2MASS Images

Photometry and Astrometry from H-compressed and Uncompressed Images

1. Summary

For stars with SNR > 20, the average difference between the IRAF and default 2MASS pipeline photometry ranges between about 0.01-0.07 mag with most values around 0.04-0.05 mag. About 0.03 mag of these offsets can be explained by aperture corrections applied to the pipeline processing but not the IRAF data. Any residual offsets, if real, may be a result of kernel smoothing the coadded images.
Not surprisingly, the effects of the saturated READ2 sources are clearly seen in the coadded image photometry. Since the saturation limits are not supplied in the image headers, care must be taken in choosing stars to compute relative photometry.
The RMS difference between the (aperture) coadd and (PSF) 2MASS V3 photometry is about 0.04-0.05 mag for sources with SNR >= 20. This is true for both the H-compressed and uncompressed images.
The RMS of the astrometric differences are about 0.10-0.13" for both right ascension and declination. In the galactic plane region in particular systematic astrometric offsets of 0.2" or less can be seen as a function of the cross-scan position. According to Gene [Kopan], this was expected since distortion corrections were not implemented in PIXPHOT.Direct comparison between the IRAF magnitudes derived from the compressed and uncompressed images show that there is increased scatter at faint magnitudes, and also systematic biases in that the compressed images yield fainter magnitudes (see Table 2b).

2. Analysis

I compared the astrometry and photometry derived from 5 coadded images with values listed in the 2MASS V3 point source catalog. The goal was to determine the accuracy in which the photometry/astrometry can be derived from the images. The 5 fields analyzed are:

Field	Date	Scan	RA	Dec
Northern polar cap	991013n	s0077	21:46:56.02	+83:26:09.3
Southern polar cap	000318s	s0033	05:47:45.29	-87:12:42.2
Intermediate lat north	000518n	s0077	19:24:01.00	+53:18:06.3
Intermediate lat south	990326s	s0065	10:20:16.03	-39:32:27.7
Galactic plane	980711s	s0027	16:55:02.00	-35:32:55.0

These were selected to cover a range of source density and time, and hemisphere. Otherwise, there is nothing special about these fields.

The analysis was performed on the H-compressed and uncompressed images.
Sources were identified in the images using daofind as implement in IRAF. The sky sigma was taken from the image headers, and a detection threshold of 5 sigma was used.
Aperture photometry was performed using the digiphot.apphot package in IRAF. I used a aperture radius of 4" and a sky annulus that extends from 14 to 20". The magnitude zero point were taken from the image headers.
The X/Y pixels determined from DAOFIND were converted to equatorial coordinates using the SAO WCS library.
Sources were matched between DAOFIND and 2MASS using a 3" search radius.

3. Figures

Table 2a contains link to plots that show the magnitude/astrometric offsets for each field and for both the Uncompressed (U) and H-compressed (C) images plotted as a function of the 2mass pipeline magnitude and the cross scan pixel in the coadd image. Table 2b compares the uncompressed and H-compressed photometry directly. The main points to note are:

There is a systematic photometric offset between the IRAF and 2MASS photometry. This is discussed further in Section 4, and can be explained partly by aperture corrections that were not applied to the IRAF photometry.
The galactic plane plot shows that the IRAF magnitudes are brighter than the 2MASS V3 magnitudes by a few tenths of a magnitude for stars brighter than about mag=7. This is clearly due to saturation in the 2MASS coadds (i.e. these are the Read 1 sources). It should be noted that the saturation limit is not in the image headers, but should roughly be about 8000 counts according to Gene.
The delta mag plots for 2MASS aperture photometry clearly show less scatter across the full magnitude range compared to the PSF results. This is expected since the aperture magnitudes were computed with essentially identical algorithms. The agreement is not exact even though I used the same aperture size and sky annulus. In the galactic plane plots in particular, the IRAF aperture magnitudes tend to be brighter at faint magnitudes. I suspect what is happening is that the photometry at faint magnitudes depends critically on estimating the sky background, and the scatter may reflect differences between the 2MASS pipeline and IRAF algorithms. (The default IRAF algorithm uses a 3 sigma pixel rejection and a intensity-weighted mean and the sky pixel histogram).
Assuming the uncompressed images as "truth", Table 2b shows that there is additional scatter and bias in the H-compressed image photometry.

**Table 2a. Astrometric/Photometric plots**
Field	vs. 2MASS Magnitude				vs. Cross Scan direction
Field	Delta mag (V3 aperture)	Delta mag (V3 default)	Delta ra*cos(dec)	Delta dec	Delta mag (V3 aperture)	Delta mag (V3 default)	Delta ra*cos(dec)	Delta dec
Northern polar cap	U C	U C	U C	U C	U C	U C	U C	U C
Southern polar cap	U C	U C	U C	U C	U C	U C	U C	U C
Intermediate glat - north	U C	U C	U C	U C	U C	U C	U C	U C
Intermediate glat - south	U C	U C	U C	U C	U C	U C	U C	U C
Galactic Plane	U C	U C	U C	U C	U C	U C	U C	U C

Table 2b. Uncompressed vs. H-compressed Magnitudes

Field Plot

Northern polar cap X

Southern polar cap X

Intermediate glat - north X

Intermediate glat - south X

Galactic Plane X

**Table 2b. Uncompressed vs. H-compressed Magnitudes**
Field	Plot
Northern polar cap	X
Southern polar cap	X
Intermediate glat - north	X
Intermediate glat - south	X
Galactic Plane	X

4. Statistics

For each field, I computed various statistics to characterize the offsets between the IRAF measurements and the pipeline processing (default magnitudes). The statistics were computed for stars with SNR >= 20 and fainter than 9 magnitude to avoid saturation. Sources with obvious discrepant magnitudes, i.e. (IRAF - 2MASS) < -0.1 mag), were excluded from the computations. The results are tabulated in Tables 3-8. The main conclusions are:

The average photometric offset (Table 3a) are positive in all fields and all bands by about 0.04 mag in the sense the IRAF photometry gives fainter magnitudes than the 2MASS processing pipeline. Most likely these offsets are due to curve-of-growth/aperture corrections that were not applied to the IRAF data. Table 3b shows the aperture corrections for these fields as applied in the 2MASS processing pipeline. In many instances these aperture corrections can explain most of the photometric offset, but there is still a residual effect of 0.02-0.04 mag in some instances. As pointed out by Gene, since the coadded images are smoothed, the actual aperture corrections for the coadds will be larger than derived from the processing pipeline. If this explains the remaining difference, this analysis suggests that the aperture correction due to kernel smoothing is about 0.03 mag or less.

Any astrometric offset is less than 0.05" in both RA and DEC with an observed dispersion of about 0.10". From the Figures in Table 2, though, there are systematic offsets as a function of cross-scan position. This is most notable in the galactic plan cross-scan plot linked in table 2. These offsets are at most 0.2" at the edge of the scans.

Table 3a: Average photometric offsets

Field Uncompressed H-compressed

J H K J H K

Northern polar cap 0.031 0.035 0.047 0.043 0.042 0.052

Southern polar cap 0.034 0.075 0.034 0.044 0.083 0.037

Intermediate lat north 0.042 0.042 0.042 0.054 0.048 0.047

Intermediate lat south 0.038 0.052 0.041 0.049 0.058 0.048

Galactic plane 0.018 0.048 0.012 0.049 0.069 0.030

Table 3b: 2MASS V3 aperture corrections

Field Uncompressed

J H K

Northern polar cap 0.024 0.033 0.030

Southern polar cap 0.031 0.043 0.031

Intermediate lat north 0.024 0.033 0.028

Intermediate lat south 0.019 0.031 0.027

Galactic plane 0.023 0.027 0.023

Table 4: RMS of Photometric offsets

Field Uncompressed H-compressed

J H K J H K

Northern polar cap 0.033 0.031 0.038 0.037 0.033 0.038

Southern polar cap 0.052 0.053 0.025 0.053 0.057 0.024

Intermediate lat north 0.035 0.055 0.034 0.041 0.061 0.035

Intermediate lat south 0.049 0.029 0.050 0.058 0.029 0.052

Galactic plane 0.057 0.055 0.051 0.085 0.073 0.064

**Table 4: RMS of Photometric offsets**
Field	Uncompressed	H-compressed
J	H	K	J	H	K
Northern polar cap	0.033	0.031	0.038	0.037	0.033	0.038
Southern polar cap	0.052	0.053	0.025	0.053	0.057	0.024
Intermediate lat north	0.035	0.055	0.034	0.041	0.061	0.035
Intermediate lat south	0.049	0.029	0.050	0.058	0.029	0.052
Galactic plane	0.057	0.055	0.051	0.085	0.073	0.064

Table 5: Average RA offsets (arcsec)

Field Uncompressed H-compressed

J H K J H K

Northern polar cap -0.011 -0.001 -0.022 -0.008 -0.004 -0.020

Southern polar cap 0.014 0.025 0.010 0.005 0.022 0.010

Intermediate lat north 0.011 0.034 0.009 0.009 0.031 0.009

Intermediate lat south 0.033 0.006 0.020 0.036 0.006 0.021

Galactic plane 0.002 0.021 0.027 0.001 0.022 0.025

**Table 5: Average RA offsets (arcsec)**
Field	Uncompressed	H-compressed
J	H	K	J	H	K
Northern polar cap	-0.011	-0.001	-0.022	-0.008	-0.004	-0.020
Southern polar cap	0.014	0.025	0.010	0.005	0.022	0.010
Intermediate lat north	0.011	0.034	0.009	0.009	0.031	0.009
Intermediate lat south	0.033	0.006	0.020	0.036	0.006	0.021
Galactic plane	0.002	0.021	0.027	0.001	0.022	0.025

Table 6: Average Dec offsets (arcsec)

Field Uncompressed H-compressed

J H K J H K

Northern polar cap -0.002 -0.011 -0.031 -0.004 -0.008 -0.031

Southern polar cap -0.001 -0.012 -0.008 0.000 -0.018 -0.004

Intermediate lat north -0.015 -0.001 0.024 -0.016 0.002 0.025

Intermediate lat south -0.008 0.007 -0.005 -0.008 0.008 -0.005

Galactic plane 0.012 0.013 0.011 0.012 0.014 0.011

**Table 6: Average Dec offsets (arcsec)**
Field	Uncompressed	H-compressed
J	H	K	J	H	K
Northern polar cap	-0.002	-0.011	-0.031	-0.004	-0.008	-0.031
Southern polar cap	-0.001	-0.012	-0.008	0.000	-0.018	-0.004
Intermediate lat north	-0.015	-0.001	0.024	-0.016	0.002	0.025
Intermediate lat south	-0.008	0.007	-0.005	-0.008	0.008	-0.005
Galactic plane	0.012	0.013	0.011	0.012	0.014	0.011

Table 7: RMS of RA offsets (arcsec)

Field Uncompressed H-compressed

J H K J H K

Northern polar cap 0.110 0.133 0.108 0.102 0.132 0.109

Southern polar cap 0.119 0.114 0.117 0.116 0.119 0.115

Intermediate lat north 0.124 0.130 0.108 0.130 0.130 0.111

Intermediate lat south 0.096 0.091 0.128 0.100 0.091 0.129

Galactic plane 0.102 0.102 0.144 0.109 0.104 0.147

**Table 7: RMS of RA offsets (arcsec)**
Field	Uncompressed	H-compressed
J	H	K	J	H	K
Northern polar cap	0.110	0.133	0.108	0.102	0.132	0.109
Southern polar cap	0.119	0.114	0.117	0.116	0.119	0.115
Intermediate lat north	0.124	0.130	0.108	0.130	0.130	0.111
Intermediate lat south	0.096	0.091	0.128	0.100	0.091	0.129
Galactic plane	0.102	0.102	0.144	0.109	0.104	0.147

Table 8: RMS of Dec offsets (arcsec)

Field Uncompressed H-compressed

J H K J H K

Northern polar cap 0.106 0.086 0.092 0.108 0.087 0.092

Southern polar cap 0.117 0.128 0.107 0.122 0.130 0.107

Intermediate lat north 0.095 0.098 0.090 0.101 0.099 0.091

Intermediate lat south 0.091 0.098 0.103 0.096 0.099 0.103

Galactic plane 0.093 0.104 0.101 0.098 0.105 0.103

**Table 8: RMS of Dec offsets (arcsec)**
Field	Uncompressed	H-compressed
J	H	K	J	H	K
Northern polar cap	0.106	0.086	0.092	0.108	0.087	0.092
Southern polar cap	0.117	0.128	0.107	0.122	0.130	0.107
Intermediate lat north	0.095	0.098	0.090	0.101	0.099	0.091
Intermediate lat south	0.091	0.098	0.103	0.096	0.099	0.103
Galactic plane	0.093	0.104	0.101	0.098	0.105	0.103

Editor's note: This study was made by John Carpenter.