Data Preview 0: The Simulated Data Set from the DESC’s DC2

In 2021, the Rubin Observatory will release Data Preview 0 (DP0), the first of three data previews during the period leading up to the start of Rubin Operations. DP0 will make simulated LSST-like data products available in the Rubin Science Platform (RSP) to Rubin Observatory staff and up to 300 individuals from the science community (“DP0 delegates”).

The dataset adopted for DP0 is the simulated data generated by the Dark Energy Science Collaboration (DESC) for their Data Challenge 2 (DC2). This post provides a brief summary of the DESC paper “The LSST DESC DC2 Simulated Sky Survey”, the data products that will be made available, and the Rubin Science Platform. Future data previews, DP1 and DP2, will be based on commissioning data.

For more information see this topic about community participation in DP0 and consider joining a virtual Info Session about DP0.

The DC2’s Realistic Truth Catalogs – One of the primary goals of the DESC’s DC2 was to provide a testbed from which to understand systematics and validate the DESC’s analysis pipelines. The DC2’s science goals required realistic simulations for many aspects, including:

  • galaxy colors, sizes, shapes, blending, spatial correlations, and photo-z
  • large-scale structure; galaxy cluster colors, luminosities, radial profiles
  • Type Ia supernovae light curves and host galaxies
  • AGN variability and time delays
  • weak lensing effects for galaxies
  • strong lensing effects for SNe Ia, AGN, and galaxies
  • simulated variability for 10% of the stars
    • periodic: RR Lyrae, Cepheid, eclipsing binaries, AM CVn
    • non-periodic: microlensing, MLT-dwarf flares
  • Milky Way dust extinction and reddening effects
  • point-spread-function systematics such as the brighter-fatter effect

What Was Not Included in DC2? – The DC2 simulation does not include any Solar System objects, nor does it include non-Ia types of supernovae or other extragalactic transients such as tidal disruption events or kilonovae. Diffuse features such as tidal streams or intracluster light are also not included.

The DC2 Simulated Survey Data Products – In total, the DC2 wide-fast-deep (WFD) survey simulation includes five years of observations of 300 square degrees of extragalactic regions in six filters (ugrizy). Realistic observing conditions and a baseline survey strategy were applied (for those familiar with the term, OpSim minion_1016 was used).

As described below, the DP0 data will be released in two stages: DP0.1 and DP0.2. For DP0.1, the image processing outputs that will be available will simulate the LSST data release (DR) data products for the wide-fast-deep (WFD) main survey at five years and are referred to as DR6-WFD (for LSST, DR1 is 0.5 years of data, DR2 is the first year, etc.). DR6-WFD includes the raw images, processed visit images (PVIs, where a visit is a single 30 second observation), source catalogs for the PVIs, coadded images for each band, and multi-band object catalogs for the coadds. It is important to note that the processing does not include difference image analysis, and that the catalog photometry is all from the direct images. It is also important to note that the source catalogs for the PVIs are unassociated and forced photometry has not been run on the visit images, which means that there are no time-domain object catalogs or light curves. Furthermore, the AGNs and strongly lensed systems were not part of the DC2 simulations for DR6-WFD (they were part of the deep drilling field, DDF, which was simulated separately from the WFD).

For DP0.2, at a minimum, the same data products as described above will be made available, but in this case processed by the Rubin operations team using an up-to-date version of the LSST Science Pipelines provided by the Rubin Observatory construction team. This might include additional data products that are produced by the pipelines as of Sep 2021. Further updates about additional DC2 data products, such as forced photometry (lightcurves) or, potentially, the one-year outputs (DR2-WFD) or difference image analysis (DIA) and DIA photometry, will be announced at a later date.

DP0 in the Rubin Science Platform (RSP) – The RSP is a set of integrated web applications and services deployed at the Data Access Centers (DACs) through which scientists will access, visualize, query, and analyse the LSST data products. The RSP has three main aspects: the portal aspect offering exploratory interfaces, query forms, and visualization tools; the notebook aspect for creating and executing Jupyter Notebooks in a curated environment; and the application programming interface (API) aspect for remote access. DP0 will use the Notebook and Portal aspects of the RSP.

Example DC2 Analyses – For readers familiar with Jupyter Notebooks, the DESC’s DC2 team has made available a selection of rendered DC2 analysis notebooks in their DC2 analysis GitHub repository. Neither knowledge of python nor experience with Jupyter Notebooks will be required to participate in DP0. Participants might use this time to learn python or how to use notebooks, or might instead focus their activities on the portal aspect of the RSP.

DP0’s Two Stages, DP0.1 and DP0.2 – The DP0.1 data set will contain the DC2 images and catalogs as processed by the DESC using the LSST Science Pipelines, and the DP0.1 data products will be available in the RSP in their original format as processed by the DESC, which is similar to that of planned LSST data products. The DP0.2 data set will contain the DC2 images and catalogs reprocessed by Rubin Observatory using a more recent version of the LSST Science Pipelines, and the data products will be available in the RSP in a format that is consistent with the planned LSST data products. The DP0.1 data products will be available in the RSP on June 30 2021, and the DP0.2 data products by March 31 2022.

There will be a few minor capability differences between the data access mechanisms available in the Rubin Science Platform for DP0.1 and DP0.2, which might affect some delegates’ planned activities. For example, in DP0.1, the notebook aspect of the RSP will offer image access via the Butler (a middleware component for persisting and retrieving image datasets), whereas in DP0.2 image access will be available via Virtual Observatory (VO) services. In DP0.1, the portal aspect of the RSP will only be available for catalogs, whereas in DP0.2 it will also be available for images.

Resources for further information.
“The LSST DESC DC2 Simulated Sky Survey”, arXiv:2010.05926
“CosmoDC2: A Synthetic Sky Catalog for Dark Energy Science with LSST”, Korytov et al. 2019
The Rubin Science Platform Vision Document
Guidelines for Community Participation in DP0 (RTN-004)


Relevant to this topic:

The data products described at the above link will also be released by Rubin Observatory as part of DP0.1. The difference is that the Rubin release will be in the Rubin Science Platform, and will include visit images, coadds, and source catalogs, in addition to the Object table in the above public release.

1 Like