Galaxy Infall Times in Projected Phase Space: TNG300 and TNG-Cluster Data Products

Galaxy Infall Times in Projected Phase Space: TNG300 and TNG-Cluster Data Products Overview This release contains the galaxy infall-time data products and accompanying Jupyter notebooks presented in Lawlor-Forsyth (2026). The products are constructed from galaxies and galaxy clusters in the IllustrisTNG TNG300 and TNG-Cluster simulations over the observation-redshift range $0 \leqslant z \leqslant 1$. For each selected galaxy cluster, the analysis follows the main progenitor history of the cluster and the main progenitor histories of its satellite galaxies. Galaxy infall is defined relative to the evolving cluster progenitor. The reported infall time is the elapsed time since a galaxy first entered within $3R_{200}$ of its observation-time host cluster while belonging to the corresponding progenitor FoF group. The release includes physical galaxy catalogues, projected phase-space catalogues along three orthogonal lines of sight, cluster selections and histories, infall-assignment statistics, and the notebooks required to generate and analyze the products. Files Data products infall_time_data_public.hdf5 The analysis-ready public data product. This is the recommended download for most users and is approximately 108 MB. It contains: file-, simulation-, and snapshot-level metadata; selected cluster samples; cluster main-progenitor histories; physical galaxy catalogues with assigned infall times; stellar masses and star-formation rates; projected phase-space catalogues; cluster-selection and infall-assignment statistics. The large intermediate halo and subhalo catalogues are omitted. This file contains all quantities required by the tutorial notebook, projected phase-space map construction, interpolation, and most independent downstream analyses. infall_time_data.hdf5 The complete pipeline data product. In addition to the analysis-ready products listed above, this file contains the trimmed FoF halo and Subfind subhalo catalogues used by the pipeline. It is intended for users who wish to inspect, reproduce, or modify earlier pipeline stages without repeatedly reading the complete IllustrisTNG group catalogues. Notebooks 01_generate_infall_time_data.ipynb The data-generation pipeline. This notebook: constructs trimmed halo and subhalo catalogues; selects galaxy clusters at each observation snapshot; follows cluster and galaxy main-progenitor histories; assigns galaxy infall times; generates physical and projected galaxy catalogues; records selection and failure statistics; exports the lightweight public data product. The complete combined file is generated in two passes. The notebook is first run with simulation = "TNG300" and then with simulation = "TNG-Cluster", using the same output file. Running this notebook requires access to the relevant IllustrisTNG simulation files and the illustris_python package. It was designed primarily for use through the IllustrisTNG JupyterLab service. 02_infall_time_data_tutorial.ipynb A tutorial for loading and analyzing infall_time_data_public.hdf5. The tutorial demonstrates how to: inspect the public HDF5 hierarchy; load physical and projected galaxy catalogues; inspect cluster-selection and infall-assignment statistics; combine TNG300 and TNG-Cluster; construct median infall-time, occupancy, and RMSE maps; examine infall-time distributions within coarse projected phase-space bins; interpolate within a single snapshot; interpolate between snapshots in lookback time; reproduce the $5 \times 5$ median-infall-time and RMSE grids; inspect cluster histories and run consistency checks. Recommended starting point Most users should download: infall_time_data_public.hdf5 02_infall_time_data_tutorial.ipynb In the tutorial configuration cell, update: PUBLIC_FILE = Path("./infall_time_data_public.hdf5") so that it points to the downloaded public HDF5 file, and then run the notebook sequentially. Public-file organization The relevant HDF5 hierarchy is: infall_time_data_public.hdf5 ├── Header ├── TNG300 │ ├── Header │ ├── Snapshot_N │ │ ├── Header │ │ ├── GalaxySample │ │ │ ├── Galaxies │ │ │ └── Projected │ │ └── Properties │ │ ├── ClusterSelection │ │ ├── ClusterHistories │ │ └── InfallStatistics │ └── ... └── TNG-Cluster └── ... GalaxySample/Galaxies contains one row per distinct physical galaxy. GalaxySample/Projected contains up to three rows per physical galaxy, corresponding to projections along the (x), (y), and (z) simulation axes. These projected rows are multiple views of the same object and should not be interpreted as independent physical galaxies. Scientific definitions Cluster selection Clusters are selected at each observation snapshot using a default mass threshold of $$M_{200\mathrm{c}} \geqslant 10^{14}~M_{\odot}.$$ The pipeline additionally applies membership and central-progenitor consistency checks described in the generation notebook and Lawlor-Forsyth (2026). Galaxy selection Satellite galaxies satisfy a default stellar-mass threshold of $$M_{*} \geqslant 10^{9}~M_{\odot},$$ belong to a selected cluster FoF group, and lie within $3R_{200}$ of the cluster centre at the observation snapshot. Infall time At each historical snapshot shared by the galaxy and cluster histories, a galaxy is considered inside the evolving cluster when: its merger-tree host group matches the FoF group of the cluster progenitor; and its three-dimensional distance from the cluster centre is no greater than $3R_{200}$. The infall snapshot is the earliest snapshot satisfying both conditions. The elapsed infall time is $$t_{\mathrm{infall}} = t_{\mathrm{lookback}}(z_{\mathrm{infall}}) - t_{\mathrm{lookback}}(z_{\mathrm{obs}}).$$ Galaxies whose merger trees cannot be loaded, for which no infall event is identified, or which later move back outside $3R_{200}$ are excluded from the final retained sample. Historical coverage Galaxy and cluster histories are followed back to snapshot 21, corresponding approximately to $z=4$. Galaxies that are already inside $3R_{200}$ at snapshot 21 are assigned the elapsed time corresponding to this boundary. Their true first entry may have occurred earlier, so these values should be interpreted as lower limits. Projected phase space Projected phase-space coordinates are calculated along each of the three Cartesian simulation axes: $$R_{2\mathrm{D}}/R_{200}$$ and $$|V_{\mathrm{LoS}}|/\sigma_{\mathrm{LoS}}.$$ The line-of-sight velocity includes both the galaxy–cluster peculiar-velocity difference and the Hubble-flow contribution associated with their physical line-of-sight separation. No projected phase-space binning is imposed in the saved data product. Users may therefore adopt alternative limits, bin widths, occupancy thresholds, interpolation schemes, or statistical models. Software requirements Tutorial notebook The tutorial uses: numpy; h5py; scipy; matplotlib; astropy. It does not require illustris_python or access to the original IllustrisTNG data. Generation notebook The generation pipeline uses: numpy; h5py; astropy; illustris_python. It additionally requires access to the IllustrisTNG snapshot headers, FoF and Subfind group catalogues, and SubLink merger trees. Citation When using these data products or notebooks, please cite both: the associated paper: Lawlor-Forsyth (2026) this Zenodo record: Galaxy Infall Times in Projected Phase Space: TNG300 and TNG-Cluster Data Products (10.5281/zenodo.21496572) Users of the underlying simulations should also cite the relevant IllustrisTNG simulation and data-release publications. License The Jupyter notebooks and associated source code are licensed under the MIT License. The HDF5 data products are licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0). Data license: CC BY 4.0 Notebook/code license: MIT Contact Questions concerning the data products or notebooks may be directed to: Cam Lawlor-Forsyth, [e-mail]