Automatic Recalibration of HST Archival DataA new paradigm for archives
|
|
Dennis Crabree *, Daniel Durand *, Norman Hill *,
Séverin Gaudet *, Benoît Pirenne **
|
*) | Canadian Astronomy Data Centre, Dominion Astrophysical Observatory,
Herzberg Institute of Astrophysics, National Research Council of
Canada, Victoria, BC
| |
**) | Space Telescope -- European Coordinating Facility,
European Southern Observatory, Garching, Germany
|
|
|
...
|
|
|
Abstract: The CADC and ST-ECF have implemented a pipeline which automatically
recalibrates data as it is requested from the archive. This process uses
the latest recalibration software and the recommended calibration reference
files, which may be different than the calibration files used in the original
calibration. The service is now available through the World-Wide Web.
|
Background Information
HST science data is automatically calibrated when it is received at
the STScI and this calibrated data is included in the archive. The
calibration software, which is contained in the IRAF/STSDAS hst_calib
package, takes as input the raw data and any necessary calibration
reference images or tables if they are already available. The software
determines which calibration steps to perform by checking the values
of the calibration switches in the header. It selects which reference
files to use in the calibration by examining the reference file
keywords. The values of these switches and keywords depend upon the
exact configuration of the instrument, the date of the observation and
any other constraints. The values are set in the headers of the raw
data in the RSDP (Routine Science Data Pipeline).
Until now, when users requested calibrated data from the HST archive,
they received the data produced by the RSDP pipeline. However, with
instrumental properties changing with time, better approaches for
calibration of some instruments have been introduced and there have
been other general improvements to the calibration of HST data. So
what is a user to do? Fortunately, the same, or improved, software
which runs in the calibration pipeline at STScI is also available in
the released version of IRAF/STSDAS. One can recalibrate data from the
archive by starting with the raw data, editing the appropriate header
keywords to reflect the new calibration files and running the
appropriate software. The Institute maintains a database which
contains the recommended calibration reference files for each
observation. However, this is not the most convenient approach for
users, and this led us to develop an automatic recalibration process
for HST data which essentially duplicates what a user would do
manually.
Automatic Recalibration of HST Data
In order to offer automatic, quasi on-line, recalibration of HST data
several steps need to be in place:
- The raw science data needs to be on-line
- The needed calibration reference files and tables need to be on-line
- A pipeline to recalibrate the data needs to be in place
Unlike STScI, we do not have the resources to store the large 12"
optical disks on-line in jukeboxes. The approach we have taken for this
recalibration effort is to store the compressed raw data for all science
observations on-line on CD-ROM. Using standard "gzip"
compression, the raw
data for a typical WFPC2 observation occupies less than 1 MB. The raw data
for all public HST data currently fit on less than 50 CDs which are
stored on-line in a Pioneer 500 CD jukebox. Data is copied to CD-ROM
as it becomes public at the CADC and a second disk is made for the
ST-ECF.
In order for the recalibration to work properly, all of the necessary
calibration files and tables need to be on-line. The CADC maintains a
directory structure containing all of the necessary files by querying
the 'bestref' tables each night to identify the distinct reference files
which are needed, comparing this list to the ones currently on-line and
then reporting which files need to be installed. Currently, there are
approximately 9 GB of reference files on-line. Often a reference file is
not available at the CADC/ST-ECF, i.e. it has not yet been distributed, and
must be retrieved from STScI using Starview. A process at the ST-ECF
copies the reference files regularly over the network to ensure that
both sites have all of the needed reference files on-line.
While there are tasks within IRAF/STSDAS tasks for changing the header
keywords associated with recalibration, they must be run manually. We
developed a small IRAF/STSDAS pipeline to automate this process and to
perform the complete 'recalibration pipeline'. One of the key steps in
this process is a Perl script which queries the database to identify
the recommended calibration reference files and tables. The output
from this Perl script is a small IRAF script which updates the
necessary IRAF parameters needed to update the raw data headers.
The process which occurs when a request is received is the following:
- The raw data is copied from CD-ROM to a staging area and
decompressed
- An IRAF process is started which runs the recalibration pipeline.
This pipeline does the following:
- reads the data into IRAF/STSDAS (converts from FITS to
internal format)
- runs the Perl script to find out the recommended reference files
- updates the raw headers to represent the new calibration information
- runs the appropriate calibration pipeline and writes the results
back as FITS files
- Copies the recalibrated data to the anonymous ftp area for retrieval
Our automatic process duplicates what a normal user would do manually. That
is, get the raw data, update the header to reflect the current information on
which reference files to use and run the standard calibration pipeline.
Future developments
There are several ways by which the system described here
can be further improved. Some of the possibilities being contemplated
at the moment include:
- inclusion of the pipelines for the new instruments (STIS and
NICMOS);
- addition of extra calibration steps for the WFPC2 (like cosmic
ray removal, co-addition of frames). This however, entails the
need to create the currently unavailable notion of "products" or
groups of observations;
- addition of more 'experimental' calibration steps with e.g.,
improved background treatment (FOS);
- the instrument physical models now being developed at
the ST-ECF for FOS and STIS.
- interactive recalibration for expert users where processing
steps can be turned off or on, and specific calibration files
used.
Other archives have also showed interest in this new archive paradigm
and one could imagine the final archive of IUE being operated in this
manner with the addition of a physical model of its instrument. The
ESO VLT archive will also consider this approach for its own data
archive.
Access to Recalibrated Data
Automatic re-calibration does not necessarily guarantee better results
than the original pipeline calibration. We expect better results in
almost all cases. It is always advisable, however, to carefully
examine the results and, when in doubt, to compare both the old and
new calibration results. We will do our best to keep checking the
re-calibration results ourselves but it would understandably be
difficult to make an exhaustive examination of all results.
Access to recalibrated HST data is offered through the CADC and ST-ECF Web
Interface to the Hubble archive. Typical processing time for a WFPC2 dataset
is 3 minutes while for the FOS and HRS this time drops to approximately 1
minute.
The relevant URLs are:
http://cadcwww.dao.nrc.ca/hst
and
http://archive.eso.org/wdb/wdb/hst/science/form.