GSE vs. LMC: reshaping of radially biased stellar haloes by satellites

Abstract

Perturbations from the Large Magellanic Cloud (LMC) of the Milky Way's stellar and dark matter haloes are well-established. However, studies have generally not considered haloes with high radial anisotropy, like debris from the Gaia Sausage-Enceladus (GSE) in the Milky Way. We run a series of test particle simulations of stellar haloes with different velocity anisotropies $β\in[0.5,0.9]$. The LMC causes these initially axisymmetric haloes to become approximately triaxial. Their major axes are aligned with its orbital plane and tilted by up to $\sim14^\circ$ with respect to a fixed Galactic disc. These effects become much more dramatic as $β$ increases, causing the halo to fractionate spatially according to anisotropy. This confirms the expectations of an analytical model, which predicts that orbits with eccentricities $e\gtrsim0.95$ should azimuthally align with the tidal field of the LMC. The reshaping of the $β=0.9$ halo creates strong overdensities of $\sim40\%$ at heliocentric distances as close as 15 kpc. These coincide with the well-known Virgo Overdensity (VOD) and Hercules-Aquila Cloud (HAC), which have previously been associated with the GSE. We propose that the HAC and VOD were created by the dynamical alignment of highly eccentric orbits by the LMC, and are not necessarily relics of the GSE merger geometry. We conclude that previous works have significantly underestimated perturbations from the LMC in the inner stellar halo by not considering sufficiently high velocity anisotropy. This effect should be corrected for when constructing equilibrium models of the GSE.

Figures (11)

• Figure 1: Geometry of our analytic model setup. Top panel: top-down view of the Galactic plane, with an illustration of an orbit around its apocentre at radius $r_\mathrm{apo}$ and azimuth $\psi$. Bottom panel: edge-on view of the Galactic plane with the location of the satellite marked, at radius $r_\mathrm{sat}$ and polar angle $\theta_\mathrm{sat}$. An example of an orbit with inclination $i$ is also shown.
• Figure 2: Integral of motion space $(L_z,E)$ coloured by the minimum dimensionless pendulum energy $k_\mathrm{min}^2\equiv k^2(\theta=0)$. The black dotted line indicates orbits with eccentricity $e=0.95$, and the contours show the GSE. A circular orbit at the Sun's radius is shown with the $\odot$ symbol. At low energies alignment by the satellite's tidal field is only possible ($k_\mathrm{min}^2<1$) on highly eccentric orbits ($e\gtrsim0.95$).
• Figure 3: Past orbit of the LMC in our simulations. Top panel: the orbit in Galactic coordinates $(l,b)$ (where $l$ increases to the left). The approximate regions of the HAC and VOD are shown for comparison. The HAC is split into two regions, north (HAC-N) and south (HAC-S) of the Galactic disc (see Section \ref{['section:overdensities']}). The Middle panel: the track of the orbit in the $(y,z)$ plane, close to the LMC's orbital plane. The current locations of the Milky Way and LMC are marked. Bottom panel: Galactocentric radius $r_\mathrm{LMC}$ of the LMC as a function of time $t$. The simulation runs from $t\approx-4$ Gyr (when the LMC is near apocentre) to the present-day at $t=0$.
• Figure 4: Projected density of our fiducial simulation (with $\beta=0.9$), in the initial (top row) and final (bottom row) snapshots. The left-hand panels show the top-down view of the Galactic plane, and the others show edge-on projections. The LMC and its past orbit are marked, and the location of the Sun is shown with a $\odot$ symbol. The LMC causes the halo to become triaxial and tilted with respect to the Galactic plane.
• Figure 5: Evolution of the axis ratios and orientations of the $\beta=0.9$ stellar halo. Top row: the Galactic longitude $l_\mathrm{long}$ and latitude $b_\mathrm{long}$ of the halo's long axis as a function of time, where we have taken $180^\circ\leq l_\mathrm{long}<360^\circ$. The vertical dashed lines mark the time at which the LMC is 200 kpc from the Milky Way. Bottom-left panel: the track of the long axis in Galactic coordinates $(l_\mathrm{long},b_\mathrm{long})$ compared to the direction of the LMC and several measurements of the stellar halo. These come from han2022, lane2023, and li2025. The orientation of our simulated halo is very close to the measurement of the GSE by lane2023. Bottom-right panel: the intermediate-to-long and short-to-long axis ratios $p$ and $q$ as a function of time. The corresponding parameters for the han2022 fit are shown with dashed lines, with $1\sigma$ uncertainties indicated by coloured bands.