Table information for 'xpparams.main'


This table is available for ADQL queries and through the TAP endpoint.

Resource Description: We present astrophysical parameters of 220 million stars, based on Gaia XP spectra and near-infrared photometry from 2MASS and WISE. Instead of using ab initio stellar models, we develop a data-driven model of Gaia XP spectra as a function of the stellar parameters, with a few straightforward built-in physical assumptions. This resource is a VO re-publication of the resulting catalog of stellar parameters. For bulk downloads, the covariances, the trained model, and more, see

For a list of all services and tables belonging to this table's resource, see Information on resource 'Parameters of 220 million stars from Gaia BP/RP (XP) spectra'

Citing this table

This table has an associated publication. If you use data from it, it may be appropriate to reference 2023arXiv230303420Z (ADS BibTeX entry for the publication) either in addition to or instead of the service reference.

To cite the table as such, we suggest the following BibTeX entry:

  title={Parameters of 220 million stars from Gaia BP/RP (XP) spectra},
  author={Zhang, Xiangyu and Green, Gregory M. and Rix, Hans-Walter},
  howpublished={{VO} resource provided by the {GAVO} Data Center}

Resource Documentation

Example: A Dust Map

As an example of what one can do with this data, consider the following ADQL query to generate an all-sky extinction map for stars at a distance between 400 pc and 600 pc:

select source_id/140737488355328 as hpx, avg(ext) as meanext
from xpparams.main
mod_parallax between 1.67 and 2.5
        and quality_flags < 8
        and mod_parallax/err_mod_parallax > 5
        and err_ext < 0.2
group by hpx

This exploits the fact that Gaia source ids can be converted to HEALPixes to produce a map and shows how to do useful quality cuts that allow relatively careless use of the data.

If you use TOPCAT to execute this on the GAVO DC TAP service (select Asynchronous mode and make sure you set Max Rows to something more than 50000, because that is how many pixels our map will have), you can do a sphere plot, then add a Healpix control. In it, select the table resulting from this query, manually set the HEALPix level to 6 and configure the Axis to Aitoff projection in the Galactic system to arrive at a plot like this:


Example: Metallicity of a Globular Cluster

Try the following query yielding data on Omega Cen and illustrating how to match with the local Gaia DR3 catalogue in order to constrain proper motions:

        g.pmra, g.pmdec, g.phot_g_mean_mag, g.phot_bp_mean_mag,
FROM xpparams.main as xpp JOIN gaia.edr3lite as g USING (source_id)
        distance(xpp.ra, xpp.dec, 201.697, -47.479472)<0.5
        AND distance(g.pmra, g.pmdec, -3.24, -6.73)<1.25
        AND xpp.quality_flags < 8
        AND xpp.mod_parallax/xpp.err_mod_parallax > 5.
        AND xpp.feh_confidence > 0.5
        AND xpp.err_fe_h < 0.2

Again in TOPCAT, try a 3d plot of RA, Dec and 1/mod_parallax. You will want to manually cut the parallax axis a bit to get rid of (presumably spurious) background stars. Make fe_h the aux axis in the “Form“ tab. If you look “from above“, you will see that the quality cuts have punched a hole into the cluster. If you look from the side, you can see that the cluster stars are (by and large) rather metal-poor compared to foreground stars:



Sorted by DB column index. [Sort alphabetically]

NameTable Head DescriptionUnitUCD
source_id Source Id Gaia DR3 unique source identifier. You can match this against gaia.dr3lite on this TAP service. N/A;meta.main
ra RA Gaia ICRS right ascension for this object. deg pos.eq.ra;meta.main
dec Dec Gaia ICRS declination for this object. deg pos.eq.dec;meta.main
teff Teff Estimated effective Temperature. Note that the raw HDF5 files released by Zhang et al. (2023) give Teff in a different unit (Kilokelvin). K phys.temperature.effective
fe_h [Fe/H] Log of Fe/H in solar units N/A phys.abund.Fe
logg log(g) Log of surface gravity in solar units N/A phys.gravity
ext Ext. Estimated extinction parameter. To convert to the extension at a particular wavelength, multiply this by that wavelength's value in the extinction curve, available at and in the footnote. [Note R] mag phys.absorption
mod_parallax Est. Par. Parallax estimated from the model. mas pos.parallax
err_teff Err. Teff Error in estimated effective temperature. Note that the raw HDF5 files released by Zhang et al. (2023) give the error in Teff in a different unit (Kilokelvin). K stat.error;phys.temperature.effective
err_fe_h Err. [Fe/H] Error in fe_h N/A stat.error;phys.abund.Fe
err_logg Err. log(g) Error in log_g N/A stat.error;phys.gravity
err_ext Err. Ext. Error in ext mag stat.error;phys.absorption
err_mod_parallax Err. Par. Error in the parallax estimated from the model. mas stat.error;pos.parallax
chi2_opt χ² χ² of the best-fit solution. Divide by 61 to obtain χ² per degree of freedom. N/A
ln_prior ln(prior) Natural log of the GMM prior on stellar type, at the location of the optimal solution. N/A
teff_confidence Conf. T A neural-network-based estimate of the confidence in the effective temperature estimate, on a scale of 0 (no confidence) to 1 (high confidence). N/A
feh_confidence Conf. [Fe/H] A neural-network-based estimate of the confidence in the [Fe/H] estimate, on a scale of 0 (no confidence) to 1 (high confidence). N/A
logg_confidence Conf. logg A neural-network-based estimate of the confidence in the log(g) estimate, on a scale of 0 (no confidence) to 1 (high confidence). N/A
quality_flags Quality Flags The three least significant bits represent whether the confidence in effective temperature, [Fe/H] and log(g) is less than 0.5, respectively. The 4th bit is set if chi2_opt/61 > 2. The 5th bit is set if ln_prior < -7.43. The 6th bit is set if our parallax estimate is more than 10 sigma from the GDR3 measurement (using reported parallax uncertainties from GDR3). The two most significant bits are always unset. We recommend a cut of quality_flags < 8 (the "basic reliability cut"), although a stricter cut of quality_flags == 0 ensures higher reliability at the cost of lower completeness. N/A meta.code.qual

Columns that are parts of indices are marked like this.


The following services may use the data contained in this table:


VO nerds may sometimes need VOResource XML for this table.


Note R

This table is originally at

wavelength extinction_curve
392.0 4.004029750823975
402.0 3.932934284210205
412.0 3.8616459369659424
422.0 3.790104627609253
432.0 3.66058611869812
442.0 3.6425302028656006
452.0 3.5208635330200195
462.0 3.42183256149292
472.0 3.3702077865600586
482.0 3.31054949760437
492.0 3.2349464893341064
502.0 3.103348970413208
512.0 3.028446674346924
522.0 2.913299798965454
532.0 2.8309452533721924
542.0 2.766111135482788
552.0 2.658158779144287
562.0 2.5784859657287598
572.0 2.5361363887786865
582.0 2.4805185794830322
592.0 2.412776470184326
602.0 2.356668472290039
612.0 2.313209056854248
622.0 2.272326707839966
632.0 2.2307708263397217
642.0 2.1915647983551025
652.0 2.156630516052246
662.0 2.1017379760742188
672.0 2.040684938430786
682.0 2.0002477169036865
692.0 1.9592469930648804
702.0 1.9099318981170654
712.0 1.8628478050231934
722.0 1.8250858783721924
732.0 1.7893791198730469
742.0 1.7572511434555054
752.0 1.7151219844818115
762.0 1.7081632614135742
772.0 1.6803325414657593
782.0 1.6145453453063965
792.0 1.566452145576477
802.0 1.5282924175262451
812.0 1.4971987009048462
822.0 1.4678758382797241
832.0 1.4417716264724731
842.0 1.4134937524795532
852.0 1.3825805187225342
862.0 1.3539363145828247
872.0 1.3050405979156494
882.0 1.2772356271743774
892.0 1.2389837503433228
902.0 1.2139592170715332
912.0 1.193802833557129
922.0 1.155206322669983
932.0 1.143110990524292
942.0 1.113704800605774
952.0 1.075573444366455
962.0 1.0743821859359741
972.0 1.0339452028274536
982.0 0.9810613989830017
992.0 1.0053104162216187
1235.0 0.6821472644805908
1662.0 0.4082775115966797
2159.0 0.2695426046848297
3352.6 0.13694290816783905
4602.8 0.09808244556188583