SKA-Mid Simulated HI image cubes

This page provides the simulated SKA-Mid HI image cubes and the corresponding source catalogues

SKA-Mid HI simulated image cubes

These data correspond to a 2000h simulated observation of a 20 square degree area with the full SKA-Mid array between 950 - 1150 MHz frequency range. The dataset has been designed to enable HI source finding and characterization and includes truth catalogues with the HI sources injected in the simulation as well a companion radio continuum image. These data were part of the SKA Science Data Challenge 2. A full description of the dataset is available in Hartley et al. (2023), arXiv:2303.07943

Datasets of increasing size are available, as listed above. All datasets include a truth catalogue and the accompanying radio continuum image

 

File description

File Size

Repository Link

Small development dataset

10 GB

Download

Development dataset

40 GB

Download

Full dataset - HI cube

1 TB

Please request this download by writing to skaoSDCsimulations@skao.int

Full dataset - truth catalogue

100 MB

Download

 

Scoring package

To evaluate one's performance in HI source finding and characterization, the scoring module used for the SKA science data challenge 2 is available as a pip-installable Python package here. Running this package requires the truth catalogues for the datasets of interest (available above). The format of the catalogue needs to strictly follow the formatting rules adopted in SDC2 for the scoring package to work correctly. 

 

Image of the HI cube
The simulated datacube, before noise and instrumental effects are added. Covering a sky area of 20 square degrees and featuring nearly a quarter of a million galaxies, the cube represents an SKA observation of neutral hydrogen — or "HI" — emission.
Last modified on 04 September 2026