1.     Pipeline Overview..

2.     Pipeline Products.

3.     Spectral Data Product Files.

4.     Source Extraction.

5.     Primary Object Catalog - Merged FUV+NUV Catalog (mcat file)

6.     Pipeline Products Gallery.

7.     GALEX Image Artifacts.

 

1.    Pipeline Overview

 

The GALEX data pipeline converts GALEX satellite telemetry data and any necessary corollary data into calibrated images and catalogs. The GALEX Science Operations Center (SOC) receives data from the satellite and ingestpipe unpacks it into time-tagged photon lists, instrument/SC housekeeping and satellite aspect information.  From these data sets, orbpipe generates images, spectra and source catalogs. 

 

Astrometric modules rta (for rectify, transform and aspect) and deltaphot correct the photon positions for detector and optical distortions and determines an aspect solution using star positions from the time-tagged photon data.  The astrometric refinement process is diagrammed below.

 

 

A photometric module simplemap accumulates the photons into count and intensity maps and sextra extracts sources from images. A spectroscopic module galexspac  uses image source catalog inputs to extract spectra of these sources from the multiple slitless grism observations. 

 

 

2.    Pipeline Products

 

Data from each single-orbit visit stored in a single directory in the GALEX file system.   The directory path has the following format:

 

<ROOT>/<proc ver>/<tile>/<obs mode>/<product>/<image>/<try>/. 

 

An example path and filename is:

 

Path: <root>/01-vsn/10330-AISCHV2_381_40554/d/00-visits/0002-img/03-try/

 

Filename: AISCHV2_381_40554_0002_sv12-xd-mcat.fits

 

This path contains the pipeline processing version number, a tile identification string, instrument observing mode (direct, grism, opaque; represented by d, g, o, in each case preceded by the wavelength, data product type (single visit, multiple visit), and the pipeline processing try. Additionally, a unique source ID tag is be generated by the pipeline which uniquely encodes both the path information and the source number from merged source and spectra catalogs.

 

All of the pipeline products are described in the table below.

 

Table 3.1­ Pipeline Products Summary

Filename Suffix	File type	Description
Low-level pipeline products		Photon and housekeeping files
-asp.rec	Binary record	Satellite attitude solution
-asprta.rec	Binary record	Refined attitude solution
-rtastar.fits	FITS binary table	Star catalog for attitude refinement
-scst.fits	FITS binary table	Spacecraft state file
-nd-dphcent.txt	ASCII record	Deltaphot attitude output
-nd-dph.fits	FITS binary table	Deltaphot photon delta output
-[n or f]d-raw6.fits	FITS binary table	Raw photon data
-[n or f]d-x.fits	FITS binary table	Extended photon record
Images		Maps and associated files
-[n or f]d-cnt.fits	FITS image	Count map (J2000)
-[n or f]d-dose.fits	FITS image	Dose map (detector frame)
-[n or f]d-exp.fits	FITS image	Exposure map (J2000)
-[n or f]d-intbgsub.fits	FITS image	Background subtracted intensity map (J2000)
-[n or f]d-int.fits	FITS image	Intensity map (J2000)
-[n or f]d-movie.fits	FITS image cube	Time-slice count maps (J2000)
-[n|f]d-rr.fits	FITS image	low res relative response (J2000)
-[n or f]d-rrhr.fits	FITS image	high-res relative response (J2000)
-[n or f]d-skybg.fits	FITS image	sky background image (J2000)
-[n or f]d-wt.fits	FITS image	weight/mask image (J2000)
Catalogs		FUV, NUV and merged catalogs
-[n or f]d-cat.fits	FITS binary table	Sextractor catalogs for images
-fd-ncat.fits	FITS binary table	FUV extractions using NUV positions
-nd-fcat.fits	FITS binary table	NUV extractions using FUV positions
-xd-mcat.fits	FITS binary table	Merged source catalog
Flags		Artifact flag information
-[n or f]d-flags.fits	FITS image	Artifact Flag image (J2000)
-[n or f]d-flagstar.fits	FITS binary table	Star catalog for flagging
-[n or f]d-flag_bright.ds9reg	ds9 Region file	Bright star flag regions
-[n or f]d-flag_dichroic.ds9reg	ds9 Region file	Dichroic reflection flag regions
-[n or f]d-flag_edge.ds9reg	ds9 Region file	Edge flag regions
-[n or f]d-flag_near.ds9reg	ds9 Region file	Near bright star flag regions
-[n or f]d-flag_window.ds9reg	ds9 Region file	Window reflection regions
Grism-specific files
-[n or f]g-gsax.fits	FITS binary table	Spectral extraction parameter information for each source
-[n or f]g-pri.fits	FITS binary table	Image strips for each source from a single visit
-[n or f]g-prc.fits	FITS binary table	Image strips (combined) for each source from multiple visits
-[n or f]g-prm.fits	FITS binary table	Image strips (median) for each source from multiple visits
-[n or f]g-gsp.fits	FITS binary table	Spectral data for each extracted source vs. wavelength
-xg-gsp.fits	FITS binary table	Spectral data for each extracted source vs.wavelength (both bands)
-[n or f]g-xsp.fits	FITS binary table	Spectral data for each extracted source vs. offset

 

 

3.    Spectral Data Product Files

The spectral data product files are stored in binary FITS format, but they are GALEX-specific.  The formats of these files is described below:

-pri.fits (not included in ERO release)

This file contains the image strips for individual exposures for each spectral-extracted source.  The photon data, response, and masking

images are recorded for each source.  Masking for the image strip is stored in the response image (neighbor masking) and the

photon data image (other masking) as negative numbers. Spectral response variations with position (column) are stored in the response

image in the final row.  These images contain the total accumulated photon data for each source for a given single visit or exposure.

Each source has its own FITS header unit.  The order of the image strips is identical to the order of sources in the extracted source catalog given

in the (root)-gsax.fits file.

 

First Header Unit: Grism angle used for these extractions, total exposure time, total number of spectrally extracted sources (ECOUNT),

field center, and other extraction parameters.

 

Second through “N" Header Unit: Accumulated image strips for photon data and response for any of N-1 sources in the field of view.

The header portion defines the image dimensions (PRI_NC,PRI_NR). The scale is given in arcseconds per pixel in both spatial and dispersion

directions.  The header key words also include the blue limit offset in arcseconds (ARCSEC1) relative to the object center (undeviated) position,

which is used to derive the wavelength scale.  The image strips typically cover multiple grism orders, e.g. 1st, 2nd, and 0th.  The image data is

stored as 2-byte integers, which can be scaled to true data values using the header key words DATZERO, DATSCALE, RSPZERO, and RSPSCALE.

-prc.fits

This file contains the combined or summed image strips for multiple exposures for each spectral-extracted source.  The format is the same as

for the pri files, except that masking has been applied before summing. That is, all masked pixels in the individual (pri) image strips were

set to zero before summing.

 

First Header Unit: Contains the filenames of all *-pri.fits which  have been combined, as well as their respective grism angles.

 

Second through ”N” Header Unit: Accumulated image strips for photon data and response for any of N-1 sources in the field of view.

-prm.fits

This file contains the medianed image strips for multiple exposures for each spectral-extracted source.  The format is the same as for the pri files, except that masked pixels have been ignored by the median value calculation.

 

First Header Unit: Contains the filenames of all *-pri.fits which have been medianed, as well as their respective grism angles.

 

Second through “N” Header Unit: Accumulated image strips for photon data and response for any of N-1 sources in the field of view.

-gsax.fits

General information (binary FITS table). This file contains parameter information, as well as a source list table including extraction

parameters for each source. This source list catalog is created on the first exposure and read in and used for each subsequent exposure. This

file also contains a table of data for each accumulated exposure.

 

First Header Unit: Header key words contain various extraction parameters for ``galaxspac" (spectral accumulation program).  These

include the flux cutoff (i.e., the brightest ``N" sources) used on the direct image source catalog, the field RA,Dec origin, the masking parameters,

band number, the total number of exposures (or orbits) and exposure time accumulated.

 

Second Header Unit: Extracted source list.  This contains a source catalog of all the sources which have been extracted.  This

includes the RA,Dec positions, global IDs, direct image FUV and NUV flux rates, direct image FWHMs, and extraction parameters (object and

background widths and length of the spectrum) for each source.

 

Third Header Unit: Accumulated exposures (orbits) list.  This contains a catalog of all the photon data sets added into the image

strips.  This includes all the *-pri.fits filenames, grism angles, field offsets used (x, y, twist (rotation) relative to the direct image source

positions), exposure times, and reduction dates.

-xsp.fits

An xsp file, labeled *-fg-xsp and *-ng-xsp, is extracted for each wavelength band.  Each file represents two sequential spectral orders as recorded  along the spectral axis on the detector, so units are photons per second per arcsecond.  Each source has its own FITS header unit,  in the same order as the data occur in the -pri.fits file.

 

First Header Unit: Contains the header key word for the total number of sources extracted (ECOUNT).

 

Second through ``N" Header Unit: Spectra for each source and its computed 1 sigma error.  Spectra are given in photons/second/pixel vs.

offset arcsecond (position relative to object center or undeviated object position).

 

-gsp.fits

There are three gsp files for each spectrum: *-fg-gsp, *-ng-gsp, and *-xg-gsp.  The first two are grism-order combined (m =1 and 2 for NUV, and 2 and 3 for FUV), flux calibrated, and rebinned onto a linear wavelength scale. The third is the conjoined spectrum of the first two. The units for these are photons s^-1 cm^-2 Angstrom^-1. The fg-gsp and ng-gsp files are for many purposes rendered obsolete by the xg-gsp file. Since the dispersion is linear, the zero point and  dispersion of the wavelength scale are given as table entries.   Two of the vector entries (flux and error) are for a simple (summation)  extraction, and two vectors are for the optimal extraction.

 

Each file consists of a short primary header (giving the number of  spectra in the file) and a single extension. In this extension each row gives spectral information (default fluxes, associated errors, optimally  extracted fluxes, associated errors) for an astronomical object. These rows are listed in order of their appearance in   the extracted source catalog in the (root)-gsax.fits file. However,  ”masked” sources appearing in the gsax file are not extracted as spectra and thus not included in the gsp file listing.

 

 

4.    Source Extraction

 

For source extraction, the GALEX pipeline utilizes the program SExtractor (Bertin & Arnouts 1996) for detection and photometry of sources in the GALEX imaging data. A general description of SExtractor is given in Bertin & Arnouts (1996) as well as in the Sextractor manual (Available from http://terapix.iap.fr/soft/sextractor/). Another useful reference is the ``SExtractor for Dummies'' manual written by Benne Holwerda {Available from his web site http://www-int.stsci.edu/~holwerda/SE/}.  The pipeline includes a module poissonbg which computes a background map for each image as well as a corresponding detection threshold image.   Typical backgrounds in high Galactic latitude GALEX fields are ~10³ and 10⁴ photons s^-1 arcsec^-2 in the FUV and NUV  bands, respectively. With such low count rates, the distribution of count rates in each image is typically quite non-Gaussian, even for  relatively long exposures. In order to deal with these low backgrounds,  uses a modified clipping algorithm which makes use of the full Poisson distribution.

 

When running SExtractor, we use two images as input: one for detection and the second for photometry. For the detection image we use the ratio of the background-subtracted data image to the detection threshold map. All pixels in this ratio map which rise above the detection threshold will have values greater than one and we have therefore set the parameters in SExtractor so that it will consider all of these pixels as possibledetections. The image used for photometry is simply the background subtracted data image. Since we subtract the background from the data before running SExtractor, the background value within SExtractor is explicitly set to zero. As a result, the error estimates computed by Sextractor do not include the uncertainty due to the background counts. In the FUV-NUV merged catalog, we have updated the magnitude and flux errors to account for the background counts. (These are the fields called NUV_MAGERR, NUV_FLUXERR, FUV_MAGERR and FUV_FLUXERR in the merged catalog). These fluxes and errors refer to the Kron magnitudes measured by SExtractor and are computed as follows:

 

where f is the flux from the source in counts/sec, s is the sky level in counts/sec/pixel, is the area over which the flux is measured

( for the Kron flux) and t is the effective exposure time in seconds. Then the corresponding magnitude error is

.

 

SExtractor makes a few different measurements of the total flux of a source. For resolved sources, the MAG_AUTO measurement is probably the most appropriate choice while MAG_APER would be better suited for measurements of unresolved sources. Despite its name, MAG_BEST is not necessarily the best choice for most applications since the measurement for all objects are not made in a consistent way.

 

 

5.      Primary Object Catalog - Merged FUV+NUV Catalog (mcat file)

The merged object catalog is the final catalog product from the GALEX data pipeline.  It contains the full complement of source extractions from both the FUV and NUV images.  The columns of the mcat file fall into the following major categories:

 

·         Observation description

o       Contains global object ID, pipeline version, image tile number, observation type, optics wheel position, image product type, pipeline re-run (try) number, band, mcat ID, subvisit number and leg number (where applicable)

·         Band merger output quantities and flags

o       Includes merged position in Equatorial and Galactic coordinates, extinction E(B-V) from Schlegel maps, and calibrated fluxes, magnitudes and errors.  “Calibrated”currently means that values have been converted to AB magnitudes.  Four additional columns contain FUV fluxes extracted at NUV detection positions, NUV fluxes extracted at FUV detection positions and their errors.  These are useful for any study that is best performed using identical aperture photometry in each band (e.g. for colors, photometric redshifts)

·         Neighbor properties

o       Includes neighbor count and total flux out to three radii (currently set to 5”, 10” and 30”), closest neighbor distance, PA and nuv and fuv mag.

·         Derived quantities and map levels at merged output position

o       Includes extractions from pipeline-generated maps at the position of the detected source.  Sky background [counts s^-1 pixel^-1], artifact [flag value], weight [exposure-relative response product], contrast [dimensionless] and masked pixel counts [number] values are given for each source.

·         Sextractor output columns from FUV image extractions

·         Sextractor output columns from NUV image extractions

o       Please consult the sextractor manual and Sextractor for Dummies for definitions of these outputs

 

Catalog Fields

 

Table 3.2  Merged Object Catalog (-mcat.fits)

Name	Format	Unit	Definition
Object ID and orbpipe info
ggoid	2J	Number	Galex Global Object ID
vsn	1J	Number	Version number
tile	1J	Number	Tile number
type	1J	Number	Obs.type (0=single,1=multi)
ow	1J	Number	Optics wheel (1=drct,2=grsm,3=opaq)
prod	1J	Number	Product number (_visits,_main,etc.)
img	1J	Number	image number (exposure# for _visits)
try	1J	Number	Try number (processing attempt number)
band	1J	Number	Band number (1=nuv,2=fuv,3=both)
id	1J	Number	Merged ID.
Subvisit	1I	Number	Sub-visit number for AIS
leg	1I	Number	Planned leg number for AIS
Merged object quantities
alpha_j2000	1D	Degrees	Merged Right Ascension(degrees).
delta_j2000	1D	Degrees	Merged Declination (degrees).
prob	1E	Probability	Chi-squared probability of match.
sep	1E	Arcseconds	Separation between band positions.
Extinction	1E	Number	E(B-V) Galactic Reddening.
glon	1E	Degrees	Galactic longitude in degrees.