Supplementary Catalogs

Click a file to download it with your browser, or copy its link and download with wget --content-disposition <THE_LINK>.

The geometry file "geometry.hdf5" contains a 3-D mask array that indicates the reconstruction volume.

To find out whether or not a given object is in the reconstruction volume, we load the mask array from the geometry file. The mask array is a floating point array whose values are either 0.0 or 1.0. We turn them into a Boolean array:

with h5py.File('supplementary/reconstruction-pipeline/geometry.hdf5', 'r') as f:
    mask = f['Masks/SDSS_reconstruction_area'][()] > 0.99
mask.shape, mask.dtype

Output:

((500, 500, 500), dtype('bool'))

l_grid = 500 / mask.shape[0]                # grid size in Mpc/h
x = np.array([
    (332.6301, 318.91406, 63.530796),       # position of Coma
    (0., 0., 0.)                            # position of box origin
])
idx_grid = (x / l_grid + 0.5).astype(int)
mask[tuple(idx_grid.T)]

Output:

array([ True, False])

To download the full tree catalogs, select a catalog and use the shell command to download it. Replace <API_KEY> with your API key.

Please refer to Yangyao Chen, et al. 2023 for the details of the extension algorithm.

To facilitate the common loading pattern of simulation data, the extended version of subhalo merger trees is organized into 125 self-consistent, individual files, each of which contains the trees in a subbox with a side length $L_{\rm subbox} = 100 h^{-1}{\rm Mpc}$. This is done by partitioning the whole simulation box (of side length $L_{\rm box} = 500 h^{-1}{\rm Mpc}$) into $5 \times 5 \times 5$ equal-size subboxes, and assign each subhalo merger tree to a subbox according to the location of the "most massive root subhalo". The extension algorithm is then applied to each subbox, producing 125 files containing the extended version of trees. The subbox is indexed in a row-major manner. For example, the subbox with subbox_id=3 covers the spatial range $0 \leqslant x < 100$, $0 \leqslant y < 100$ and $300 \leqslant z < 400$ (in $h^{-1}{\rm Mpc}$).

Available fields for each subhalo are listed below.

In a given subbox, trees are stored one after one, and subhalos in a given tree are stored in depth-first searching (DFS) order. That means the first progenitor of a subhalo is stored next to it. Thus, by using the main_leaf index, the main branch of a subhalo can be easily selected by a Python "slice".

Extended subhalos at given redshift: We also provide catalogs of subhalos at given redshifts. For example, the catalog at $z=0$ (snapshot 99) for subbox "15" is obtained by selecting all subhalos with snap = 99 in the file of that subbox, and outputing all fields of the selected subhalos into a new file named "subhalos.snap99.subbox15.hdf5". Note that in the output file, subhalo_id is no longer contiguous, and hence the look-up of, e.g. f_in_grp, n_in_grp, etc., has to be performed with a hash table (e.g. Python dict).

Assembly histories: For the subhalos at a given redshift, we also provide their history information. This is recorded in a separate file named "history.snap{snap_id}.hdf5". Subhalos in a given subbox are stored in a data group named "Subboxes/{subbox_id}" under that file. The number of subhalos and their storage order are exactly the same as those in the file "subhalos.snap{snap_id}.subbox{subbox_id}.hdf5" The following fields are available for each subhalo:

Subhalos matching to hydrodynamical simulations: The extension algorithm matches subhalos in the target, low-resolution simulation (S_targ, ELUCID here, with extension) to those in a reference, high-resolution, dark-matter-only simulation (S_ref_dmo, TNG100-1-Dark here) to get assigned physical quantities and positions. In addition, a hydrodynamical simulation with baryonic physics (S_ref_baryonic, TNG100-1 here) can be run under the same initial conditions as S_ref_dmo and populate subhalos in S_ref_dmo with galaxy properties. The combination of these two procedures lights up each subhalo in S_targ with a galaxy modeled by S_ref_baryonic.

We provide a catalog for all subhalos in S_targ with matched indices in S_ref_dmo and S_ref_baryonic. This relies on the matching of subhalos in S_ref_dmo and S_ref_baryonic (see TNG site, the LHaloTree variant). A small number of low-mass subhalos do not have baryonic matching, whose matched indices are set to -1. The file named "tng_100_1_matched.snap{snap_id}.hdf5" have data groups named "Subboxes/{subbox_id}" for each subbox of ELUCID, and under each data group the follow datasets are stored:

See the Jupyter notebook extended_subhalo_samples.ipynb.

Semi-analytical galaxies are available. These galaxies are modeled by following the extended (subhalo) merger tress produced by the DMO runs. Each file named sams.snap{snap_id}.hdf5 records the galaxies in one snapshot.

Available fields for each galaxy are listed below, where $n_{\rm g, snap}$ denotes the total number of galaxies at a snapshot.

Other fields, such as x_ext, v_ext, m_crit200, m_tophat, v_max, ..., are also dumped, with their values taken from the host subhalos (see the list of extended merger tress).

See the python scripts example.py.