The level-1 specification for photometric precision for extended sources is better than 10% (0.109 mag) for sources at H < 13.8 mag. The 2MASS data appear to meet this requirement as stated and mostly also as extended to other bands.
This spec sets out standards for realized photometry including all sources of error, and is meant to indicate the accuracy of photometry for moderately bright sources that might be used for Tully-Fisher studies, for example. The spec is interpreted to mean the error of sources in the half magnitude bin from 13.3 < H < 13.8. The spec was based on assumption before the survey began that H-band was to be preferred for Tully-Fisher work, but since subsequent indications are that J and K are probably at least as good, the spec could be extended to the precision of J and K magnitudes in a similar magnitude interval relative to the sensitivity limits.
We can test this longer-term accuracy fairly well by comparing data in the overlaps that was taken at well-separated times. We show below the uncertainty as calculated in the analysis of extended source sensitivity (previous section), but for sources in the appropriate magnitude intervals where the duplicate measurements were made at least one month apart. Between 2700 and 5000 sources were available for each test.
Note that all of the magnitudes have better than 10% errors. Even if we consider the contribution of a potential 0.044 mag uncertainty in the uniformity suggested in the next section, this source of error (added quadratically) would only push the K-band isophotal magnitude slightly beyond the 10% level.K-band 20 mag/sq arcsec 7" circular isophotal magnitude aperture magnitude uncertainty uncertainty 14.1 < J < 14.6 0.087 0.064 13.3 < H < 13.8 0.099 0.080 12.5 < K < 13.0 0.108 0.067
We can also test the consistency between northern and southern telescopes by the same procedure. In this case between 300 and 600 comparisons were available in each case.
Again all of the magnitudes have better than 10% errors, and are highly consistent with the comparisons at a single telescope. Since these comparisons involve many additional variables (airmass, season, arrays, etc.), this appears to be good confirmation that the survey has met the photometric precision specification.K-band 20 mag/sq arcsec 7" circular isophotal magnitude aperture magnitude uncertainty uncertainty 14.1 < J < 14.6 0.090 0.068 13.3 < H < 13.8 0.098 0.075 12.5 < K < 13.0 0.106 0.068