Nibauer & Bonaca (2025) made the first fully data-driven measurement of the Galactic acceleration field along a stellar stream, using the GD-1 stream's phase-space track. The measured accelerations agree with previous enclosed-mass and flattening estimates but deviate from axisymmetry at about 2 sigma, and are fit by a dark matter halo that is triaxial and tilted out of the Galactic plane — with two degenerate solutions: a near-prolate configuration tilted ~18 degrees toward the Sun, or an oblate-like one tilted ~56 degrees. Whether this tilted, triaxial configuration is a global property of the Milky Way's inner dark matter halo — rather than a feature local to GD-1's region, a residue of time-dependent perturbation by the Large Magellanic Cloud, or an artifact of distance and stream-fitting systematics — remains undetermined, as does which of the two admissible geometries describes the Galaxy. No published independent acceleration measurement from a spatially separated stream exists, no reported constraint breaks the two-mode degeneracy, and no self-consistent time-dependent reanalysis with realistic LMC histories has appeared; adjacent results (equilibrium models of the tilted stellar halo, a local 3D potential reconstruction near the Sun, DESI DR2 stellar-halo mapping) corroborate a tilted, aspherical halo but are not decisive.

Background

Stellar streams are ribbons of stars tidally stripped from dissolving star clusters; because their stars orbit in the Galactic gravitational potential, the stream's shape and motion encode the potential's mass distribution. GD-1 is one of the longest and coldest (i.e., dynamically narrow) such streams in the Milky Way, a remnant of a globular cluster spanning about 100 degrees on the sky, with member stars at Galactocentric distances of roughly 7-14 kpc. Modern surveys supply its full six-dimensional phase-space position: sky positions and proper motions (angular sky motions) from Gaia, line-of-sight radial velocities from SDSS, LAMOST and DESI spectroscopy, and distances from fitting stars to the subgiant branch (a well-calibrated evolutionary stage of the color-magnitude diagram). A cold stream does not trace a single orbit, but each small segment of it follows a short piece of an orbit. Nibauer, Belokurov, Cranmer, Goodman & Ho (2022) exploited this: fitting a smooth spline curve through the stream's track in position-velocity space and differentiating it along the track yields the local gravitational acceleration, without assuming any parameterized form for the Milky Way's potential. Nibauer & Bonaca (2025) applied this method to GD-1, obtaining, at the location (R,z,ϕ)=(11.9 kpc,7.3 kpc,171.1∘)(R,z,\phi)=(11.9~\mathrm{kpc}, 7.3~\mathrm{kpc}, 171.1^\circ) in cylindrical Galactic coordinates, acceleration components aR=−2.5a_R=-2.5, az=−1.8a_z=-1.8, and aϕ=0.2 km s−1 Myr−1a_\phi=0.2~\mathrm{km\,s^{-1}\,Myr^{-1}}, an enclosed mass of 1.4±0.1×1011M⊙1.4\pm0.1\times10^{11}M_\odot within 14 kpc, and a halo density flattening (short-to-long axis ratio) qρ,z=0.81q_{\rho,z}=0.81 — the latter two consistent with previous estimates. The surprise is the non-zero azimuthal component aϕa_\phi (reaching about 0.65 km s−1 Myr−10.65~\mathrm{km\,s^{-1}\,Myr^{-1}} over part of the stream): a static, axisymmetric halo (one invariant under rotation about the Galactic z-axis) produces no azimuthal acceleration, so the data deviate from axisymmetry at about 2 sigma. Fitting global triaxial halo models (with three unequal principal axes, described by axis ratios a:b:c and by a tilt of the long axis out of the Galactic plane and a yaw angle within it) yields two configurations that both reproduce the data: Mode 1, with axis ratios 1:0.75:0.70, long axis tilted roughly 18 degrees above the Galactic plane in the direction of the Sun; and Mode 2, with axis ratios 1:0.95:0.65, tilted 56 degrees. Mode 1 agrees with independent measurements of the tilted stellar halo from Gaia and the H3 survey and with cosmological simulations (in which cold dark matter halos are generically triaxial and tilted relative to the galactic disk), while Mode 2 does not. The authors tested alternative explanations — the gravitational pull of the Large Magellanic Cloud (LMC), a massive satellite galaxy whose infall demonstrably perturbs the outer halo, and spiral-arm perturbations — using simplified rigid or static models, finding the LMC incompatible with the measured aϕa_\phi at only the 2-3 sigma level. Known systematics include the assumed stream distance track (assigned a 0.5 kpc systematic uncertainty, the spread among literature models) and the spline-fitting choices (knot number, intrinsic-scatter treatment). Because GD-1 occupies a limited spatial region, the measurement constrains the acceleration field only along one stream, and the authors explicitly identify the required follow-up work: study additional streams independently to test whether the constraint represents the total inner halo or is local to GD-1, and obtain additional constraints to decide between the two tilted-halo modes. Conclusively excluding external-perturber explanations — for instance through fully self-consistent time-dependent modeling including realistic LMC histories — remains open, since the authors' rigid and static tests disfavored the LMC only at the 2-3 sigma level.

Problem Statement

Determine whether the Milky Way's inner (roughly within 20 kpc) dark matter halo is globally tilted and triaxial, as suggested by the GD-1 stream acceleration measurement of Nibauer & Bonaca (2025), or whether that signal is local to GD-1's region of the Galaxy, produced by time-dependent perturbers, or dominated by systematics. Concretely, the problem — posed as necessary follow-up work by the source authors — comprises two linked requirements in service of this single objective: (1) obtain independent measurements of the Galactic acceleration field from one or more stellar streams (or equivalent dynamical tracers) spatially separated from GD-1, with uncertainties propagated from distance-track, stream-track-fitting, and intrinsic-scatter systematics, sufficient to test whether the non-axisymmetric signal and the preferred tilted-triaxial configuration are representative of the global inner halo rather than a local feature; and (2) acquire constraints that break the degeneracy between the two admissible halo geometries — a prolate-like configuration with axis ratios about 1:0.75:0.70 and long axis tilted about 18 degrees above the Galactic plane toward the Sun, versus an oblate-like configuration with axis ratios about 1:0.95:0.65 tilted about 56 degrees — deciding which, if either, describes the Milky Way (including the possibility that the halo is axisymmetric or that neither mode is adequate). Because the authors' rigid and static perturber tests disfavored the Large Magellanic Cloud explanation at only the 2-3 sigma level, conclusively excluding external-perturber explanations remains part of the open question. The problem is solved when a quantitative, reproducible analysis supports a decisive statement about the global shape and orientation of the Milky Way's inner dark matter halo (including rejection of the tilted-triaxial hypothesis in favor of an axisymmetric or perturber-driven explanation) at a confidence level that accounts for the stated systematic uncertainties.

References

  1. Nibauer, J., & Bonaca, A. 2025, 'Galactic Accelerations from the GD-1 Stream Suggest a Tilted Dark Matter Halo', The Astrophysical Journal Letters 985, L22, DOI 10.3847/2041-8213/add0a9, https://doi.org/10.3847/2041-8213/add0a9 (preprint arXiv:2504.07187, https://arxiv.org/abs/2504.07187)
  2. Nibauer, J., Belokurov, V., Cranmer, M., Goodman, J., & Ho, S. 2022, 'Charting Galactic Accelerations with Stellar Streams and Machine Learning', The Astrophysical Journal 940, 22, DOI 10.3847/1538-4357/ac93ee, https://doi.org/10.3847/1538-4357/ac93ee
  3. Dillamore, A. M., & Sanders, J. L. 2026, 'Geometry of the Milky Way's dark matter from dynamical models of the tilted stellar halo', Monthly Notices of the Royal Astronomical Society 546, stag226, DOI 10.1093/mnras/stag226, https://doi.org/10.1093/mnras/stag226 (preprint arXiv:2510.00095, https://arxiv.org/abs/2510.00095)
  4. Putney, E., Shih, D., Lim, S. H., & Buckley, M. R. 2025, 'ClearPotential: Revealing Local Dark Matter in Three Dimensions', arXiv:2512.09989, https://arxiv.org/abs/2512.09989
  5. Li, S., Wang, W., Koposov, S. E., Amarante, J. A. S., Deason, A. J., et al. 2026, 'The Milky Way Stellar Halo Is Twisted and Doubly Broken: Insights from DESI DR2 Milky Way Survey Observation', The Astrophysical Journal 999, 108, DOI 10.3847/1538-4357/ae41b9, https://doi.org/10.3847/1538-4357/ae41b9
  6. Li, S., Wang, W., Koposov, S. E., Amarante, J. A. S., Deason, A. J., et al. 2026, 'The velocity field of our Milky Way outer stellar halo based on DESI DR2', The Astrophysical Journal 1005, 19, DOI 10.3847/1538-4357/ae7346, https://doi.org/10.3847/1538-4357/ae7346
  7. Liu, H., & Du, C. 2026, 'Chemical Properties and Sagittarius-induced Dynamical Perturbations of the GD-1 Stream', The Astrophysical Journal 1000, 40, DOI 10.3847/1538-4357/ae486f, https://doi.org/10.3847/1538-4357/ae486f
  8. He, Z., Wang, L., Zhou, Z., Huang, Y., & Vasiliev, E. 2026, 'Disentangling the Morphology of Palomar 5: Effects of the Bar, Spiral Arms, LMC, and Halo Flattening', The Astrophysical Journal 1007, 132, DOI 10.3847/1538-4357/ae8b98, https://doi.org/10.3847/1538-4357/ae8b98
  9. Viterbo, G., & Buck, T. 2026, 'The dynamical memory of tidal stellar streams: Joint inference of the Galactic potential and the progenitor of GD-1 with flow matching', Astronomy & Astrophysics 707, A363, DOI 10.1051/0004-6361/202558358, https://doi.org/10.1051/0004-6361/202558358 (preprint arXiv:2512.04600, https://arxiv.org/abs/2512.04600)