Images of two Latte simulations of Milky Way-like galaxies (m12i and m12f) at z = 0, from the Latte simulation suite. Left image is zoomed out, showing a disk-dominated galaxy at center, a realistic population of satellite dwarf galaxies, and a diffuse stellar halo, including tidal streams and shells from disrupted satellites. Right image zooms in the Milky Way-like galaxy at center.


Mock galactic (Aitoff) projection within a Latte simulation of a Milky Way-like galaxy, as seen from a star 10 kpc from the center.


The Latte suite of FIRE-2 simulations, part of the FIRE project, aims to push the resolution frontier by simulating Milky Way-like galaxies at unprecedented resolution, self-consistently modeling the stellar disk, stellar halo, dark-matter halo, and satellites dwarf galaxies of Milky Way-like systems. The combination of ultra-high resolution and the state-of-the-art FIRE model makes the Latte simulation suite an unprecedented theoretical tool for understanding the formation of the Milky Way (and M31 and similar-mass galaxies) in a cosmological context.

The Latte simulations are a suite of individual (isolated) Milky Way-mass halos that are simulated with the same resolution, cosmology, and FIRE-2 physics model.
They are selected from a cosmological volume of periodic box length 85.5 Mpc with ΛCDM cosmology: Ω_Λ = 0.728, Ω_matter = 0.272, Ω_baryon = 0.0455, h = 0.702, σ_8 = 0.807, and n_s = 0.961. From this volume, we select halos at z = 0 based only on their mass, M_200m = 1 – 2 x 10^12 M_sun, and an isolation criterion (no neighboring halos of similar mass within at least 5 x R_200m) to limit computational cost. In particular, our selection is agnostic to any halo properties beyond mass and isolation, including formation history, concentration, spin, or subhalo population. The particle mass resolution is 35,000 M_sun for dark-matter particles and 7070 M_sun for gas and star particles. The spatial (force) resolution is: 40 pc for dark matter, 4 pc for stars, and 1 pc (minimum) for gas.

The Latte suite of similation was introduced in Wetzel et al. (2016). We have shown that the Latte simulations produce galaxies that reasonably agree with the observed properties of the MW, M31, and similar-mass galaxies at z = 0, without any `fine-tuning’ of our galaxy-formation model, including:

  • realistic stellar-to-halo mass relation (Hopkins et al 2018)
  • realistic stellar thin + thick disk morphology and metallicity gradients (Ma et al 2017)
  • realistic HI gas kinematics (El-Badry et al 2018)
  • circum-galactic medium (halo gas) measurements of HI and OVI that agree with the COS-Halos survey (Hummels et al, in prep)
  • realistic satellite dwarf galaxies that do not suffer from the `missing satellites’ or `too-big-to-fail’ problems in ΛCDM cosmology (Wetzel et al 2016, Garrison-Kimmel et al 2018)
  • satellite dwarf galaxies with realistic metallicity distributions (Escala et al 2018)
  • realistic stellar halos (Sanderson et al 2017, Bonaca et al 2017)
  • giant molecular clouds (Lakhlani et al, in prep)

Full (raw) snapshot files at z = 0 for three Latte simulations (m12f, m12i, m12m), including all dark matter, gas, and star particles, as well as corresponding synthetic Gaia DR2-like surveys for each simulation, are available at