B. Stroozas, K. McDonald, B. Antia, J. McDonald, A. Wiercigroch
                       Center for EUV Astrophysics,
               2150 Kittredge St., University of California,
                     Berkeley, California 94720, USA


   Initial results for bright extreme ultraviolet sources discovered during
the EUVE all-sky (50-740 A) and deep ecliptic (65-360 A) surveys have been
published as a Bright Source List (BSL) and released to the astronomical
community with the recent NASA research announcement (NRA 93-OSS-02, Appendix
F).  This paper describes the data processing software, the EUVE survey data
set, and the production of the BSL at the Center for EUV Astrophysics.  The
contents, format, and selection criteria for sources, the data processing
strategy, some problems encountered, and a summary of the BSL results are


   The Bright Source List (BSL) is a detailed list of bright extreme ultra-
violet (EUV) sources detected during the survey phase of the EUVE mission
(calibration targets are also included).  Two distinct surveys, the all-sky
and deep surveys, were conducted by the four EUVE telescopes during the first
six months of the mission.
   The all-sky survey was conducted by three scanning telescopes in four dis-
tinct filters (Lexan/B, Al/Ti/C, Ti/Sb/Ti/Al or "Dagwood", and Sn/SiO) covering
the wavelength bandpass 50-740 A.  In a period of six months, these telescopes
mapped the entire sky with exposures ranging from ~400 seconds along the eclip-
tic equator to ~20000 seconds at the ecliptic poles.  Concurrently, the Deep
Survey telescope (DS) conducted the "deep" survey (more sensitive by a factor
of ten) of a  band along the ecliptic in two filters (Lexan/B and Al/C) cover-
ing the bandpass 65-360 A.
   The EUVE Pipeline team was mainly responsible for developing the BSL.  This
team consists of five staff research associates who are responsible for proces-
sing the EUVE telemetry through the pipeline subset of the End-to-End System
(EES) software that was developed at the Center for EUV Astrophysics (CEA).
   The BSL is available to the astronomical community from the EUVE Science
Archive.  It can be obtained via anonymous ftp from
or by sending electronic mail to


   The processing of the EUVE telemetry uses the EES Pipeline software, which
consists of two parts:  the ESOC (EUVE Science Operations Center) and SDAF
(Science Data Analysis Facility) Pipelines [1].  The ESOC Pipeline is the
front-end of the system and is run on every telemetry file (~50 per day) re-
ceived from EUVE.  It separates the engineering information from the photon
information and processes each in different ways.  With the engineering infor-
mation, it plots stripcharts of engineering monitors and satellite aspect, re-
ports on detector characteristics and prints logs of command echoes and engi-
neering limit transitions.  Using the photon information, it generates preli-
minary science products for further processing.  These preliminary science
products are of two types:  "pigeonhole" and skymap files.  A pigeonhole is a
file that contains photon information from a circular area of the sky (~10 arc
minute radius) centered on a position where an EUV source is expected to be
seen (e.g. a known bright white dwarf or late-type star).  A master pigeonhole
catalog of such potential EUV sources (~6000 entries) was prepared before
launch and was used in the processing.  The skymap files are tiny maps of the
sky for each telemetry file.
   An entire day's worth of these preliminary products are then merged together
and passed on to the SDAF Pipeline for the back-end processing.  The SDAF
Pipeline is used daily to append these merged products to the existing pro-
cessed data pool.  Periodically, the source detection and flux calculation
routines are run to search for sources and report on their intensities.  The
end products are in the form of complete skymaps, pigeonholes for master cata-
log sources, and a source detection log.  When potential new sources (not in
the master catalog) are detected, the raw telemetry is reprocessed through the
entire system to build pigeonholes for them.  This is a fairly complicated and
time-consuming process through which all the telemetry must flow.
   The source detections were derived from the skymaps, while the pigeonholes,
which provided more detailed information, were used in the source verification.
Because of the complicated nature of the DS instrumental point-spread function,
the skymaps alone were used in the deep survey source verifications (provided
by Dr. J. Vallerga and B. Antia).


   Although the EES Pipeline software was fully tested before launch via mul-
tiple simulations, some software problems were encountered in orbit.  There
was a small timing error, some UV leaks when bright B stars passed over tiny
pinholes in the filters and a problem with the source detection software (see
section 4.3).  The software was patched to solve or compensate for these prob-
lems, yielding the first scientifically validated version of the EES Pipeline,
which was released on 27 August 1992, and used to process the entire survey
data set.
   As the processing progressed, the survey data set was divided into three
distinct batches:

 o 22 July (beginning of survey) to 28 August 1992
 o 29 August to 12 October 1992
 o 13 October 1992 to 10 January 1993 (approximate end of survey)

   The processed survey data set has some gaps because of calibrations and
processing problems (see section 4.3).  Due to time constraints, these gaps
were not reprocessed for inclusion in the BSL but are presently being filled.


   The Pipeline team planned the overall strategy, identifying and scheduling
necessary tasks.  The major questions to be addressed were:

 (1) How and in what format would the BSL be presented and what criteria would
	be used to select the sources contained therein?
 (2) What would be the data processing strategy for production of the BSL?
 (3) What would be the schedule for producing and releasing the BSL?

4.1 Content, Format, and Selection Criteria

   In answer to question (1), updated and revised internal versions of the BSL
were issued to allow time for the CEA scientists to review the sources before
the external BSL publication with the NASA research announcement (NRA).  The
actual source list contains the following information:
 o Name of the EUV source.  The adopted format was EUVE Jhhmm+/-dd.d (according
	to the established convention of the International Astronomical Union),
	which designated an EUVE source located at the approximate position of
	"hh" and "mm", hours and minutes of Right Ascension, and "dd.d", deci-
	mal degrees of Declination, in Epoch 2000 coordinates.
 o Possible optical counterpart identification as taken from comparison with
	astronomical catalogs, including the SIMBAD and NED databases.
 o Spectral type of the possible optical counterpart (if known).
 o Source RA/DEC coordinate (Epoch 2000).  These positions are generally accu-
	rate to one arc minute or less.
 o Count rate per filter in the scanners and/or DS.
 o ROSAT Wide Field Camera Bright Source Catalog designation (when applicable).
 o Notes and comments regarding specific sources.  General comment types in-
	clude such things as in which survey(s) the source was detected and
	which sources were pointed (calibration) observations.

   A major effort has been made to provide a BSL with no spurious source.  All
included sources simultaneously satisfied the following criteria:

 o The detection significance of the source was S>=6*sigma in any single filter.
 o The visual examination of the pigeonhole data, convolved with the telescope
	PSF, confirmed the existence of the source and the image shape was not
 o No ultraviolet (UV) pinhole leak (see below) was revealed by a careful anal-
	ysis of the scanning of the source in detector coordinates.

   In addition,

 o Sources with a detection significance 3*sigma=4*sigma were included.

   Two filter features were accounted for during BSL production.  The first was
tiny pinholes in certain filters which allowed non-EUV flux to be transmitted
when a bright source scanned directly across the pinhole.  These pinhole-leak
sources were easily weeded out of the BSL in the visual verification process.
The second problem was a known UV leak in Lexan/B, which was inherent in the
filter design.  A pure Lexan filter transmits ("leaks") UV flux near 1600 A
and longward of 2200 A.  To attenuate this UV flux, boron was added to the EUVE
Lexan filters.  At some point, too much boron begins to attenuate the EUV flux,
so the EUVE Lexan/B filters were designed such that a UV leak would occur only
for bright B stars (m_V<5), of which ~100 were expected to be observed during
the surveys.  A calibration of these bright B stars is now in progress to allow
for subtraction of the UV flux contribution.
   In order to be of maximum use to prospective EUVE guest observers (GOs),
the BSL has been made as error-free and complete as possible for the sources
included.  In this conservative process, a number of real sources may have
been omitted (specifically, some 200 sources with significance between
3*sigma  S  6*sigma, with no clear counterpart and with no subsidiary studies
to confirm the reality of the source, have been excluded from the BSL).

4.2 Data Processing Strategy

   In answer to question 2 posed at the beginning of this section, after deci-
ding to focus on completeness instead of sheer number of sources, we decided
that the strategy for BSL production was as follows:

 (1) The BSL was released internally to CEA in updated and revised versions.
	This would allow for more internal review and comments from the CEA
	scientists before final publication of the BSL with the NRA.
 (2) Initial focus was on the master pigeonhole catalog sources.
 (3) Final focus was on the newly detected sources.

   The work to be done was identified, broken down into detailed tasks and
scheduled and tracked for progress.  This project was completed ahead of
schedule and internal versions of the BSL were more complete than originally
planned.  The internal release schedule was as follows:

 o Version 1 was released internally to CEA on 1 March 1993 and contained a
	limited amount of information on all master pigeonhole catalog sources
	and new sources processed through 12 October 1992.
 o Version 2 was released internally on 1 April 1993 and contained preliminary
	complete information for all survey BSL sources.
 o Version 2a was released on 9 April 1993 for distribution on the initial EUVE
	CD-ROM.  This release included only a sample subset of the version 2
	BSL, with limited information on each source.
 o Version 3 was released internally on 23 April 1993 and contained the best-
	effort complete information on all survey BSL sources.  This included
	improved position and flux information and additional follow-ups on
	unidentified sources.  Internal release on this date allowed the CEA
	scientists sufficient time to review the BSL before NRA publication.
 o Due to a delay at NASA in released the EUVE NRA, Version 4 was released on
	6 June 1993 in a CEA preprint at the 182nd meeting of the American
	Astronomical Society in Berkeley, CA.  This version contained complete
	information on 341 EUV sources.
 o Version 5 was completed on 21 June 1993 and released as an appendix to the
	second EUVE NRA.  This version supplemented the version 4 results with
	15 additional objects, yielding a total of 356 EUV sources.
 o A version 6 is in the planning stages for publication in a scientific jour-
	nal.  This version will contain all the in-depth science analysis for
	the sources in the list.

4.3 Problems

   Many software problems were encountered.  Although most were minor, there
were major problems with the engineering history database system, a 512 milli-
second timing error and the UV leaks through the tiny filter pinholes.  The
minor problems caused delays in the data processing.  The major ones were se-
rious enough to warrant starting the entire survey processing anew.
   Also, there were some gaps in the survey data set.  Although the majority of
these gaps were  due to calibration pointings (~29 days, which will be filled
in during the first six months of the GO phase), other days were missing from
the science products.  Four different days of data (two in October and one each
in November and December) were missing owing to corrupted merged products from
the ESOC Pipeline.  Two additional days were missing owing to a loss of aspect
knowledge affecting four orbits' worth of data near the change of GMT day on
17/18 September 1992.  The nature of the problems with all six of these days
required that each day be reprocessed in full and this was not feasible due to
time constraints.  Finally, the last three days of the survey were not included
because a new version of the EES processing software was installed.
   There was also a problem with spurious source detections.  In areas of the
sky that were not fully scanned, had low exposure or were contaminated by poin-
ted data (from calibrations), the source detection software found hundreds of
spurious sources, which required visual review to weed them out.
   Another major set of problems were computer-related, causing delays in the
data processing.  The sophisticated computer hardware used and the great de-
mands from the many data processing and analysis efforts were very taxing to
the system and led to failures and delays in processing.
   The problems mentioned above should not indicate to the reader a lack of
progress.  Initial processing for the survey skymaps was completed within one
month after the end of the survey and have been completely rebuilt since.  The
Pipeline team was also able to complete the BSL ahead of schedule in order to
meet the NRA release date.


   The detailed BSL results were published with the NRA and contained a total
of 356 sources, 39 of which were detected in the deep survey.  Of these, 263
were also in the ROSAT WFC Bright Source Catalog.  Only pointlike sources have
been included in the BSL; extended sources were omitted (although the Moon and
the Vela and Cygnus supernova remnants were seen in the skymaps).  A breakdown
of the BSL sources by classification (based on possible counterpart informa-
tion) and filter are shown in Tables 1 and 2, respectively.
   It should be noted that this class breakdown of possible optical counterpart
identifications is based on existing catalogs, which often are biased towards
or against specific classes of objects.  The distribution of the Table 1 sour-
ces is displayed in fig. 1.  An example exposure map for the Lexan/B and
Al/Ti/C filters are shown in fig. 2.

	     TABLE 1:  Source Breakdown by Classification 
	|  Number of Sources        Classification              |
	|        173              Late-type stars (F,G,K,M)     |
	|        113              White dwarfs                  |
	|         23              Unidentified (non-ROSAT WFC)  |
	|         14              Cataclysmic variables         |
	|         10              other galactic                |
	|          7              B stars                       |
	|          6              A stars                       |
	|          6              Active galactic nuclei        |
	|          4              BL Lac objects                |
	|        356              Total BSL Sources             |

	        TABLE 2:  Source Breakdown by Filter  
	|                  Detections per Filter                 |
	|  Lexan/B   Al/Ti/C   Dag   Tin   DS Lexan/B   DS Al/C  |
	|    327        92      17    11        39           6   |
	|              Detections in Multiple Filters            |
	|         One   Two   Three   Four   Five   Six          |
	|         251    85     10      9      1     0           |


1. B. Antia, JBIS, this issue (1993).


Figure 1:  Distribution of the 356 BSL sources in celestial (RA/DEC) coordinates.
	The center of the map represents zero hours RA and zero degrees DEC.
	RA increases toward the left edge of the map (+12h), wraps to the right
	edge and continues back toward the center.  DEC is mapped vertically
	(+90 deg at top, -90 deg at bottom).

Figure 2:  Example exposure contour maps for the EUVE all-sky survey Lexan/B
	and Al/Ti/C filters.  The exposure time for the regions of the sky
	that were mapped is represented by several contour levels.  The maps
	show areas of the sky (within the contours for 100 sec exposure) where
	no data have yet been obtained; these gaps are currently being re-ob-
	served, but the data from these areas were not available for the BSL
Page created by
Last modified 10/14/98