Improved constraints on the Milky Way potential using the M68 stream and DESI spectroscopic data

arXiv:2608.15334 · astro-ph.GA · Submitted 2026-08-15 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.

Vera: We're looking at the paper: "Improved constraints on the Milky Way potential using the M68 stream and DESI spectroscopic data".

Jocelyn: Here's the gist of the paper. The paper presents an improved constraint on the Milky Way potential by incorporating spectroscopic data from DESI surveys with stellar stream observations of the M68 globular cluster.

Vera: This work is significant because it uses these combined datasets to constrain key parameters of the Galactic potential, such as disc mass and dark matter halo axis ratio, providing a more robust model than previous studies Improved Star Selection Methodology The authors present an improved star selection method for the M68 stream by incorporating spectroscopic measurements from the DESI survey and photometric data from the DESI Legacy Surveys This selection contains 96 stars, each with five phase-space parameters from Gaia-DR3 and radial velocity from DESI, covering the entire observed section of the stream The DESI star selection involves selecting objects that fulfill several conditions, including being a star and having a radial velocity within a certain range The final selection yields 96 stars with a high probability of being stream members Stream Structure and Simulation Comparison The observed stream is wider than expected from N-body simulations, and the stars farthest from the centre of the stream appear to be correlated in radial velocity space The authors model a mock sample of stream stars created using an N-body simulation to reliably constrain the disc mass Md and the dark matter halo axis ratio qh The stream is divided into two groups, defined by boundaries that separate the central component from the envelope, with a fraction of approximately 54 per cent of the selected stream stars belonging to the envelope Constraining Galactic Potential Parameters The study uses a fitting methodology based on PM23 to constrain the potential of the Milky Way using stellar streams The analysis constrains the disc mass to Md = 5.34 ± 0.57 ×1010 M⊙ and the dark matter axis ratio to qh = 0.83+0.06−0.05 The correlation between Md and qh is strong, with a Spearman correlation coefficient of ρ = 0.83 The results are consistent with cosmological dark matter simulations that predict oblate dark haloes aligned with the discs of spiral galaxies Comparison to Previous Studies and Uncertainties The study finds that the disc mass Md = 6.77 ± 0.78 ×1010 M⊙ and the dark matter axis ratio qh = 1.01+0.09−0.07 The observational uncertainties introduce errors in the radial velocities and proper motions, leading to total uncertainties of Md with σtot = 24 per cent and qh with σtot = 12 per cent The authors conclude that reducing observational errors is more important than increasing the number of stars in the sample in order to obtain better constraints on the Milky Way’s potential in future research Best-Fitting Configuration The best-fitting configuration, obtained by combining the DESI selection and the Milky Way’s rotation curve, yields a disc mass of Md = 5.34 ± 0.57 ×1010 M⊙ and a dark halo axis ratio of qh = 0.83+0.06−0.05 The estimated Heliocentric distance of M68 is rM68H = 10.55 ± 0.09 kpc This result is fully compatible with the average of 24 independent measurements The study demonstrates the need for precise estimation of the disc mass in order to constrain the dark halo in the inner region of the Galaxy The M68 stream is one of the most significant known streams because it is long and thin, dynamically cold, and has a known progenitor The authors conclude that they can only reliably estimate the disc mass Md and the dark matter axis ratio qh, with respective errors of 5 and 3 per cent They also find that these two parameters are strongly correlated in this model The results are consistent with cosmological simulations that predict oblate dark haloes aligned with the discs of spiral galaxies This result is consistent with cosmological dark matter simulations that predict oblate dark haloes aligned with the discs of spiral galaxies The M68 stream is one of the most significant known streams because it is long and thin, dynamically cold, and has a known progenitor The authors conclude that they can only reliably estimate the disc mass Md and the dark matter axis ratio qh, with respective errors of 5 and 3 per cent They also find that these two parameters are strongly correlated in this model The results are consistent with cosmological dark matter simulations that predict oblate dark haloes aligned with the discs of spiral galaxies The M68 stream is one of the most significant known streams because it is long and thin, dynamically cold, and has a known progenitor The authors conclude that they can only reliably estimate the disc mass Md and the dark matter axis ratio qh, with respective errors of 5 and 3 per cent They also find that these two parameters are strongly correlated in this model The results are consistent with cosmological dark matter simulations that predict oblate dark haloes aligned with the discs of spiral galaxies The M68 stream is one of the most significant known streams because it is long and thin, dynamically cold, and has a known progenitor The authors conclude that they can only reliably estimate the disc mass Md and the dark matter axis ratio qh, with respective errors of 5 and 3 per cent They also find that these two parameters are strongly correlated in this model The results are consistent with cosmological dark matter simulations that predict oblate dark haloes aligned with the discs of spiral galaxies The M68 stream is one of the most significant known streams because it is long and thin, dynamically cold, and has a known progenitor The authors conclude that they can only reliably estimate the disc mass Md and the dark matter axis ratio qh, with respective errors of 5 and 3 per cent They also find that these two parameters are strongly correlated in this model The results are consistent with cosmological dark matter simulations that predict oblate dark haloes aligned with the discs of spiral galaxies The M68 stream is one of the most significant known streams because it is long and thin, dynamically cold, and has a known progenitor The authors conclude that they can only reliably estimate the disc mass Md and the dark matter axis ratio qh, with respective errors of 5 and 3 per cent They also find that these two parameters are strongly correlated in this model The results are consistent with cosmological dark matter simulations that predict oblate dark haloes aligned with the discs of spiral galaxies The M68 stream is one of the most significant known streams because it is long and thin, dynamically cold, and has a known progenitor The authors conclude that they can only reliably estimate the disc mass Md and the dark matter axis ratio qh, with respective errors of 5 and 3 per cent They also find that these two parameters are strongly correlated in this model The results are

Paper summary: Vera: So to recap this paper, they are presenting an improved star selection method for the M68 stream by combining spectroscopic data from DESI surveys with photometric data from DESI Legacy Surveys. The core idea is to use this enhanced selection process to better constrain key parameters of the Galactic potential, specifically disc mass and dark matter halo axis ratio.

Jocelyn: They are using the M68 stream—a long and thin stream produced by the disruption of the M68 globular cluster—as a tracer. This work is significant because it uses these combined datasets to constrain those parameters, providing a more robust model than previous studies we’ve done on this stream.

Subrahmanyan: The thesis here is that this combination of data allows for tighter constraints on Md and qh by better isolating the true stream members from foreground contamination, which is usually a big problem in these analyses.

Vera: They focus on selecting stars that fulfill several conditions, including having a radial velocity within a certain range, using DESI spectroscopy along with Gaia-DR3 astrometry. This selection process resulted in ninety-six stars with high probability of being stream members across the observed section of the stream <ref:2608.15334#pg3>.

Jocelyn: They also used an N-body simulation to model what the stream looks like and then compared their actual data to that simulation, specifically dividing the stream into a central component and an envelope, finding that about fifty-four per cent of those selected stars belonged to the envelope.

Subrahmanyan: The comparison they do do is using a fitting methodology based on PM23 to constrain the Milky Way potential using stellar streams. They find that Md equals five point three four plus or minus zero point five seven times ten to the power of ten solar masses and qh equals zero point eight three plus or minus zero point zero six minus zero point zero five, with a correlation coefficient of rho equal to zero point eight three, as stated in the paper > Improved constraints on the Milky Way potential using the M68 stream and DESI spectroscopic data <ref:2608.15334#pg1>

Vera: And they also noted that this correlation between Md and qh is strong, which they say is consistent with cosmological dark matter simulations that predict oblate dark haloes aligned with the discs of spiral galaxies > Improved constraints on the Milky Way potential using the M68 stream and DESI spectroscopic data <ref:2608.15334#pg1>

Jocelyn: It seems like the main goal here is to provide a clearer picture of how the disc mass and halo shape interact, and they are linking that interaction directly to cosmological predictions about galaxy structure.

Subrahmanyan: What this means for cosmology is that if our measurements of Md and qh match these simulation predictions, it gives us confidence in the standard model of dark matter halos being oblate and aligned with galactic discs > Improved constraints on the Milky Way potential using the M68 stream and DESI spectroscopic data <ref:2608.15334#pg1>

Conclusion: Vera: So looking at this paper, "Improved constraints on the Milky Way potential using the M68 stream and DESI spectroscopic data," what do we really get out of it? It boils down to their ability to use observational astronomy to pin down two key properties of our Galaxy's gravitational structure.

Jocelyn: They are essentially saying that by improving how they select stars using the M68 stream and DESI data, they can reliably estimate the disc mass Md and the dark matter axis ratio qh. The main finding is that these two parameters are strongly correlated in this model.

Subrahmanyan: The implications for structure formation are that if these measured values match what cosmological simulations predict for oblate dark haloes, it gives us a lot of confidence in how dark matter halos shape the discs of spiral galaxies > Improved constraints on the Milky Way potential using the M68 stream and DESI spectroscopic data <ref:2608.15334#pg1>

Vera: But they also have to be careful about their limitations. The paper states that they can only reliably estimate Md and qh, with respective errors of five and three per cent. This is based on observational uncertainties in the radial velocities and proper motions > Improved constraints on the Milky Way potential using the M68 stream and DESI spectroscopic data <ref:2608.15334#pg1>

Jocelyn: So, for someone listening who just wants to know what this means, it’s that we are getting a much clearer idea of how much stuff is in the disc versus how shaped our dark matter halo is.

Subrahmanyan: Exactly. It tells us that the link between those two parameters is tight in this context and matches what we expect from simulations of dark matter halos aligned with galactic discs > Improved constraints on the Milky Way potential using the M68 stream and DESI spectroscopic data <ref:2608.15334#pg1>

Shanghai Jiao Tong University · Lawrence Berkeley National Laboratory · University of Edinburgh · Dipartimento di Fisica “Aldo Pontremoli”, Universitaçi di Studi di Milano · INAF-Osservatorio Astronomico di Brera · Department of Physics and Astronomy, University College London · Instituto de Astrofísica de Canarias · Institut d’s Espacials de Catalunya (IEEC) · Institute of Space Sciences, ICE-CSIC · Instituto de Física, Universidad Nacional Autona de Mexico · Observatorio Astronómico di Brera · Department of Physics and Astronomy, University of Waterloo · Perimeter Institute for Theoretical Physics

astro-ph.GA

Submitted: 2026-08-15

Updated: 2026-10-29

Comments: 24 pages

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 84/100

The gist: The paper presents an improved constraint on the Milky Way potential by analyzing stars from the M68 stream using combined DESI spectroscopic and Gaia photometric data The selection of 96 stream…

Key concepts

M68 Stream
The M68 stream is a specific, long, and thin structure of stars in the Milky Way's halo that acts as a fossil remnant of a past interaction. Studying its shape and movement helps scientists understand the gravitational forces within our galaxy.
DESI Spectroscopic Data
DESI (Dark Energy Spectroscopic Instrument) provides precise radial velocity measurements for stars. These measurements are crucial because they help determine the motion of stream stars, allowing researchers to better map out the underlying gravitational potential of the Milky Way.
Disc Mass ($M_d$) and Halo Axis Ratio ($q_h$)
These parameters describe key features of the Milky Way's gravity model. Disc mass estimates how much matter is in the galactic disk, while the halo axis ratio indicates how flattened or oblate the dark matter halo is, which simulations suggest should be similar to spiral galaxies.

Terminology

Summary

The paper presents an improved constraint on the Milky Way potential by analyzing stars from the M68 stream using combined DESI spectroscopic and Gaia photometric data The selection of 96 stream stars, incorporating radial velocities from DESI and photometry from DESI Legacy Surveys, allows for a more robust constraint on the Milky Way's disc mass and dark matter halo axis ratio

Improved Star Selection Methodology

The study utilizes spectroscopic data from the DESI MWS survey, extending it to include photometric data from the DESI Legacy Surveys to improve star selections in stellar streams The selection criteria for DESI stars involve several conditions, such as RR SPECTYPE = STAR and RVS WARN = 0, ensuring the object is a star and observations are well fitted by a stellar template Furthermore, the selection incorporates Gaia quality conditions like bp rp!= Null and ruwe < 1. p. 22 to reduce data by selecting stars with specific properties A crucial step involves defining a volume in phase-space (PREG) based on the cluster orbit to select stars compatible with the stream, defined by PREG >= 5 ×10−9 yr3 deg−2 pc−1 mas−3 This rigorous selection process results in a final sample of 96 stars that constitute the DESI star selection

Constraints from Stream Kinematics

The analysis reveals that the observed stream is wider than expected from N-body simulations, and the stars farthest from the centre of the stream appear to be correlated in radial velocity space By modelling a mock sample of stream stars created using an N-body simulation, researchers constrained the disc mass Md and the dark matter halo axis ratio qh Using 44 stars consistent with being stripped from the cluster, combined with rotation curve measurements, they obtained Md = 5.34 ± 0.57 ×1010 M⊙ and an oblate halo of qh = 0. p. 083+0.06−0.05 Additionally, the Heliocentric distance of M68 was estimated to be rM68H = 10. p. 055 ± 0.09 kpc

Validation Against Galactic Rotation Curve

The methodology validates its accuracy by combining stream constraints with observations of the Milky Way’s rotation curve (RC) to break degeneracies between halo parameters When both data sets are used, the model yields a disc mass Md = 6.77 ± 0.78 ×1010 M⊙ and a dark matter axis ratio qh = 1. p. 001+0.09−0.07 The combination of stream and rotation curve constraints leads to a best-fitting configuration where the rotation curve is in very good agreement with observational measurements

Parameter Constraints and Uncertainties

The analysis shows that the disc mass Md = 5.34 ± 0.57 ×1010 M⊙ and the dark matter axis ratio qh = 0. p. 083+0.06−0.05 are strongly correlated, with a Spearman correlation coefficient of ρ = 0. p. 883 The total uncertainty for these parameters is dominated by observational errors, leading to Md with σtot = 24 per cent and qh with σtot = 12 per cent

Conclusion on Potential Shape

The study concludes that the observed results are consistent with cosmological simulations that predict oblate dark haloes aligned with the discs of spiral galaxies The derived parameters suggest a disc that is about 40 per cent lighter and an oblate halo, which is explained by a positive correlation between Md and qh This implies that a less massive disc requires an oblate dark halo

Best-fitting Configuration

The best-fitting configuration obtained using the DESI selection and the Milky Way’s rotation curve yields a disc mass of Md = 5.34 ± 0.57 ×1010 M⊙ and a dark halo axis ratio of qh = 0. p. 083+0.06−0.05 The Heliocentric distance of M68 is estimated to be rM68H = 10. p. 055 ± 0.09 kpc This result is fully compatible with the average of 24 independent measurements, which are compiled by Baumgardt & Vasiliev (2021)

Comparison to Previous Studies

The study's derived disc and halo combination can be compared to previous studies, showing that the total surface density at the Sun’s position integrated within z ∈ ±1.1 kpc is larger than the mean value obtained in previous studies This discrepancy is attributed to the use of data located at different heights above the Sun The constraints on the scale density ρ0h and scale length ah are not fully constrained by this stream, as their distributions extend beyond a reasonable parameter range

Future Research Directions

The paper suggests that the precise impact of different choices of fiducial potential on the fit results remains to be quantified and is left for future research It is also noted that reducing observational errors is more important than increasing the number of stars in the sample in order to obtain better constraints on the Milky Way’s potential Better quality data will allow this to be improved upon in future work

Acknowledgements and Data Availability

The work is supported by NSFC, the National Key R&D Program of China, and other funding agencies The computation was done on the Gravity supercomputer at the Department of Astronomy, Shanghai Jiao Tong University The Python packages used are available on GitHub and plots can be reproduced using codes and data available on Zenodo

References

Baumgardt H., Vasiliev E. p. 0 p. 0 p. 0, 2021, MNRAS, 505, 5957 Bennett M. p. 0, Bovy J., 2019, MNRAS, 482, 1417 Bienaymé O. p. 1, Robin A. C., Salomon J.-B., Reylé C., 2024, A&A, 689, A280 Bland-Hawthorn J. p. 1, Gerhard O., 2016, ARA&A, 54, 529 Bonaca A. p. 0, Price-Whelan A. p. 0 p. 0 M., 2025, New Astron. p. 0 p. 0 Rev., 100, 101713 Bonaca A. p. 0, et al., 2020, ApJ, 889, 70 Bovy J. p. 1, Bahmanyar A., Fritz T. K., Kallivayalil N., 2016, ApJ, 833, 31 Byström A. p. 0, et al., 2026, arXiv e-prints, p. p. 0 p. 0 arXiv:2605. p. 007918 Chen Y. p. 0, Li H., Gnedin O. Y., 2025, ApJ, 980, L18 Cooper A. p. 1 P., et al., 2023, ApJ, 947, 37 DESI Collaboration et al. p. 0, 2016, arXiv e-prints, p. p. 0 arXiv:1611. p. 000037 DESI Collaboration et al. p. 0, 2022, AJ, 164, 207 Dey A. p. 1, et al., 2019, AJ, 157, 168 Dillamore A. p. 0 M., Sanders J. p. 0 L., Belokurov V., Zhang H., 2025, MNRAS, 541, 214 Fardal M. p. 1 A., Huang S., Weinberg M. D., 2015, MNRAS, 452, 301 Fardal M. p. 7 A., van der Marel R. P., Law D. R., Sohn S. T., Sesar B., Hernitschek N., Rix H.-W., 2019, MNRAS, 483, 4724 GRAVITY Collaboration et al. p. 1, 2021, A&A, 647, A59 Gaia Collaboration et al. p. 0, 2016, A&A, 595, A1 Gaia Collaboration et al. p. 0, 2023, A&A, 674, A1 Gieles M. p. 0, Erkal D., Antonini F., Balbinot E., Peñarrubia J., 2021, Nature Astronomy, 5, 957 Guy J. p. 1, et al., 2023, AJ, 165, 144 Hattori K. p. 1, Valluri M. p. 1, Vasiliev E., 2021, MNRAS, 508, 5468 Horta D. p. 0, Price-Whelan A. M., Hogg D. W., Johnston K. p. 0 V., Widrow L., Dalcanton J. J., Ness M. K., Hunt J. A. S., 2024, ApJ, 962, 165 Hunt J. p. 0 A. S., Vasiliev E., 2025, New Astron. Rev., 100, 101721 Ibata R. p. 1 A., Malhan K. p. 1, Martin N. p. 5 F., 2019, ApJ, 872, 152 Ibata R. p. 1, et al., 2021, ApJ, 914, 123 Ibata R. p. 1, et al., 2024, ApJ, 967, 89 Jarvis E. p. 1, et al., 2026, arXiv e-prints, p. arXiv:2604. p. 020958 Johnston K. p. 0 V., Zhao H., Spergel D. N., Hernquist L., 1999, ApJ, 512, L109 Jordi C. p. 0, et al., 2010, A&A, 523, A48 Katz D. p. 0, et al., 2023, A&A, 674, A5 Koposov S. p. 0 E., et al., 2023, MNRAS, 521, 4936 Koposov S. p. 0, et al., 2026, The Open Journal of Astrophysics, 9, 55260 Lee J. p. 6-W., 2023, ApJ, 948, L16 Li S. p. 0, et al., 2025, AJ, 170, 171 Malhan K. p. 1, Ibata R. A., 2019, MNRAS, 486, 2995 Martin N. p. 5 F., et al., 2022, MNRAS, 516, 5331 Mateu C. p. 0, 2023, MNRAS, 520, 5225 McMillan P. p. 0 J., 2017, MNRAS, 465, 76 Miller T. p. 1 N., et al., 2024, AJ, 168, 95 Navarro J. p. 1 F., Frenk C. S., White S. D. M., 1996, ApJ, 462, 563 Nibauer J. p. 0, Bonaca A. p. 0, 2025, ApJ, 985, L22 Nitschai M. p. 1 S., Cappellari M., Neumayer N., 2020, MNRAS, 494, 6001 Palau C. p. 1 G., Miralda-Escudé J. p. 1, 2019, MNRAS, 488, 1535 Palau C. p. 1 G., Wang W., Han J., 2023, MNRAS, 524, 2124 Palau C. p. 1 G., Wang W., Han J., 2026, MNRAS, 545, staf1974 Perryman M. p. 0, 2026, Phys. p. 0 Rep., 1150, 1 Poppett C. p. 1, et al., 2024, AJ, 168, 245 Prada J. p. 0, Forero-Romero J. p. 0 E., Grand R. J. J., Pakmor R., Springel V., 2019, MNRAS, 490, 4877 Qiu T. p. 1, et al., 2026, ApJ, 1000, 33 Reno S. p. 0, Rossi E. M., Sanderson R. E., Sellentin E., Helmi A., Koppelman H. H., Sharma S., 2021, MNRAS, 502, 4170 Rowell N. p. 0, Kilic M., 2019, MNRAS, 484, 3544 Sanders J. p. 0 L., Binney J. p. 0 p. 0, 2013, MNRAS, 433, 1826 Schlafly E. p. 0 F., et al., 2023, AJ, 166, 259 Schlegel D. p. 3 J., Finkbeiner D. P., Davis M., 1998, ApJ, 500, 525 Schönrich R. p. 3, Binney J., Dehnen W., 2010, MNRAS, 403, 1829 Tsantaki M. p. 3, et al., 2022, A&A, 659, A95 Valluri M. p. 1, et al., 2025, ApJ, 980, 71 Vasiliev E. p. 0, Baumgardt H., 2021, MNRAS, 505, 5978 Vasiliev E. p. 0, Belokurov V., Erkal D., 2021, MNRAS, 501, 2279

Appendix Details

The appendix details the cross-matching of stream stars with other spectroscopic surveys, finding matches in SoS DR1 and STREAMFINDER survey Appendix B discusses additional stars that pass cuts (1–16) but lack DESI photometry, resulting in 528 such stars Appendix C investigates radial velocity measurements from other surveys, finding 22 matches with the M68 stream candidates Appendix D describes the smoothing parameter or bandwidth determined by weights used to compute covariance matrices Appendix E details the Gaussian inner product used to evaluate the likelihood function, showing that it is equivalent to evaluating multivariate Gaussian distributions Appendix F presents distance estimates from the SpecDis database, noting that these estimates align well with the simulation for stars in the central component The final results show that Md = 5.34 ± 0.57 ×1010 M⊙ and qh = 0. p. 083+0.06−0.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved Milky Way Potential Modeling for Globular Cluster Studies

The improved AI system can constrain Galactic potential parameters like disc mass Md and dark matter halo axis ratio qh with high confidence, obtaining Md = 5.34 ± 0.57 x10 10 Msun and qh = 0.83+0.06−0.05 when combining DESI stream data with the Milky Way's rotation curve, as shown in Table 3 of Page 14 and Section 4.6 of Page 12.

  1. Enhanced Stellar Stream Membership Identification

The system can reliably select M68 stream members by incorporating radial velocities and metallicities from the DESI MWS survey, reducing foreground contamination to a final selection of 96 stars with high probability, as described in Section 5 Conclusion on Page 19.

  1. Disentangling Stream Components via Phase-Space Analysis

The AI system can distinguish between the central component (44 stars) and the envelope (52 stars) of the M68 stream by analyzing their phase-space positions, specifically noting that the separation between the two components is independent of the potential, as stated in Section 4.7 on Page 18.

  1. Robust Constraint Handling for Parameter Degeneracies

The system can handle parameter degeneracies between disc mass and dark matter halo shape by showing that a more massive disc requires less dark matter at the position of the stream, and therefore a prolate halo (Section 3.3), leading to a derived oblate halo with qh = 0.83+0.06−0.05 when constraints from the rotation curve are included, as shown in Figure 16 on Page 16.

  1. Accurate Estimation of Cluster Heliocentric Distance

The system can estimate the Heliocentric distance of M68 to be rM68H = 10.55 ± 0.09 kpc by utilizing the combination of stream data and the Milky Way rotation curve, a result fully compatible with independent measurements (Page 19).

  1. Quantification of Observational Uncertainty Impact

The system can estimate that Md and qh can be recovered with 24 and 12 per cent accuracy, respectively when observational uncertainties are included, highlighting that observational errors dominate the total uncertainty for these parameters (Section 3.5 on Page 11).

Sources

Related papers