill & \hfill C[8]\hfill & \hfill \mathrm{\cdots}\hfill & \hfill C[S(n3)]\hfill \\ \hfill \hfill & \hfill \hfill & \hfill \hfill & \hfill 1\hfill & \hfill C[9]\hfill & \hfill \mathrm{\cdots}\hfill & \hfill C[S(n4)]\hfill \\ \hfill \hfill & \hfill \hfill & \hfill \hfill & \hfill \hfill & \hfill \mathrm{\ddots}\hfill & \hfill \mathrm{\ddots}\hfill & \hfill \mathrm{\vdots}\hfill \\ \hfill \hfill & \hfill \hfill & \hfill \hfill & \hfill \hfill & \hfill \hfill & \hfill 1\hfill & \hfill C[S1]\hfill \\ \hfill \hfill & \hfill \hfill & \hfill \hfill & \hfill \hfill & \hfill \hfill & \hfill \hfill & \hfill 1\hfill \end{array}\right]$
host_galaxy_flag : Flag indicative of processing or scientific quality for the morphological parameters fitting (byte)
This flag provides information about the processing or scientific quality of the results of the Quasar morphology analysis chain for the source of interest.
The flag coding is the following:

1:
Host Galaxy measured with circular Sersic profile

2:
Host Galaxy measured with elliptical Sersic profile

3:
No host galaxy detected

4:
Poor solution measured with elliptical Sersic profile

5:
No convergence of fitting but host galaxy detected

6:
No convergence of fitting or doubtful solution due to presence of a secondary source at d $$ 5 arcsec
source_selection_flags : Bit indicative of whether the input data from a given module met the source list eligibility criteria for the source of interest (int)
Bit indicative of whether the input data from a given module met the source list eligibility criteria for the source of interest.
The bit is coded as follows:

•
bit 0: The source belongs to the GaiaCRF3 (agn_cross_id table)

•
bit 1: The source belongs to the frame_rotator_source table with either
used_for_reference_frame_orientation or used_for_reference_frame_spin set to True. 
•
bit 2: The source meets the eligibility criteria for the output of the Surface brightness analysis module (see Chapter 9).

•
bit 3: The source belongs to the qso_catalogue_name table.

•
bit 4: The source meets the eligibility criteria for the output of the classification module based on photometric lightcurves (see Chapter 10.3).

•
bit 5: The source meets the eligibility criteria for the output of the SOS AGN module (see Chapter 10.4)

•
bit 6: The source meets the eligibility criteria for the output of the DSC module (see Chapter 11.3.2).

•
bit 7: The source meets the eligibility criteria for the redshift determined by the QSOC module (see Chapter 11.3.14).

•
bit 8: The source meets the eligibility criteria for the classification output of the QSOC module (not yet applicable to DR3).