WISE acquired approximately 12 independent Single-exposure images covering each point on the Ecliptic plane during each sky coverage epoch. The per epoch coverage increases towards the Ecliptic poles because the scan paths converge, reaching ~3000 directly at the north and south Ecliptic poles. As a result, the net exposure time per AllWISE Catalog source varies significantly around the sky, as will the corresponding sensitivity and completeness of the Catalog. The frequency distribution of source depth-of-coverage in the AllWISE Catalog, compiled from w?m parameter, is shown in Figure 1. Sources near the Ecliptic have a median coverage of ~24 in W1 and W2 because of the addition of the two epochs for those bands in AllWISE. The maximum Catalog coverage is ~300 for W1 and W2 and ~180 for W3 and W4 because of the artificial coverage attenuation used when processing the 73 Tiles closest to the Ecliptic poles (see V.3.a.iv).
|Figure 1 - The frequency distribution of source depth-of-coverage, per band, in the AllWISE Source Catalog. The blue, green, magenta and red lines denote the W1, W2, W3 and W4 coverage, respectively. The secondary peak near 160 is caused by the artificial coverage attenuation used when processing the Tiles closest to the Ecliptic poles.|
As illustrated in Figure 1 above, the sky coverage depth for sources in the AllWISE Catalog and Reject Table is approximately twice as large in W1 and W2 as it is in W3 and W4. AllWISE combined W1 and W2 Single-exposure images from the WISE 4-Band Cryo, 3-Band Cryo and NEOWISE Post-Cryo survey phases, and W3 and W4 images from the 4-Band Cryo phase only. The additional epoch of W1 and W2 coverage accenuates the weight of those two bands in determining source properties such as position and motion. The discrepant time coverage between W1/W2 and W3/W4 can also result in aberrant source motion measurement (see II.6.e.iv).
Sky coverage for AllWISE is discussed in IV.2. The addition of the W1 and W2 images from the second complete sky coverage epoch essentially fills in the few low coverage gaps that were present in those bands in the All-Sky Release Catalog. However, low coverage regions in W3 and W4 remain approximately the same as in the All-Sky Release Catalog because there are no additional images used in those bands.
The sensitivity improvement resulting from the doubling of the
coverage depth in W1 and sometimes W2 is not the full 0.4 mag
where confusion noise limits source detection. At high galactic latitudes,
this occurs in W1 for coverage >60-70, and slightly larger in
W2, >80-90. These depths are nominally reached at Ecliptic latitudes
of ~60° and ~75°, respectively. The AllWISE Catalog
is confusion limited at much brighter flux levels in the
very high source density regions in the Galactic Plane, even when the
depth-of-coverage is much lower (e.g.
It was not possible to the full AllWISE source extraction processing to the full available depth-of-coverage for Tiles close to the Ecliptic poles because of memory usage and runtime limitations. Therefore, the coverage depth in those Tiles is artificially attenuated to ~165 frames. The AllWISE Catalog and Reject Table entries in those Tiles are therefore capped. This accounts for the second peak in coverage frequency distributions shown in Figure 1 above. This is well beyond confusion-limited coverage depth for in W1 and W2, but not W3 and W4. Full-depth Atlas Images are available for these Tiles, but not the corresponding full-depth source extractions.
The AllWISE Source Catalog is not a "Point Source" Catalog. It contains detections of point-like objects, such as stars and unresolved galaxies, as well as resolved sources such as close multiple stars, galaxies, and detections of sections of large nearby galaxies and clumps or filaments in Galactic nebulosity, as long as they meet the Catalog selection criteria described in V.4.
AllWISE Source Catalog and Reject Table entries that may be resolved relative to the WISE PSF in any band are identified by having an ext_flg value >0. The profile-fitting (w?mpro) and curve-of-growth corrected standard aperture photometry (w?mag) are optimized for point sources and may underestimate the true brightness of extended sources. See II.1.d.i for a discussion of alternative measurements for resolved objects.
Relatively conversative criteria were used to select the deep extractions from AllWISE Multiframe Pipeline processing to be in the Source Catalog to strive for high reliability. However, as discussed in II.4.b, the reliability at the SNR=5 level that was required for the Catalog is still generally >90% in most regions. Therefore, a significant number of real, faint source detections may be found in the AllWISE Reject Table.
To be included in the AllWISE Source Catalog, a source extraction was required to have a reliable flux measurement with SNR≥5 in at least one band (see V.4.c). Sources with SNR≥5 in one band frequently have SNR<5 fluxes in other bands.
Very bright sources effectively mask the sky in their vicinity. Detections of fainter sources are suppressed because the local background levels are elevated from the scattered light halos, and because of confusion with diffractions spikes, latent images, and optical ghost images from the bright sources. Furthermore, even when fainter sources are detected near very bright objects, they may be flagged as contaminated and inadvertently filtered from the Catalog. The impact of masking by bright sources is discussed in IV.2.c.ii.
Extraction of very faint detections near very bright sources is additionally suppressed because of an error in the profile-fit source extraction process that effectively disabled subtraction of bright source fluxs prior to fitting fainter nearby detections. As a result, the fits to the fainter detections were attempted with flux from the nearby brighter source still present. The chi-squared minimization naturally pulled the position of the faint detections towards the photo-center of the bright sources, so the faint detection was erroneously extracted at nearly the same position and with nearly the same flux as the original brighter source. This is the origin of the small-separation, same Tile (SSST) sources that are removed during the duplicate resolution step in Catalog Generation (see V.4).
The reliability of AllWISE Source Catalog is estimated to be >99.9% for sources brighter than SNR=20 in unconfused regions of the sky (II.4.b). The fractional reliability decreases for fainter objects and in regions where there is less coverage. Even at this very high level of reliability, the size of the AllWISE Catalog means that it may contain thousands of entries that do not correspond to an actual source of infrared radiation on on the sky.
The source selection criteria used to construct the Source Catalog (V.4) were designed to select reliable extractions from the superset in the Working Database produced by Multiframe Pipeline processing. These criteria we intended to exploit the WISE survey strategy of obtaining multiple, independent exposures of each region of the sky, conservative detection limits and flagging of artifacts in processing pipeline.
When in doubt about the reliability of any entry in the AllWISE Source Catalog, the very best way to verify that is is real is to look at its position in the AllWISE Atlas Images.
Here are some tips for selecting preferentially reliable entries in the AllWISE Source Catalog.
If the associated measurement uncertainty in a band is null for any of the different AllWISE Catalog photometry, then the quoted magnitude is computed from the 2σ flux upper limit. Care must be taken to not mis-interpret upper limits as detections. A special exception to this is the case of missing W1 flux uncertainties for some sources observed during the early 3-Band Cryo phase.
As described in II.1.d.i, several different types of photometric measurements are presented in the AllWISE Source Catalog and Reject Table. The best photometry for the great majority of objects is the deep detection profile-fit measurements that are made by minimizing the PSF template chi-squared on all Single-exposures and in all bands covering a source, simultaneously. The few exceptions to this are noted below.
The fluxes for sources that are brighter than W1<8 mag and W2<7 mag, the nominal saturation limits, and that were observed during the NEOWISE Post-Cryo phase (89.4°<λ<221.8° or 280.6°<λ<48.1°), have much larger uncertainties and exhibit more scatter in the AllWISE extraction processing as the measurements from just the 4-Band Cryo data that were used to produce the WISE All-Sky Release Catalog. This is mostly likely because of untracked changes in the wings of the W1 and W2 PSF as temperatures increased during the Post-Cryo phase.
As described in V.1.b.ii, W1 images taken during the early part of the 3-Band Cryo survey phase began to exhibit an increasing number of hard-saturated pixels because of the rising temperatures. Interaction between the saturated pixels that did not correlate with objects on the sky and saturated pixel masking in the source extraction process resulted in the profile-fit W1 flux uncertainties for some sources being set to zero and sometime being elevated above the level expected for a given flux level. The flux measurements themselves were generally okay, but when the flux uncertainties were zero, the measurements were interpreted to have SNR=0 and therefore the sources appear to be non-detections in W1.
Cases for which the W1 flux uncertainty is missing are easily identified in the AllWISE Catalog and Reject Table by having W1 photometric quality flag of ph_qual='Z'. For these cases, the profile-fit magnitude w1mpro is usually valid, and the averaged Single-exposure profile-fit photometry W1 magnitude uncertainty, w1sigp2 can be used as a proxy for the deep detection profile-fit magnitude uncertainty.
Cases in which the W1 profile-fit photometry flux uncertainty is merely elevated cannot be uniquely identified. See V.1.b.iii for a general description of their characteristics. In general, these objects will have one or more of the following symptions, but these symptions do not guarantee that the W1 flux uncertainty will be affected:
The large, brightness-dependent flux underestimation that affected sources in the All-Sky Release Source Catalog (see VI.3.c.i.2 in the All-Sky Release Explanatory Supplement) has been largely corrected in AllWISE because of the recalibration of the Single-exposure image pixels. However, a small residual flux underestimation bias remains in AllWISE measurements because of background overestimation caused by contamination of nearby sources in the background measurement annulus. The error that disabled the source flux subtraction in profile-fit measurement processing also negated the effect of the recursive background estimation that was intended to address this component of the background measurement problem.
The small remaining flux underestimation bias in AllWISE photometry is brightness dependent, and grows to approximately 0.1 mag in W1 at the faint detection limits. There is little if any bias in W2 in lower source density regions. The strength of the bias in all bands increases as local source densities increase, though. In very high source density fields, the maximum bias may be as large as 0.2-0.3 mag at the detection limit.
The differences between AllWISE W1 and W2 magnitudes and Spitzer IRAC band 1 and band 2 magnitudes for the same objects in the Chandra Deep Field - South (CDF-S), are plotted as a function of IRAC magnitude in Figure 2. The same comparison with Spitzer using the WISE All-Sky Release Catalog photometry is shown in Figure 3, which shows a much larger flux underestimation bias.
|Figure 2 - The difference between AllWISE Catalog W1 (top) and W2 (bottom) profile-fit magnitudes and Spitzer IRAC band 1 and band 2 aperture magnitudes for the same sources in the CDF-S field plotted as a function of the IRAC magnitudes. The sample plotted is limited to objects with IRAC colors in the range -0.1<[IRAC1-IRAC2]<0.05 mag to avoid biases due color terms between the WISE and IRAC photometric systems. Small white points are individual sources, blue triangles are the trimmed average magnitude differences in 0.25 mag wide bins, and the magenta dashed line shows the mean quoted WISE magnitude uncertainty as a function of magnitude.||Figure 3 -The same comparison as in Figure 2, but for WISE All-Sky Release Catalog photometry. Note the much larger flux underestimation bias in the older WISE photometry.|
The profile-fitting source extraction for AllWISE attempts to actively deblend a single source detection into two components if the chi-squared goodness-of-fit to a single PSF is >1.5. If the resulting two-component fit reduces the chi-squared from the initial one, and the two-component chi-squared is <3 (from an initial value >3), the two sources are reported. In AllWISE processing, active deblending was limited to adding at most one component. Actively deblended sources in the AllWISE Source Catalog and Reject Table can be identified by having the parameter na=1.
The profile-fit photometry measurement model includes a constant term that accounts for the uncertainty in the PSF template. The PSF uncertainty term dominates the measurement precision at high SNR levels where photon noise is negligible. As a result, Catalog and Reject Table w?snr values do not exceed ~60 even for extremely bright but non-saturated sources.
The aperture photometry measurements in the AllWISE source extraction have no compensation for saturated pixels. Consequently, the aperture photometry measurements will systematically underestimate true flux for sources brighter than the nominal saturation limits of 8, 7, 3.8 and 0.4 mag in W1, W2, W3 and W4, respectively. The systematic and increasing flux underestimation with increasing brightness is illustrate in Figures 5a-d in II.1.d.i.
The aperture photometry measurement quality flags are provided to assess if one of more pixels in the measurement apertures may be contaminated for the reasons cited above. The flags are provided for all of the different type of aperture measurements, w?flg, w?flg_[1-8], w?gflg.
AllWISE measures apparent motion on the sky, not proper motion. Because WISE observes near 90° solar elongation, objects are observed at their maximum parallax factors. So an object observed in two epochs separated by six months will have a motion that is half of its yearly proper motion and twice its parallax. For nearby objects, parallaxes can therefore alter the apparent motion significantly. See II.6.b.i for a detailed discussion.
As with flux and position measurements, the accuracy of the motion measurement is highest for the brightest, non-saturated sources, and decreasing with decreasing SNR. WISE observed most of the sky two complete times, but 20% was observed three times. Objects in the three-epoch coverage region will have better motion accuracy.
Known issues that can affect the accuracy of the apparent motions reported in the AllWISE Source Catalog and Reject Table are described in II.6.e. Any factor that affects the precision of source position determination, such as confusion or source blending, extended sources, lower depth-of-coverage will also tend to degrade the quality of the measured motion. In addition, extremely red sources whose W3 and W4 emission is dominant can adversely affect motion measurements because the W3 and W4 data come entirely from one measurement epoch.
As described in V.3.v.ii.3, the finite motion radial step size of 0.0275 arcsec/year used in the profile-fit chi-squared minimization procedure leads to quantization of the resulting motion measurements. The quantization level is smaller than the typical measurement uncertainty.
The profile-fit chi-squared minimization process allows a maximum of 250 iterations. The step size in motion space is 0.0275 arcsec/year. Large, spurious motion values will therefore tend to pile-up at 250*0.0275 = 6.875 arcsec/year. There can be larger motions reported. See V.3.v.ii.4 for more details.
When single deep source detections are deblended during the profile-fit source extraction in AllWISE processing (i.e. active deblending), an apparent motion measurement is possible only for the original, or primary source detection. This limitation occurs because the information for added source component is output before it can come up in the normal order for the source fitting. Hence, it is common to find that Catalog and Reject Table entries that have the active deblend parameter, na=1, to have null values in the various motion-fit parameter columns (i.e. ra_pm, dec_pm, pmra, pmdec, etc.).
Early in the WISE survey, the spacecrafts' magnetic torque rods were enabled to dump accumulated momentum when scans approached within 45° of the Ecliptic poles. Activating the torque rods resulted in a small jump in the telescope pointing and smearing of the resulting images. Most of the smeared images were given poor quality scores and were excluded from AllWISE Multiframe Pipeline processing. However, a few images with slight smearing were included, and the resulting scatter in source positions for some of the images produced poorer Multiframe position reconstruction. (see V.2.c in the All-Sky Release Explanatory Supplement).
Because of a small residual band-to-band offset that was not removed by the Multiframe Position Reconstruction improvements for AllWISE, the reconstructed position of rare, very red sources that are detected primarily in W3 and/or W4 may be offset from bluer sources by up to ~70 mas in the in-scan (ecliptic longitude) direction. The sign of the offset is in the sense that the ecliptic longitude positions of very red sources will be slightly larger than bluer sources. See II.5.d.i for more details.
The identification and flagging of flagging of spurious and contaminated sources in the AllWISE Multiframe processing was both over- and under-aggressive in some circumstances, and inaccurate because of imperfect knowledge of the geometry of artifacts at large distances from very bright objects. In cases were the flagging was overly aggressive because of overestimation of the parent source brightness, fainter nearby objects may have been flagged as spurious detections of artifacts and omitted from the Catalog. These will be in the Reject Table, though. In cases were the parent source source brightness was underestimate, or the position of the parent source was not known accurately, then spurious detections of artifacts may have gone unflagged, and wound up in the AllWISE Catalog.
A case where the flagging is inaccurate because of uncertainties in the position angle behavior of diffraction spike very far from the parent source is illustrated in Figure 4. AllWISE images are shown for a WISEA J231943.74+173202.6, which is a source, that is clearly visible and detected in W1 and W2, but that is superimposed on a diffractions spike from a very bright object, RAFGL 3068 that lies 21.8 arcmin to the southwest. This object has >2σ flux detections in W1, W2 and W3, but the W3 detection is really a measurement of the flux from the diffraction spike. This object has cc_flags='0000'.
|Figure 4 - 600x600 arcsec region of AllWISE Atlas Image at the position of WISEA J231943.74+173202.6. This is a real object that is detected in W1 and W2 but that has a spurious W3 "detection" because it is superimposed on a diffraction spike from the very red, bright source RAFGL 3068 that lies 21.8 arcmin to the southwest.|
The accuracy of artifact identification and in the AllWISE processing depends on the knowledge of the flux and position of the bright "parent" source that produces the image artifacts. Most inner solar system objects that move significantly between WISE Single-exposures are suppressed in the coadded Atlas Images, so usually do not trigger detections that persist into the Catalog. However, a few of the very bright moving objects (and their artifacts) such as Mars, Jupiter, Saturn and the brightest few main belt asteroids are not completely filtered out by the pixel outlier rejection and may cause spurious source extractions. This is especially true in W3 and W4 where these objects are brightest, and where they do not benefit from a second coverage epoch when the objects had moved out of the field of view.
Detections of residual artifacts caused by Mars are illustrated in Figure 5 where the spatial distribution of AllWISE Source Catalog entries in a 2°x2° region around the position where the planet was observed during the WISE 4-band Cryo phase is shown. The diagonal pattern is caused by Mars' motion between Single-exposures.
The mean locations at which Mars, Jupiter and Saturn were observed during the 4-Band Cryo survey phase is given in Table 1, along with the AllWISE Atlas Tile in which those locations are found.
|Planet||RA (deg)||Dec. (deg)||Atlas Tile|
|Figure 5 - The distribution of AllWISE Source Catalog objects in a 2°x2° region around the location of Mars when it was observed during the WISE 4-Band Cryo phase. The pattern of excess Catalog sources is caused by spurious detections of artifacts that persist into the AllWISE coadded Atlas Images. The cross-shaped region near α,δ=134.166,+19.849 with missing sources is caused by the bright star T Cnc.|
The Profile-fit (w?mpro) and standard aperture (w?mag) photometry in the Source Catalog are optimized for point-source characterization. They both systematically underestimate the brightness of sources that are extended relative to the WISE PSF. Catalog sources suspected of being resolved are identified by having ext_flg values >0.
Large aperture photometry provided for Catalog entries may provide better brightness estimates for extended objects. In addition, for WISE Catalog sources that are associated with objects in the 2MASS Extended Source Catalog (XSC), photometric measurements made using elliptical apertures scaled from the object size and shape in the 2MASS XSC are provided in the Catalog (c.f. II.1.d.i) .
The source detection and extraction steps in AllWISE processing may split large, extended objects such as large galaxies and galactic nebulae into multiple pieces. Extractions of any and all pieces that that satisfy the source selection criteria will appear in the Catalog. The Catalog ext_flg column can be used to identify AllWISE sources that may be sections of large extended objects.
The brightness measurements given in the AllWISE Catalog and Reject Table entries represent the most likely flux levels in all of the Single-exposure observations that cover the sources. Objects whose flux levels varied during the time interval covered by the individual WISE observations will therefore have a mean brightness reported in the Catalog, and may have larger than expected flux uncertainties or profile-fit chi-squared goodness of fit values. The probability of flux variability in each band is tabulated in var_flg parameter. A detailed description of the algorithm used to estimate the probability of variability and known limitations of the estimation are given in V.3.b.vi. Examples of how to use the var_flg parameter in Catalog queries are given in II.3.c.
Objects that are resolved with respect to the WISE PSF or that are contaminated by image artifacts will exhibit more intrinsic scatter in repeated Single-exposure measurements than clean, isolated point sources. As a result, the probability of variability encoded in the var_flg may be overestimated for extended and/or contaminated objects. Extended sources can be identified by having ext_flg>0. Source contaminated by image artifacts have cc_flags!='0' in any band that is contaminated.
As described in section VIII.3.e.iv of the NEOWISE Post-Cryo Release Explanatory Supplement, the Post-Cryo W1 and W2 profile-fit flux measurements of sources become systematically brighter than those the measurements of the same sources on 4-Band and 3-Band Cryo phase images beginning at the nominal saturation limits of W1<8 and W2<7 mag. The systematic bias between Cryo and Post-Cryo Single-exposure measurements can give a false indication of flux variability. For this reason, the Post-Cryo Single-exposure measurements for sources brighter than the saturation limit are are not used when assessing the probability of the flux variability for those objects.
The AllWISE Source Catalog and Reject Table entries include photometry for the closest entry from the 2MASS Point Source Catalog (PSC) that falls within 3" of the AllWISE source position (II.1.f.i). These are associations not identifications. Although the position accuracy of the two Catalogs is very good, there is a non-zero probability of chance associations between physically unrelated objects, as well as missed associations between the Catalogs. This is particularly true for objects with high proper motions, and those in regions with large foreground extinction. You should always confirm associations by reviewing carefully the entries from both Catalogs.
For the AllWISE-2MASS cross-matching that produces the association information in the AllWISE Catalog and Reject Table, possible proper motion of objects in the 11 years separating the epochs of the two surveys is not taken into account. Therefore, associations for objects that move at a rate >0.3 arcsec/yr may be missed.
As described in V.2.b, proper motion of the 2MASS PSC astrometric reference stars was taken into account when refining the Single-exposure image position reconstruction solutions for AllWISE source astrometry.
As a rule of thumb, you should assume that every entry in the Reject Table is spurious until you have verified its reliability by visually examining its position on the Atlas Images.
As the name implies, the AllWISE Reject Table contains primarily spurious detections of low SNR noise excursions, image artifacts produced by bright sources and transient single frame events such as cosmic rays and satellite trails. However, the Reject Table also contains detections of real, low SNR sources, reliable measurements of brighter sources that were rejected from the Catalog because of overly conservative artifact flagging or because they fall in very low frame depth-of-coverage regions. Used with caution, the Reject Table can be a useful database that augments the AllWISE Source Catalog, particularly when following-up sources that appear on the Atlas Images but are not found in the Catalog.
These may be Tile overlap duplicates that occur when a source lies in the overlap region between adjacent Tiles, or small-separation, same-Tile (SSST) duplicates that occur when a faint, usually spurious noise detection is pulled towards the location of a much brighter nearby source.
One unique member of each duplicate extraction was selected during the duplicate resolution process for AllWISE Catalog Generation. The other, redundant extractions remain in the Reject Table. Reject Table entries that are duplicates of sources that are in the Catalog have use_src=0. Duplicates in the Reject Table that were rejected because they are spurious SSST extractions have rel="r".
Last Updated: 7 October 2015