arXiv subjects

Robyn E. Sanderson

Publications and source records attributed to Robyn E. Sanderson.

At least 37 records · Page 2Linked to original sources

arXiv2022

Applying the metallicity-dependent binary fraction to double white dwarf formation: Implications for LISA

Short-period double white dwarf (DWD) binaries will be the most prolific source of gravitational waves (GWs) for the Laser Interferometer Space Antenna (LISA). DWDs with GW frequencies below $\sim1$ mHz will be the dominant contributor to a stochastic foreground caused by overlapping GW signals. Population modeling of Galactic DWDs typically assumes a binary fraction of 50% and a log-uniform Zero Age Main Sequence (ZAMS) orbital period distribution. However, recent observations have shown that the binary fraction of close, solar-type stars exhibits a strong anti-correlation with metallicity which modulates the ZAMS orbital period distribution below $10^4$ days. In this study we perform the first simulation of the Galactic DWD population observable by LISA which incorporates an empirically-derived metallicity-dependent binary fraction, using the binary population synthesis suite COSMIC and a metallicity-dependent star formation history. We compare two models: one which assumes a metallicity-dependent binary fraction, and one with a binary fraction of 50%. We repeat our analysis for three different assumptions for Roche-lobe overflow interactions. We find that while metallicity impacts the evolution and intrinsic properties of our simulated DWD progenitor binaries, the LISA-resolvable populations of the two models remain roughly indistinguishable. However, the size of the total Galactic DWD population orbiting in the LISA frequency band is reduced by more than half when accounting for a metallicity-dependent binary fraction for two of our four variations, which also lowers the effective foreground. The LISA population remains unchanged in number for two variations, highlighting the sensitivity of the population to binary evolution prescriptions.

astro-ph.HE↗

arXiv2022

Orbital dynamics and histories of satellite galaxies around Milky Way-mass galaxies in the FIRE simulations

The orbits of satellite galaxies encode rich information about their histories. We investigate the orbital dynamics and histories of satellite galaxies around Milky Way (MW)-mass host galaxies using the FIRE-2 cosmological simulations, which, as previous works have shown, produce satellite mass functions and spatial distributions that broadly agree with observations. We first examine trends in orbital dynamics at z = 0, including total velocity, specific angular momentum, and specific total energy: the time of infall into the MW-mass halo primarily determines these orbital properties. We then examine orbital histories, focusing on the lookback time of first infall into a host halo and pericenter distances, times, and counts. Roughly 37 per cent of galaxies with Mstar < 10^7 Msun were `pre-processed' as a satellite in a lower-mass group, typically ~2.7 Gyr before falling into the MW-mass halo. Half of all satellites at z = 0 experienced multiple pericenters about their MW-mass host. Remarkably, for most (67 per cent) of these satellites, their most recent pericenter was not their minimum pericenter: the minimum typically was ~40 per cent smaller and occurred ~6 Gyr earlier. These satellites with growing pericenters appear to have multiple origins: for about half, their specific angular momentum gradually increased over time, while for the other half, most rapidly increased near their first apocenter, suggesting that a combination of a time-dependent MW-mass halo potential and dynamical perturbations in the outer halo caused these satellites' pericenters to grow. Our results highlight the limitations of idealized, static orbit modeling, especially for pericenter histories.

astro-ph.GA↗

arXiv2022

The observable properties of galaxy accretion events in Milky Way-like galaxies in the FIRE-2 cosmological simulations

In the $Λ$-Cold Dark Matter model of the Universe, galaxies form in part through accreting satellite systems. Previous work have built an understanding of the signatures of these processes contained within galactic stellar halos. This work revisits that picture using seven Milky Way-like galaxies in the \textit{Latte} suite of FIRE-2 cosmological simulations. The resolution of these simulations allows a comparison of contributions from satellites above M$_{*}$$\gtrsim$10$\times$$^{7}$M$_{\odot}$, enabling the analysis of observable properties for disrupted satellites in a fully self-consistent and cosmological context. Our results show that, the time of accretion and the stellar mass of an accreted satellite are fundamental parameters that in partnership dictate the resulting spatial distribution, orbital energy, and [$α$/Fe]-[Fe/H] compositions of the stellar debris of such mergers $at$ $present$ $day$. These parameters also govern the resulting dynamical state of an accreted galaxy at $z=0$, leading to the expectation that the inner regions of the stellar halo (R$_{\mathrm{GC}}$ $\lesssim$30 kpc) should contain fully phase-mixed debris from both lower and higher mass satellites. In addition, we find that a significant fraction of the lower mass satellites accreted at early times deposit debris in the outer halo (R$_{\mathrm{GC}}$ $>$50 kpc) that are $not$ fully phased-mixed, indicating that they could be identified in kinematic surveys. Our results suggest that, as future surveys become increasingly able to map the outer halo of our Galaxy, they may reveal the remnants of long-dead dwarf galaxies whose counterparts are too faint to be seen $in$ $situ$ in higher redshift surveys.

astro-ph.GA↗

arXiv2022

Constraining the Tilt of the Milky Way's Dark Matter Halo with the Sagittarius Stream

Recent studies have suggested that the Milky Way (MW)'s Dark Matter (DM) halo may be significantly tilted with respect to its central stellar disk, a feature that might be linked to its formation history. In this work, we demonstrate a method of constraining the orientation of the minor axis of the DM halo using the angle and frequency variables. This method is complementary to other traditional techniques, such as orbit fitting. We first test the method using a simulated tidal stream evolving in a realistic environment inside an MW-mass host from the FIRE cosmological simulation, showing that the theoretical description of a stream in the action-angle-frequency formalism still holds for a realistic dwarf galaxy stream in a cosmological potential. Utilizing the slopes of the line in angle and frequency space, we show that the correct rotation frame yields a minimal slope difference, allowing us to put a constraint on the minor axis location. Finally, we apply this method to the Sagittarius stream's leading arm. We report that the MW's DM halo is oblate with the flattening parameter in the potential $q\sim0.7-0.9$ and the minor axis pointing toward $(\ell,b) = (42^{o},48^{o})$. Our constraint on the minor axis location is weak and disagrees with the estimates from other works; we argue that the inconsistency can be attributed in part to the observational uncertainties and in part to the influence of the Large Magellanic Cloud.

astro-ph.GA↗

arXiv2022

On the stability of tidal streams in action space

In the Gaia era it is increasingly apparent that traditional static, parameterized models are insufficient to describe the mass distribution of our complex, dynamically evolving Milky Way (MW). In this work, we compare different time-evolving and time-independent representations of the gravitational potentials of simulated MW-mass galaxies from the FIRE-2 suite of cosmological baryonic simulations. Using these potentials, we calculate actions for star particles in tidal streams around three galaxies with varying merger histories at each snapshot from 7 Gyr ago to the present day. We determine the action-space coherence preserved by each model using the Kullback-Leibler Divergence to gauge the degree of clustering in actions and the relative stability of the clusters over time. We find that all models produce a clustered action space for simulations with no significant mergers. However, a massive (mass ratio prior to infall more similar than 1:8) interacting galaxy not present in the model will result in mischaracterized orbits for stars most affected by the interaction. The locations of the action space clusters (i.e. the orbits of the stream stars) are only preserved by the time-evolving model, while the time-independent models can lose significant amounts of information as soon as 0.5--1 Gyr ago, even if the system does not undergo a significant merger. Our results imply that reverse-integration of stream orbits in the MW using a fixed potential is likely to give incorrect results if integrated longer than 0.5 Gyr into the past.

astro-ph.GA↗

arXiv2022

The In-situ Origins of Dwarf Stellar Outskirts in FIRE-2

Extended, old, and round stellar halos appear to be ubiquitous around high-mass dwarf galaxies ($10^{8.5}<M_\star/M_\odot<10^{9.6}$) in the observed universe. However, it is unlikely that these dwarfs have undergone a sufficient number of minor mergers to form stellar halos that are composed of predominantly accreted stars. Here, we demonstrate that FIRE-2 (Feedback in Realistic Environments) cosmological zoom-in simulations are capable of producing dwarf galaxies with realistic structure, including both a thick disk and round stellar halo. Crucially, these stellar halos are formed in-situ, largely via the outward migration of disk stars. However, there also exists a large population of "non-disky" dwarfs in FIRE-2 that lack a well-defined disk/halo and do not resemble the observed dwarf population. These non-disky dwarfs tend to be either more gas poor or to have burstier recent star formation histories than the disky dwarfs, suggesting that star formation feedback may be preventing disk formation. Both classes of dwarfs underscore the power of a galaxy's intrinsic shape -- which is a direct quantification of the distribution of the galaxy's stellar content -- to interrogate the feedback implementation in simulated galaxies.

astro-ph.GA↗

arXiv2022

Kinematics and Metallicity of Red Giant Branch Stars in the Northeast Shelf of M31

We obtained Keck/DEIMOS spectra of 556 individual red giant branch stars in 4 spectroscopic fields spanning $13-31$ projected kpc along the Northeast (NE) shelf of M31. We present the first detection of a complete wedge pattern in the space of projected M31-centric radial distance versus line-of-sight velocity for this feature, which includes the returning stream component of the shelf. This wedge pattern agrees with expectations of a tidal shell formed in a radial merger and provides strong evidence in favor of predictions of Giant Stellar Stream (GSS) formation models in which the NE shelf originates from the second orbital wrap of the tidal debris. The observed concentric wedge patterns of the NE, West (W), and Southeast (SE) shelves corroborate this interpretation independently of the models. We do not detect a kinematical signature in the NE shelf region corresponding to an intact progenitor core, favoring GSS formation models in which the progenitor is completely disrupted. The shelf's photometric metallicity distribution implies that it is dominated by tidal material, as opposed to the phase-mixed stellar halo or the disk. The metallicity distribution ([Fe/H]$_{\rm phot}$ = $-0.42$ $\pm$ $0.01$) also matches the GSS, and consequently the W and SE shelves, further supporting a direct physical association between the tidal features.

astro-ph.GA↗

arXiv2022

Gaia May Detect Hundreds of Well-characterised Stellar Black Holes

Detection of black holes (BHs) with detached luminous companions (LCs) can be instrumental in connecting the BH properties with their progenitors' since the latter can be inferred from the observable properties of the LC. Past studies showed the promise of Gaia astrometry in detecting BH-LC binaries. We build upon these studies by: 1) initialising the zero-age binary properties based on realistic, metallicity-dependent star-formation history in the Milky Way (MW), 2) evolving these binaries to current epoch to generate realistic MW populations of BH-LC binaries, 3) distributing these binaries in the MW preserving the complex age-metallicity-Galactic position correlations, 4) accounting for extinction and reddening using three-dimensional dust maps, 5) examining the extended Gaia mission's ability to resolve BH-LC binaries. We restrict ourselves to detached BH-LC binaries with orbital period <10 yr such that Gaia can observe at least one full orbit. We find: 1) the extended Gaia mission can astrometrically resolve 30-300 detached BH-LC binaries depending on our assumptions of supernova physics and astrometric detection threshold; 2) Gaia's astrometry alone can indicate BH candidates for 10-100 BH-LC binaries by constraining the dark primary mass >3 Msun; 3) distributions of observables including orbital periods, eccentricities, and component masses are sensitive to the adopted binary evolution model, hence can directly inform binary evolution models. Finally, we comment on the potential to further characterise these BH binaries through radial velocity measurements and observation of X-ray counterparts.

astro-ph.GA↗

arXiv2021

Orbital phase-driven biases in Galactic mass constraints from stellar streams

One of the most promising tracers of the Galactic potential in the halo region are stellar streams. However, individual stream fits can be limited by systematic biases. To study these individual stream systematics, we fit streams in Milky Way-like galaxies from FIRE cosmological galaxy formation simulations with an analytic gravitational potential by maximizing the clustering of stream stars in action space. We show that for coherent streams the quality of the constraints depends on the orbital phase of the observed stream stars, despite the fact that the phase information is discarded in action-clustering methods. Streams on intermediate phases give the most accurate results, whereas pericentre streams can be highly biased. This behaviour is tied to the amount of correlation present between positions and momenta in each stream's data: weak correlation in pericentre streams prohibits efficient differentiation between potentials, while strong correlation in intermediate streams promotes it. Although simultaneous fitting of multiple streams is generally prescribed as the remedy to combat individual stream biases, we find that combining multiple pericentric streams is not enough to yield a bias-free result. We finally show that adopting the two-component Stäckel model does not fundamentally induce a biased mass estimate. With our full data set of two multi-wrap streams, we recovered the true rotation curve of the simulated galaxy within $12\%$ over the entire range of radii covered by our set of stars (10 - 176 kpc) and within $6.5\%$ between the 5 and 95-percentile distance range (23 - 109 kpc).

astro-ph.GA↗

arXiv2021

Reading the CARDs: the Imprint of Accretion History in the Chemical Abundances of the Milky Way's Stellar Halo

In the era of large-scale spectroscopic surveys in the Local Group (LG), we can explore using chemical abundances of halo stars to study the star formation and chemical enrichment histories of the dwarf galaxy progenitors of the Milky Way (MW) and M31 stellar halos. In this paper, we investigate using the Chemical Abundance Ratio Distributions (CARDs) of seven stellar halos from the Latte suite of FIRE-2 simulations. We attempt to infer galaxies' assembly histories by modelling the CARDs of the stellar halos of the Latte galaxies as a linear combination of template CARDs from disrupted dwarfs, with different stellar masses $M_{\star}$ and quenching times $t_{100}$. We present a method for constructing these templates using present-day dwarf galaxies. For four of the seven Latte halos studied in this work, we recover the mass spectrum of accreted dwarfs to a precision of $<10\%$. For the fraction of mass accreted as a function of $t_{100}$, we find residuals of $20-30\%$ for five of the seven simulations. We discuss the failure modes of this method, which arise from the diversity of star formation and chemical enrichment histories dwarf galaxies can take. These failure cases can be robustly identified by the high model residuals. Though the CARDs modeling method does not successfully infer the assembly histories in these cases, the CARDs of these disrupted dwarfs contain signatures of their unusual formation histories. Our results are promising for using CARDs to learn more about the histories of the progenitors of the MW and M31 stellar halos.

astro-ph.GA↗

arXiv2021

The origin of metal-poor stars on prograde disk orbits in FIRE simulations of Milky Way-mass galaxies

In hierarchical structure formation, metal-poor stars in and around the Milky Way (MW) originate primarily from mergers of lower-mass galaxies. A common expectation is therefore that metal-poor stars should have isotropic, dispersion-dominated orbits that do not correlate strongly with the MW disk. However, recent observations of stars in the MW show that metal-poor ([Fe/H] < -2) stars are preferentially on prograde orbits with respect to the disk. Using the FIRE-2 suite of cosmological zoom-in simulations of MW/M31-mass galaxies, we investigate the prevalence and origin of prograde metal-poor stars. Almost all (11 of 12) of our simulations have metal-poor stars on preferentially prograde orbits today and throughout most of their history: we thus predict that this is a generic feature of MW/M31-mass galaxies. The typical prograde-to-retrograde ratio is ~2:1, which depends weakly on stellar metallicity at Fe/H] < -1. These trends predicted by our simulations agree well with MW observations. Prograde metal-poor stars originate largely from a single LMC/SMC-mass gas-rich merger 7 - 12.5 Gyr ago, which deposited existing metal-poor stars and significant gas on an orbital vector that sparked the formation of and/or shaped the orientation of a long-lived stellar disk, giving rise to a prograde bias for all low-metallicity stars. We find sub-dominant contributions from in-situ stars formed in the host galaxy before this merger, and in some cases, additional massive mergers. We find few clear correlations between any properties of our MW/M31-mass galaxies at z = 0 and the degree of this prograde bias as a result of diverse merger scenarios.

astro-ph.GA↗

arXiv2021

Using Action Space Clustering to Constrain the Accretion History of Milky Way like Galaxies

In the currently favored cosmological paradigm galaxies form hierarchically through the accretion of numerous satellite galaxies. Since the satellites are much less massive than the host halo, they occupy a small fraction of the volume in action space defined by the potential of the host halo. Since actions are conserved when the potential of the host halo changes adiabatically, stars from an accreted satellite are expected to remain clustered in action space as the host halo evolves. In this paper, we identify accreted satellites in three Milky Way like disk galaxies from the cosmological baryonic FIRE-2 simulations by tracking satellite galaxies through simulation snapshots. We then try to recover these satellites by applying the cluster analysis algorithm Enlink to the orbital actions of accreted star particles in the present-day snapshot. We define several metrics to quantify the success of the clustering algorithm and use these metrics to identify well-recovered and poorly-recovered satellites. We plot these satellites in the infall time-progenitor mass (or stellar mass) space, and determine the boundaries between the well-recovered and poorly-recovered satellites in these two spaces with classification tree method. The groups found by Enlink are more likely to correspond to a real satellite if they have high significance, a quantity assigned by Enlink. Since cosmological simulations predict that most stellar halos have a population of insitu stars, we test the ability of Enlink to recover satellites when the sample is contaminated by 10-50% of insitu star particles, and show that most of the satellites well-recovered by Enlink in the absence of insitu stars, stay well-recovered even with 50% contamination. We thus expect that, in the future, cluster analysis in action space will be useful in upcoming data sets (e.g. Gaia) for identifying accreted satellites in the Milky Way.

astro-ph.GA↗

arXiv2021

The Galaxy Progenitors of Stellar Streams around Milky Way-mass Galaxies in the FIRE Cosmological Simulations

Stellar streams record the accretion history of their host galaxy. We present a set of simulated streams from disrupted dwarf galaxies in 13 cosmological simulations of Milky Way (MW)-mass galaxies from the FIRE-2 suite at $z=0$, including 7 isolated Milky Way-mass systems and 6 hosts resembling the MW-M31 pair (full dataset at: https://flathub.flatironinstitute.org/sapfire). In total, we identify 106 simulated stellar streams, with no significant differences in the number of streams and masses of their progenitors between the isolated and paired environments. We resolve simulated streams with stellar masses ranging from $\sim 5\times10^5$ up to $\sim 10^{9} M_\odot$, similar to the mass range between the Orphan and Sagittarius streams in the MW. We confirm that present-day simulated satellite galaxies are good proxies for stellar stream progenitors, with similar properties including their stellar mass function, velocity dispersion, [Fe/H] and [$α$/H] evolution tracks, and orbital distribution with respect to the galactic disk plane. Each progenitor's lifetime is marked by several important timescales: its infall, star-formation quenching, and stream-formation times. We show that the ordering of these timescales is different between progenitors with stellar masses higher and lower than $\sim 2\times10^6 M_\odot$. Finally, we show that the main factor controlling the rate of phase-mixing, and therefore fading, of tidal streams from satellite galaxies in MW-mass hosts is non-adiabatic evolution of the host potential. Other factors commonly used to predict phase-mixing timescales, such as progenitor mass and orbital circularity, show virtually no correlation with the number of dynamical times required for a stream to become phase-mixed.

astro-ph.GA↗

arXiv2021

New families in our Solar neighborhood: applying Gaussian Mixture models for objective classification of structures in the Milky Way and in simulations

The standard picture of galaxy formation motivates the decomposition of the Milky Way into 3--4 stellar populations with distinct kinematic and elemental abundance distributions: the thin disk, thick disk, bulge, and stellar halo. To test this idea, we construct a Gaussian mixture model (GMM) for both simulated and observed stars in the Solar neighborhood, using measured velocities and iron abundances (i.e., an augmented Toomre diagram) as the distributions to be decomposed. We compare results for the Gaia-APOGEE DR16 crossmatch catalog of the Solar neighborhood with those from a suite of synthetic Gaia-APOGEE crossmatches constructed from FIRE-2 cosmological simulations of Milky Way-mass galaxies. We find that in both the synthetic and real data, the best-fit GMM uses five independent components, some of whose properties resemble the standard populations predicted by galaxy formation theory. Two components can be identified unambiguously as the thin disk and another as the halo. However, instead of a single counterpart to the thick disk, there are three intermediate components with different age and alpha abundance distributions (although these data are not used to construct the model). We use decompositions of the synthetic data to show that the classified components indeed correspond to stars with different origins. By analogy with the simulated data, we show that our mixture model of the real Gaia-APOGEE crossmatch distinguishes the following components: (1) a classic thin disk of young stars on circular orbits (46%), (2) thin disk stars heated by interactions with satellites (22%), (3, 4) two components representing the velocity asymmetry of the alpha-enhanced thick disk (27%), and (5) a stellar halo consistent with early, massive accretion (4%).

astro-ph.GA↗

arXiv2021

Galactic potential constraints from clustering in action space of combined stellar stream data

Stream stars removed by tides from their progenitor satellite galaxy or globular cluster act as a group of test particles on neighboring orbits, probing the gravitational field of the Milky Way. While constraints from individual streams have been shown to be susceptible to biases, combining several streams from orbits with various distances reduces these biases. We fit a common gravitational potential to multiple stellar streams simultaneously by maximizing the clustering of the stream stars in action space. We apply this technique to members of the GD-1, Pal 5, Orphan and Helmi streams, exploiting both the individual and combined data sets. We describe the Galactic potential with a Stäckel model, and vary up to five parameters simultaneously. We find that we can only constrain the enclosed mass, and that the strongest constraints come from the GD-1, Pal 5 and Orphan streams whose combined data set yields $M(< 20\ \mathrm{kpc}) = 2.96^{+0.25}_{-0.26} \times 10^{11} \ M_{\odot}$. When including the Helmi stream in the data set, the mass uncertainty increases to $M(< 20\ \mathrm{kpc}) = 3.12^{+3.21}_{-0.46} \times 10^{11} \ M_{\odot}$.

astro-ph.GA↗

arXiv2020

Quantifying the Stellar Halo's Response to the LMC's Infall with Spherical Harmonics

The vast majority of the mass in the Milky Way (MW) is in dark matter (DM); we therefore cannot directly observe the MW mass distribution, and have to use tracer populations in order to infer properties of the MW DM halo. However, MW halo tracers do not only feel the gravitational influence of the MW itself. Tracers can also be affected by MW satellites; Garavito-Camargo et al. (2019) (hereafter GC19) demonstrate that the Large Magellanic Cloud (LMC) induces a density wake in the MW DM, resulting in large scale kinematic patterns in the MW stellar halo. In this work, we use spherical harmonic expansion (SHE) of the velocity fields of simulated stellar halos in an effort to disentangle perturbations on large scales (e.g., due to the LMC itself as well as the LMC-induced DM wake) and small scales (due to substructure). Using the GC19 simulations, we demonstrate how the different terms in the SHE of the stellar velocity field reflect the different wake components, and show that these signatures are a strong function of the LMC mass. An exploration of model halos built from accreted dwarfs Bullock & Johnston (2005) suggests that stellar debris from massive, recent accretion events can produce much more power in the velocity angular power spectra than the perturbation from the LMC-induced wake. We therefore consider two models for the Sagittarius (Sgr) stream -- the most recent, massive accretion event in the MW apart from the LMC -- and find that the angular power on large scales is generally dominated by the LMC-induced wake, even when Sgr is included. We conclude that SHE of the MW stellar halo velocity field may therefore be a useful tool in quantifying the response of the MW DM halo to the LMC's infall.

astro-ph.GA↗

arXiv2020

Beyond Gaia: Asteroseismic Distances of M giants using Ground-Based Transient Surveys

Evolved stars near the tip of the red giant branch (TRGB) show solar-like oscillations with periods spanning hours to months and amplitudes ranging from $\sim$1 mmag to $\sim$100 mmag. The systematic detection of the resulting photometric variations with ground-based telescopes would enable the application of asteroseismology to a much larger and more distant sample of stars than is currently accessible with space-based telescopes such as \textit{Kepler} or the ongoing Transiting Exoplanet Survey Satellite (\textit{TESS}) mission. We present an asteroseismic analysis of 493 M giants using data from two ground-based surveys: the Asteroid Terrestrial-impact Last Alert System (ATLAS) and the All-Sky Automated Survey for Supernovae (ASAS-SN). By comparing the extracted frequencies with constraints from \textit{Kepler}, the Sloan Digital Sky Survey Apache Point Observatory Galaxy Evolution Experiment (APOGEE), and Gaia we demonstrate that ground-based transient surveys allow accurate distance measurements to oscillating M giants with a precision of $\sim$15$\%$. Using stellar population synthesis models we predict that ATLAS and ASAS-SN can provide asteroseismic distances to $\sim$2$\times$10$^{6}$ galactic M giants out to typical distances of $20-50 \; \rm{kpc}$, vastly improving the reach of Gaia and providing critical constraints for Galactic archaeology and galactic dynamics.

astro-ph.SR↗

arXiv2020

Cataloging Accreted Stars within Gaia DR2 using Deep Learning

The goal of this study is to present the development of a machine learning based approach that utilizes phase space alone to separate the Gaia DR2 stars into two categories: those accreted onto the Milky Way from those that are in situ. Traditional selection methods that have been used to identify accreted stars typically rely on full 3D velocity, metallicity information, or both, which significantly reduces the number of classifiable stars. The approach advocated here is applicable to a much larger portion of Gaia DR2. A method known as "transfer learning" is shown to be effective through extensive testing on a set of mock Gaia catalogs that are based on the FIRE cosmological zoom-in hydrodynamic simulations of Milky Way-mass galaxies. The machine is first trained on simulated data using only 5D kinematics as inputs and is then further trained on a cross-matched Gaia/RAVE data set, which improves sensitivity to properties of the real Milky Way. The result is a catalog that identifies around 767,000 accreted stars within Gaia DR2. This catalog can yield empirical insights into the merger history of the Milky Way and could be used to infer properties of the dark matter distribution.

astro-ph.GA↗