Aquarius-A: Benchmark MW Dark Matter Halo

• Aquarius-A is a designation for a benchmark Milky Way–mass dark matter halo from the Aquarius simulations, used to study halo morphology and substructure.
• High-resolution realizations like Aq-A-2 and Aq-A-4 enable detailed analyses of halo shape, tagged stellar haloes, and stream–subhalo interaction rates.
• The term ‘Aquarius-A’ exemplifies context specificity, differing in usage across cosmology, radiology QA, and local group dwarf galaxy research.

Aquarius-A is a context-dependent designation whose most technically developed usage is Aq-A, one of the Milky Way–mass dark-matter haloes in the Aquarius Project. In that literature, Aq-A serves as a high-resolution benchmark for studies of halo shape, subhalo structure, tagged stellar haloes, tidal streams, and direct stream–subhalo encounter statistics, especially in the Aq-A-2 and Aq-A-4 realizations (Sanderson et al., 2016). The same name family also appears in other domains, including the AQUARIUS radiology quality-assurance framework, the Aquarius stellar stream or co-moving group, and the isolated dwarf irregular galaxy Aquarius (DDO 210); this suggests that the term must be interpreted strictly from disciplinary context (Wismueller et al., 2022).

1. Context and nomenclature

In cosmological simulation work, Aquarius-A denotes halo Aq-A, one of the six Milky Way–mass haloes of the Aquarius Project, a set of zoom-in, dark-matter-only simulations re-simulated from the Millennium II cosmological run (Sanderson et al., 2016). The Aq-A label is resolution-qualified when necessary, most commonly as Aq-A-2, Aq-A-3, or Aq-A-4.

Other literatures use closely related names without referring to the Aquarius Project halo. In radiology QA, the framework is consistently named AQUARIUS, expanded as Artificial Intelligence-Based QUality Assurance by Restricted Investigation of Unequal Scores; the paper does not introduce an explicit “Aquarius-A” variant, so any equation of “Aquarius-A” with that framework is interpretive rather than in-paper nomenclature (Wismueller et al., 2022). In Galactic archaeology, Aquarius refers to a stellar stream or co-moving group first identified in RAVE and later debated as either disrupted globular-cluster debris or a Galactic dynamical structure (Boer et al., 2012). In Local Group dwarf-galaxy work, Aquarius denotes the dwarf irregular DDO 210, not the Aquarius simulation halo (Kirby et al., 2016).

2. Aq-A as a Milky Way–mass benchmark halo

The Aquarius simulations are high-resolution zoom-in $\Lambda$CDM simulations of Milky Way–mass haloes. In the stream–subhalo interaction study, the specific realization is Aq-A-2, with virial mass $M_{200} = 1.842 \times 10^{12}\,M_\odot$ and particle mass $m_p = 1.37 \times 10^4\,M_\odot$ (Sanderson et al., 2016). At that resolution, Aquarius-A resolves subhalos down to $\sim 10^6\,M_\odot$ and characteristic sizes down to $\sim 0.3$ kpc, which is why it is used to probe the dynamically relevant subhalo mass range for thin stellar streams, $\sim 10^6$–$10^8\,M_\odot$ (Sanderson et al., 2016).

A complementary characterization appears in the stellar-halo stream analysis at Aq-2 resolution, which uses a flat $\Lambda$CDM cosmology with $\Omega_m = 0.25$, $\Omega_\Lambda = 0.75$, $M_{200} = 1.842 \times 10^{12}\,M_\odot$0, $M_{200} = 1.842 \times 10^{12}\,M_\odot$1, and $M_{200} = 1.842 \times 10^{12}\,M_\odot$2 (Gómez et al., 2013). That work specifies a Plummer-equivalent softening length of 65.8 pc and 128 snapshots from $M_{200} = 1.842 \times 10^{12}\,M_\odot$3 to $M_{200} = 1.842 \times 10^{12}\,M_\odot$4 (Gómez et al., 2013). In the halo-shape study, Aq-A-4 is used for convergence analysis, with $M_{200} = 1.842 \times 10^{12}\,M_\odot$5, $M_{200} = 1.842 \times 10^{12}\,M_\odot$6, $M_{200} = 1.842 \times 10^{12}\,M_\odot$7, $M_{200} = 1.842 \times 10^{12}\,M_\odot$8, $M_{200} = 1.842 \times 10^{12}\,M_\odot$9, and $m_p = 1.37 \times 10^4\,M_\odot$0 (Vera-Ciro et al., 2011).

The Aquarius literature repeatedly treats Aq-A as a representative Milky Way–mass CDM halo, but the representativeness is always conditional on model assumptions: the simulations are collisionless, lack a live disk or bulge, and encode baryonic effects only through post-processing when stellar tracers are added (Sanderson et al., 2016).

3. Halo shape, anisotropy, and subhalo structure

Aquarius-A is a central case in studies of halo morphology. Measured at the instantaneous virial radius, Aquarius haloes evolve from typically prolate configurations at early times to more triaxial/oblate geometries at the present day, and this evolution correlates with the angular distribution of infalling material: narrow-filament accretion produces prolate haloes, whereas more isotropic accretion produces triaxial/oblate haloes (Vera-Ciro et al., 2011). At $m_p = 1.37 \times 10^4\,M_\odot$1, the radial structure of Aq-A preserves that history: inner regions are more prolate, outer regions more triaxial/oblate, and the paper identifies a prolate–oblate transition radius for Aq-A of roughly $m_p = 1.37 \times 10^4\,M_\odot$2 (Vera-Ciro et al., 2011).

A second line of work analyzes spatial and velocity anisotropy in mock stellar haloes built on Aquarius. In Aq-A, the whole-sky spatial anisotropy rises with radius when bound satellites are included, but after satellite removal it increases only to $m_p = 1.37 \times 10^4\,M_\odot$3 kpc and then saturates, indicating that the inner-halo anisotropy is driven by diffuse substructure plus halo shape, whereas the outer-halo anisotropy is dominated by surviving satellites (Mondal et al., 2023). The same study finds that Aq-A shows a prominent dip in the velocity-anisotropy profile around $m_p = 1.37 \times 10^4\,M_\odot$4 kpc; removing bound satellites removes that dip, linking it to a massive, tangentially dominated bound satellite or satellite group (Mondal et al., 2023).

Subhalo shape work further sharpens the picture. Across Aquarius, the triaxiality of field haloes increases with halo mass, and the smallest haloes are about 40–50% rounder than Milky Way–like objects at $m_p = 1.37 \times 10^4\,M_\odot$5 (Vera-Ciro et al., 2014). For subhaloes likely to host luminous satellites comparable to the classical dwarf spheroidals, the mean axis ratios are $m_p = 1.37 \times 10^4\,M_\odot$6 and $m_p = 1.37 \times 10^4\,M_\odot$7 at $m_p = 1.37 \times 10^4\,M_\odot$8 kpc, increasing with radius; their velocity ellipsoids become strongly tangentially biased in the outskirts as a consequence of tidal stripping (Vera-Ciro et al., 2014).

4. Tagged stellar halo and stream population in Aq-A

The stellar-halo literature on Aquarius-A is based on particle tagging rather than hydrodynamics. In one implementation, GALFORM is run on Aquarius merger trees and the 1% most bound dark-matter particles in subhaloes are tagged with stellar populations over time (Gómez et al., 2013). In the stream–subhalo interaction study, the retagging is simpler: at the infall snapshot, the most bound 1% of dark-matter particles are tagged as stars, and the total stellar mass assigned by the semi-analytic model of Starkenburg et al. (2013) is divided evenly over those tagged particles (Sanderson et al., 2016). This simplified “1% at infall” tagging was chosen partly because it makes density variations and gaps easier to interpret (Sanderson et al., 2016).

At 8 kpc along the major axis, Aq-A-2 has local accreted stellar-halo density

$m_p = 1.37 \times 10^4\,M_\odot$9

and velocity ellipsoid $\sim 10^6\,M_\odot$0 (Gómez et al., 2013). In the same local “solar-neighbourhood-like” sphere, the number of contributing satellites is $\sim 10^6\,M_\odot$1, the number of tagged star particles is $\sim 10^6\,M_\odot$2, the fraction of star particles in resolved streams is $\sim 10^6\,M_\odot$3, the fraction of stellar mass in resolved streams is $\sim 10^6\,M_\odot$4, and the number of resolved streams is $\sim 10^6\,M_\odot$5 (Gómez et al., 2013). That makes Aq-A-2 the poorest-resolved local volume in that sample, and the paper argues that the low resolved stream fraction is primarily a resolution effect rather than strong chaotic mixing (Gómez et al., 2013).

A broader sky-projection study finds that Aq-A contains >4×10⁵ tagged dark-matter tracer particles carrying stars and, after resampling, about $\sim 10^6\,M_\odot$6 MSTO stars between 1 and 50 kpc (Helmi et al., 2011). In those mock observations, Aq-A exhibits a rich mixture of broad overdensities and thin low-surface-brightness streams, including explicit Sagittarius-like and Orphan-like analogues (Helmi et al., 2011).

Action–angle analyses of Aq-A streams show that many streams still align along relatively straight lines in approximate angle and frequency spaces, even when computed in a spherical NFW potential that is only an approximation to the true halo (Buist et al., 2015). However, Aq-A is more triaxial than Aq-D and its circular-velocity curve contains a bump, so its streams display stronger wiggles, greater frequency-space thickness, and larger angle–frequency misalignments than in cleaner test-particle cases (Buist et al., 2015). The paper attributes most of these deviations to using an incorrect potential, while suggesting that the remaining noisy and patchy morphology is likely due to interactions with the large number of dark-matter subhalos present in the cosmological simulation (Buist et al., 2015).

5. Stream–subhalo interaction rates measured in Aq-A

The most direct “Aquarius-A” result in this corpus is the first self-consistent measurement of stream–subhalo interaction rates in a cosmological Milky Way–mass halo (Sanderson et al., 2016). Using retagged Aq-A-2, the authors selected 18 thin streams at $\sim 10^6\,M_\odot$7, followed them from progenitor infall, and recorded every snapshot in which a dark-matter subhalo passed within fixed impact-parameter thresholds of 1, 2, or 5 kpc, or within one or two times the subhalo half-mass radius $\sim 10^6\,M_\odot$8 (Sanderson et al., 2016).

For a stream of present-day length $\sim 10^6\,M_\odot$9, age $\sim 0.3$0, and recorded encounter count $\sim 0.3$1, the interaction rate is defined as

$\sim 0.3$2

with units of $\sim 0.3$3 (Sanderson et al., 2016). Expressed as encounters per 10 kpc per 10 Gyr, the median Aq-A rates across the 18-stream sample are: $\sim 0.3$4

$\sim 0.3$5

$\sim 0.3$6

(Sanderson et al., 2016). Interpreted literally, a typical thin Aq-A stream 10 kpc long and 10 Gyr old experiences of order one encounter within 1 kpc, about nine within 2 kpc, and about sixty-two within 5 kpc, but the paper explicitly treats these as lower limits because both temporal and particle resolution lower the measured rates (Sanderson et al., 2016).

The mass dependence is also notable. For fixed impact parameters, the encounter distribution broadly tracks the subhalo mass function, so low-mass subhalos dominate weak or distant interactions. For thresholds tied to subhalo size, such as $\sim 0.3$7, or for encounters strong enough to satisfy the paper’s $\sim 0.3$8 criterion, the mass distribution becomes much flatter, and subhalos from $\sim 0.3$9 to $\sim 10^6$0 are roughly equally represented (Sanderson et al., 2016). This suggests that close, potentially gap-opening interactions in Aq-A are not controlled exclusively by the most massive perturbers.

Comparison with analytic estimates is systematically one-sided. Yoon, Johnston & Hogg (2011) predict encounter counts about an order of magnitude higher than the Aq-A measurements at $\sim 10^6$1 kpc, with better agreement at 5 kpc (Sanderson et al., 2016). Carlberg’s gap-opening rates likewise overpredict the number of effective encounters in Aq-A by factors of 10–50 in the 1 kpc mass-limited interpretation, and remain high under a $\sim 10^6$2 threshold interpretation (Sanderson et al., 2016). The paper treats the Aq-A counts as conservative because snapshot spacing is $\sim 10^6$3 Gyr, a typical subhalo travels $\sim 10^6$4 kpc between snapshots, streams contain only a few hundred to $\sim 10^6$5 tagged particles, and repeated encounters by the same subhalo are likely undercounted (Sanderson et al., 2016).

6. Other usages and controversies surrounding the name

Outside the Aquarius Project, the most visible astronomical use of the name is the Aquarius stream or Aquarius group. A high-resolution abundance study of six stars argued that the stream was chemically coherent, with $\sim 10^6$6 to $\sim 10^6$7, mean $\sim 10^6$8, dispersion $\sim 10^6$9 dex, and abundance patterns consistent with disrupted globular-cluster debris rather than a dwarf spheroidal galaxy (Boer et al., 2012). A later MIKE study of five Aquarius stars reached the opposite conclusion, finding $10^8\,M_\odot$0 to $10^8\,M_\odot$1, mean $10^8\,M_\odot$2, dispersion $10^8\,M_\odot$3 dex, no Na–O anti-correlation, and chemistry largely indistinguishable from Milky Way field stars apart from one likely $10^8\,M_\odot$4 Cen debris star; it therefore argued for the term Aquarius group and a Galactic dynamical origin rather than an accreted classical globular cluster (Casey et al., 2013). This is an explicit controversy in the Aquarius nomenclature.

A different usage appears in Local Group dwarf-galaxy work, where Aquarius means the isolated dwarf irregular DDO 210. In that context the galaxy has distance $10^8\,M_\odot$5 kpc, stellar mass $10^8\,M_\odot$6, H I mass $10^8\,M_\odot$7, stellar velocity dispersion $10^8\,M_\odot$8, and mean metallicity $10^8\,M_\odot$9 (Kirby et al., 2016). That Aquarius is unrelated to Aq-A.

The same caution applies outside astronomy. In radiology QA, AQUARIUS is a hybrid human–machine framework that compares AI image analysis with NLP-derived report labels and sends only discordant cases for expert review. In an intracranial hemorrhage study on 1936 head CT scans, expert review of only 29 discordant cases reduced human QA effort by 98.5% and identified six non-reported true ICH-positive cases (Wismueller et al., 2022). The framework is explicitly named AQUARIUS, not Aquarius-A (Wismueller et al., 2022).

Taken together, these literatures establish that “Aquarius-A” is not a universal object name but a context-bound identifier. In current technical usage, its most precise and best-specified meaning remains Aq-A of the Aquarius simulations, especially when discussing halo morphology, tagged stellar haloes, and the cosmological baseline for stream–subhalo interactions (Sanderson et al., 2016).