The Level 1 specification for photometric uniformity is better than 10% over the sky. The 2MASS data appear to meet this requirement, although further tests are needed.
Testing the uniformity is not straightforward, because gradual drifts in, for example, calibration or airmass effects are difficult to isolate. While external checks are useful, none are available which are themselves immune from these problems, and using these introduces further uncertainties because they are typically limited to small regions of the sky and the measurements are carried out using different methods.
One internal check in 2MASS is galaxy colors. Because the galaxy population exhibits a fairly narrow range of (J-Ks) colors at a fixed magnitude, we expect large samples of galaxies with the same magnitudes to have the same mean color. [(J-H) colors have very little dependence on redshift, so they do not provide a similar test.] Extinction by interstellar dust reddens the colors, so we limit the analysis here to sources at galactic latitudes |b|>30°. Two obvious areas for concern are differences between the north and the south, and airmass effects at the celestial poles.
We selected a set of galaxies somewhat arbitrarily by requiring the 7´´ circular aperture J magnitudes to be between 14 and 15 mag. (This range was chosen to give small errors of the mean and large magnitude-dependent color changes below.) Comparing the mean (J-Ks) colors for the northern and southern galaxies, no significant differences were apparent:
|mean value||north-south||error of the mean|
|(J-Ks) in 7´´ aperture||1.11916||0.00056||0.00320|
The errors of the mean listed here is based just on the uncertainties in calculating the mean values for north and south given the rms variations of the values.
The significance of this is that the (J-Ks) color differences are very small compared to the K-correction (redshift effect): the mean color for galaxies 0.1 mag fainter or brighter (i.e., between 13.9 and 14.9 mag, or between 14.1 and 15.1 mag) shift by 0.00722 mag redder or bluer. Thus the lack of (J-Ks) color change implies that the southern and northern galaxies truly represent objects from the same magnitude range to within 0.044 mag [=(0.00320/0.00722)×0.1 mag].
We can also test for color effects near the celestial poles. Here we are limited in the current accuracy of the test because of insufficient number statistics, and the poles are themselves at fairly low galactic latitude. Nevertheless, comparing galaxies within 20° of the pole with those that pass within 20° of the zenith (and requiring all to be between galactic latitudes of 20° and 35°), we find the following colors:
|mean value||zenith-pole||error of the mean|
|(J-Ks) in 7´´ aperture||1.13101||-0.00139||0.00783|
|(J-H) in 7´´ aperture||0.70171||-0.00014||0.00725|
This comparison shows no significant difference in zenith and pole color measurements. Note though that the mean (J-Ks) colors do show evidence of reddening relative to the high galactic latitude measurements earlier. The (J-H) colors here are particularly useful for showing the lack of a color effect, since mean (J-H) color of galaxies is nearly constant as a function of magnitude.
[Modified 2000 Sep 12 by S. Van Dyk.]