Shaken and stirred: the Milky Way's dark substructures

MNRAS 467, 4383–4400 (2017) doi:10.1093/mnras/stx360 Advance Access publication 2017 February 11 Shaken and stirred: the Milky Way’s dark substructures Till Sawala,1‹ Pauli Pihajoki,1 Peter H. Johansson,1 Carlos S. Frenk,2 Julio F. Navarro,3 † Kyle A. Oman3 and Simon D. M. White4 1 Department of Physics, University of Helsinki, Gustaf Hällströmin katu 2a, FI-00014 Helsinki, Finland for Computational Cosmology, Department of Physics, University of Durham, South Road, Durham DH1 3LE, UK 3 Department of Physics and Astronomy, University of Victoria, 3800 Finnerty Road, Victoria, BC V8P 5C2, Canada 4 Max-Planck Institute for Astrophysics, Karl-Schwarzschild-Straße 1, D-85741 Garching, Germany 2 Institute ABSTRACT The predicted abundance and properties of the low-mass substructures embedded inside larger dark matter haloes differ sharply among alternative dark matter models. Too small to host galaxies themselves, these subhaloes may still be detected via gravitational lensing or via perturbations of the Milky Way’s globular cluster streams and its stellar disc. Here, we use the APOSTLE cosmological simulations to predict the abundance and the spatial and velocity distributions of subhaloes in the range 106.5 –108.5 M inside haloes of mass ∼1012 M in  cold dark matter. Although these subhaloes are themselves devoid of baryons, we find that baryonic effects are important. Compared to corresponding dark matter only simulations, the loss of baryons from subhaloes and stronger tidal disruption due to the presence of baryons near the centre of the main halo reduce the number of subhaloes by ∼1/4 to 1/2, independently of subhalo mass, but increasingly towards the host halo centre. We also find that subhaloes have non-Maxwellian orbital velocity distributions, with centrally rising velocity anisotropy and positive velocity bias that reduces the number of low-velocity subhaloes, particularly near the halo centre. We parametrize the predicted population of subhaloes in terms of mass, galactocentric distance and velocities. We discuss implications of our results for the prospects of detecting dark matter substructures and for possible inferences about the nature of dark matter. Key words: Local Group – cosmology: theory – dark matter. 1 I N T RO D U C T I O N The  cold dark matter (hereafter CDM) model explains many large-scale observations, from the anisotropy of the microwave background radiation (e.g. Wright et al. 1992) to the distribution of galaxies in the cosmic web (Davis et al. 1985), but inferences about the particle nature of dark matter or its possible (self)-interactions require observations on far smaller scales. Warm dark matter (WDM) particles, such as sterile neutrinos with masses of a few keV, have free-streaming scales of less than 100 kpc, and differ from CDM in terms of the halo mass functions at mass scales on the order of 109 M and below (e.g. Avila-Reese et al. 2001; Bose et al. 2016), while weak self-interactions would produce shallow cores of the order of several kpc in the centre of dark matter haloes (e.g. Spergel & Steinhardt 2000). In principle, there is no shortage of observations that probe these small scales. They include the structures seen  E-mail: † Senior CIfAR Fellow in the Lyman α forest (e.g. Croft et al. 2002; Viel et al. 2013), the abundance of dwarf galaxies in deep H I surveys (Tikhonov & Klypin 2009; Papastergis et al. 2011) and the abundance (e.g. Klypin et al. 1999; Boylan-Kolchin, Bullock & Kaplinghat 2011; Lovell et al. 2012; Kennedy et al. 2014) as well as internal kinematics that probe the density profiles (e.g. Walker & Peñarrubia 2011; Strigari, Frenk & White 2014) of Local Group dwarf galaxies. While these studies have progressively narrowed the parameter space of viable dark matter candidates, inferences about the nonbaryonic nature of dark matter from observations of the Universe’s baryonic components are inherently limited by uncertainties in our understanding of complex astrophysical processes, such as radiative hydrodynamics, gas cooling, star formation, metal-enrichment, stellar winds, supernova and AGN feedback and cosmic reionization. For simple number counts, the effects of baryons in suppressing the formation of dwarf galaxies in CDM can be degenerate with the effects of WDM (e.g. Sawala et al. 2013). As of 2016, a plethora of studies have also offered baryonic solutions to the various problems for CDM that had previously been identified in dark matter only (hereafter DMO) simulations (e.g. Okamoto, Gao & Theuns 2008;  C 2017 The Authors Published by Oxford University Press on behalf of the Royal Astronomical Society Accepted 2017 February 8. Received 2017 February 8; in original form 2016 September 7 4384 T. Sawala et al. Governato et al. 2010; Zolotov et al. 2012; Brooks et al. 2013; Arraki et al. 2014; Chan et al. 2015; Sawala et al. 2015; Dutton et al. 2016). In addition, in the CDM cosmological model, the majority of low-mass substructures that would most easily discriminate between different dark matter models are predicted to be completely dark (Bullock, Kravtsov & Weinberg 2000; Benson et al. 2002; Okamoto et al. 2008; Ocvirk et al. 2016; Sawala et al. 2016a), and hence unobservable through starlight. Fortunately, alternative methods exist that can reveal small structures and substructures purely through their gravitational effect and detect even pure dark matter haloes, thereby potentially breaking the degeneracy with baryonic physics. MNRAS 467, 4383–4400 (2017) While the above phenomena have a gravitational origin, they still fall short of providing a complete census of dark matter substructures. Instead, inferences about dark matter models based on the number of detected perturbations must be made statistically and, in each case, require an accurate prediction of the abundance, properties and distribution of dark matter substructures inside the central ∼10–20 kpc of galaxy or group-sized dark matter haloes. Previous work has relied on very high-resolution DMO simulations such as VIA LACTEA II (Diemand, Kuhlen & Madau 2007) and AQUARIUS (Springel et al. 2008). These have shown that tidal stripping reduces the mass fraction of dark matter contained in self-bound substructures towards the halo centre (e.g. Springel et al. 2008). It has also been argued that the presence of a stellar disc and adiabatic contraction of the halo can lead to enhanced tidal disruption of substructures. Based on DMO simulations with an additional massive disc-like potential, D’Onghia et al. (2010) quantified the disruption of substructures through tidal stripping due to the smooth halo, tidal stirring near pericentre and ‘disc shocking’ by the passage of a substructure through the dense stellar disc. For their parameters, this led to a depletion of substructures by up to a factor of 3 for a subhaloes of mass 107 M . Similarly, Yurin & Springel (2015) imposed a less massive disc inside a DMO simulation, (...truncated)