The 2MASS data processing began in mid-April 1998. As of mid-May, most of the processed data came from a set of 5 nights chosen as an RTB dataset, covering 184 sq. deg., 0.5% of the sky outside of high source density areas. We have visually examined all the ~9,000 sources identified as extended by the data processing, and classified them into galaxies, double stars (two nearby stars just under 2MASS resolution ability), etc.
Analysis of these data show that:
All magnitudes referenced in this paper are circular apertures with radius=7", except for the plot of galaxy sizes.
There are two major independent ways to distinguish galaxies from other extended sources such as double stars with > 99% reliability: using one or more two-dimensional image size parameters, or a simple color selection.
"Wedge shape" is an image size parameter that very effectively discriminates against double stars, the largest other category of sources in the Extended Source DB over most of the sky. "Wedge shape" is the minimum average radial size after excluding a 45° wedge placed at all possible orientations centered on the source. Wedge shape is normalized by subtracting the wedge shape of stellar sources to remove the PSF to first order.
Fig. 1. Histogram of galaxies and non-galaxies versus normalized "wedge shape" (see text). Solid line: sources verified to be galaxies. Dashed red line: Other sources, mainly double stars.
"Color_score" is a measure of the distance of the observed galaxy colors from the closest point source color (see Jarrett et al poster 55.15 this session):
color_score = (J-K) - 0.9 + ( (H-K)>0.3 ) * ( (H-K)-0.3 )
Fig. 2. Histogram of galaxies and non-galaxies versus color_score. Solid line: sources verified to be galaxies. Dashed red line: Other sources, mainly double stars. Dashed yellow line: sources with uncertain classification, probably galaxies. Dashed blue line: triple stars.
Simply adding the "wedge shape" and the "color-score" effectively isolates the galaxies:
Fig. 3. Histogram of galaxies and non-galaxies versus {wedge shape plus color_score}. Solid line: sources verified to be galaxies. Dashed red line: Other sources, mainly double stars. Dashed yellow line: sources with uncertain classification, probably galaxies. Dashed blue line: triple stars.
The Differential Log N / Log S plot shows that the database is essentially complete down to J~15, H~14.3, and Ks~13.9 mag, where the Ks completeness is being pushed deeper by galaxies detected as extended at J. (A Ks-only database, consisting of galaxies classified as extended only in the K band, would be complete to around 13.5 mag.) The density of detected galaxies is ~34 per square degree, consistent with the estimates of Glazebrook et al 1994 (MNRAS, 266:65).
Fig. 4. Differential Log N / Log S for galaxies. Error bars are shown as dashed lines. The aqua line gives the K counts of Glazebrook et al 1994 for K > 13 mag and a no-evolution SIGMA_0 = 0.1 extrapolation from 12.0 < K < 13.5 mag.
For the >90% of the sky outside high source density areas, this implies 1.3 million true galaxies down to those completeness levels. There will be less than another 0.1 million true galaxies in the 10% of the sky with high source density, due to increased thresholds due to confusion noise.
Most 2MASS galaxies have redshifts between z=0 and z=0.2, derived from observed colors, NED redshifts, and our previous work (see Schneider et al 1997, 1997 DENIS Conference).
2MASS galaxy colors (J-H and H-K) are consistent with observing a distribution of galaxies with normal galaxy colors that have k-corrections between z=0 and z=0.2 (see also Jarrett et al poster 55.15 this session). Fig. 5 shows the color-color plot for 2MASS galaxies with Ks < 13 mag so that the color uncertainty is less than 0.10 mag. The database is complete to nearly 1 full mag fainter, corresponding to a redshift 60% greater than the limit of that plot.
Fig. 5. Color-color Plot for sources with Ks < 13 mag. Color uncertainties are less than 0.10 mag. Green lines are stellar loci. Pink and gray lines are model-derived K corrections supplied by Roc Cutri using data from Brian McLeod. The few outliers in this plot come from galaxies confused with nearby larger galaxies.
We have searched NED for redshifts for these galaxies. Fewer than 2% of the 2MASS galaxies have redshifts in NED, so the results must be interpreted with caution. The following redshift distribution is consistent with expectations that 2MASS galaxies have redshifts between z=0 and z=0.2, recognizing that NED is very incomplete for the larger redshifts:
Fig. 6. Redshift distribution of 2MASS galaxies with previously-known redshifts.
Finally, the sizes of 2MASS galaxies, as measured by the radii to the circular 20th Ks mag per sq. arcsecond isophotal contour are typically ~20" at Ks20=11 mag, ~10" at Ks20 = 12 mag and ~7" at Ks20 = 13 mag. Radii below 7" cannot be reliably determined from the 2MASS data, and hence smaller radii default to 7".
Fig. 7. Circular radii to 20th Ks mag per sq. arcsecond contour versus Ks20.
http://spider.ipac.caltech.edu/staff/tchester/2mass/analysis/galaxies/first_results/index.html
Comments and feedback: Tom Chester
Last update: 5 June 1998 (Glazebrook counts note added 16 August 2000)
