These simulations include self-consistent cosmic ray injection, transport, and feedback in gas. This is the 'single bin' model that represents cosmic rays with a single energy density, roughly the energy of cosmic-ray protons near the peak of the spectrum at 1-10 GeV. These simulations use the cosmic-ray diffusion coefficient that best fits gamma-ray observations, kappa = 3e29 cm^2 g^-1 (or 700 in Gizmo's code units). This value of kappa also produces the most meaningful differences from the simulations with only MHD+.

These simulations also include MHD+, magnetohydrodynamics as well as fully-anisotropic Spitzer-Braginskii conduction and viscosity.

We ran these simulations after we fixed the cosmic-ray heating bug in Gizmo (after 2018-4-26). That bug erroneously applied a heating term from cosmic rays to all gas in the simulation at z >~ 8, which prevented gas cooling and star formation at those early times, and which affects other FIRE-2 simulations.

Most of these MHD+ and Cosmic Ray simulations have a subset of 61 out of 601 snapshots available, specifically, every tenth snapshot. However, all 601 snapshots are available for m12i, m12f, and m12m, and for them we also provide halo/galaxy catalogs, merger trees, and particle tracking. In addition, the following simulations are missing some snapshots (see their notes.txt):
m09, m10v, m11b, m11d, m11h, m12m

For m12f, m12i, and m12m, within track/ we provide pointer indices between particles, as well as host galaxy coordinates, but only for the available snapshots. We also provide star formation coordinates (specifically, the coordinates of a star particle at the first snapshots after it formed). Because (except for m12f, m12i, m12m) we did not have all 601 snapshots when we ran particle tracking for these simulations, these 'formation' coordinates could be up to ~250 Myr after a star particle formed, so use caution in interpreting them.