) with a cut of 6% of the total
exposure time. For further details see Bocchino, Barbera, &
Sciortino (1995).
Next: Modelling the 2-D Response of Cameras and Spectrographs with Synphot
Previous: New AIPS Task FRMAP for Mapping the Spectral Line Radio Sources
Table of Contents --- Search ---
PS reprint
G. Mackie, G. Fabbiano, F. R. Harnden, Jr., D.-W. Kim
Smithsonian Astrophysical Observatory, 60 Garden St, Cambridge, MA 02138, U.S.A.
M. Barbera, F. Bocchino, F. Damiani, A. Maggio, G. Micela, S. Sciortino
Istituto e Osservatorio Astronomico ``G.S. Vaiana'', Palermo, Italy
P. Ciliegi
Osservatorio Astronomico di Brera, Milano, Italy
A systematic analysis of astronomical archive data can reveal important information on previously known objects as well as serendipitous discoveries. In the following we describe an automated pipeline that operates on ROSAT X-ray PSPC data, using new algorithms designed for PSPC data. A brief example of database implementation is given.
Summary: The pipeline consists primarily of IRAF Command Language (CL) scripts, organized into modules, that comprise related science tasks. Input data is taken from the US ROSAT Public Data Archive, comprising 1) photon events list, 2) exposure map, 3) event rates file, and 4) corrected aspect file for each pointed observation sequence. The pipeline has run on Suns (Sun OS and Solaris) at SAO and on DEC Ultrix at Palermo.
ROSAT PSPC observations are characterized by intervals with
significant background count-rate enhancements mainly due to the
particle or solar scattered radiation contaminations. These enhancements are not screened out by the standard
processing software (SASS). We have developed an algorithm for automated
time screening in order to reduce the effect of such contaminations on
the measured signal-to-noise ratio (SNR) of pointed sources in the field
of view.
Our approach differs from previous recipes suggested by Snowden et al.
(1992, 1993) because it is model independent. We do not attempt to
model the physical origin of the contribution, but we try to eliminate
them by performing a temporal cut which maximizes the efficiency
function, defined as the ratio of a faint source SNR measured in the
screened and the unscreened observation. The efficiency function
depends on the amount of the screened out time and it represents
the trade-off between time rejection for background enhancement
elimination and time retaining to preserve the counting statistics.
The algorithm has been tested on real and simulated data in order to
find the parameters that optimize its functionality. In the case of a
crowded PSPC observation pointed to the Pleiades open cluster, a typical
SNR gain of 12 % with respect of the unscreened case is obtained for
faint sources () with a cut of 6% of the total
exposure time. For further details see Bocchino, Barbera, &
Sciortino (1995).
A wavelet source detection algorithm specially designed for ROSAT use is used in the pipeline and is described fully by Damiani et al. 1996 (this conference).
Sources are flagged as extended if
1) the wavelet detection extension
indicator (ratio of scale of maximum SNR to 
(PSF) x correction
factor) is >1.3, and 2) the ratio of the azimuthally averaged surface
brightness profile data points to normalized PSF profile data points
is >3. Individual surface brightness profiles, contour plots,
STScI Digitized Sky Survey subimages, broad
(0.2--2.0 keV), hard (0.4--2.0 keV) and soft (0.2--0.4 keV) subimages,
and exposure maps are then produced for these flagged sources.
Spectral extraction is performed on sources that possess net counts greater than or equal to a minimum total count value (typically set to 200--300 counts). Color and hardness ratios are derived from the extracted 34 channel spectra. Preliminary spectral modelling inside the pipeline via XSPEC is also available, assuming single component (Raymond-Smith thermal plasma, thermal bremsstrahlung or power law) models.
For each data sequence a source list of known (catalogued) sources occurring in the PSPC field is derived (hereafter Minicat). This is done via SYBASE (SAO) or INGRES (Palermo) scripts that search a Master Catalogue in a 1 degree radius around the central pointing position. The Master Catalogue is derived from 43 separate stellar and extragalactic optical, radio, infrared and X-ray catalogues. The Minicat is then used as a source list to perform flux and upper limit measurements. During each pipeline sequence run a source id list is created by a matching process performed with IRAF table tasks. This list is derived by comparing PSPC source positions with source positions from the Minicat. The source-id matching criteria takes into account the varying width of the PSF of the PSPC with offaxis distance.
The pipeline
produces a variety of data products including reduced
broad, hard and soft images (corrected for data screening and exposure
variation), the screened exposure map, full field contour maps with
detected sources and Minicat sources labelled, full field STScI CD-ROM
DSS images, as well as values of the global background and 
for broad,
hard and soft images. Data products from extended sources are
mentioned above.
Database input files are also produced by the pipeline.
These files contain source information such as detected counts,
extraction radii, extension indicators, and flux values, as well as
sequence information such as total integration time, screeened time
and date of observation.
The pipeline is now operational at both SAO and Palermo. Initial results from pipeline operation at SAO were presented by Mackie et al. (1995) which concentrated on analysis and results of early-type galaxies derived from 35 PSPC sequences. For the complete archive set we plan to run the pipeline simultaneously at SAO and Palermo on different data sets. The output data files will be archived to DAT tape, and the pipeline database output files will be combined to produce separate databases at both sites. Current elapsed times per sequence at SAO on a Sun 1000 server are 2.5--3.5 hrs, and we hope to reduce this to ~2hrs per sequence. We anticipate that full pipeline production will begin shortly at both sites.
The pipeline produces 22 files that are formatted for database ingestion via a bulk copy mode. Procedures have been written to extract information from these database files. For example, the SYBASE procedure below, p-ngcic-flux-detect-early, is used to derive a list of X-ray fluxes for NGC and IC galaxies that have been detected and are early-type galaxies.
The SYBASE procedure p-ngcic-flux-detect-early is given in type, with
descriptive comments in italics:
create procedure ngcic-flux-detect-early (set procedure name)
as
select wmid.id, detect-flux.flux-ray, ngcic-space.b-mag,
ngcic-space.vel, ngcic-space.T-type (search these database file.columns)
into out-ngcic-flux-detect-early (the output file)
from ngcic-space, wmid, detect-flux (from these database files)
where detect-flux.ror=wmid.ror (condition 1: same sequence number)
and detect-flux.wsnum=wmid.wsnum (condition 2: same source number)
and wmid.id=ngcic-space.name (condition 3: same source id.)
and ngcic-space.T-type < 1 (condition 4: galaxies are early-type)
go (do the procedure creation)
execute ngcic-flux-detect-early (execute procedure statement)
go (do the execution)
and a sample of the output (in database file
out-ngcic-flux-detect-early) is:
This work has been supported by NASA LTSA NAGW-2681. We wish to acknowledge K. Rhode and D. Alexander who made significant contributions to pipeline software development.
Damiani, et al. 1996, this volume
Mackie, et al. 1995, in Proceedings of Röntgenstrahlung from the Universe, Wurzburg, Germany
Rhode, K., Fabbiano, G., & Mackie, G. 1994, in Astronomical Data Analysis Software and Systems III, A.S.P. Conf. Ser., Vol. 61, eds. Dennis R. Crabtree, R. J. Hanisch & Jeannette Barnes (San Fransicso: ASP), 371
Snowden, S. L., Plucinsky, P. P., Briel, U., Hasinger, G., & Pfeffermann, E. 1992, ApJ, 393, 819
Snowden, S. L., & Freiberg, M. J. 1993, ApJ, 404, 403