2MASS All Sky Release:
Comparison with Level 1 Specifications

The 2MASS project has prepared an all-sky release very similar in characteristics to previous incremental releases. This release increases the integrated sky coverage in the public domain from 47% to virtually 100% with improved processing of all sources. The All-sky Release data products include a Point Source Release containing positions and brightness information for over 470 million objects, an Extended Source Catalog containing positions, magnitudes and basic shape information for nearly 1.6 million resolved sources, and an Image Atlas containing nearly 5 million J, H and Ks images covering the survey area.

In constructing this document, the 2MASS team has drawn upon and summarized both internal and Explanatory Supplement documentation. To provide the most complete information, many of the links here point to the original detailed analyses. In all cases, however, the 2MASS team has arranged this page so that all primary information is available in summary in the first level below this front page.

The 2MASS All Sky Release includes improved processing of data from the area contained in the Incremental Data Releases, and is intended to supercede these earlier releases. This summary outlines the major improvements in the final data processing pipeline.

The 47% of the sky delivered in the Second Incremental Release has been in the public domain since March 2000 and has led to hundreds of publications -- with most now coming from outside the 2MASS collaboration. This rigorous practical testing of the preliminary product has yielded only minor comments on the data's shortcomings. All evidence points to the final release being of better quality than its predecessor. For reference, this link documents the reports associated with the development and validation of the final pipeline.

A release preview page supplies the details of sky coverage and source counts. The Explanatory Supplement (currently under construction) will be available as web pages, as was the case for the previous release. Section I.6 is of particular interest. It contains a compilation of caveats that will be of interest to users of 2MASS data products. The relevant subsections are:

Below is a summary of science team analysis of the release characteristics in the context of the Survey's Level 1 specifications.

2MASS photometry has a dynamic range of >20 magnitudes owing to extraction software that address three different exposure regimes. The Survey's nominal frame exposure time was 1.3 seconds obtained by differencing two reads of the NICMOS3 array separated by that time interval. Both reads were independently recorded and saved for future processing. The first read (often referred to as "Read1" in subsequent documents) occurred 51 milliseconds after array reset and represents an effective 51 millisecond exposure image. The second read ("Read2") occurred 1.351 seconds after reset. The point source database includes a rd_flg flag which distinguishes between different techniques for populating the "default magnitude" column for sources depending on their level of saturation (all XSC photometry derives from the 1.3 second Read2-Read1 difference). Since different techniques can be used in different wavelength bands for a given source, the rd_flg is a three-character flag with an independent value for each band revealing the source of the default magnitude (i.e. j_m, h_m, or k_m) column in the point source release database.

  • rd_flg=2: This photometry derives from PSF profile-fitting applied to the 1.3 second integration time Read2-Read1 exposure difference and represents the bulk of the default photometry in the 2MASS point source release. Subsequent documentation may cite either "Read2-Read1" or "Read2" photometry in referring to source extractions from the 1.3 second exposures. Although the rd_flg=2 default magnitude is based on the psf fitting algorithm, aperture photometry is also provided in separate columns because aperture magnitudes can be superior for bright stars and in circumstances where the psf fit is poor. Typically, rd_flg=2 sources have magnitudes ranging from about 9th magnitude to the faint detection limit at around 16th magnitude. The saturation threshold is band specific with Ks-band tending to saturate for sources nearly one-half to one magnitude brighter than for the J and H bands.

  • rd_flg=1: This photometry derives from aperture photometry operating on the 51 millisecond image obtained immediately following array reset. The primary objective was to extract reliable flux for sources saturated in the 1.3 second Read2-Read1 exposure. Subsequent documentation may cite "Read1" photometry in referring to unsaturated source extractions from the 51 millisecond exposure. Magnitudes for "Read 1" extractions typically range from 4th to 9th magnitude.

  • rd_flg=3: This photometry derives from estimating source flux from the wings of a stellar image completely saturated in even the 51 millisecond Read1 exposure. 2MASS successfully photometered sources as bright as -4 magnitude using this technique. In step with the jargon above, photometry extracted with this method is often referred to as "Read3" photometry when, in reality, there was no third readout of the system.

  • The photometry of sources is carried out in each band (JHK) independently. A process called band merging combines the bands into a single database row with common coordinates. The band merging process may give inconsistent results when close pairs are inconsistently deblended by the software in each band. This sort of confused band merge is indicated by rd_flg=6. Band merging can occasionally fail outright in which case two (or even three) distinct sources with rd_flg=1,2 or 3 in one or two bands lies spatially adjacent to another source with rd_flg=1,2, or 3 in two or one band.

This release includes reliable sources fainter than the Level 1 specification 10-sigma flux limits (as was also the case for the Second incremental release). In order to distinguish between "catalog" sources and the fainter content of the release point source database, the 2MASS project has created a new column/flag -- photometric quality (ph_qual). This flag contains three characters, one for each band. The ph_qual designation, as the name suggests, represents the photometric uncertainty in the measurement -- with "A" representing measurements with photometric uncertainty consistent with the Level 1 catalog specifications. "B","C","D", and "E" represent successively poorer measurements. "Catalog" sources include

  • Sources receiving photometric quality of "A" (SNR>10 and sigma<0.109) in any one of the three wavelength bands.

  • All "Read1" sources (i.e. saturated in the 1.3s "Read2-Read1" exposure) successfully extracted by the pipeline (ph_qual = A, B, C, D, or E) and not identified as artifacts. Read1 sources which do not receive an "A" in any of the three ph_qual entries technically do not meet the Level 1 requirements individually. The Science Team concluded it was more important to include these 14173 bright reliable sources at the "catalog" level for the sake of completeness.

  • All "Read3" sources which completely saturated even the 51ms "Read1" exposure. The uncertainties for Read3 extractions are substantially larger than the Level 1 uncertainty requirements for bright sources, however the Level 1 specifications place no requirements on these thoroughly saturated sources. Once again, the Science Team opted to include these sources at the "catalog" level since they were highly reliable and improve completeness at the brightest magnitudes.


Category Requirement Performance
Point Source Catalog
Photometric sensitivity 10-sigma at 15.8, 15.1, 14.3 mag at J, H, Ks respectively for abs(b)>10o met for full unconfused sky;
exceeded for most of sky
Photometric uniformity <4% maximum bias in photometric zeropoint around the sky <2% achieved
Photometric precision <5% 1-sigma for bright stars unsaturated in the 1.3s exposure <3% achieved
Astrometric accuracy <0.5" 1-sigma relative to the reference frame <0.1" achieved
Completeness >0.99 at 10-sigma sensitivity limits met
Reliability >0.9995 0.9997 demonstrated in test areas; no known source of unreliability
in excess of specification
Bright star photometry

(Read1 / rd_flg=1)

 <2% bias at Read1 saturation limit (Ks~4.0 mag)
<5% 1-sigma repeatability at Read1 faint limit (Ks~8.0 mag)
<10% 1-sigma repeatability at Read1 saturation limit (Ks~4.0 mag)		 <4% at worst

2% achieved

2% achieved
Extended Source Catalog
Photometric sensitivity 10-sigma at J<15.0 mag, H<14.2 mag, Ks<13.5 mag met
Photometric precision <10% 1-sigma repeatability for H<13.8 mag 7-10% achieved
Photometric uniformity <10% maximum bias around sky <4% achieved
Completeness >0.90 for |b|>30 degrees met
Reliability >0.80 for 10<|b|<20 degrees

>0.99 for |b|>20 degrees

achieved 0.93-0.95

achieved 0.992-0.995
General Survey
Sky coverage>95% 99.998% imaged;
99.5% coverage in point sources;
98% coverage in extended source
Other Information
Galactic Plane performance No Requirements