Data Specifications

The page contains a full description of the data products accompanying the public release of the Auriga Project data: the snapshots containing raw particle/cell data; the group catalogues that hold information about halos and subhalos for each snapshot; the merger trees; and supplementary data.

Snapshots

globus path structure: level?/<suite_name>/halo_Z/snapdir_X/snapshot_X.Y.hdf5

Each Auriga simulation stores a number of snapshots containing the raw data of each particle and cell. Each snapshot (denoted by number X) is split across Y files within a snapdir_X directory. The number of snapshots varies between the different simulation sets (see the table on the data download page). In total, there are 4 different output lists used. The collapsible tables below list the scale factor and redshift values of snapshots for each of these output lists.

The particle/cell data in any given snapshot is organized according to their group/subgroup membership, as per the FoF or Subfind algorithms. For each particle type, the sort order is: GroupNumber, SubgroupNumber, BindingEnergy, where particles belonging to the group but not to any of its subgroups ("fuzz") are included after the last subgroup. The following figure provides a schematic view of the particle organization within a snapshot, for one particle type. Note that the truncation of a snapshot in chunks is arbitrary, thus halos may happen to be stored across multiple, subsequent chunks. Similarly, the different particle types of a halo can be stored in different sets of chunks.

Every HDF5 snapshot contains several groups: "Header", "Parameters", "Configuration", and five "PartTypeX" groups, for the following particle types (the DM only runs have a single PartType1 group):

• gas cells: PartType0
• high-resolution dark matter particles: PartType1
• low-resolution dark matter particles: PartType2 & PartType3
• star particles & wind particles: PartType4
• black holes: PartType5
• tracer particles: PartType6

The Header information can be viewed easily with the following command:

where the snapshot file corresponds to the block 0 of snapshot 127.

The most important information held in the Header group is:

The data fields/attributes belonging to each PartTypeX group can be similarly checked with command line interface hdf5 script. For example, the following command shows the fields for high resolution dark matter particles:

$ h5ls snapshot_127.0.hdf5/PartType1
Coordinates              Dataset {1311009, 3} 
ParticleIDs              Dataset {1311009} 
Potential                Dataset {1311009} 
Velocities               Dataset {1311009, 3} 
        

The system of units used in Auriga is:

UnitLength          3.085678e24        ;  1.0 Mpc
            UnitMass            1.989e43           ;  1.0e10 solar masses
            UnitVelocity        1e5                   ;  1 km/sec
        

The following table lists the most important data fields in the snapshot for each of the PartTypeX groups, including dimensions, units and descriptions. For an extensive list, we refer the reader to IllustrisTNG website (but note that the UnitLength is 1 kiloparsec for those simulations, not 1 Megaparsec as used in Auriga).

In addition to the snapshots described above, there exist high-cadence (5 Myr) snapshots containing only star particle information for the dwarf galaxy simulation suites: Halos_1e11msol/4, Halos_1e11msol/3, and Halos_1e10msol/3.

Note that the python analysis package (available at this bitbucket repository) structures vector arrays, such as coordinates and velocities, as 2D arrays of the shape (N,3), where the second dimension indices [0,1,2] correspond to [Z,Y,X] directions, respectively.

Merger Trees

globus path structure: level?/<suite_name>/mergertrees/halo_Z/trees_sf1_X.Y.hdf5

The merger trees for the Auriga simulations are produced in post-processing using the code developed by Springel et al. 2005, and are very similar to the LHaloTree data product of the IllustrisTNG simulations. They provide useful linking indices that connect individual objects catalogued by FOF and SUBFIND both within the same snapshot and over different snapshots. This is important, because subhalos are not always listed in the same order between snapshots: as they are ordered by mass, 2 objects of similar mass can overtake each other over time and consequently their subhalo index can change.

The schematic diagram below illustrates how the linking indices given in the merger trees work. The most important linking indices are the 'FirstProgenitor', 'Descendent', and 'NextProgenitor' ones. In particular, 'Descendent' allows one to walk foward in time and identify the same object at the next output. On the other hand, 'FirstProgenitor' allows one to talk backward in time by identifying an object's most massive progenitor at preceding outputs. Finally, 'NextProgenitor' moves sideways to identify other progenitors in the preceding output that have the same descendant as each other (if this is more than one, a merger has occurred). Traversing the tree in any direction (forward, backward, or sideways) can be done recursively until an index of -1 is returned, which indicates the end of the branch. We provide an example of walking the tree on the Analysis Code & Examples page.

The mergertree output is split across several HDF5 files called trees_sf1_127.[chunknum].hdf5. In each file there are a number of groups named TreeX, where X is an integer which increases from zero to the number of tree groups in that file chunk. For the Auriga zoom-in simulations, Tree0 is the relevant tree group. The following tables summarise the data fields held in the tree.

In addition to the above, many of the fields stored in the Group and subhalo catalogues are also stored in the tree for convenience.

It is useful to note that the python analysis package (available at this bitbucket repository) automatically converts the units of many snapshot, group catalogue, and merger tree fields to physical units upon loading the data.

Accreted Particle lists

globus path structure: level?/<suite_name>/accretedlists/halo_Z/halo_ZstarID_accreted_all_newmtree_X.hdf5

We provide additional information about the star particles found within the $R_{200}$ of the main halo at redshift zero. This information is designed to be particularly useful for understanding the origin and accretion history of star particles that were born elsewhere and later accreted into the main halo, and that exist in the final snapshot as either disrupted debris or part of a satellite galaxy. These data can be structured in a series of files, one for each snapshot of the simulation that contains only the star particles born in that snapshot. Thus, all files should be loaded to get a complete list of all the star particles found within the $R_{200}$ of the halo at redshift zero; an example of how to do this can be found on the Analysis Code & Examples page. The data fields are described in the table below.

Note that, for a minority of star particles, some of the ex-situ fields can have negative values if not applicable: for example, BirthSubhaloIndex is negative if a star particle is not associated with a subhalo at its time of birth.

Published high level data

High level data are the outcome of analysis of the simulations, such as various galaxy properties, and are often tabulated in publications. In this section, we make high level data available in a convenient form, such as .txt files that can be easily read by, for example, numpy.

i) Milky Way galaxy main properties [download]

Table 1 of Grand et al. (2017) that lists properties such as the mass and scale lengths of the stellar disc and bulge of the main galaxy in the Milky Way mass simulations at redshift zero. Quantities and units are listed in the downloadable file. A citation to this paper is requested if these data are used and/or referenced.

ii) Milky Way galaxy HI gas disc properties [download]

Table 1 of Marinacci et al. (2017) that lists the mass and sizes of the HI gas discs of the main galaxy in the Milky Way mass simulations at redshift zero. Quantities and units are listed in the downloadable file. A citation to this paper is requested if these data are used and/or referenced.

iii) Milky Way satellite & subhalo properties [download]

Table 1 of Simpson et al. (2018) that lists the mass, maximum rotation velocity, number of subhalos, and their fraction of quenched and HI-poor satellite galaxies and subhalos of Milky Way mass hosts at redshift zero. Quantities and units are listed in the downloadable file. A citation to this paper is requested if these data are used and/or referenced.

iv) Milky Way bar properties [download]

Table 1 of Fragkoudi et al. (2024) that lists various properties of the stellar bar of the main galaxy in the Milky Way mass simulations, including the bar formation time, and its strength and length at formation time and redshift zero. The disc mass and size at these times are also given. Quantities and units are listed in the downloadable file. A citation to this paper is requested if these data are used and/or referenced.

v) Milky Way stream properties [download]

Table B1 of Riley et al. (2024) and Table A1 of Shipp et al. (2024) list various properties of accretion events onto Milky Way mass hosts, including morphology at redshift zero, orbital quantities, and matches in higher resolution runs. Quantities and units are listed in the downloadable file. Citations to both papers are requested if these data are used and/or referenced.