IRAS SMALL SCALE STRUCTURE CATALOG Catalog Prepared by George Helou and D. W. Walker Under the supervision of the Joint IRAS Science Working Group (JISWG) Notes about this computer-readable documentation: The following text is an excerpt from the original printed documentation, including Chapter I "Introduction" and Chapter VII "Formats", which was keyed-in at the Astronomical Data Center. Slight modifications may have been made to the text in order to describe the machine-readable version of the catalog. The text makes references to other chapters in the printed documentation as well as to the IRAS Explanatory Supplement. These references are in the form: XX.XX.XX, for chapter, section and subsection, respectively. For example, IV.B.3a refers to chapter IV, section B, subsection 3a, and Suppl.V.H.9 refers to chapter V, section H, subsection 9 in the IRAS Explanatory Supplement. The reader is directed to the printed versions of the IRAS documentation for these references. Because of the restrictions of the ASCII character set, some notational differences appear between the printed documentation and this text. When appropriate, greek letters are written out as their english names, except the word "micron" appears instead of the term "Mu-meters" or "um". In general, subscripts appear in parentheses, for example, flux per unit frequency, "f subscript Nu", appears as f(Nu) and can be pronounced, "f sub Nu". Terms given in italics in the original text are enclosed in double quotation marks. At the end of this computer-readable text is a discussion of the IRAS Small Scale Structure Catalog as it appears in FITS-table format. This was authored at the ADC, and does not appear in the original documentation. Please direct any reports of typographical errors to: Lee Brotzman Code 930.3 NASA/GSFC Greenbelt, MD 20771 (301) 286-6953 BITNET: ZMLEB@SCFVM SPAN: CHAMP::BROTZMAN Internet: zmleb@scfvm.gsfc.nasa.gov or zmleb%scfvm@dftsrv.gsfc.nasa.gov I. INTRODUCTION The primary mission of the Infrared Astronomical Satellite (IRAS) was to conduct a sensitive and unbiased survey of the sky in four wavelength bands at 12, 25, 60, and 100 microns. Launched in January 1983, IRAS ceased operations in November 1983 after having successfully surveyed 96% of the sky. The IRAS mission, the data processing, and the products are described in detail in the "Explanatory Supplement to the IRAS Catalogs and Atlases" (Vol. I of this series), with which the reader is assumed to be familiar. References to chapters and sections in the "Supplement" are given here prefixed with Suppl., e.g., Suppl.I.A. Results from the IRAS survey are presented in various forms depending primarily on the angular scale of interest. They include a catalog of infrared point sources (the Point Source Catalog), the present Small Scale Structure Catalog (hereafter the SSS Catalog or the Catalog), and an atlas of absolute surface brightness images of the infrared sky (the Sky Brightness Images). The Point Source Catalog contains some 250,000 entries resulting from a search for spatially unresolved sources in the survey data. At the other extreme, the Sky Brightness Images display the large scale appearance of the sky with a resolution of 4'. The SSS Catalog attempts to fill the gap by listing sources resolved in any band up to a maximum size of 8', aiming in particular at an adequate description of the infrared emission from objects such as galaxies, planetary nebulae or compact HII regions. Given the complexity of the infrared sky and the infinite variety of source shapes, the scope of a SSS Catalog remains by necessity vaguely defined, quite apart from the technical problems posed by the specifics of the telescope design and operation. This introduction describes the Catalog and the manner in which it was constructed, discusses the uncertainties in the various quantities listed, and gives a statistical overview of the sources in the Catalog. The Small Scale Structure Catalog contains 16,740 sources which point out directions on the sky where IRAS resolved emission on the scale of 2' to 8', in at least one of the four bands, in a reliable and unconfused manner. Although the objective was to find intuitively acceptable "small extended sources", the Catalog entries are in fact defined by excursions in sky brightness detectable above the local baseline on spatial scales smaller than 8', excluding point sources. The Catalog lists for each entry and in each band separately: a position accurate to about one arcminute (rms); a flux density accurate to 50% (rms) overall, and better at high signal to noise ratios; a rough indicative size if a point source is present at the same location; and a variety of warning and processing flags. Completeness in the Catalog is not estimated since it varies with position on the sky as a function of source density and number of survey coverages. Fewer than 3% of the Catalog entries are unreliable. Almost all of these are weak point sources broadened by radiation hits or detector noise. In spite of every effort to select only resolved, well-defined, and isolated sources for inclusion in the Catalog, a few problems remain. At low signal to noise ratios, the processor may have erred in deciding whether a source is resolved or not. Many entries, mostly at 60 to 100 microns, refer only to emission features that are details in larger structures known as Galactic cirrus. Naturally, some close double and multiple point sources appear in the Catalog. There is also some evidence for flux densities at 12 and 25 microns being overestimated by 0.5 to 1 Jy. The greatest amount of caution perhaps is required in interpreting the ABSENCE of an entry in this Catalog, in one or in all bands. Incompleteness is a complex function of band, brightness, and location so an absence does not necessarily imply the lack of detectable extended emission in any band. While Small Scale Structure cannot be defined or measured with the same finality as point sources, Catalog entries always represent correlated and confirmed events in the detector output, making them valuable pointers to the IRAS images and raw data. The SSS processor has distilled from the survey a wide variety of objects and provided reasonable estimates of their parameters, but if a detailed shape description, or more accurate position, flux, or size are needed, it is necessary to go back to the fundamental reference, the raw detector ouptut. The algorithms used to detect and measure the entries in this Catalog are described in Suppl.V.E, and summarized in Chapter II. Chapter III examines the detection processing to assess the types of sources detected and the influence of noise on their reliability. Chapter IV describes the strategy used to select the most reliable sources for the Catalog from the larger pool of confirmed detections. Chapter V provides an overview of the statistial properties of the Catalog, a comparison with other IRAS data products, and estimates for the accuray of the measured quantities. Chapter VI discusses reliability and completeness. Chapter VII details the formats of the printed and machine-readable versions of the Catalog. VII. FORMATS The Small Scale Structure Catalog is available in three media, on tape, in print, and in machine-readable text. On tape the Catalog is contained in one file ordered by right ascension and by declination in case of conflict. The printed version has essentially identical contents to the tape copy, but is organized and formatted differently and shows only one association. In both versions and for each source, data are shown only for those bands where detections have been found and have survived processing through "weeks confirmation". No upper limits are given where no detections were found; none are given where detections were dropped because of confusion ("cluster analysis"). Absence of data from this Catalog at a given position in a given band does not imply absence of emission, but simply lack of information (see Chapter VI) for a discussion of completeness). On the other hand, flags NEARPS, SES1 and HD, which depend only on position in the sky, are always estimated for all four bands, at the mean position of each source. A. Tape Version In general, the data for each source is organized in three 80-byte records followed by a variable number of association records; Table VII.A.1 shows detailed organization. The first two records contain basic parameters and warning flags referring to the band-merged source, whereas the third record is made up of four 20-byte components containing position and quality flags for each of the bands. When present, the 12 micron data is inserted into the first 20 bytes of the record; the 25, 60, and 100 micron data go into the second, third, and last block of 20 bytes in the record, respectively. Quarters corresponding to bands absent from the source are filled with blanks. The fourth and subsequent records contain data on positional associations between the Catalog source and objects in a variety of astronomical catalogs listed in Suppl.V.H.9. Each record holds data on two associations, in two 40-byte blocks. All data entries for each source are described below in their order of appearance in Table VII.A.1 B. Machine-Readable Version The machine-readable version of the catalog is divided into two files. The first is a 240 byte record file containing the 16,740 sources. The second is a 58 byte record file containing the 5,178 associations. The file formats are described below in Table VII.A.1 and Table VII.A.1a. Source Name: NAME ------------------ The letter "X" prefixes all names in this Catalog to distinguish them from Point Source names. The name is then derived from position by combining the hours and minutes from the right ascension with the sign, degrees, and decimal fraction of degrees from the declination. In case of name duplication the letters "A", "B", etc. are appended to the name in order of increasing right ascension, and of increasing declination in case of a tie. "Band Merging" Flag: BMFLG --------------------------- The "band merging" flag gives the number of bands in which the source has an entry in the third record (regardless of flux quality in these bands), together with indications on the "band merging" processing history for that source. BMFLG=C or D indicates 3 or 4 mutually confirming (Suppl.V.E.6) components in the source. BMFLG=I, J, K or L indicates 1, 2, 3 or 4 components in a source having experienced "band merging" complications and having survived the modified "final selection" discussed in IV.B.3.a. Position: RAHR, RAMIN, RASEC, DSIGN, DECDEG, DECMIN, DECSEC ------------------------------------------------------------ Source position is the simple mean of the positions of all the individual band components in the source regardless of the individual flux quality flags. It is given in equatorial coordinates for the 1950 equinox. Positional accuracy is discussed in V.C.2. Number of Sightings: NH(4) --------------------------- The number of survey coverages contributing a sighting to the weeks confirmed source is given for each band as a single character. The order for this and similar arrays in the Catalog is from 12 microns (subscript = 1) to 100 microns (subscript = 4). Flux Density: FLUX(4) ---------------------- An estimate of the spatially integrated flux density from the source is given in Jansky for each band, with three digit precision. This is transformed from the fluxes in the IRAS bands assuming a spectrum with Nu x f(Nu) = constant (see Suppl.VI.C). To obtain the actual flux density at the nominal wavelength for each band, FLUX must be corrected according to the prescription in Suppl.VI.C, which is reproduced at the end of this volume. The estimation algorithm is described in Suppl.V.E.3, 4 and 5; calibration in II.C. Accuracy is discussed in V.C.1. Optical Cross Talk Flag: XTALK(4) ---------------------------------- This quantity indicates the extent and kind of cross-talk for which each component in the source was flagged (the printed version gives only a summary described under "FCAT" below). As discussed in IV.B.2a above, the processor accumulated the number NXT of detections that might have been caused simply by optical cross-talk, along with the number NS of detections. If NXT=0 then XTALK=0 for that band. If NXT > 0, and either (i) (NS-NXT) > 2, or (ii) (NXT/NS) < 2/3 then XTALK=1 for moderate cross-talk. If NXT > 0, and neither (i) or (ii) is true then XTALK=2 for severe cross-talk. If in addition a band component was flagged for cross-talk at the level of "final source selection", then XTALK is increased by 4. Note that the component is given a low quality rating if XTALK > 0. Near-by Point Sources: NEARPS(4) --------------------------------- This flag gives a count of all weeks-confirmed point sources (regardless of whether or not they are included in the Point Source Catalog) in each band within a 9' radius of the mean position of the SSS Catalog source. NEARPS=1 indicates most often that the emission has been detected as a point source as well as here. Larger values are a warning of possible confusion affecting the source, or an indication that this source combines emission from two or more barely resolved point sources (see DBLPS below). NEARPS is given as a single character per band, and denoted by a letter when it exceeds 9 ("A" for 10, "B" for 11, etc.). Near-by Intermediate Small Extended Sources: SES1(4) ----------------------------------------------------- This flag gives a count in each band of all hours-confirmed entries in FISES (Suppl.V.E.3) within a 9' radius of the mean position of the source. In a clean sky, SES1 would be equal to NH; larger values implying a higher density of detections, either spurious or due to complex structure, point to areas where "cluster analysis" was active. Large values of SES1 for a band not represented in the source may indicate reliable extended detections discarded in subsequent processing. SES1 appears as a single character per band in the same notation as NEARPS. Cirrus Indicator: CIR ---------------------- This cirrus flag gives the number of point source detections at 100 microns only, hours confirmed but not necessarily weeks confirmed, within a 30' radius of the SSS source. Cirrus is rich in structure on all scales, and may supply a 100 micron or a 60 micron component to a source detected at other wavelengths, or may combine with the 100 micron or 60 micron emission from a source and severely compromise its flux measurement. Values of CIR above six invite caution in interpreting both 60 and 100 micron detections. High Source Density Flag: HD ----------------------------- This single character flag denotes whether the source falls in a region of high source density in the sense of Point Source clean-up processing (Suppl.V.H.6). This flag was not used in SSS processing, but warns against confusion by pointing out areas where point source density may exceed the resolving capability of the instrument. When the flag is written out in binary notation the least significant bit refers to the 25 micron band, and so on. Values greater than 9 are denoted by letters as for NEARPS (See Table Suppl.X.B.2). Double Point Source Flag: DBLPS -------------------------------- This flag indicates the possibility that this Catalog source is the combination of two point sources (see IV.B.2.e for definition). The indication is given for each band, encoded as for HD, with the bit set to one when the possibility exists for the corresponding band. This flag is estimated only for bands represented in the source, regardless of the flux quality in these bands. Coincident Point Source: PTSRC ------------------------------- A small extended source will often trigger the point source detection processor, either because it is still small enough to produce an acceptable fit to the template (bright point sources with low correlation coefficients are examples of this), or because it actually contains one or more point-like components. When an entry from the Point Source Catalog lies within a distance D of the SSS Catalog source position it is indentified as a point source counterpart to the SSS Catalog source, and its IRAS name is entered in this column. The distance D used is half the largest value of UNC (V.C.3) among the bands in the source. When two or more counterpart candidates are found, the nearest one is chosen, and an asterisk precedes PTSRC as a warning. Clearly, the source descriptions here and in the Point Source Catalog are complementary, and neither can be ignored. Inferred Source Size: PSIZ(4) ------------------------------ An estimate of the source size is given in deci-arcminutes for those bands where the point Source counterpart is detected. The estimate is obtained from 1/2 PSIZ(i) = FWHM(i) x (SSSFLUX(i)/PTSRCFLUX(i)) where FWHM(i) is the full width at half maximum of a point source in band i, namely 0.82', 0.84', 1.44', and 3.14', respectively, at 12, 25, 60, and 100 microns; SSSFLUX and PTSRCFLUX are the fluxes in the two catalogs. See V.B.1 and V.C.3 for details and a discussion of accuracy. Positional Associations: NID, IDTYPE ------------------------------------- Positional associations between objects in the Small Scale Structure Catalog and objects in astronomical catalogs were obtained following the same strategy used for Point Sources (Suppl.V.H.9). The only difference is that the search radius around sources in this Catalog is never less than 120". NID is the total number of matches found. IDTYPE ranges from 1 to 4 to indicate an association found in an extraglactic catalog (1), a stellar catalog (2), other catalogs (3), or matches in multiple types of catalogs (4). Data on each associated object appear in a 40 byte block whose format is detailed further below. Each record starting with the fourth data record for the source contains two of these blocks. The format for the association is almost identical to the Point Source Catalog (see Suppl.X.B.1). Flux Quality Flag: FQLT ------------------------ This is the quality class assigned to the band component by the scheme described in Chapter IV and Table IV.A.1. High, intermediate and low qualities are denoted, respectively, by A, B, and F. Starting with this flag, data for each band appear separately in the corresponding quarter of the third data record for the source. Final Catalog Selection Flags: FCAT ------------------------------------ The results of the three main stages of "final selection" (Figure IV.B.1) are presented in this flag which appears in the tape version as a single character ranging from 0 to V, and is decomposed in the printed version into three parts (VII.B). When FCAT is written out in binary notation, the least significant bit indicates the result of the flux test (IV.B.2.d), and is set if failed; the next bit refers to the detection count test (IV.B.2.c), and is set if failed. The next two bits refer to the repeatability test (IV.B.2.b), and signal one of four possible outcomes: 00 for intermediate values of N/M except N/M = 2/2; 01 for low N/M; 10 for high N/M; and 11 for N/M = 2/2 (Table IV.B.3). The most significant bit in FCAT is set if the detection is flagged for cross-talk of any kind (IV.B.2.a). Table VII.A.2 shows all possible values of FCAT, their representation both on tape and in print, and their meaning. Right Ascension Offset: DRA ---------------------------- This gives in seconds the amount to add to the mean right ascension of the source (shown in the first record) to obtain the right ascension measured for the emission in this band. Declination Offset: DDEC ------------------------- This gives in arcseconds the amount to add to the mean declination of the source (shown in the first record) to obtain the declination measured for the emission in this band. Positional Uncertainty from Intensity Distribution: UNC -------------------------------------------------------- The second moments about major and minor axes of the roughly modeled intensity distribution are estimated as explained in Suppl.V.E.3, and their geometric mean is given here as a diameter in deci-arcminutes. Because of the coarse grid used in the definition, UNC is best interpreted as a positional uncertainty, namely the 95% confidence diameter for the position in this band (see V.C.3 for a discussion). Number of Detections: NS ------------------------- This total number of seconds-confirmed detections in the source is accumulated through "cluster analysis" and "weeks confirmation", and includes detections that failed to seconds-confirm because of a failed detector. Association File: CATNO,SOURCE,TYPE,RADIUS,POS,FIELD1,FIELD2,FIELD3 -------------------------------------------------------------------- For each match in an astronomical catalog, CATNO is the number identifying that catalog in Tables Suppl.V.H.1 and Suppl.X.B.4. SOURCE is the name of the object in that catalog, and TYPE its character or spectral or morphological type if available. RADIUS is the distance in arcseconds from the IRAS position to the associated object. POS is the position angle measured in degrees East of North of the direction from the IRAS source to the associated object. FIELD1-3 carry values depending on the catalog in question (Table Suppl.X.B.4). Typically FIELD1 and FIELD2 contain magnitudes in decimag, and FIELD3 a size in arcseconds. TABLE VII.A.1 Format of the Small Scale Structure Sources File Start Byte Name Description Units Format 1 NAME IRAS source name --- 10A1 11 BMFLG Number of bands in source --- A1 [1] and "band merging" warning flag 12 RAHR Right ascension (1950) Hours I2 14 RAMIN Right ascension (1950) Minutes I2 16 RASEC Right ascension (1950) Seconds F4.1 20 DSIGN Declination (1950) --- A1 21 DECDEG Declination (1950) Degrees I2 23 DECMIN Declination (1950) Arcminutes I2 25 DECSEC Declination (1950) Arcseconds I2 27 NH(4) Number of hours-confirmed --- 4A1 sightings 31 FLUX(4) Averaged, spatially Jy 4E8.2 [1] integrated flux density (no color correction) 63 XTALK(4) Cross-talk flag --- 4A1 [1] 67 NEARPS(4) Number of near-by weeks- --- 4A1 confirmed point sources 71 SES1(4) Number of near-by FISES entries --- 4A1 75 CIR Number of hours-confirmed, --- I2 100 micron only point sources 77 BLANK Four spare bytes --- 4A1 81 HD High source density flag --- A1 (encoded, one bit per band) 82 DBLPS Possibility this is a double --- A1 point source (1 bit per band) TABLE VII.A.1 Format of the Small Scale Structure Catalog Sources File Start Byte Name Description Units Format 83 PTSRC Name of IRAS Point Source --- 12A1 counterpart and conflict flag 95 PSIZ(4) Size estimate from comparison Deciarc- 4I3 of FLUX with PTSRC flux minutes 107 NID Number of associations --- I2 109 IDTYPE Type of objects associated --- I4 113 BLANK 48 spare bytes --- 48A1 161-180 (blank if no 12 micron component) 161 FQLT 12 micron flux quality class --- A1 [1] 162 FCAT 12 micron "final selection" flags --- A1 [1] 163 DRA Right ascension offset from Seconds F6.1 mean position to 12 micron 169 DDEC Declination offset from mean Arcseconds I4 position to 12 micron 173 UNC 95% confidence diameter for Deciarc- I3 position at 12 micron minutes 176 NS Number of individual detections --- I3 in 12 micron component 179 BLANK Two spare bytes --- 2A1 181-200 25 micron equivalent of bytes 161-180 (blank if no 25 micron component) 201-220 60 micron equivalent of bytes 161-180 (blank if no 60 micron component) 221-240 100 micron equivalent of bytes 161-180 (blank if no 100 micron component) TABLE VII.A.1a Format of the Small Scale Structure Catalog Associations File Start Byte Name Description Units Format 1 NAME IRAS Source Name --- A10 12 RECNO Rec. Num. of source in data file --- I6 19 CATNO Catalog identifier --- I2 21 SOURCE Object ID in that catalog --- 15A1 36 TYPE Object type in that catalog --- 5A1 41 RADIUS Distance from IRAS source Arcseconds I3 to associated object 44 POS Position angle from IRAS Degrees I3 [1] source to object E of N 47 FIELD1 Object data field Nr 1 [2] I4 [1] (magnitude or other) 51 FIELD2 Object data field Nr 2 [2] I4 [1] (magnitude or other) 55 FIELD3 Object data field Nr 3 [2] I4 [1] (magnitude or other) [1] This quantity appears with a different format or representation, or is omitted in the printed version of the catalog. [2] The definition and units of the quantities in FIELD1-3 depend on the individual catalog in which the association is found. See Table Suppl.X.B.4. TABLE VII.A.2 Interpretation of Final Source Selection Flag Optical N/M Detection Flux Tape Printed Binary Cross- Test Count Threshold Version FCAT Notation Talk [1] FQLT FCAT XEI FCAT ----------------------------------------------------------------------------- NO MED* PASS PASS B 0 00 00000 NO MED* PASS FAIL F 1 01 00001 NO MED* FAIL PASS F 2 02 00010 NO MED* FAIL FAIL F 3 03 00011 NO LOW PASS PASS F 4 10 00100 NO LOW PASS FAIL F 5 11 00101 NO LOW FAIL PASS F 6 12 00110 NO LOW FAIL FAIL F 7 13 00111 NO HIGH PASS PASS A 8 20 01000 NO HIGH PASS FAIL B 9 21 01001 NO HIGH FAIL PASS B A 22 01010 NO HIGH FAIL FAIL B B 23 01011 NO 2/2 PASS PASS B C 30 01100 NO 2/2 PASS FAIL F D 31 01101 NO 2/2 FAIL PASS F E 32 01110 NO 2/2 FAIL FAIL F F 33 01111 YES MED* PASS PASS F G X00 10000 YES MED* PASS FAIL F H X01 10001 YES MED* FAIL PASS F I X02 10010 YES MED* FAIL FAIL F J X03 10011 YES LOW PASS PASS F K X10 10100 YES LOW PASS FAIL F L X11 10101 YES LOW FAIL PASS F M X12 10110 YES LOW FAIL FAIL F N X13 10111 YES HIGH PASS PASS F O X20 11000 YES HIGH PASS FAIL F P X21 11001 YES HIGH FAIL PASS F Q X22 11010 YES HIGH FAIL FAIL F R X23 11011 YES 2/2 PASS PASS F S X30 11100 YES 2/2 PASS FAIL F T X31 11101 YES 2/2 FAIL PASS F U X32 11110 YES 2/2 FAIL FAIL F V X33 11111 [1] In the repeatability test, "MED*" indicates that N/M is in the intermediate range (Table IV.B.2) excluding N/M = 2/2; this latter case in indicated by "2/2". FITS-Table-Formatted Version of the SSS Catalog The IRAS Small Scale Structure Catalog, as described above, is not in a format that translates easily into FITS Tables format (see Harten, R.H. Grosbol, P., Greisen, E.W. and Wells, D.C. 1988, A. & A. Suppl. 73, 365). The number of association fields varies for each object, and this runs counter to the stipulation that all records in a FITS table be of fixed length and have a single uniform format. In order to process the SSS Catalog into FITS format, the first three 80-byte records which always appear for each source were contatenated to form a single 240-byte record with a uniform format (in other words, NAXIS1=240 for the main data table). Then, the associations were removed from the main body of the catalog and written to a second file in a uniform format. Each association record was prefixed with (a) the SSS Catalog source name, and (b) the record number in the main data file where that SSS source could be found. If the FITS table formatted version of the catalog were loaded into some database managagement system, a join operation should be applied to "re-associate" the associations with the records of the main data file, either by using the source name as a key, or the main data file record numbers. Please note that the separation of the association fields from the main data file is only for the purposes of placing the catalog in FITS tables format. The original, multi-record format as described above is preserved in copies of the catalog that are not FITS formatted. -- Lee E. Brotzman, Astronomical Data Center