Refining open cluster parameters with Gaia XP metallicities

TL;DR

This work tests whether metallicities derived from Gaia XP (low-resolution) spectra can break the age–metallicity degeneracy in open-cluster parameter estimation using Gaia DR3 photometry. By applying a synthetic CMD method via ASteCA with Approximate Bayesian Computation, the authors derive age, distance modulus, and extinction for 20 open clusters, using XP-based metallicities from ARC, ZGR, and FS as priors (or treating metallicity as a free parameter). They find that XP-derived [Fe/H] values are offset from high-resolution spectroscopy by about 0.1–0.15 dex, but, when used as priors, yield ages comparable to high-resolution results and more precise than neural-network–based photometric methods, especially for clusters lacking a populated RGB. The study shows Gaia XP metallicities can significantly improve cluster parameter estimates while avoiding traditional spectroscopy, though extinctions and distances remain challenging in some cases. Overall, Gaia data, with careful quality cuts and priors, enable accurate, large-scale Galactic cluster analyses and underline the potential for XP-based metallicities in future work.

Abstract

The precision of cluster parameter determination has significantly improved with the availability of homogeneous photometric Gaia data, however, challenges such as age-metallicity degeneracy and lack of spectroscopic observations remain. In this paper we investigate whether metallicities derived from low-resolution Gaia XP spectra can be effectively used to break degeneracies and improve the accuracy of OC parameter determinations. We analysed 20 OCs using isochrone fitting methods on Gaia DR3 photometry and metallicity estimates from several Gaia XP-based catalogues. We derived age, distance modulus, and extinction using the Approximate Bayesian Computation (ABC). We compared the parameter estimates to the values obtained in other works through isochrone fitting with spectroscopically constrained metallicities or through neural network techniques applied only to the photometry. We found the systematic difference between Gaia XP derived metallicities and those obtained from high-resolution spectroscopy to be 0.1-0.15 dex. We found a systematic age difference of <0.03 +- 0.13 dex compared to isochrone fitting using high-resolution spectroscopy, and <0.08 +- 0.21 dex compared to neural network-based methods, and a median individual error of ~0.065 dex. Despite their low resolution, Gaia XP metallicities effectively constrain parameters of clusters lacking a well-populated RGB. When used with stringent quality cuts and incorporated as priors, they allow to determine ages comparable in precision to those based on high-resolution spectroscopy and more precise than photometry-only neural network methods. These results highlight the potential of Gaia data for accurate cluster parameter analysis and detailed Galactic studies without relying on traditional spectroscopy.

Figures (14)

• Figure 1: CMD of the open cluster IC 4756 from Gaia DR3 photometry. The isochrones correspond to the derived parameters based on the metallicity prior from the respective Gaia XP-based stellar parameters estimates catalogue (see Section \ref{['sect:metallicity']} for the description of the catalogues), namely ARC (red), ZGR (blue) and FS (yellow). The grey isochrone corresponds to the fit performed with metallicity as a free parameter.
• Figure 2: Metallicities determined from the catalogues ARC, ZGR and FS of the open cluster NGC 3532 as a function of Gaia $G_\mathrm{BP}-G_\mathrm{RP}$ colour. Empty circles show the estimates that were discarded after the respective cuts, whereas filled circles show the values that satisfied the cuts' criteria. The black line and the shaded area show the [Fe/H] and $\sigma$[Fe/H] determined with high-resolution spectroscopy by Netopil16 and used by B19, whereas the points with error bars show the weighted mean [Fe/H] and its dispersion of the respective catalogue. The percentage indicates the fraction of the stars that were accepted and used for [Fe/H] determination after the respective cuts.
• Figure 3: Top panels: Metallicity of individual OCs of the sample derived from three Gaia XP-based catalogues (on the Y axis) presented against that from three high-resolution spectroscopic surveys (on the X axes of the left, middle and right plots). Bottom: Difference between the two values for each cluster. The same colour-coding scheme applies as in Fig. \ref{['fig:cmd']}. The 1:1 relation is shown by the diagonal black line in the top plot and, respectively, by the horizontal line indicating zero difference in the bottom plot. The number in the top corner indicates the median difference and the root mean square difference (RMSD) of the respective catalogue.
• Figure 4: Iron abundance dispersion $\sigma \text{[Fe/H]}$ against metallicity [Fe/H] for individual open clusters. The same colour-coding scheme applies as in Fig. \ref{['fig:cmd']}.
• Figure 5: Metallicity of individual OCs of the HRS sample determined as a free parameter in the procedure. The grey colour denotes this work; the green colour shows the predictions of the ANN of cavallo24. Figure layout is the same as in Fig. \ref{['fig:met_errors']}: X axis shows the [Fe/H] from high-resolution spectroscopy, the bottom plot shows the difference between the values, and the black line indicates 1:1 relation on the top plot and difference of zero in the bottom plot. Clusters with three or fewer red giant branch (RGB) stars are marked with an X symbol. The number in the bottom left corner indicates the median difference and RMSD.