I. Introduction

6. Cautionary Notes

b. Point Source Catalog (PSC)

i. Origin and Quality of Photometry

Photometry of point sources is performed several different ways during 2MASS data processing, and the "default magnitudes" listed for each source in the PSC can have different origins depending on source brightness and environment. The source and quality of the "default magnitude" fields in the Second Incremental Release PSC are summarized by the "rd_flg", "bl_flg" and "cc_flg" fields. It is essential that users refer to these flags when interpreting photometry for any source in the Catalog. Each of these flags is comprised of three characters, each corresponding to one band; the first character is the J-band value, the second is the H value, and the third is the Ks value.

Faint Stars

The majority of sources in the 2MASS PSC have default magnitudes obtained using profile-fit photometry performed simultaneously on the combination of all six individual 1.3-s "Read 2-Read 1" (R2-R1) exposures. These sources are indicated with a rd_flg value of "2" in the appropriate bands. Occasionally, the profile-fitting photometry routines will fail for sources in crowded environments, or that lie in regions with complex backgrounds. If a valid aperture photometry magnitude is available it will be listed in the default magnitude field in the appropriate band. However, such magnitudes are highly uncertain. These objects have a rd_flg value of "4" in the affected bands.

Bright Stars

Sources brighter than 7-8 magnitudes will saturate in the 1.3-s R2-R1 exposures. These objects have default magnitudes from aperture photometry performed on the 51-ms R1 frames. Such sources have rd_flg=1 in the appropriate band.

Saturated Read 1 (R1)

Stars brighter than approximately fifth magnitude will saturate in even the 51-ms exposures. The default magnitude assigned for these objects is -99.999, since there is currently no useful photometric information available for these objects, and the rd_flg value for the appropriate band is "3".

Bright Star Filler Entries

The very brightest stars in the near-infrared sky saturate the detectors so heavily that they may not be unambiguously detected during processing. Placeholders for 18,122 bright stars have been included in the 2MASS Second Incremental Release PSC. The default magnitudes for these objects is always -99.999, and the rd_flg value is "8" in all three bands.

The following table summarizes the rd_flg values and their significance with respect to the origin of the "default magnitudes" in the PSC records. The "rd_flg" is a 3-character string, where the first character refers to the J-band, the second to H and the third to Ks. It is not uncommon for sources near the R1 or R2-R1 saturation levels or sources of extreme color to have different "rd_flg" values in different bands.

rd_flg Value (1 per band)
Source of Photometry
Not detected in that band - quoted magnitude is a 95% confidence upper limit
R1 aperture photometry
R2-R1 profile-fit photometry
Saturated in R1 - default mag=-99.999
R2-R1 aperture photometry - extremely uncertain!
Very bright star - default Mag=-99.999; position provided from external catalogs for informational purposes only

Blended Sources

The blend flag (bl_flg) included with each point source record indicates the number of components fit simultaneously in profile fit photometry (rd_flg="2"). This provides a measure of source density and possible confusion

bl_flg Value (1 per band)
Source not detected in that band
Single profile fit to isolated source
Multiple sources detected in small region and fit simultaneously. Value indicates the number of components fit

Bright Star Artifacts

The cc_flg also encodes several other conditions that indicate challenges to point source photometry, such as confusion during bandmerging, and during the duplicate source rectification procedure. The following table summarizes the possible values in cc_flg, and shows the number of sources in the Second Incremental Release PSC having each cc_flg value in each band.

cc_flg Value (1 per band)
Nature of Artifact or Confusion
Source is unaffected by artifacts, or source not detected in that band
Persistence Image from Bright Star
Nearby Diffraction Spike
Horizontal "Stripe" due to Bright Star
Confusion with nearby source, or during duplicate source rectification
Confusion in Bandmerging
Photometric Normalization problem
Unreliable source - probable meteor trail or noise detection

ii. Out-of-Field Bright Source Artifacts

The brightest stars can cast artifacts into adjacent Survey Tiles. Diffraction and scattered light masking has been carried out across Tile boundaries within the combined database of all Tiles in the Second Incremental Release, using positions of bright 2MASS detected stars and near-infrared bright stars drawn from several sources (cf. IV.7; some of which may have been too bright for 2MASS to detect as point sources). The external artifact "seed" list likely contains objects which, either due to an errant flux in the external catalogs or variability, was not a bright source in the 2MASS observations. This will yield an occasional mask on the sky which has no apparent associated extremely bright source.

iii. Meteor Trail Detections and Other Unreliable Sources

The source selection criteria used to generate the PSC are designed to minimize the number of spurious detections along meteor trails. A relatively small number of residual meteor trail detections persist in the the Catalog, though, because they satisfy the formal Catalog selection requirements. These residual sources can be efficiently identified as having no optical counterpart, a relatively high 2 value (>7) from the profile-fit photometry in one or more bands, and are detected in only one of the six 2MASS frames that cover each point on the sky. These criteria also describe the characteristics of other very faint "sources" which are likely spurious detections of noise. Sources that are such suspected unreliable detections are indicated in the PSC by having a cc_flg value of "U" in one or more bands.

iv. Photometric Uncertainties

Two measures of photometric uncertainty are provided for each 2MASS "default magnitude". The first, "<band>_msig" is the pure measurement error returned by the photometry algorithm (where <band> is the 2MASS band, either J, H, or Ks). Because of statistical fluctuations, faint sources will occasionally have unphysically low measurement errors. The second photometric uncertainty quoted in the PSC is "<band>_msigcom", which is the root-sum-square combination of "<band>_msig" with the uncertainty in the nightly photometric zero-point offset, the estimated flat-fielding residual (0.005 mag), and the R1 photometric normalization uncertainty, for bright sources. The "<band>_msigcom" provide a better estimate of total photometric errors.

5% of the bright H-band sources have approximately 50% larger than typical uncertainties due to an as yet unidentified problem with one of the production PSF's used to fit the point source profile for a particular image size. The fluxes for these sources appear not to be biased in any significant way relative to the rest of the survey sources. Similarly about 15% of the Ks-band sources appear to have about 30% smaller than typical uncertainties due to a similar phenomenon.

v. Cross-scan Photometric Bias

The profile-fit photometric algorithm is sensitive to small variations in the shape of the PSF across the focal plane. As a result, fluxes of sources in the Second Incremental Release PSC at the extreme east-west edges of the arrays can be biased as much as a few percent relative to the center of the scan.

vi. High source density regions

In crowded regions the source extraction threshold is automatically raised in response to elevated confusion noise so that extracted source flux limits are brighter than those that prevail at high galactic latitude. Analysis has shown that the PSF extraction uncertainties are a reasonable representation of the confusion noise in these dense regions.

In areas of high source density, the probability is high that an image artifact from a bright sources will fall on top of one or more real sources. The process of artifact identification will result in many real sources being filtered out of the Catalogs, or at least flagged as being affected by artifacts. This is especially apparent in the cores of globular clusters, and near the galactic center.

Aperture photometry in regions with source densities in excess of ~65,000 deg-2 can be corrupted due to confusion in both the source apertures and sky annuli. This can result in errors in the empirical aperture curve-of-growth derivation in a Tile that produce photometric biases in one or more bands and therefore color biases of up to ~0.2 mags. This bias is spatially correlated within a Tile because the curve-of-growth corrections are derived for and applied to all sources having the same seeing shape, and the seeing is estimated on spatial frequencies no smaller than the scale of an Atlas Image (17 arcminutes in declination). Users performing statistical analyses of source colors in within 10°-15° of the Galactic center, or any other region where the source density can exceed ~65,000 deg-2 over a significant area, should be aware that such biases can be present. See Section IV.4c for a detailed discussion of this bias.

vii. H-band Detection Thresholds and Atmospheric OH Airglow

The point source extraction threshold algorithm that responds to confusion noise in high source density regions also responded to elevated noise due to low frequency structure in the image backgrounds due to OH airglow. This results in an effective loss in sensitivity in the H-band relative to J and Ks in severe airglow conditions, leading to a number of J and Ks detected sources.

An improved noise estimator that is not sensitive to the large scale structure in the airglow backgrounds was introduced in pipeline processing in early March of 1999. Northern data acquired after 21 September 1998 UT and southern data after 13 January 1999 UT have been processed using the new noise estimator and should be less sensitive to large scale structure. During the reprocessing phase after the end of survey observations, all data will be rerun with the new noise estimator.

viii. Solar System Objects

Sources which are at the positions of known asteroids, comets, planets or planetary satellites at the time of the observations are flagged in the "mp_flg" column. These are positional associations, and not necessarily identifications. Therefore, they remain in the PSC. Some fraction of the putative detections are not valid, but are chance superpositions of the predicted positions with background sources. This is particularly likely at low Galactic latitudes where the density of background stars is large. See Section IV.9 for more details.

ix. Extremely faint sources

Sources with SNR>7 fluxes in any one band were included in the PSC. Fluxes in the remaining two bands may lie well below this threshold, may be entirely dominated by noise, or may be reported as upper limits. The magnitudes in these remaining bands can be unphysically faint, or will have associated uncertainties that imply a non-detection (i.e. 95% confidence upper limits).

x. Optical Associations

Every 2MASS source was positionally correlated with the ACT or USNO-A optical catalogs. At high latitudes approximately 90% of the sources have optical counterparts within 5´´. For convenience, the 2MASS PSC contains optical B- and R-band magnitudes from USNO-A or B- and V-band magnitudes from ACT, and positional offset information from the detected 2MASS source. These are positional associations, however, and not necessarily identifications.

xi. Position Reconstruction

2MASS positions are tied to the International Celestial Reference System (ICRS) via the ACT Reference Catalog on a tile-by-tile basis. Typically there is an abundance of ACT stars in a Tile and the astrometric solution yields positions accurate to <0.2´´ for sources with SNR>20. A small fraction of Tiles contain very few astrometric reference stars, or the reference stars may cover only a part of the Tile. In these cases, the positional solution may random walk as much as 1´´ from the ICRS frame. The quoted astrometric uncertainties in the PSC should reflect this random walk.

There is also known to be a ~0.09´´ bias in the reconstructed declinations of the bright R1-only-measured stars relative to fainter stars. No measurable bias exists in right ascension.

[Last Updated: 2000 June 20; by R. Cutri and M. Skrutskie]

Previous page. Next page.
Return to Explanatory Supplement TOC Page.