C.2 Hubble Spectroscopic Legacy Archive (HSLA)

The Hubble Spectroscopic Legacy Archive (HSLA) identifies individual targets across all COS and STIS science spectra, and for each target provides a classification, a set of spectral data products and a human-readable metadata file. The spectra consist of science quality coadds, which use the same code as HASP for combining individual spectra for a given instrument and grating (see section C.1.2), and an abutted “quick-look” spectrum that spans the full wavelength range observed by combining data from both instruments and multiple gratings. In the regions where there are wavelength overlaps, a prioritized list of gratings is used to determine which of them are included in the abutment. More details are available at the HSLA website and in COS ISR 2025-18.

C.2.1 Target Association

The HST archive contains over 60,000 COS and STIS exposures distributed over a few thousand observing programs.  The task of target association was to (i) identify all the unique sources observed, (ii) determine the complete set of calibrated exposures associated with each source, and (iii) label each source with a unique name and provide its best possible coordinates.

Targets were associated based on an iterative cross-matching of coordinates. All observations lying within 2.0 arcsec (corrected for proper motions to a common epoch) were considered part of the same association.  The names were taken either from SIMBAD, or NED if that wasn’t available, or if absent from both catalogs from the Phase II proposals. For each target, an HSLA Catalog name was constructed by modifying the target name and appending a unique numerical identifier.

Once defined, target names (including their numeric identifiers) remain static in the HSLA. New data on an already observed target will be identified by the coordinate match and added to the association, while previously unobserved sources will be assigned new names and HSLA Catalog entries.

C.2.2 Target Classification

A key component of HSLA is the creation of a three-tier classification scheme. The vocabulary and hierarchical structure of the classification were determined from the most relevant descriptions of the targets in SIMBAD, NED and the Phase II proposals. At the Tier 1 level there are three classes: Galaxy, ISM and Star, which have 13, 11 and 20 classes at the Tier 2 level, respectively. The Tier 3 list of classes is extensive, and each of them is mapped to an entry in the Unified Astronomy Thesaurus. The entire list is available on the HSLA Classifications website, which also provides a link to the full schema as a JSON file.

The target classifications are assigned automatically to existing and new data. An important consideration was to make the scheme maximally useful for archival research. The archive can be searched for HSLA data using any of the classification terms, so, for example, one may retrieve a list of all the Galaxy observations or the subset of Dwarf Galaxies from a Tier 2 search, or a Dwarf Spheroidal Galaxy using a Tier 3 term. It is to be noted that the scheme was designed to be sufficiently broad that not all of the Tier 2 classes are populated with COS or STIS observations at the current time.

C.2.3 Data Products

HSLA produces a human-readable metadata file for each target.  The naming convention is:

hst_<HSLA_CAT>_metadata.txt

where HSLA_CAT is the target catalog name, which consists of a standard designation followed by a unique numerical identifier, separated by a double-dash.  For example, the CALSPEC standard white dwarf star GD71 has the HSLA catalog name, ‘gd71—0002’, and Supernova 1987A is ‘sn1987a—0713’.  The catalog names are included in the metadata files.

Target-level coadded spectra are produced for each instrument and grating, while cenwaves are combined. In the case of COS, the coadds are also separated by lifetime position (LP) to preserve the spectral resolutions, which vary significantly across LPs.

The naming convention is as follows for COS and STIS coadded spectra:

hst_cos_<HSLA_CAT>_<GRATING>-<LP>_cspec.fits
hst_stis_<HSLA_CAT>_<GRATING>_cspec.fits

and 
hst_<HSLA_CAT>_aspec.fits 

for the abutted spectrum.

In addition to these, HSLA produces preview plots (in PNG format) and trailer files for each coadded spectrum.

C.2.4 Data Access

The HSLA data may be accessed in several ways, each of which has its particular advantages.  In all cases the three-tier classification terms can be used for the searches.

The HST search form in MAST provides access to HSLA data using a form that will be familiar to many COS and STIS users. It is the best option when the needed targets or the program IDs are known.

Astroquery is a streamlined python interface and can be used for generalized searches and bulk downloads. The “Introduction to the HSLA Data Products and Tools” jupyter notebook provides an example query.

The MAST Portal will also be familiar to many users. The default query returns data from multiple telescopes and instruments. In order to retrieve only HSLA data the “Advanced Search” option may be used with “HSLA” as the Project.

The “Tier Query Links” page provides shortcuts to MAST Portal searches that return all the HSLA datasets for a given Tier 1 or Tier 2 classification.