Samir Nepal
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Samir Nepal has a total of 326 co-authors. Here are the top co-authors by number of shared publications:
- Cristina Chiappini
Leibniz-Institut für Astrophysik Potsdam (AIP)
13 shared publications - Matthias Steinmetz
Leibniz-Institut für Astrophysik Potsdam (AIP)
9 shared publications - Friedrich Anders
Universitat de Barcelona (UB)
6 shared publications G. Guiglion
Max Planck Institute for Astronomy
6 shared publications- Guillaume Guiglion
Zentrum für Astronomie der Universität Heidelberg
6 shared publications - Gerard Gilmore
University of Cambridge
5 shared publications
Latest Publications
- Open Access
- 21 September 2026
- Published by Oxford University Press (OUP) in Monthly Notices of the Royal Astronomical Society
- Vol. 552 (3)
NGC 1901: a new benchmark cluster for gyrochronology in the TESS Southern continuous viewing zone
We use Gaia DR3 astrometry and TESS time-series photometry to study NGC 1901, a sparse open cluster in the TESS Southern continuous viewing zone, that has been neglected due to its proximity on the sky to the Large Magellanic Cloud. We refine its membership list and measure rotation periods for 32 high-confidence rotating stars. By applying the gyro-interp empirical model, we calculate a gyrochronal age of 830 $\pm$ 50 Myr. Because NGC 1901 possesses a sub-solar metallicity of [Fe/H]$\approx {-0.11}$, its excellent agreement with empirical models suggests that temperature-period relations mitigate metallicity effects for Sun-like stars. Our results place NGC 1901 in the gap between the benchmark open clusters Praesepe (670 Myr) and NGC 6811 (1 Gyr). We find that the rotation sequence exhibits significant spin-down stalling in late K dwarfs, supporting models of core–envelope recoupling. Furthermore, our findings provide constraints for the cluster’s structural evolution, including its recently identified tidal tail and corona. This study establishes NGC 1901 as a new benchmark for calibrating the next generation of gyrochronal models.- 21 September 2026
- Published by Oxford University Press (OUP) in Monthly Notices of the Royal Astronomical Society
- Vol. 552 (3)
Open AccessNGC 1901: a new benchmark cluster for gyrochronology in the TESS Southern continuous viewing zone
We use Gaia DR3 astrometry and TESS time-series photometry to study NGC 1901, a sparse open cluster in the TESS Southern continuous viewing zone, that has been neglected due to its proximity on the sky to the Large Magellanic Cloud. We refine its membership list and measure rotation periods for 32 high-confidence rotating stars. By applying the gyro-interp empirical model, we calculate a gyrochronal age of 830 $\pm$ 50 Myr. Because NGC 1901 possesses a sub-solar metallicity of [Fe/H]$\approx {-0.11}$, its excellent agreement with empirical models suggests that temperature-period relations mitigate metallicity effects for Sun-like stars. Our results place NGC 1901 in the gap between the benchmark open clusters Praesepe (670 Myr) and NGC 6811 (1 Gyr). We find that the rotation sequence exhibits significant spin-down stalling in late K dwarfs, supporting models of core–envelope recoupling. Furthermore, our findings provide constraints for the cluster’s structural evolution, including its recently identified tidal tail and corona. This study establishes NGC 1901 as a new benchmark for calibrating the next generation of gyrochronal models. - Open Access
- 11 August 2026
- Published by Cambridge University Press (CUP) in Proceedings of the International Astronomical Union
- Vol. 20 (S395) , 201-205
New insights on MW disc history with machine learning, precise ages and Gaia DR3
Recent observations of high-redshift galaxies with JWST and ALMA suggest that disc galaxies can form early, while studies of metal-poor stars in disc orbits in the Milky Way (MW) imply a similar early disc formation scenario. Using a chrono-chemo-dynamical analysis of the Gaia -DR3 RVS stars, with high-quality stellar parameters provided by the hybrid-CNN analysis and precise distances and stellar ages from StarHorse, we explore the MW’s early history. We confirm the existence of old metal-poor stars in thin-disc orbits. These metal-poor stars in disc orbits form a readily visible tail of the MW’s old disc, spanning a wide metallicity range from [Fe/H]† < -1.0 to super-solar values. We calculate the vertical velocity dispersion ( σV z ) for the high- [ α/ Fe] thick disc, finding 35 ± 0.6 km.s −1 , while the thin disc within the same old-age range has a σ Vz that is lower by 10 to 15 km.s −1 . Our findings align the MW with JWST and ALMA observations of thin galactic discs at high redshifts and provide new constraints for reshaping our understanding of galaxy evolution.
- 11 August 2026
- Published by Cambridge University Press (CUP) in Proceedings of the International Astronomical Union
- Vol. 20 (S395) , 201-205
Open AccessNew insights on MW disc history with machine learning, precise ages and Gaia DR3
Recent observations of high-redshift galaxies with JWST and ALMA suggest that disc galaxies can form early, while studies of metal-poor stars in disc orbits in the Milky Way (MW) imply a similar early disc formation scenario. Using a chrono-chemo-dynamical analysis of the Gaia -DR3 RVS stars, with high-quality stellar parameters provided by the hybrid-CNN analysis and precise distances and stellar ages from StarHorse, we explore the MW’s early history. We confirm the existence of old metal-poor stars in thin-disc orbits. These metal-poor stars in disc orbits form a readily visible tail of the MW’s old disc, spanning a wide metallicity range from [Fe/H]† < -1.0 to super-solar values. We calculate the vertical velocity dispersion ( σV z ) for the high- [ α/ Fe] thick disc, finding 35 ± 0.6 km.s −1 , while the thin disc within the same old-age range has a σ Vz that is lower by 10 to 15 km.s −1 . Our findings align the MW with JWST and ALMA observations of thin galactic discs at high redshifts and provide new constraints for reshaping our understanding of galaxy evolution.
- Top Cited PapersOpen Access
- 24 June 2026
- Published by American Astronomical Society in The Astrophysical Journal Supplement Series
- Vol. 285 (1)
The Nineteenth Data Release of the Sloan Digital Sky Survey
Mapping the local and distant Universe is key to our understanding of it. For decades, the Sloan Digital Sky Survey (SDSS) has made a concerted effort to map millions of celestial objects to constrain the physical processes that govern our Universe. The most recent and fifth generation of SDSS (SDSS-V) is organized into three scientific “mappers”: the Milky Way Mapper, which aims to chart the various components of the Milky Way and constrain its formation and assembly; the Black Hole Mapper, which focuses on understanding supermassive black holes in distant galaxies across the Universe; and the Local Volume Mapper, which uses integral field spectroscopy to map the ionized interstellar medium in the Local Group. This paper describes the scope and content for the nineteenth data release (DR19) of SDSS, which is the most substantial to date in SDSS-V. DR19 is the first to contain data from all three mappers. Additionally, we also describe nine value-added catalogs that enhance the science that can be conducted with the SDSS-V data. Finally, we discuss how to access SDSS DR19 and provide illustrative examples and tutorials.
- 24 June 2026
- Published by American Astronomical Society in The Astrophysical Journal Supplement Series
- Vol. 285 (1)
Top Cited PapersOpen AccessThe Nineteenth Data Release of the Sloan Digital Sky Survey
Mapping the local and distant Universe is key to our understanding of it. For decades, the Sloan Digital Sky Survey (SDSS) has made a concerted effort to map millions of celestial objects to constrain the physical processes that govern our Universe. The most recent and fifth generation of SDSS (SDSS-V) is organized into three scientific “mappers”: the Milky Way Mapper, which aims to chart the various components of the Milky Way and constrain its formation and assembly; the Black Hole Mapper, which focuses on understanding supermassive black holes in distant galaxies across the Universe; and the Local Volume Mapper, which uses integral field spectroscopy to map the ionized interstellar medium in the Local Group. This paper describes the scope and content for the nineteenth data release (DR19) of SDSS, which is the most substantial to date in SDSS-V. DR19 is the first to contain data from all three mappers. Additionally, we also describe nine value-added catalogs that enhance the science that can be conducted with the SDSS-V data. Finally, we discuss how to access SDSS DR19 and provide illustrative examples and tutorials.
- Open Access
- 10 June 2026
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 710, A173
TwinSpecNet: Extending APOGEE’s chemical reach to low-signal-to-noise spectra via empirical paired learning
Context . Large spectroscopic surveys rely on automated pipelines to deliver homogeneous stellar labels; however, a substantial fraction of observations are carried out at a low signal-to-noise ratio (S/N) where label estimates become imprecise or are omitted. In APOGEE, these low-S/N spectra visits tend to sample faint and distant populations (i.e. the bulge, outer halo, and satellite systems), while still encoding recoverable chemical information. Aims . We present TwinSpecNet (TSN), a paired-learning framework that exploits APOGEE’s multi-visit observing strategy: by training on empirical low- and high-S/N spectral twins of the same stars, TSN learns to suppress stochastic noise while preserving the ASPCAP label scale. Methods . TSN employs a Vision Transformer encoder with dual objectives: reconstructing high-S/N flux from low-S/N visits and predicting stellar parameters and abundances with calibrated uncertainties. Results . TSN reduces label scatter relative to visit-level ASPCAP for S/N < 60 visits. It reproduces the ASPCAP scale with residual scatters of σ ≃ 19 K in T eff , σ ≃ 0.06 dex in log g , and σ ≃ 0.03 dex in [Fe/H]. TSN tightens intra-cluster abundance dispersions, recovers cleaner chemical sequences in inner-disk and bulge and satellite samples, and improves C/N-based age precision for APOKASC giants from 1.70 to 1.49 Gyr. Conclusions . By learning survey-specific noise patterns from repeated observations, TSN demonstrates how empirical paired learning can extend the chemical reach of existing spectroscopic data, providing a template that is applicable to other multi-visit surveys.
- 10 June 2026
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 710, A173
Open AccessTwinSpecNet: Extending APOGEE’s chemical reach to low-signal-to-noise spectra via empirical paired learning
Context . Large spectroscopic surveys rely on automated pipelines to deliver homogeneous stellar labels; however, a substantial fraction of observations are carried out at a low signal-to-noise ratio (S/N) where label estimates become imprecise or are omitted. In APOGEE, these low-S/N spectra visits tend to sample faint and distant populations (i.e. the bulge, outer halo, and satellite systems), while still encoding recoverable chemical information. Aims . We present TwinSpecNet (TSN), a paired-learning framework that exploits APOGEE’s multi-visit observing strategy: by training on empirical low- and high-S/N spectral twins of the same stars, TSN learns to suppress stochastic noise while preserving the ASPCAP label scale. Methods . TSN employs a Vision Transformer encoder with dual objectives: reconstructing high-S/N flux from low-S/N visits and predicting stellar parameters and abundances with calibrated uncertainties. Results . TSN reduces label scatter relative to visit-level ASPCAP for S/N < 60 visits. It reproduces the ASPCAP scale with residual scatters of σ ≃ 19 K in T eff , σ ≃ 0.06 dex in log g , and σ ≃ 0.03 dex in [Fe/H]. TSN tightens intra-cluster abundance dispersions, recovers cleaner chemical sequences in inner-disk and bulge and satellite samples, and improves C/N-based age precision for APOKASC giants from 1.70 to 1.49 Gyr. Conclusions . By learning survey-specific noise patterns from repeated observations, TSN demonstrates how empirical paired learning can extend the chemical reach of existing spectroscopic data, providing a template that is applicable to other multi-visit surveys.
- Open Access
- 13 March 2026
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 707, A190
The spheroidal bulge of the Milky Way: Chemodynamically distinct from the inner-thick disc and bar
Studying the composition and origin of the inner region of our Galaxy—the “Galactic bulge”—is crucial for understanding the formation and evolution of the Milky Way and other galaxies. We present new observational constraints based on a sample of around 18 000 stars in the inner Galaxy, combining Gaia DR3 RVS and APOGEE DR17 spectroscopy. Gaia -RVS complements APOGEE by improving the sampling of the metallicity, [Fe/H], in the −2.0 to −0.5 dex range. This work marks the first application of Gaia -RVS spectroscopy to the bulge region, enabled by a novel machine learning approach (hybrid-CNN) that derives stellar parameters from intermediate-resolution spectra with precision comparable to APOGEE’s infrared data. We performed full orbit integrations using a barred Galactic potential and applied orbital frequency analysis to disentangle the stellar populations in the inner Milky Way. For the first time, we are able to robustly identify the long-sought pressure-supported bulge traced by the field stars. We show this stellar population to be chemically and kinematically distinct from the other main components co-existing in the same region. The spheroidal bulge has a metallicity distribution function (MDF) peak at around −0.70 dex extending to solar values. It is dominated by a high-[α∕Fe] population with almost no dependency on metallicity, consistent with very rapid early formation predating the thick disc and the bar. We find evidence that the bar has influenced the dynamics of the spheroidal bulge, introducing a mild triaxiality and radial extension. We identify a group of stars on X 4 orbits, likely native to the early spheroid, as this population mimics the chemistry of the spheroidal bulge, with a minor contamination from the more metal-poor ([Fe∕H]< −1.0) halo. We find the inner thick disc to be kinematically hotter ( V φ ≈ 125 km s −1 ) than the local thick disc. The disc, chemically distinct from the spheroidal bulge and bar, is predominantly metal-poor with an MDF peak at [Fe/H] ≈ −0.45 dex and includes a high fraction of stars with sub-solar [Fe/H] and intermediate [α/Fe]. In contrast to the spheroidal bulge, the [α/Fe] disc shows a steeper decline with [Fe/H], consistent with smaller star formation efficiency than that of the spheroidal bulge. Both the thick disc and the spheroidal bulge present vertical metallicity gradients. We find that the Galactic bar contains both metal-rich and metal-poor stars, as well as high and low [α/Fe] in nearly equal measure. However, their relative contributions vary significantly across different orbital families. The bar shows no strong metallicity trends in orbital extent or velocity dispersions and maintains a consistent elongated shape across all metallicities, indicating that it is a well-mixed dynamical structure. Despite their spatial overlap, the spheroidal bulge, thick disc, and bar occupy distinct regions in both phase space and chemical abundance, indicating separate formation pathways. The stars with [Fe/H]< −1.0 and crossing the Galactic bulge are comprised by accreted populations primarily (70%) belonging to the Gaia-Enceladus/Sausage (GES) merger with an MDF peak at −1.30 dex and possibly a secondary merger remnant with an MDF peak at −1.80 dex.
- 13 March 2026
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 707, A190
Open AccessThe spheroidal bulge of the Milky Way: Chemodynamically distinct from the inner-thick disc and bar
Studying the composition and origin of the inner region of our Galaxy—the “Galactic bulge”—is crucial for understanding the formation and evolution of the Milky Way and other galaxies. We present new observational constraints based on a sample of around 18 000 stars in the inner Galaxy, combining Gaia DR3 RVS and APOGEE DR17 spectroscopy. Gaia -RVS complements APOGEE by improving the sampling of the metallicity, [Fe/H], in the −2.0 to −0.5 dex range. This work marks the first application of Gaia -RVS spectroscopy to the bulge region, enabled by a novel machine learning approach (hybrid-CNN) that derives stellar parameters from intermediate-resolution spectra with precision comparable to APOGEE’s infrared data. We performed full orbit integrations using a barred Galactic potential and applied orbital frequency analysis to disentangle the stellar populations in the inner Milky Way. For the first time, we are able to robustly identify the long-sought pressure-supported bulge traced by the field stars. We show this stellar population to be chemically and kinematically distinct from the other main components co-existing in the same region. The spheroidal bulge has a metallicity distribution function (MDF) peak at around −0.70 dex extending to solar values. It is dominated by a high-[α∕Fe] population with almost no dependency on metallicity, consistent with very rapid early formation predating the thick disc and the bar. We find evidence that the bar has influenced the dynamics of the spheroidal bulge, introducing a mild triaxiality and radial extension. We identify a group of stars on X 4 orbits, likely native to the early spheroid, as this population mimics the chemistry of the spheroidal bulge, with a minor contamination from the more metal-poor ([Fe∕H]< −1.0) halo. We find the inner thick disc to be kinematically hotter ( V φ ≈ 125 km s −1 ) than the local thick disc. The disc, chemically distinct from the spheroidal bulge and bar, is predominantly metal-poor with an MDF peak at [Fe/H] ≈ −0.45 dex and includes a high fraction of stars with sub-solar [Fe/H] and intermediate [α/Fe]. In contrast to the spheroidal bulge, the [α/Fe] disc shows a steeper decline with [Fe/H], consistent with smaller star formation efficiency than that of the spheroidal bulge. Both the thick disc and the spheroidal bulge present vertical metallicity gradients. We find that the Galactic bar contains both metal-rich and metal-poor stars, as well as high and low [α/Fe] in nearly equal measure. However, their relative contributions vary significantly across different orbital families. The bar shows no strong metallicity trends in orbital extent or velocity dispersions and maintains a consistent elongated shape across all metallicities, indicating that it is a well-mixed dynamical structure. Despite their spatial overlap, the spheroidal bulge, thick disc, and bar occupy distinct regions in both phase space and chemical abundance, indicating separate formation pathways. The stars with [Fe/H]< −1.0 and crossing the Galactic bulge are comprised by accreted populations primarily (70%) belonging to the Gaia-Enceladus/Sausage (GES) merger with an MDF peak at −1.30 dex and possibly a secondary merger remnant with an MDF peak at −1.80 dex.
- Open Access
- 5 March 2026
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 707, A111
The oldest Milky Way stars: New constraints on the age of the Universe and the Hubble constant
Aims. In this work, we exploit the most robust, old, and cosmology-independent age estimates of individual stars from Gaia DR3 to place a lower bound on the age of the Universe, t U . These constraints can be used as an anchor point for any cosmological model, thus providing an upper limit to the Hubble constant H 0 . Methods. Our primary stellar age catalog comprises 3000 of the oldest and most robustly measured main-sequence turn-off (MSTO) and subgiant branch (SGB) stars, with ages older than 12.5 Gyr and associated uncertainty below 1 Gyr. Stellar ages are derived via isochrone fitting using the Bayesian code StarHorse , spanning the uniform range 0−20 Gyr, not considering any cosmological prior knowledge on t U . By applying a conservative cut in the Kiel diagram and strict quality cuts on both stellar parameters and posterior probability distribution shapes, and filtering out potential contaminants, we isolated a final sample of 160 bona fide stars, the most numerous sample of precise and reliable MSTO and SGB stars ages available to date. Results. The age distribution of the final sample peaks at 13.6 ± 1.0 (stat) ± 1.4 (syst) Gyr. Assuming a maximum formation redshift for these stars of z f = 20, corresponding to a formation delay of ∼0.2 Gyr, we obtain a lower bound on t U of t U ≥ 13.8 ± 1.0 (stat) ± 1.4 (syst) Gyr. Considering the 10th percentile of the posterior probability distributions of the individual stars, we find that, at 90% confidence level, 70 stars favour t U > 13 Gyr, while none exceeds 14.1 Gyr. For this upper envelope to fall below 13 Gyr, a shift of nearly the full systematic error budget would be required, indicating that such low values are only attainable under very peculiar assumptions. Conclusions. This work presents the first statistically significant use of individual stellar ages as cosmic clocks, opening a new independent approach for cosmological studies. While this analysis already represents a significant step forward, future Gaia data releases will enable even larger and more precise stellar samples, further strengthening these constraints.
- 5 March 2026
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 707, A111
Open AccessThe oldest Milky Way stars: New constraints on the age of the Universe and the Hubble constant
Aims. In this work, we exploit the most robust, old, and cosmology-independent age estimates of individual stars from Gaia DR3 to place a lower bound on the age of the Universe, t U . These constraints can be used as an anchor point for any cosmological model, thus providing an upper limit to the Hubble constant H 0 . Methods. Our primary stellar age catalog comprises 3000 of the oldest and most robustly measured main-sequence turn-off (MSTO) and subgiant branch (SGB) stars, with ages older than 12.5 Gyr and associated uncertainty below 1 Gyr. Stellar ages are derived via isochrone fitting using the Bayesian code StarHorse , spanning the uniform range 0−20 Gyr, not considering any cosmological prior knowledge on t U . By applying a conservative cut in the Kiel diagram and strict quality cuts on both stellar parameters and posterior probability distribution shapes, and filtering out potential contaminants, we isolated a final sample of 160 bona fide stars, the most numerous sample of precise and reliable MSTO and SGB stars ages available to date. Results. The age distribution of the final sample peaks at 13.6 ± 1.0 (stat) ± 1.4 (syst) Gyr. Assuming a maximum formation redshift for these stars of z f = 20, corresponding to a formation delay of ∼0.2 Gyr, we obtain a lower bound on t U of t U ≥ 13.8 ± 1.0 (stat) ± 1.4 (syst) Gyr. Considering the 10th percentile of the posterior probability distributions of the individual stars, we find that, at 90% confidence level, 70 stars favour t U > 13 Gyr, while none exceeds 14.1 Gyr. For this upper envelope to fall below 13 Gyr, a shift of nearly the full systematic error budget would be required, indicating that such low values are only attainable under very peculiar assumptions. Conclusions. This work presents the first statistically significant use of individual stellar ages as cosmic clocks, opening a new independent approach for cosmological studies. While this analysis already represents a significant step forward, future Gaia data releases will enable even larger and more precise stellar samples, further strengthening these constraints.
- Open Access
- 31 October 2024
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 691, A98
Transferring spectroscopic stellar labels to 217 million Gaia DR3 XP stars with SHBoost
With Gaia Data Release 3 (DR3), new and improved astrometric, photometric, and spectroscopic measurements for 1.8 billion stars have become available. Alongside this wealth of new data, however, there are challenges in finding efficient and accurate computational methods for their analysis. In this paper, we explore the feasibility of using machine learning regression as a method of extracting basic stellar parameters and line-of-sight extinctions from spectro-photometric data. To this end, we built a stable gradient-boosted random-forest regressor (xgboost), trained on spectroscopic data, capable of producing output parameters with reliable uncertainties from Gaia DR3 data (most notably the low-resolution XP spectra), without ground-based spectroscopic observations. Using Shapley additive explanations, we interpret how the predictions for each star are influenced by each data feature. For the training and testing of the network, we used high-quality parameters obtained from the StarHorse code for a sample of around eight million stars observed by major spectroscopic stellar surveys, complemented by curated samples of hot stars, very metal-poor stars, white dwarfs, and hot sub-dwarfs. The training data cover the whole sky, all Galactic components, and almost the full magnitude range of the Gaia DR3 XP sample of more than 217 million objects that also have reported parallaxes. We have achieved median uncertainties of 0.20 mag in V-band extinction, 0.01 dex in logarithmic effective temperature, 0.20 dex in surface gravity, 0.18 dex in metallicity, and 12% in mass (over the full Gaia DR3 XP sample, with considerable variations in precision as a function of magnitude and stellar type). We succeeded in predicting competitive results based on Gaia DR3 XP spectra compared to classical isochrone or spectral-energy distribution fitting methods we employed in earlier works, especially for parameters AV and Teff, along with the metallicity values. Finally, we showcase some potential applications of this new catalogue, including extinction maps, metallicity trends in the Milky Way, and extended maps of young massive stars, metal-poor stars, and metal-rich stars.
- 31 October 2024
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 691, A98
Open AccessTransferring spectroscopic stellar labels to 217 million Gaia DR3 XP stars with SHBoost
With Gaia Data Release 3 (DR3), new and improved astrometric, photometric, and spectroscopic measurements for 1.8 billion stars have become available. Alongside this wealth of new data, however, there are challenges in finding efficient and accurate computational methods for their analysis. In this paper, we explore the feasibility of using machine learning regression as a method of extracting basic stellar parameters and line-of-sight extinctions from spectro-photometric data. To this end, we built a stable gradient-boosted random-forest regressor (xgboost), trained on spectroscopic data, capable of producing output parameters with reliable uncertainties from Gaia DR3 data (most notably the low-resolution XP spectra), without ground-based spectroscopic observations. Using Shapley additive explanations, we interpret how the predictions for each star are influenced by each data feature. For the training and testing of the network, we used high-quality parameters obtained from the StarHorse code for a sample of around eight million stars observed by major spectroscopic stellar surveys, complemented by curated samples of hot stars, very metal-poor stars, white dwarfs, and hot sub-dwarfs. The training data cover the whole sky, all Galactic components, and almost the full magnitude range of the Gaia DR3 XP sample of more than 217 million objects that also have reported parallaxes. We have achieved median uncertainties of 0.20 mag in V-band extinction, 0.01 dex in logarithmic effective temperature, 0.20 dex in surface gravity, 0.18 dex in metallicity, and 12% in mass (over the full Gaia DR3 XP sample, with considerable variations in precision as a function of magnitude and stellar type). We succeeded in predicting competitive results based on Gaia DR3 XP spectra compared to classical isochrone or spectral-energy distribution fitting methods we employed in earlier works, especially for parameters AV and Teff, along with the metallicity values. Finally, we showcase some potential applications of this new catalogue, including extinction maps, metallicity trends in the Milky Way, and extended maps of young massive stars, metal-poor stars, and metal-rich stars.
- Open Access
- 15 August 2024
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 688, A167
Discovery of the local counterpart of disc galaxies at z > 4: The oldest thin disc of the Milky Way usingGaia-RVS
Context.JWST recently detected numerous disc galaxies at high redshifts, and there have been observations of cold disc galaxies at z > 4 with ALMA. In the Milky Way (MW), recent studies highlight the presence of metal-poor stars in cold-disc orbits, suggesting an ancient disc. This prompts two fundamental questions. The first refers to the time of formation of the MW disc, and the second to whether it originated as the thin disc or the larger velocity dispersion thick disc.Aims.We carried out a chrono-chemo-dynamical study of a large sample of stars with precise stellar parameters, focusing on the oldest stars in order to decipher the assembly history of the MW discs.Methods.We investigated a sample of 565 606 stars with 6D phase space information and high-quality stellar parameters coming from thehybrid-CNNanalysis of theGaia-DR3 RVS stars. The sample contains 8500 stars with [Fe/H] < −1. For a subset of ∼200 000 main sequence turn-off (MSTO) and subgiant branch (SGB) stars, we computed distances and ages using theStarHorsecode, with a mean precision of 1% and 12%, respectively.Results.First, we confirm the existence of metal-poor stars in thin-disc orbits. The majority of these stars are predominantly old (> 10 Gyr), with over 50% being older than 13 Gyr. Second, we report the discovery of the oldest thin disc of the Milky Way, which extends across a wide range of metallicities, from metal-poor to super-solar stars. The metal-poor stars in disc orbits manifest as a readily visible tail of the metallicity distribution. We calculate the vertical velocity dispersion (σVz) for the high-[α/Fe] thick disc, finding 35 ± 0.6 km s−1, while the thin disc within the same age range has aσVzthat is lower by 10–15 km s−1. Our old thin discσVzappears similar to those estimated for the high-z disc galaxies. Third, as a verification ofStarHorseages, we extend the [Y/Mg] chemical clock to the oldest ages and estimate a slope of −0.038 dex ⋅ Gyr−1. Finally, we confirm our discovery of the old thin disc by showing that the ‘splash’ population includes high- and low-[α/Fe] populations that are both old, and extends to a wider [Fe/H] range, reaching supersolar [Fe/H]. We find that about 6–10% of the old thin disc was heated to thick-disc orbits. The youngest ‘splashed’ stars appear at 9–10 Gyr and may suggest a Gaia-Sausage/Enceladus (GSE) merger at this period.Conclusions.The Milky Way thin disc formed less than 1 billion years after the Big Bang and continuously built up in an inside-out manner – this finding precedes the earlier estimates of the time at which the MW thin disc began to form (around 8–9 Gyr) by about 4–5 billion years. We find that the metal-poor stars in disc orbits reported by previous studies belong to this old thin disc. Considering a massive merger event such as the GSE, a splash is expected – we find a portion of the old thin disc is heated to thick disc velocities and the splash extends to supersolar [Fe/H] regimes.
- 15 August 2024
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 688, A167
Open AccessDiscovery of the local counterpart of disc galaxies at z > 4: The oldest thin disc of the Milky Way usingGaia-RVS
Context.JWST recently detected numerous disc galaxies at high redshifts, and there have been observations of cold disc galaxies at z > 4 with ALMA. In the Milky Way (MW), recent studies highlight the presence of metal-poor stars in cold-disc orbits, suggesting an ancient disc. This prompts two fundamental questions. The first refers to the time of formation of the MW disc, and the second to whether it originated as the thin disc or the larger velocity dispersion thick disc.Aims.We carried out a chrono-chemo-dynamical study of a large sample of stars with precise stellar parameters, focusing on the oldest stars in order to decipher the assembly history of the MW discs.Methods.We investigated a sample of 565 606 stars with 6D phase space information and high-quality stellar parameters coming from thehybrid-CNNanalysis of theGaia-DR3 RVS stars. The sample contains 8500 stars with [Fe/H] < −1. For a subset of ∼200 000 main sequence turn-off (MSTO) and subgiant branch (SGB) stars, we computed distances and ages using theStarHorsecode, with a mean precision of 1% and 12%, respectively.Results.First, we confirm the existence of metal-poor stars in thin-disc orbits. The majority of these stars are predominantly old (> 10 Gyr), with over 50% being older than 13 Gyr. Second, we report the discovery of the oldest thin disc of the Milky Way, which extends across a wide range of metallicities, from metal-poor to super-solar stars. The metal-poor stars in disc orbits manifest as a readily visible tail of the metallicity distribution. We calculate the vertical velocity dispersion (σVz) for the high-[α/Fe] thick disc, finding 35 ± 0.6 km s−1, while the thin disc within the same age range has aσVzthat is lower by 10–15 km s−1. Our old thin discσVzappears similar to those estimated for the high-z disc galaxies. Third, as a verification ofStarHorseages, we extend the [Y/Mg] chemical clock to the oldest ages and estimate a slope of −0.038 dex ⋅ Gyr−1. Finally, we confirm our discovery of the old thin disc by showing that the ‘splash’ population includes high- and low-[α/Fe] populations that are both old, and extends to a wider [Fe/H] range, reaching supersolar [Fe/H]. We find that about 6–10% of the old thin disc was heated to thick-disc orbits. The youngest ‘splashed’ stars appear at 9–10 Gyr and may suggest a Gaia-Sausage/Enceladus (GSE) merger at this period.Conclusions.The Milky Way thin disc formed less than 1 billion years after the Big Bang and continuously built up in an inside-out manner – this finding precedes the earlier estimates of the time at which the MW thin disc began to form (around 8–9 Gyr) by about 4–5 billion years. We find that the metal-poor stars in disc orbits reported by previous studies belong to this old thin disc. Considering a massive merger event such as the GSE, a splash is expected – we find a portion of the old thin disc is heated to thick disc velocities and the splash extends to supersolar [Fe/H] regimes.
- Open Access
- 26 January 2024
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 682, A9
BeyondGaiaDR3: Tracing the [α/M] – [M/H] bimodality from the inner to the outer Milky Way disc withGaia-RVS and convolutional neural networks
Context.In June 2022,GaiaDR3 provided the astronomy community with about one million spectra from the Radial Velocity Spectrometer (RVS) covering the CaII triplet region. In the nextGaiadata releases, we anticipate the number of RVS spectra to successively increase from several 10 million spectra to eventually more than 200 million spectra. Thus, stellar spectra are projected to be produced on an ‘industrial scale’, with numbers well above those for current and anticipated ground-based surveys. However, one-third of the published spectra have 15 ≤S /N≤ 25 per pixel such that they pose problems for classical spectral analysis pipelines, and therefore, alternative ways to tap into these large datasets need to be devised.Aims.We aim to leverage the versatility and capabilities of machine learning techniques for supercharged stellar parametrisation by combiningGaia-RVS spectra with the full set ofGaiaproducts and high-resolution, high-quality ground-based spectroscopic reference datasets.Methods.We developed a hybrid convolutional neural network (CNN) that combines theGaiaDR3 RVS spectra, photometry (G, G_BP, G_RP), parallaxes, and XP coefficients to derive atmospheric parameters (Teff, log(g) as well as overall [M/H]) and chemical abundances ([Fe/H] and [α/M]). We trained the CNN with a high-quality training sample based on APOGEE DR17 labels.Results.With this CNN, we derived homogeneous atmospheric parameters and abundances for 886 080 RVS stars that show remarkable precision and accuracy compared to external datasets (such as GALAH and asteroseismology). The CNN is robust against noise in the RVS data, and we derive very precise labels down to S/N =15. We managed to characterise the [α/M] - [M/H] bimodality from the inner regions to the outer parts of the Milky Way, which has never been done using RVS spectra or similar datasets.Conclusions.This work is the first to combine machine learning with such diverse datasets and paves the way for large-scale machine learning analysis ofGaia-RVS spectra from future data releases. Large, high-quality datasets can be optimally combined thanks to the CNN, thereby realising the full power of spectroscopy, astrometry, and photometry.
- 26 January 2024
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 682, A9
Open AccessBeyondGaiaDR3: Tracing the [α/M] – [M/H] bimodality from the inner to the outer Milky Way disc withGaia-RVS and convolutional neural networks
Context.In June 2022,GaiaDR3 provided the astronomy community with about one million spectra from the Radial Velocity Spectrometer (RVS) covering the CaII triplet region. In the nextGaiadata releases, we anticipate the number of RVS spectra to successively increase from several 10 million spectra to eventually more than 200 million spectra. Thus, stellar spectra are projected to be produced on an ‘industrial scale’, with numbers well above those for current and anticipated ground-based surveys. However, one-third of the published spectra have 15 ≤S /N≤ 25 per pixel such that they pose problems for classical spectral analysis pipelines, and therefore, alternative ways to tap into these large datasets need to be devised.Aims.We aim to leverage the versatility and capabilities of machine learning techniques for supercharged stellar parametrisation by combiningGaia-RVS spectra with the full set ofGaiaproducts and high-resolution, high-quality ground-based spectroscopic reference datasets.Methods.We developed a hybrid convolutional neural network (CNN) that combines theGaiaDR3 RVS spectra, photometry (G, G_BP, G_RP), parallaxes, and XP coefficients to derive atmospheric parameters (Teff, log(g) as well as overall [M/H]) and chemical abundances ([Fe/H] and [α/M]). We trained the CNN with a high-quality training sample based on APOGEE DR17 labels.Results.With this CNN, we derived homogeneous atmospheric parameters and abundances for 886 080 RVS stars that show remarkable precision and accuracy compared to external datasets (such as GALAH and asteroseismology). The CNN is robust against noise in the RVS data, and we derive very precise labels down to S/N =15. We managed to characterise the [α/M] - [M/H] bimodality from the inner regions to the outer parts of the Milky Way, which has never been done using RVS spectra or similar datasets.Conclusions.This work is the first to combine machine learning with such diverse datasets and paves the way for large-scale machine learning analysis ofGaia-RVS spectra from future data releases. Large, high-quality datasets can be optimally combined thanks to the CNN, thereby realising the full power of spectroscopy, astrometry, and photometry.
- Open Access
- 4 January 2024
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 681, L8
Insights from super-metal-rich stars: Is the Milky Way bar young?
Context. Bar formation and merger events can contribute to the rearrangement of stars within the Galaxy in addition to triggering star formation (SF) epochs. Super-metal-rich (SMR) stars found in the solar neighbourhood (SNd) can be used as tracers of such events as they are expected to originate only in the inner Galaxy and to have definitely migrated.Aims. We study a homogeneous and large sample of SMR stars in the SNd to provide tighter constraints on the epoch of the bar formation and its impact on the Milky Way (MW) disc stellar populations.Methods. We investigated a sample of 169 701 main sequence turnoff (MSTO) and subgiant branch (SGB) stars with 6D phase space information and high-quality stellar parameters coming from thehybrid-CNNanalysis of theGaia-DR3 RVS stars. We computed distances and ages using theStarHorsecode with a mean precision of 1% and 11%, respectively. Of these stars, 11 848 have metallicity ([Fe/H]) above 0.15 dex.Results. We report a metallicity dependence of spatial distribution of stellar orbits shown by the bimodal distribution in the guiding radius (Rg) at 6.9 and 7.9 kpc, first appearing at [Fe/H] ~ 0.1 dex, becoming very pronounced at higher [Fe/H]. In addition, we observe a trend where the most metal-rich stars, with [Fe/H] ~ 0.4 dex, are predominantly old (9–12 Gyr), but show a gradual decline in [Fe/H] with age, reaching approximately 0.25 dex about 4 Gyr ago, followed by a sharp drop around 3 Gyr ago. Furthermore, our full dataset reveals a clear peak in the age–metallicity relationship during the same period, indicating a SF burst around 3–4 Gyr ago with slightly sub-solar [Fe/H] and enhanced [α/Fe].Conclusions. We show that the SMR stars are good tracers of bar activity.We interpret the steep decrease in the number of SMR stars at around 3 Gyr as the end of the bar formation epoch. In this scenario the peak of bar activity also coincides with a peak in the SF activity in the disc. Although the SF burst around 3 Gyr ago has been reported previously, its origin was unclear. Here we suggest that the SF burst was triggered by the high bar activity, 3–4 Gyr ago. According to these results and interpretation, the MW bar could be young.
- 4 January 2024
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 681, L8
Open AccessInsights from super-metal-rich stars: Is the Milky Way bar young?
Context. Bar formation and merger events can contribute to the rearrangement of stars within the Galaxy in addition to triggering star formation (SF) epochs. Super-metal-rich (SMR) stars found in the solar neighbourhood (SNd) can be used as tracers of such events as they are expected to originate only in the inner Galaxy and to have definitely migrated.Aims. We study a homogeneous and large sample of SMR stars in the SNd to provide tighter constraints on the epoch of the bar formation and its impact on the Milky Way (MW) disc stellar populations.Methods. We investigated a sample of 169 701 main sequence turnoff (MSTO) and subgiant branch (SGB) stars with 6D phase space information and high-quality stellar parameters coming from thehybrid-CNNanalysis of theGaia-DR3 RVS stars. We computed distances and ages using theStarHorsecode with a mean precision of 1% and 11%, respectively. Of these stars, 11 848 have metallicity ([Fe/H]) above 0.15 dex.Results. We report a metallicity dependence of spatial distribution of stellar orbits shown by the bimodal distribution in the guiding radius (Rg) at 6.9 and 7.9 kpc, first appearing at [Fe/H] ~ 0.1 dex, becoming very pronounced at higher [Fe/H]. In addition, we observe a trend where the most metal-rich stars, with [Fe/H] ~ 0.4 dex, are predominantly old (9–12 Gyr), but show a gradual decline in [Fe/H] with age, reaching approximately 0.25 dex about 4 Gyr ago, followed by a sharp drop around 3 Gyr ago. Furthermore, our full dataset reveals a clear peak in the age–metallicity relationship during the same period, indicating a SF burst around 3–4 Gyr ago with slightly sub-solar [Fe/H] and enhanced [α/Fe].Conclusions. We show that the SMR stars are good tracers of bar activity.We interpret the steep decrease in the number of SMR stars at around 3 Gyr as the end of the bar formation epoch. In this scenario the peak of bar activity also coincides with a peak in the SF activity in the disc. Although the SF burst around 3 Gyr ago has been reported previously, its origin was unclear. Here we suggest that the SF burst was triggered by the high bar activity, 3–4 Gyr ago. According to these results and interpretation, the MW bar could be young.