Version 2 of the Hubble Source Catalog

Steve Lubow, lubow[at], and the HSC Team

In its 25-year history, the Hubble mission has accumulated over one million exposures that cover tens of thousands of astronomical targets, involving about a dozen instruments, and hundreds of filters. Hubble is not typically used as a survey instrument, however. Targets are covered by up to thousands of overlapping images from several proposals with different science objectives. These images are taken with a variety of detectors, filters, exposure times, and orientations. The goal of the Hubble Source Catalog (HSC) is to provide a comprehensive single point of access for Hubble sources. Due to the complexity of the target coverage, creating such a catalog is a considerable challenge. But given the sensitivity of Hubble, there is great potential for scientific discovery through such a catalog.

The HSC describes the properties of astronomical objects that result from cross-matching sources across Hubble images. The catalog provides temporal, spectral, and morphological information about objects, as well as information about undetected objects. The HSC currently contains sources detected in three of the main imaging instruments: the Advanced Camera for Surveys (ACS) for the Wide Field Camera detector, Wide Field Planetary Camera 2 (WFPC2), and Wide Field Camera 3 (WFC3) for both the infrared (IR) and ultraviolet/visible (UVIS) detectors. However, due to image quality problems and other issues, about 5% of these images are not included.

There are several user interfaces to access the HSC: a graphical user interface called the Discovery Portal, a MAST forms interface, the Hubble Legacy Archive image viewer that provides HSC source positions overlaid on the Hubble images, and an SQL query interface through CASJobs.

A major benefit of the HSC is that it provides high-quality relative astrometry between Hubble images, typically at better than 10 mas accuracy. The relative astrometric accuracy is considerably better than what is currently available in the Hubble data headers (see Fig. 1). The image and source positions provided by the HSC user interfaces are reported in the improved coordinates. The World Coordinate System parameters for images are available through the HSC CASJobs interface. We hope to incorporate that information in the Hubble Legacy Archive image headers in the near future. The improved astrometry can be used to overlay images and generate mosaics.

Figure 1: Distribution of relative astrometric errors in milliarcseconds for sources in Hubble images before and after the Budavari & Lubow (2012) astrometric correction algorithm is applied. The original distribution corresponds to the current Hubble astrometry in data headers. The areas under the two curves in the top panel are the same when extended to sufficiently large offsets. This illustrates that the median astrometric errors are greatly reduced with the HSC-corrected astrometry. The bottom panel is plotted with a logarithmic vertical scale that shows the original Hubble astrometry error distribution clearly has an extended tail.

The HSC began as a NASA Advanced Information Systems Research Program funded grant to the Johns Hopkins University and the Institute that resulted in the first Beta version (Budavari & Lubow 2012). The project, now supported through the Institute, produced Version 1 in early 2015 (Whitmore et al. 2016).

Version 2 of the Hubble Source Catalog was released in September 2016. The main new capabilities of this version are as follows:

  • Addition of approximately four more years' worth of ACS data and one more year of WFC3 data (i.e., using data public as of June 2015). The ACS source detections extend about one magnitude deeper than in Version 1. As a result, the HSC now has 3× as many sources as in Version 1. It now contains about 92 million objects with over 300 million source detections.
  • Spectroscopic cross matching between the Cosmic Origins Spectrograph, Faint Object Spectrograph, Goddard High Resolution Spectrograph, ACS grism data, and HSC sources.
  • An improved determination of the concentration index (CI) through a filter-based normalization.
  • Availability of magnitude values ("magauto") that are accurate for galaxies through the MAST Discovery Portal user interface. The maximum number of sources displayed by the Discovery Portal has increased from 10,000 to 50,000.
  • Gaia source positions overlaid and compared with HSC source positions in the Hubble Legacy Archive image viewer.

The HSC has great potential for discovery. Of the 92 million objects in the HSC, over 10 million objects have time coverage of more than 1 year. Over 1 million objects are detected more than 25 times (see Fig. 2). Over 1 million objects are detected in 5 or more filters.

Figure 2: Distribution of the number of source detections for each HSC object (match).

Care in using the catalog is required due to some artifacts, such as the presence of spurious sources due to diffraction spikes in images (see Fig. 3). Some of these artifacts can be removed by adjusting search parameters in ways that we describe in the HSC documentation. We plan to correct some of the issues in future versions of the HSC.

Figure 3: The HSC sources in M4 that lie in an arc around the brightest saturated source are due to diffraction spikes in several contributing images that have different orientations. Note the spurious detection just above the small red cross in the center of the field. In this field, the spurious sources can be filtered out by placing limits on the CI of the sources.

The scientific exploration of the catalog is just getting started. The HSC documentation provides several use cases that cover a range of scientific objectives. The example in Figure 4 illustrates the potential of the catalog. It would take a considerable amount of time and effort for a user to carry out this analysis by starting with a set of Hubble images. It can be carried out in the HSC in just minutes.

Figure 4: White dwarfs (WD) in cluster M4 are blueward of the main band of stars. The left panel is a color magnitude diagram constructed from the HSC using the sources plotted in the right panel. Of the eight objects in the encircled WD area, seven are in Richer et al. (1997). One object is not in the WFPC2 field of view (green) used by Richer et al., and hence is a new white dwarf candidate.

Spectroscopic cross-matching to HSC sources is a new feature of Version 2. You can view the Hubble spectrum and corresponding image cutout in the Discover Portal (see Fig. 5).

Figure 5: Results of a spectroscopic search in the Discovery Portal that provides the spectrum and corresponding image in this case for an elliptical galaxy.

Due to its broad time coverage, the HSC provides the potential for the discovery of time variable objects. The Hubble Catalog of Variables project, based in Athens, Greece, is pursuing that objective.

The HSC will continue to improve the quality and depth of the catalog, as well as providing new features and data products, and will benefit from the advent of Panoramic Survey Telescope and Rapid Response System (Pan-STARRS), Gaia, and the Large Synoptic Survey Telescope in this era of digital astronomy. We hope you will explore the HSC and provide us with feedback.


Budavari, T., & Lubow, S. H. 2012, ApJ, 761, 188

Richer, H. B., et al. 1997, ApJ, 484, 741

Whitmore, B. C., et al. 2016, AJ, 151, 134