EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 677 (September 2023) A&A, 677 (2023) A162 Full HTML
Open Access
Issue
A&A
Volume 677, September 2023
Article Number A162
Number of page(s) 8
Section Galactic structure, stellar clusters and populations
DOI https://doi.org/10.1051/0004-6361/202346790
Published online 21 September 2023

© The Authors 2023

Licence Creative CommonsOpen Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This article is published in open access under the Subscribe to Open model.

Open access funding provided by Max Planck Society.

1. Introduction

Precisely determined stellar parameters are key to many astrophysical investigations. The Gaia astrophysical parameters inference system (Apsis, Bailer-Jones et al. 2013) was devised to extend the scientific content and potential of Gaia beyond its core astrometric mission of providing precise measurements of positions, parallaxes, and proper motions for more than 109 sources. Gaia Data Release 3 (DR3) includes astrophysical parameters of up to 470 million sources. The Apsis estimates are based on averaged low-resolution spectra in the Gaia BP and RP bands with a resolution of R = 20–60 combined with Gaia photometry and astrometry, and on data from the Radial Velocity Spectrograph (RVS), which provides R ≈ 11 500 spectra in the wavelength range from 846 to 870 nm for stars with G ≲ 15.2 mag. This is used to infer stellar astrophysical parameters such as effective temperature, age, mass, surface gravity log g, global metallicity [M/H], and elemental abundances (Creevey et al. 2023; Fouesneau et al. 2023; Delchambre et al. 2023).

Gaia Apsis promises to unlock vastly improved insights into a variety of research topics, such as the solar neighbourhood and young nearby moving groups, stellar structure and evolution, open and globular clusters, and the overall structure of the Milky Way with its distinct streams, bar, and spiral arms as well as its more diffuse disc and halo components. Precise stellar astrophysical parameters are also essential for the study of exoplanets, whose properties are in general determined relative to the properties of the stellar host (e.g., Magrini et al. 2022; Berger et al. 2023).

As pointed out by Bailer-Jones et al. (2013), the veracity of the Apsis methods can only be assessed on real data. This provides our motivation to use independently determined properties of bona fide single stars in the Hyades and Pleiades open clusters as external validations of, and benchmarks for, Gaia DR3 Apsis estimates.

The structure of the paper is as follows. In Sect. 2, we review the compilation of the single-star sequence of the Hyades open cluster, and apply the same methodology to the Pleiades to validate PARSEC isochrones. In Sect. 3, we compare cluster ensembles and individual stellar astrophysical properties with Apsis estimates. In Sect. 4, we discuss the findings and potential explanations for some of the issues identified. We conclude in Sect. 5 with recommendations as to the utilisation of DR3 Apsis estimates and an outlook.

2. Single-star sequences in the Hyades and Pleiades

At average distances of ≈45 pc (van Leeuwen 2009; Röser et al. 2011; Gaia Collaboration 2018) and ≈135 pc (e.g., Percival et al. 2005; Heyl et al. 2022), respectively, the Hyades and the Pleiades are among the most nearby open clusters. The vicinity ensures high-quality Gaia astrometric, photometric, and spectroscopic measurements over a wide range of spectral types and stellar masses. Age estimates for the Hyades derived from isochrone fitting are in the range of 500–800 Myr (Brandner et al. 2023b). For the Pleiades, the age estimates derived from isochrone fitting centre around 100–125 Myr (Table 1), which is in good agreement with age estimates based on the lithium-depletion boundary (e.g., Basri et al. 1996; Stauffer et al. 1998; Dahm 2015; Galindo-Guil et al. 2022).

Table 1.

Compilation of abundance, αML, and age estimates for the Pleiades open cluster based on isochrone fitting.

The stellar samples of candidate members of the open clusters are based on Gaia Collaboration (2021b) for the Hyades, and on Heyl et al. (2022) for the Pleiades. In Brandner et al. (2023b), we processed the Hyades sample by identifying and removing stars with dubious photometry in at least one of the Gaia DR3 G, BP, or RP bands. We then used the Gaia renormalised unit weight error (RUWE) parameter to identify likely astrometric binaries. In a final step, likely photometric (i.e., blended) binaries were identified based on their distance from the median stellar sequence in colour–magnitude space, taking into account the uncertainties in absolute brightness and colour. This resulted in a sample of about 600 bona fide single stars in the Hyades open cluster. In Brandner et al. (2023a), we derived the astrophysical parameters for these stars by comparison with the best-fitting family of isochrones from the PAdova and TRieste Stellar Evolution Code (PARSEC, Bressan et al. 2012; Chen et al. 2014).

Here we apply the same methodology to the Pleiades. The sample of approximately 1300 stars within 10 pc of the nominal centre of the Pleiades open cluster as defined by Heyl et al. (2022) includes 1191 sources with −0.2 ≤ BP − RP < 3.9 mag and −2.0 ≤ Gabs ≤ 13.5 mag. After rejection of 112 stars with dubious Gaia photometry, and the identification of 171 stars as likely astrometric or photometric binaries, we retain a sample of 908 bona fide single stars. We employ the PARSEC CMD 3.7 web interface1 to obtain version 1.2S grids of non-rotating isochrones in the Gaia EDR3 photometric system for a range of ages and [M/H], and for a visual extinction of AV = 0.12 mag (Meynet et al. 1993; van Leeuwen 2009, and assuming AV = 3.1 * E(B − V))2. Absolute Gabs magnitudes are based on the photo-geometric distance estimates provided by Bailer-Jones et al. (2021). χ2 minimisation is used to find the best-fitting isochrones to the 235 brightest (Gabs ≤ 8.0 mag) bona fide single upper-main sequence stars. This yields lg(age [yr]) = 8.10 ± 0.07 and [M/H] = 0.07 ± 0.03 for the Pleiades open cluster.

The colour–magnitude diagram (CMD) of ≈900 bona fide single stars in the Pleiades open cluster, and the family of best-fitting PARSEC isochrones is shown in Fig. 1. Similar to the Hyades, PARSEC isochrones provide an excellent fit to the observed sequence for Gabs ⪅ 9.0 mag. For 9.0 < Gabs < 11.0 mag, the isochrones predict fainter G magnitudes by up to 0.15 mag and bluer BP − RP colours. For still fainter, mid- to late M-dwarfs, the isochrones tend to predict overly bright G magnitudes and redder BP − RP colours (Fig. 2).

thumbnail Fig. 1.

Colour–absolute magnitude diagram of bona fide single stars in the Pleiades open cluster based on Gaia DR3. Overplotted is the family of the best-fitting PARSEC isochrones.

thumbnail Fig. 2.

Minimum distance in colour–magnitude space of each bona fide single star in the Pleiades from the best-fitting isochrone plotted against the absolute magnitude (lower abscissa) and colour (upper abscissa) on the isochrone. A positive residual corresponds to stars that are brighter (redder) than predicted by the isochrone. A negative residual corresponds to stars that are fainter (bluer) than the isochrone. The uncertainty for each source in Gabs corresponds to the Gaia uncertainty in absolute G magnitude, and the uncertainty in the residual corresponds to the quadratically added uncertainties in absolute G magnitude and BP − RP colour.

For each star, we calculate its distances to the best-fitting interpolated isochrone in colour–magnitude space to assign Teff, log g, log L, and (present-day) mass. The uncertainties in astrophysical parameters take into account the uncertainties in the Gaia measurements and also those in the [M/H] and age estimates of the isochrone. The derived properties of the bona fide single stars in the Pleiades are listed in Table 2. The corresponding stellar properties for the Hyades open cluster are presented in Table 1 in Brandner et al. (2023a).

Table 2.

Astrophysical parameters of bona fide single stars in the Pleiades open cluster.

3. Comparison with Gaia Apsis

3.1. Cluster ensemble properties: Age and metallicity

A defining characteristic of stellar populations in open clusters is their small spread in age and abundances (Parmentier et al. 2014; Bovy 2016; Kopytova et al. 2016; Magrini et al. 2017; Jeffries 2017; Krumholz & McKee 2020). In order to test how well Apsis recovers these essential properties, we compare the estimates for individual stars with the ensemble average for each cluster as derived by isochrone fitting. Apsis age estimates originate in the Final Luminosity Age and Mass Estimator (FLAME) module (Bailer-Jones et al. 2013; Creevey et al. 2023). FLAME in turn relies on input both from the Generalized Stellar Parametrizer – Photometry (GSP-Phot) and the Generalized Stellar Parametrizer – Spectroscopy (GSP-Spec).

Figure 3 shows FLAME mean age estimates along with the lower 16th and 84th quantiles for 109 stars in the Hyades and 186 stars in the Pleiades against their BP − RP colour. FLAME lower and upper age estimates are capped between 0.2 and 13.5 Gyr3. For both clusters, the FLAME age estimates reveal a strong correlation with colour for BP − RP ≤ 0.8 mag. For BP − RP > ;1.0 mag, the majority of age estimates tend to scatter around ages of ≈11 Gyr. Weighted-mean uncertainties yield 〈ageFLAME〉wm = 1.73 ± 0.04 Gyr for the Hyades and 〈ageFLAME〉wm = 0.26 ± 0.01 Gyr for the Pleiades.

thumbnail Fig. 3.

Gaia DR3 Apsis age estimates as a function of the BP − RP colour for single stars in the Hyades (left) and Pleiades (right). The blue horizontal dashed-dotted lines mark the range of ages assigned to the open clusters by isochrone fitting. For both clusters, Apsis age estimates show a strong correlation with the colour of the star for BP − RP ≤ 0.8 mag.

The Apsis GSP-Phot and GSP-Spec modules independently provide [M/H] estimates. In order to test their consistency, we identified all stars that have [M/H] estimates from both modules. Figure 4 shows the [M/H] estimates along with diagonal dashed lines marking a 1:1 correspondence. GSP-Phot [M/H] are the original published Apsis values. A recalibration using Code Version 1.6 of the gdr3apcal Python package4 according to Andrae et al. (2023) results in revised [M/H] estimates for Pleiades stars with a wider intrinsic spread; on average these are 1.0 dex lower than the original Apsis estimates. As a low average [M/H] ≈ −1.1 appears non-physical for member stars of the Pleiades (Funayama et al. 2009; Soderblom et al. 2009), we reject the recalibration of the GSP-Phot [M/H]. GSP-Spec [M/H] estimates are recalibrated according to Recio-Blanco et al. (2023), and using the polynomial coefficients applicable to open clusters. This results in an average correction of ≈0.09 dex towards higher [M/H] values.

thumbnail Fig. 4.

[M/H] of single stars in the Hyades (left) and Pleiades (right), which have both DR3 Apsis GSP-Phot and recalibrated GSP-Spec estimates. Marked as a black dashed line is the diagonal along which both [M/H] estimates agree. The blue open squares mark the range of [M/H] estimates derived from the best-fitting PARSEC isochrones. Not shown are stars with Apsis estimates of [M/H] < −0.7.

3.2. Individual stellar properties: Teff, log g, and mass

The top panel of Fig. 5 shows the Apsis deduced effective temperatures for stars in the Hyades as a function of their BP − RP colour. On the left we show the GSP-Phot derived Teff, which reveals a monotonic and relatively smooth decline with increasing colour up to BP − RP ⪅ 1.35 mag. In the range 1.35 < BP − RP < 3.0 mag, the GSP-Phot temperature estimates bifurcate. Overplotted in blue is the Teff vs. colour relation according to the best-fitting PARSEC isochrone. Compared to the isochrone, GSP-Phot Teff estimates are systematically lower by 150–250 K for BP − RP ⪅ 1.35 mag, while the bifurcated GSP-Phot Teff estimates for BP − RP > ;1.35 mag seemingly bracket the temperature according to the isochrone. For BP − RP > ;3.0 mag, the GSP-Phot Teff estimates are systematically higher by ≈50 K than implied by the isochrone. On the right, we show GSP-Spec Teff, which also reveals a bifurcation for the colour range 1.35 < BP − RP < 1.65 mag. For reference, the corresponding stellar masses according to the isochrone are indicated on the top abscissa.

thumbnail Fig. 5.

Effective temperature estimates for single stars in the Hyades (top) and Pleiades (bottom) based on GSP-Phot (left) and GSP-Spec (right) as a function of BP − RP colour. Overplotted in blue is the best-fitting PARSEC isochrone.

The bottom panel of Fig. 5 shows effective temperature vs. colour for the Pleiades. There appears to be a noticeable gap in GSP-Phot derived Teff in the range of 8500–9500 K with the majority of the stars with BP − RP < 0.3 mag being assigned Teff> 9500 K. For redder colours, that is, 0.3 < BP − RP < 1.3 mag, GSP-Phot Teff systematically falls a few hundred Kelvin below the colour–temperature relation of the best-fitting PARSEC isochrone, and for BP − RP > ;1.7 mag it systematically falls a few hundred Kelvin above this relation. We note that a few stars in the BP-RP colour range from 1.0 to 2.0 mag have a GSP-Phot temperature that falls significantly above the isochrone relation. The GSP-Spec module was only able to deduce effective temperature for 115 of the bona fide single stars in the Pleiades. While the majority closely follow the Teff–colour relation of the best-fitting isochrone, a few stars stand out by having significantly lower GSP-Spec temperatures.

The top panel of Fig. 6 shows the surface gravity for stars in the Hyades as a function of their BP − RP colour as deduced by Apsis, and the left panel shows the log g derived by GSP-Phot. Similar to the Teff estimates, log g follows a smooth, monotonic relation with colour for BP − RP ⪅ 1.35 mag. In the colour range 1.35 < BP − RP < 3.0 mag, GSP-Phot log g estimates bifurcate. For BP − RP < 3.0 mag, GSP-Phot log g [cm s−2] estimates are typically 0.1–0.2 lower than suggested by the isochrone, while they fall above the isochrone value by ≈0.03 for redder colours. The right panel shows the recalibrated GSP-Spec log g estimates according to Recio-Blanco et al. (2023). On average, the recalibration lowers log g [cm s−2] by ≈0.3 compared to the original DR3 Apsis estimates (e.g., from 5.0 to 4.7). The log g [cm s−2] values stay almost constant at 4.7 for 0.7 < BP − RP < 2.2 mag, with the exception of a dip in the range of 1.35 < BP − RP < 1.70 mag.

thumbnail Fig. 6.

Surface gravity estimates for single stars in the Hyades (top) and Pleiades (bottom) based on GSP-Phot (left) and recalibrated GSP-Spec (right). Not shown are stars with Apsis log g [cm s−2] estimates below 3.8. Overplotted in blue is the best-fitting PARSEC isochrone.

Shown in the bottom of Fig. 6 are the corresponding plots for the Pleiades. GSP-Phot log g [cm s−2] estimates for BP − RP < 1.7 mag fall below the isochrone by up to 0.2. In the colour range 1.9 < BP − RP < 3.0 mag, GSP-Phot log g estimates bifurcate. For red colours (BP − RP > ;3.0 mag), the GSP-Phot log g estimates are systematically higher than suggested by the best-fitting isochrone. Akin to the single stars in the Hyades, the recalibrated GSP-Spec log g [cm s−2] estimates for Pleiades members tend to cluster around 4.7.

4. Discussion of benchmarking results

4.1. Age and FLAME

The Apsis FLAME module is tuned using supervised learning (Rybizki et al. 2020) based on templates, which include a model of the Milky Way. Stars get assigned a population identification (popID) derived from the age bins defined by Robin et al. (2003) for the different Galactic populations, namely thin disc (popID 0–6), thick disc (popID 7), halo (popID 8), plus additional popIDs for bulge stars and stars belonging to the Magellanic Clouds or to open clusters. The apparent age–colour relation in Fig. 3 suggests that the Apsis age inference is dominated by the model of the underlying Galactic population. FLAME seems to assign stars a popID, and hence an ‘age’ primarily based on their effective temperature.

FLAME uses a solar [M/H] prior (Creevey et al. 2023), which should not be a major limitation to its applicability to the near-solar [M/H] of the Hyades ([M/H] = 0.18 ± 0.03) and Pleiades ([M/H] = 0.07 ± 0.03) open clusters. We note a potential mismatch between different stellar evolutionary models employed by GSP-Phot – which provides inputs to FLAME – and FLAME itself. According to Creevey et al. (2023), GSP-Phot employs PARSEC 1.2S Colibri S37 models (Tang et al. 2014; Chen et al. 2015; Pastorelli et al. 2020), which is the same set of models we use for the determination of the stellar parameters of the single stars in the Hyades (Brandner et al. 2023a) and Pleiades. However, FLAME itself employs Bag of Stellar Tracks and Isochrones (BaSTI, Hidalgo et al. 2018). Figure 12 in Hidalgo et al. (2018) highlights the close match of BaSTI and MIST isochrones (Choi et al. 2016), and the mismatch between BaSTI and PARSEC isochrones for young late-type stars of approximately solar abundance.

As discussed by Brandner et al. (2023b), MIST models tend to systematically predict fainter absolute magnitudes Gabs and bluer BP − RP colours for bona fide single stars in the Hyades with Teff ≤ 4700 K compared to those observed by Gaia DR3. PARSEC isochrones provide a considerably better fit to the observed single-star sequence (Brandner et al. 2023a).

A further limitation of Gaia DR3 FLAME is the apparent lower age limit of 200 Myr. This precludes a proper age dating of Pleiades members. In summary, individual FLAME age estimates seem of very limited informative power.

4.2. Metallicity and GSP-Phot/GSP-Spec

In addition to the low-resolution spectroscopic information in the BP and RP bands, GSP-Phot [M/H] estimates also incorporate Gaia parallax measurements, and use both theoretical model spectra and isochrones (Andrae et al. 2023). Fouesneau et al. (2023) caution against the use of GSP-Phot [M/H] estimates without additional investigation, and state that [M/H] estimates appear to be 0.2 dex lower on average than literature data for sources with [M/H] > ;−1 dex. This is in agreement with our findings. For both the Hyades and Pleiades samples, we find that the weighted mean [M/H] estimates of the bona fide single stars with GSP-Phot [M/H] estimates fall below the estimates derived from families of best-fitting isochrones by 0.15–0.2 dex (Fig. 4).

However, GSP-Phot [M/H] estimates of individual stars exhibit a much larger scatter than supported by the 16% and 84% quantiles of the Markov chain Monte Carlo results reported in Apsis, suggesting that the uncertainties in the parameter estimates for individual stars are underestimated by a factor of between 10 and 20. A recalculation of [M/H] using the Python package gdr3apcal (Code Version 1.6) makes the discrepancy worse. We therefore agree with the recommendation by Fouesneau et al. (2023) to not use GSP-Phot [M/H] estimates for individual stars.

GSP-Spec [M/H] estimates are based on Gaia RVS and grids of model spectra spanning the full range of Galactic stellar populations. The inherent estimation bias of this module has been quantified by Recio-Blanco et al. (2023), who also provide a prescription for correction. The correction moves the ensemble [M/H] estimates of Apsis to within ≈0.1 dex of the [M/H] values suggested from the isochrone fitting to the Hyades and Pleiades. For individual stars, in particular in the Hyades, GSP-Spec [M/H] (minimal) quantile uncertainties appear to be underestimated, which is an indication that the actual noise floor of the high-S/N estimates is considerably higher.

While the GSP-Phot and GSP-Spec estimates show a weak correlation overall, individual estimates tend to disagree by substantially more than suggested by the assigned uncertainties. In general, neither GSP-Phot nor GSP-Spec is able to reproduce the [M/H] estimates derived from the isochrone fitting, which are marked by blue open squares. The apparent close match of the weighted mean 〈[M/H]GSP-Phot〉wm = 0.1720 ± 0.001 for the Hyades is accidental, as it is the result of a selection of stars with both GSP-Phot and GSP-Spec estimates. A sample of 213 stars in the Hyades with GSP-Phot [M/H] estimates, and including 96 stars without GSP-Spec estimates, yields 〈[M/H]GSP-Phot〉wm = 0.034 ± 0.001. Overall, both GSP-Phot and GSP-Spec tend to systematically underestimate [M/H] of the stars in the Hyades and Pleiades.

4.3. Effective temperature and GSP-Phot/GSP-Spec

As the spectral characteristics of stars are primarily determined by their effective temperature, it is no surprise that this is also the most robust of the DR3 Apsis parameter estimates. Nevertheless, GSP-Phot Teff estimates seem to systematically fall by 150 K–250 K below the Teff–colour relation suggested by the best-fitting isochrones for BP − RP < 1.35 mag. We find the apparent bifurcation of Teff estimates for stars in the Hyades with 1.35 mag < BP − RP < 3.0 mag and the apparent dearth of stars with 8000 K < Teff < 9500 K in the Pleiades to be surprising. Apart from the systematic offsets, the GSP-Phot Teff quantile uncertainties appear plausible (Fig. 5, left). Overall, GSP-Spec estimates are in better agreement with the Teff–colour relation according to the PARSEC isochrone than GSP-Phot estimates, though estimates start to diverge for BP − RP > ;1.35 mag (Fig. 5, right).

4.4. Surface gravity and GSP-Phot/GSP-Spec

GSP-Phot log g estimates overall follow the log g–colour relation suggested by the best-fitting isochrones. We note the bifurcation in the log g estimates in the colour range of 2.0 mag < BP − RP < 3.0 mag, and the systematic underestimation of log g for BP − RP < 1.5 mag (Fig. 6, left).

GSP-Spec log g [cm s−2] estimates, which have been recalibrated according to the prescription provided by Recio-Blanco et al. (2023) are almost constant at 4.7 over a wide colour range, with a strange dip to ≈4.4 at BP − RP ≈ 1.4 mag for stars in the Hyades. We suspect that the lack of a dimension for rotational line broadening in the GSP-Spec model grid might force the parameter estimates for fast rotating stars in the Pleiades in particular to higher apparent log g values (Fig. 6, right). The restriction to stars with GSP-Spec quality flags vbroadG and vradG equal to 0 (see Recio-Blanco et al. 2023) removes outliers in particular at the faint end. The systematics in the DR3 Apsis GSP-Spec estimates, though, are still present.

4.5. Limitations of our method

While the isochrone fits and the deduction of stellar astrophysical parameters from the isochrones provide robust estimates of stellar properties, there are also some inherent limitations to our approach. Firstly, the χ2 isochrone fitting is based on the assumption of uniform extinction to all stars in either of the clusters5. Apsis GSP-Phot, on the other hand, fits individual line-of-sight extinction estimates for each star. While this potentially provides fits of greater accuracy, it also opens an additional dimension for fitting errors due to a strong correlation between the uncertainties in Teff and AV estimates (Andrae et al. 2023).

Secondly, our bona fide single star sequences are still contaminated by sources with hidden binarity. These are typically close binary systems with mass ratios and/or brightness ratios of < 1:2. We expect the contamination to be more severe for the Pleiades sample due to its approximately three times larger distance, which reduces the sensitivity to binary detection and rejection based on the RUWE value. Still, the bias induced by the fainter and redder secondary on the parameter estimation in such systems should be modest, and cannot explain the systematic errors encountered in the Apsis estimates.

Thirdly, we employ non-rotating isochrones. Stellar rotation affects the upper main sequence and main sequence turn-off region in the Hyades and Pleiades in particular, which are also the regions that are the most sensitive to the isochronal ages. However, for the majority of the stars in our sample, the effect on the determined astrophysical parameters should be minimal.

Finally, the PARSEC models follow a fixed Y(Z) abundance relation. However, this affects both our estimates and Apsis. We note that a deviation of the [Y/Z] abundance ratio from the relation would in particular change the energy-production rate of the low-mass stars (Baraffe & Chabrier 2018).

5. Recommendations and outlook

While Gaia offers unprecedented quality and precision in its astrometric, photometric, and spectroscopic measurements, we identify tension between Gaia DR3 Apsis estimates and astrophysical properties deduced by us for ≈1500 bona fide single stars in the Hyades and Pleiades open clusters.

Apsis provides very good to fair estimates of the effective temperature of stars, and potentially of the ensemble [M/H] of homogeneous stellar groups. However, both GSP-Phot and GSP-Spec seem to underestimate the ensemble [M/H] of two nearby open clusters by 0.1–0.2 dex. As recommended by Andrae et al. (2023), one should not use the GSP-Phot [M/H] for individual stars. GSP-Phot log g in general is sufficiently precise to deduce luminosity classes, while GSP-spec log g [cm s−2] for stars in the open clusters converge around 4.7 over a larger range of spectral types. Our assessment of Gaia DR3 Apsis and the resulting recommendations are summarised in Table 3.

Table 3.

Gaia DR3 Apsis quality assessment for main sequence stars with 11 500 K ≥ Teff ≥ 3000 K and [M/H] ≈ 0.

We also identified a few potential issues in the underlying assumptions and training sets used for Gaia DR3 Apsis. Firstly, GSP-Phot employs PARSEC isochrones. FLAME relies on BaSTI models, whose isochrones closely resemble MESA isochrones but disagree with PARSEC isochrones for solar metallicity stars with masses of ≲0.7 M⊙. This model mismatch might bias FLAME estimates. Secondly, the GSP-Spec spectral model grid lacks a dimension for rotational broadening, and therefore fast-rotating stars might be assigned excessive log g estimates. Finally, FLAME age estimates reproduce the training model. The ages start at 100 Myr according to the online documentation, while the youngest (16th quantile) ages assigned to Pleiades members are 200 Myr. FLAME lacks the ability to classify the youngest (age ≲150 Myr) Galactic population (popID = 0 according to Rybizki et al. (2020)).

Addressing these issues might help to improve the quality of Apsis estimates. Overall, DR3 Apsis seems to assign overly small uncertainties to its estimates. Therefore, Gaia DR3 Apsis cannot replace more dedicated efforts to determine stellar properties (e.g., Anders et al. 2022; Berger et al. 2023), or for example provide the level of accuracy required for the characterisation of exoplanet host stars.

We expect that the sample of 1500 bona fide single stars in the Hyades (Brandner et al. 2023a) and Pleiades open clusters with homogeneous astrophysical parameter estimates derived from PARSEC evolutionary models might also be highly useful for future benchmarks of theoretical and observational studies.

2

We attribute the higher AV = 0.20 mag reported by Heyl et al. (2022) due to their use of strictly solar metallicity [M/H] = 0 isochrones.

3

We note that this contradicts the lower limit of 0.1 Gyr quoted in the Gaia DR3 documentation release version 1.2 (7. Feb. 2023), Sect. 11.3.6.

5

We note that due to their vicinity, the line-of-sight extinction is very small towards the Hyades (AV = 3 mmag), and modest for the Pleiades (AV = 120 mmag).

Acknowledgments

This work has made use of data from the European Space Agency (ESA) mission Gaia (https://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, https://www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement. We acknowledge the use of TOPCAT (Taylor 2005, 2011).

References

  1. Anders, F., Khalatyan, A., Queiroz, A. B. A., et al. 2022, A&A, 658, A91 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  2. Andrae, R., Fouesneau, M., Sordo, R., et al. 2023, A&A, 674, A27 [CrossRef] [EDP Sciences] [Google Scholar]
  3. Bailer-Jones, C. A. L., Andrae, R., Arcay, B., et al. 2013, A&A, 559, A74 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  4. Bailer-Jones, C. A. L., Rybizki, J., Fouesneau, M., Demleitner, M., & Andrae, R. 2021, AJ, 161, 147 [Google Scholar]
  5. Baraffe, I., & Chabrier, G. 2018, A&A, 619, A177 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  6. Basri, G., Marcy, G. W., & Graham, J. R. 1996, ApJ, 458, 600 [NASA ADS] [CrossRef] [Google Scholar]
  7. Berger, T. A., Schlieder, J. E., & Huber, D. 2023, AJ, submitted [arXiv:2301.11338] [Google Scholar]
  8. Bouy, H., Bertin, E., Moraux, E., et al. 2013, A&A, 554, A101 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  9. Bouy, H., Bertin, E., Sarro, L. M., et al. 2015, A&A, 577, A148 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  10. Bovy, J. 2016, ApJ, 817, 49 [NASA ADS] [CrossRef] [Google Scholar]
  11. Brandner, W., Calissendorff, P., & Kopytova, T. 2023a, AJ, 165, 108 [NASA ADS] [CrossRef] [Google Scholar]
  12. Brandner, W., Calissendorff, P., & Kopytova, T. 2023b, MNRAS, 518, 662 [Google Scholar]
  13. Brandt, T. D., & Huang, C. X. 2015, ApJ, 807, 58 [NASA ADS] [CrossRef] [Google Scholar]
  14. Bressan, A., Marigo, P., Girardi, L., et al. 2012, MNRAS, 427, 127 [NASA ADS] [CrossRef] [Google Scholar]
  15. Chen, Y., Girardi, L., Bressan, A., et al. 2014, MNRAS, 444, 2525 [Google Scholar]
  16. Chen, Y., Bressan, A., Girardi, L., et al. 2015, MNRAS, 452, 1068 [Google Scholar]
  17. Choi, J., Dotter, A., Conroy, C., et al. 2016, ApJ, 823, 102 [Google Scholar]
  18. Creevey, O. L., Sordo, R., Pailler, F., et al. 2023, A&A, 674, A26 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  19. Cutri, R. M., Skrutskie, M. F., van Dyk, S., et al. 2003, VizieR Online Data Catalog: II/246 [Google Scholar]
  20. Dahm, S. E. 2015, ApJ, 813, 108 [Google Scholar]
  21. Delchambre, L., Bailer-Jones, C. A. L., Bellas-Velidis, I., et al. 2023, A&A, 674, A31 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  22. Fouesneau, M., Frémat, Y., Andrae, R., et al. 2023, A&A, 674, A28 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  23. Funayama, H., Itoh, Y., Oasa, Y., et al. 2009, PASJ, 61, 931 [NASA ADS] [CrossRef] [Google Scholar]
  24. Gaia Collaboration (Babusiaux, C., et al.) 2018, A&A, 616, A10 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  25. Gaia Collaboration (Brown, A. G. A., et al.) 2021a, A&A, 649, A1 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  26. Gaia Collaboration (Smart, R. L., et al.) 2021b, A&A, 649, A6 [EDP Sciences] [Google Scholar]
  27. Galindo-Guil, F. J., Barrado, D., Bouy, H., et al. 2022, A&A, 664, A70 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  28. Gossage, S., Conroy, C., Dotter, A., et al. 2018, ApJ, 863, 67 [Google Scholar]
  29. Heyl, J., Caiazzo, I., & Richer, H. B. 2022, ApJ, 926, 132 [NASA ADS] [CrossRef] [Google Scholar]
  30. Hidalgo, S. L., Pietrinferni, A., Cassisi, S., et al. 2018, ApJ, 856, 125 [Google Scholar]
  31. Høg, E., Fabricius, C., Makarov, V. V., et al. 2000, A&A, 355, L27 [Google Scholar]
  32. Jeffries, R. D. 2017, Mem. Soc. Astron. Ital., 88, 637 [Google Scholar]
  33. Kopytova, T. G., Brandner, W., Tognelli, E., et al. 2016, A&A, 585, A7 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  34. Krumholz, M. R., & McKee, C. F. 2020, MNRAS, 494, 624 [NASA ADS] [CrossRef] [Google Scholar]
  35. Magrini, L., Randich, S., Kordopatis, G., et al. 2017, A&A, 603, A2 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  36. Magrini, L., Danielski, C., Bossini, D., et al. 2022, A&A, 663, A161 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  37. Mermilliod, J. C. 1992, Bulletin d’Information du Centre de Donnees Stellaires, 40, 115 [Google Scholar]
  38. Meynet, G., Mermilliod, J. C., & Maeder, A. 1993, A&AS, 98, 477 [Google Scholar]
  39. Parmentier, G., Pfalzner, S., & Grebel, E. K. 2014, ApJ, 791, 132 [NASA ADS] [CrossRef] [Google Scholar]
  40. Pastorelli, G., Marigo, P., Girardi, L., et al. 2020, MNRAS, 498, 3283 [Google Scholar]
  41. Percival, S. M., Salaris, M., & Groenewegen, M. A. T. 2005, A&A, 429, 887 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  42. Recio-Blanco, A., de Laverny, P., Palicio, P. A., et al. 2023, A&A, 674, A29 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  43. Robin, A. C., Reylé, C., Derrière, S., & Picaud, S. 2003, A&A, 409, 523 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  44. Röser, S., Schilbach, E., Piskunov, A. E., Kharchenko, N. V., & Scholz, R. D. 2011, A&A, 531, A92 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  45. Rybizki, J., Demleitner, M., Bailer-Jones, C., et al. 2020, PASP, 132, 074501P [NASA ADS] [CrossRef] [Google Scholar]
  46. Soderblom, D. R., Laskar, T., Valenti, J. A., Stauffer, J. R., & Rebull, L. M. 2009, AJ, 138, 1292 [NASA ADS] [CrossRef] [Google Scholar]
  47. Stauffer, J. R., Schultz, G., & Kirkpatrick, J. D. 1998, ApJ, 499, L199 [NASA ADS] [CrossRef] [Google Scholar]
  48. Tang, J., Bressan, A., Rosenfield, P., et al. 2014, MNRAS, 445, 4287 [NASA ADS] [CrossRef] [Google Scholar]
  49. Taylor, M. B. 2005, in Astronomical Data Analysis Software and Systems XIV, eds. P. Shopbell, M. Britton, & R. Ebert, ASP Conf. Ser., 347, 29 [Google Scholar]
  50. Taylor, M. 2011, Astrophysics Source Code Library [record ascl:1101.010] [Google Scholar]
  51. van Leeuwen, F. 2009, A&A, 497, 209 [CrossRef] [EDP Sciences] [Google Scholar]

All Tables

Table 1.

Compilation of abundance, αML, and age estimates for the Pleiades open cluster based on isochrone fitting.

In the text
Table 2.

Astrophysical parameters of bona fide single stars in the Pleiades open cluster.

In the text
Table 3.

Gaia DR3 Apsis quality assessment for main sequence stars with 11 500 K ≥ Teff ≥ 3000 K and [M/H] ≈ 0.

In the text

All Figures

thumbnail Fig. 1.

Colour–absolute magnitude diagram of bona fide single stars in the Pleiades open cluster based on Gaia DR3. Overplotted is the family of the best-fitting PARSEC isochrones.
In the text
thumbnail Fig. 2.

Minimum distance in colour–magnitude space of each bona fide single star in the Pleiades from the best-fitting isochrone plotted against the absolute magnitude (lower abscissa) and colour (upper abscissa) on the isochrone. A positive residual corresponds to stars that are brighter (redder) than predicted by the isochrone. A negative residual corresponds to stars that are fainter (bluer) than the isochrone. The uncertainty for each source in Gabs corresponds to the Gaia uncertainty in absolute G magnitude, and the uncertainty in the residual corresponds to the quadratically added uncertainties in absolute G magnitude and BP − RP colour.
In the text
thumbnail Fig. 3.

Gaia DR3 Apsis age estimates as a function of the BP − RP colour for single stars in the Hyades (left) and Pleiades (right). The blue horizontal dashed-dotted lines mark the range of ages assigned to the open clusters by isochrone fitting. For both clusters, Apsis age estimates show a strong correlation with the colour of the star for BP − RP ≤ 0.8 mag.
In the text
thumbnail Fig. 4.

[M/H] of single stars in the Hyades (left) and Pleiades (right), which have both DR3 Apsis GSP-Phot and recalibrated GSP-Spec estimates. Marked as a black dashed line is the diagonal along which both [M/H] estimates agree. The blue open squares mark the range of [M/H] estimates derived from the best-fitting PARSEC isochrones. Not shown are stars with Apsis estimates of [M/H] < −0.7.
In the text
thumbnail Fig. 5.

Effective temperature estimates for single stars in the Hyades (top) and Pleiades (bottom) based on GSP-Phot (left) and GSP-Spec (right) as a function of BP − RP colour. Overplotted in blue is the best-fitting PARSEC isochrone.
In the text
thumbnail Fig. 6.

Surface gravity estimates for single stars in the Hyades (top) and Pleiades (bottom) based on GSP-Phot (left) and recalibrated GSP-Spec (right). Not shown are stars with Apsis log g [cm s−2] estimates below 3.8. Overplotted in blue is the best-fitting PARSEC isochrone.
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.