X.B.3 The Working Survey Data Base

IRAS Explanatory Supplement
X. The Formats of the IRAS Catalogs and Atlases
B. Point Sources
B.3 The Working Survey Data Base


Chapter Contents | Authors | References
Table of Contents | Index | Previous Section | Next Section

The most complete observational and data reduction history for point sources is contained within the Working Survey Data Base (WSDB). The WSDB is broken up into 20 files, one for each range of ecliptic longitudes called a Lune. Each file is preceded by a header file containing a single 80-character ASCII record giving the date and version of the WSDB (Table X.A.1). Except as noted below, the entries are similar to those listed above for the catalog tape. Table X.B.3a lists the entries, their variable type and length. A second file, called the Ancillary file (Table X.B.3b) contains additional flags and derived quantities obtained during final product generation. Most of the information in the ancillary file is the same as that in the catalog tape described in Section X.B.1. Character variables are in ASCII format. All arithmetic quantities are in integer format with the high order byte first.

It is important to note that the WSDB and Ancillary files are in binary format with variable length, blocked records (due to the variable number of hours-confirmed sightings or associations). Extra information has been added to the WSDB and Ancillary file tapes to make it possible to read in these variable length records. Each physical block on the tape begins with a Block Control Word of length I*4. Its high order two bytes give the length of the physical block in bytes; this length includes the 4-byte length of the Block Control Word itself. In addition, each logical record within the physical block is preceded by a Segment Control Word of length I*4. Its high order two bytes give the length of the logical record in bytes; again, this length includes the 4-byte length of the Segment Control Word itself. The inclusion of this information should make it possible for computers that can work with Variable Block formats to easily read the tape. In no case does the information for a single source span two physical blocks.

The remainder of this section describes WSDB and ancillary file entries not discussed above.

SDAS Lune Number: LUNE

In the data reduction the sky was divided into twenty "lunes" based on ecliptic coordinates. Lune 1 comprises that part of the sky with (X) > 60°; Lune 2 comprises that sky with (X) < -60°. Lunes 3-20 comprise that part of the sky with |(X)|   60° and ecliptic longitudes, (X) in 20° wide blocks. Lune 3 extends from 0° (X) < 20°, lune 4 from 20° < (X) < 40°. etc.

Ecliptic Bin Number: BIN

To aid the data reduction the entire sky was divided into some 40,000 1° sq. bins. The bin structure is quite simple in ecliptic coordinates (Fig. X.Ap1.1) and an algorithm for generating bin numbers is given in Appendix X. 1.

Ecliptic Coordinates: ELAT, ELONG

Ecliptic coordinates are given in units of 10-8 radians in the equinox 1950.

Scan Angle: SCAN

After weeks-confirmation SCAN gives the average scan angle of the focal plane with respect to the south-going local ecliptic meridian.

Positional Uncertainty: SIGY, LZ, SIGZ

As discussed in detail in Section V.D, the position refinement describes the source positional uncertainty in terms of 1(X) in-scan and cross-scan gaussian uncertainties (SIGY and SIGZ) and a uniform uncertainty (of half-width LZ) whose size depends on the exact paths of the source across the focal plane.

Number of LRS Extractions: LRSX

Each time a source with a signal-to-noise ratio greater than 25 at either 12 or 25 µm transversed the focal plane, the data reduction software automatically triggered a request to extract a spectra from the low-resolution spectrometer data (Chapter IX). The threshold was purposefully set quite low so that sources with weak continuua but strong lines could be detected. Many sources with LRSX > 0 will fail to have meaningful spectra and will thus have NLRS = 0.

Known Source ID: KSID

The positions and predicted brightnesses of some 32,000 point sources including SAO stars, IRC objects and asteroids were incorporated into the data reduction software to provide a check on the positional and photometric accuracy of the IRAS sources. Table X.B.5 lists the range of KSID values assigned to sources of various types.

The following values are given for each hours-confirmed sighting:

Flux and Flux Uncertainty: FLUX(4),SIGF(4)

Flux and flux uncertainty measurements are given in units of 10-16 W m-2 for each hours-confirmed sighting. Note that an instrumental flux, not the flux density, is given. In order to convert the instrumental flux to flux density, one must divide the instrumental flux by 13.48, 5.16, 2.58, 1.00 × 1012 Hz at 12, 25, 60 and 100 µm respectively. The derivation of flux uncertainties and the averaging of the individual hour-confirmed fluxes to give the average value (see AVGFLUX, below) quoted in the catalog is discussed in Section V.H.5.

Signal-to-Noise Ratio: TSNR(4)

If a source is detected (but not necessarily even seconds-confirmed) in a given band, then a value equal to ten times the maximum signal-to-noise ratio (SNR) observed on any detector in that band in the hours-confirmed sighting is retained. In relatively simple parts of the sky the noise estimator used to derive SNR gives reasonable values. In more complex regions near the Galactic plane (see Section V.C.2) the utility of SNR is very limited.

Correlation Coefficient: CORR

The maximum point source correlation coefficient (see Section V.C.4) obtained during a hours-confirmed sighting is retained for each band. The values for each band, expressed as percentages up to a maximum of 100, are encoded into a single integer according to the following algorithm:

CORR = CC(1) × 224 + CC(2) × 216 + CC(3) × 28 + CC(4)

(X.B.1)

Flux Status: FSTAT

In each band there is a hierarchy of measurement quality depending on how many times a given source is observed within a given hours-confirmed sighting. FSTAT plays a crucial role in determining whether a source is included in the catalog at all, whether a flux is of high, medium or low quality and how the flux averaging was performed (
Section V.H.5). for the meaning of each FSTAT value refer to Section V.D.8.

The values of FSTAT for the four bands are compressed into a single integer according to the following algorithm:

FSTAT = FSTAT(1) × 212 + FSTAT(2) × 28 + FSTAT(3) × 24 + FSTAT(4)

(X.B.2)

Detector ID's: DETID

The identities of all detectors observing a given source on the sightings (up to a maximum of 3 sightings) comprising a given hours-confirmed observation are recorded in the array DETID(I, J). The four bands run from I = 1 to 4 while the three sightings run from J = 1 to 3.

Each value of DETID contains an integer that must be decoded according to the following algorithm to obtain the detectors observing the source:

DETID(I,J) = D1 × 210 + D2 × 25 + D3
(X.B.3)
where detector numbers, D1, D2 and D3, range from 1 to 16 within each band.
Table X.B.6 lists the correspondence between detector number within a band to the true detector number. The order of detector number is significant as described in Chapter V.D.5. It should be emphasized that when three detectors are named in a sighting, indicating the presence of an edge detection, the weakest of the edge detections may not have played any part in the assignment of flux or position.

Detector Name: DNAM, TNAM

Throughout the course of the data processing each hours-confirmed sighting is known by a combination of the first detector measuring the object (DNAM) and the time (in deci-UTCS since 1981, January 1, 0h UT) of that sighting (TNAM).

Confusion Status: CSTAT, PNEARH, PNEARW

At various stages in the reconstruction of a point source attempts are made to recognize (and remedy) the effects of confusion between nearby sources. The confusion status word CSTAT plays an important role in selecting sources to be treated by the "clean-up" processor (Section V.H.2) and in determining which sources to keep in regions of high source density (see Section V.H.6). For a detailed discussion of confusion processing see Section V.D.2. Values of CSTAT and brief descriptions of their meaning are given in Section V.D.8.

The values of CSTAT for each band are encoded into a single integer according to the following algorithm:

CSTAT = CSTAT(1) × 224 + CSTAT(2) × 216 + STAT(3) × 28 + CSTAT(4)

(X.B.4)

The number of hours- and weeks-confirmed point sources within 6' × 4.5' (half-widths) of the quoted source, PNEARH and PNEARW, are encoded into a single byte:

PNEAR = PNEARW × 24 + PNEARN

(X.B.5)

Cirrus Flags:. CIRRUS, CIRR1, CIRR2, CIRR3

The three flags denoting the presence of extended 100 µm emission ("cirrus") as discussed in
Section V.H.4, are encoded in two bytes according to the algorithm:

CIRRUS = CIRR3 × 2 8 + CIRR1 × 2 4 + CIRR2

(X.B.6)

Values of CIRR2 = 0 and CIRR3 = 255 mean no data were available.

Small Extended Source Flags: SES1, SES2

The two small extended source flags for each band discussed in Section V.E.1 are encoded into two integers according to the following algorithm:

SES1 = SES1(1) × 212 + SES1(2) × 28 + SES1(3) × 24 + SES1(4)

(X.B.7)

and
SES2 = SES2(1) × 212 + SES2(2) × 28 + SES2(3) × 24 + SES2(4)

(X.B.8)

Clean up Processors: CLEAN, BRIGHT, ACCEPT, HSDPROC, MISC

In creating the WSDB from the weeks-confirmed data, several processors were applied in order to fix various known problems (see Section V.H). These processors set various flags that allow the user to understand what processing occurred. The clean-up processor allowed the weeks-confirmation of various WSDB sources that did not previously have an opportunity to weeks-confirm for purely technical reasons or which were incorrectly split asunder during band merging. The flags set by that processor are described in Table X.B.7a, and include flags indicating saturated fluxes. Optical crosstalk from certain bright objects such as Saturn and IRC+10216 produced a few spurious sources. These were marked for deletion. The byte BRIGHT (Table X.B.7b) notes these cases and also contains flags denoting final acceptance (ACCEPT) or rejection of the source in each band. The flags from the High Source Density processor (HSDPROC, see Section V.H.6) are given in 2 bytes per band (Table X.B.7c). Miscellaneous flags are set in MISC (Table X.B.7d) and include the presence of flux discrepancies in each band, whether the in-scan positional uncertainties needed to be increased (Section VII.C) and whether the source was accepted in the catalog.

The Catalog Working Survey Data Base (WSDB)
Table X.B.3a.
ByteNameDescriptionUnitsType
0*LUNESDAS Lune number---I*4
4BINEcliptic Bin Number---I*4
8ELONGEcliptic longitude 1950. 10-8 radI*4
12ELATEcliptic latitude 1950. 10-8 radI*4
16SCANaverage scan angle with respect to ecliptic meridian milli-radI*2
18SIGYIn-scan Gaussian position uncertainty µ-radI*2
20LZ Cross-scan Uniform position uncertainty µ-radI*2
22SIGZCross-scan Gaussian position uncertaintyµ-radI*2
24LRSXTotal number of LRS extraction requests ---I*2
26KSIDKnown Source ID---I*2
28NHCONNumber of Hours ---I*4
(<25) confirmed sightings
The following values repeat for each hours-confirmed sighting:
32FLUXFLUX(I)=In-band power10-16W m-24I*4
48SIGFSIGF(I)=Uncertainty in FLUX(I)10-16W m-24I*4
64TSNRTSNR(I)=10x(max SNR)---4I*2
72CORRMaximum correlation coefficient (1 value per band, Eq. X.B.1)---I*4
76FSTATFlux status word (1 value per band, Eq.X.B.2) ---I*2
88DETIDDetector ID Array (4,3) for 4 bands (Eq.X.B.3)---12I*2
102LRSXNONumber of LRS extraction requests---I*1
103DNAMdetector part of source name ---I*1
104TNAMdeci-UTCS part of source name deci-secI*4
108CSTATConfusion status flags (Eq.X.B.4) ---I*4
112--- 191repeats bytes 36-115
192--- 271for subsequent hours etc... confirmed sightings.

Ancillary WSDB File
Table X.B.3b.
ByteNameDescriptionUnitsType
0PNEARNearby hours and weeks-confirmed neighbors---I*1
1CLEANClean up Processor flags---I*1
2SES1 Number of nearby unconfirmed SES (4 bands encoded)---I*2
4SES2 Number of nearby weeks-confirmed SES (4 bands encoded) ---I*2
6CIRR1Number of 100 µm only WSDB sources,
spatially filtered 100 µm emission.
---I*2
CIRR2
CIRR3value of 100 µm half-degree beam total intensity MJy sr-1
8AVGFLUXAVGFLUX(I) Averaged flux (in-band power) in band I.10-16W m-24I*4
24AVGUNCAVGUNC(I) Uncertainty in averaged (in-band) flux in band I.10-16W m-24I*4
40HSDPROCHSDPROC(I). Flags set by the high source density processor in band I---4I*2
48RARight ascension 195010-5"I*4
52DECDeclination10-5" I*4
56NAME source name---A*12
68NLRS Number of meaningful LRS spectra---I*2
70LRSCHARcharacterization of LRS spectra ---A*2
72BRIGHTBright Source Clean up flag for catalog sources accepted in the catalog --- I*1
ACCEPT
73VARPercent likelihood of Variability---I*2
74FQUALFLUX Quality flags (one per band)
BIT 0-1, 12 µm
BIT 2-3, 25 µm, etc.
---I*1
75MISC Miscellaneous Status bits, incl. DISC, ACCEPT,SIGY
See Table X.B.7d
---I*1
76LUNE SDAS Lune number---I*4
80BINEcliptic Bin Number---I*4
84ELONGEcliptic longitude 10-8 rad I*4
88ELAT Ecliptic Latitude10-8 rad I*4
92NID 2Number of Associations---I*2
94IDTYPEType of association---I*2
96CATNOCatalog Number ---I*2
98SOURCESource ID---A*15
113 TYPE Source Type/Spectral Class---A*5
118 RADIUSPosition difference(") I*2
120 POSPosition Angle' E of NI*2
122 FIELD1object field #1---I*2
124 FIELD2object field #2---I*2
126 FIELD3object field #3---I*2
128-160  continuation of association 3 in blocks of 32 bytes  
  1. See X.B.3 for discussion of blcck and segment control words
  2. If NID=0, one blank association field of 32 bytes is written.
  3. CATNO,SOURCE,TYPE,RADIUS,POS,FIELD 1-3 are repeated in blocks of 32 bytes, 2 per logical record, as necessary. The definition and formats of FIELD1-3 depend on the individual catalog in which the association is found. See Table X.B.4.


Chapter Contents | Authors | References
Table of Contents | Index | Previous Section | Next Section