The SAGE Data Description: Delivery 1
Prepared by Margaret Meixner (PI, STScI) for the SAGE team
1. Executive Summary ..................................................................................................... 1
2. Overview of the Spitzer SAGE Survey of the LMC.................................................... 2
3. SAGEcatalogIRACepoch1 and SAGEarchiveIRACepoch1 ....................................... 3
3.1 Bandmerging to Produce Sourcelists ..................................................................... 4
3.2 Catalog and Archive Criteria ................................................................................. 4
3.3 Astrometry ............................................................................................................. 6
3.4 Precision and Accuracy of the Photometry ............................................................ 7
3.5 Completeness ......................................................................................................... 9
4. SAGEcatalogMIPS24epoch1 ..................................................................................... 14
4.1 Catalog Criteria .................................................................................................... 14
4.2 Astrometry ........................................................................................................... 15
4.3 Precision and Accuracy of the photometry .......................................................... 16
4.4 Completeness ....................................................................................................... 18
5. Appendix of SAGEcatalogIRACepoch1 and SAGEarchiveIRACepoch1 Formats 21
6. Appendix: SAGEcatalogMIPS24epoch1 Formats .................................................... 25
7. Acknowledgements .................................................................................................... 27
8. References .................................................................................................................. 28
Contributing authors: Margaret Meixner (STScI), Uma Vijh (STScI), Marilyn Meade
(Univ. of Wisconsin), Barb Whitney (Space Science Institute), Karl Gordon (Univ. of
Arizona), Chad Engelbracht (Univ. of Arizona), Sundar Srinivasan (STScI), Bernie
Shiao (STScI), Remy Indebetouw (Univ. of Virgina), Joe Hora (Harvard/CfA), Steve
Bracker (Univ. of Wisconsin), Brian Babler (Univ. of Wisconsin), Martin Cohen (UC
Berkeley)
1. Executive Summary
This document describes the first SAGE delivery, made in support of Spitzer Cycle 4,
that includes SAGEcatalogIRACepoch1, an epoch 1 IRAC catalog bandmerged with
2MASS; SAGEarchiveIRACepoch1, an epoch 1 IRAC archive bandmerged with
2MASS; and SAGEcatalogMIPS24epoch1, an epoch 1 MIPS 24 micron catalog. These
epoch 1 source lists include ~4 million IRAC and ~60,000 MIPS24 sources from the
entire LMC, and were constructed to be highly reliable, but because they were extracted
from only half of the data they are incomplete particularly at the faint end. Later SAGE
deliveries to the SSC will include more complete catalogs and archives. This document
provides a brief overview of the project and data processing. The data quality is assessed
at a preliminary level. A guide to contents of the SAGE catalog and archive columns is
provided.
1
2. Overview of the Spitzer SAGE Survey of the LMC
The SAGE project is a Cycle 2 legacy program on the Spitzer Space Telescope, entitled,
“Spitzer Survey of the Large Magellanic Cloud: Surveying the Agents of a Galaxy’s
Evolution (SAGE)”, with Margaret Meixner (STScI) as the PI. The project overview and
initial results are described in a paper by Meixner et al. (2006) and the main
characteristics of the survey are listed in Table 1. The LMC was mapped at two different
epochs dubbed epochs 1 and 2 separated by 3 months, because it provides a 90-degree
roll angle in the orientation of the detectors, which optimally removes the "striping"
artifacts in MIPS and artifacts along columns and rows in the IRAC image data. In
addition, these two epochs are useful constraints of source variability expected for
evolved stars and some young stellar objects (YSOs). Note, the predicted sensitivities
for the complete survey are listed in Table 1 are for the whole survey, whereas this first
delivery uses only half the data and thus has less sensitivity (see Meixner et al. 2006 for
more details). Not all sources in the SAGE catalog belong to the LMC. Meixner et al.
(2006) derived preliminary estimates of 18% and 12% for Milky Way foreground and
background galaxy contributions to the SAGE catalog, respectively.
Table 1: Principal Characteristics for SAGE Survey, Spitzer program ID 202031
Characteristic
Nominal Center point
survey area
AOR size, grid size
Total time (hrs)
 (m)
pixel size at 
angular resolution at 
Exposure time/ pixel at  (s)
Predicted point source
sensitivity for combined
epoch data, 5  at  (mJy)
Predicted point source
sensitivity for combined
epoch data, 5  at  (mag.)
Saturation limits (Jy) at 
Saturation limits (mag) at 
Surface brightness limits
(MJy/sr) 5  at  (mag.)
Epoch 1
Epoch 2
1
IRAC Value
RA(2000): 5h 18m 48s
Dec(2000): -68 34 12
MIPS Value
RA(2000): 5h 18m 48s
Dec(2000): -68 34 12
7.17.1
1.11.1, 77
7.87.8
254, 192
290.65
3.6, 4.5, 5.8 and 8
1.2, 1.2, 1.2,1.2
1.7, 1.7, 1.9, 2
43, 43, 43, 43
0.0051, 0.0072, 0.041,
0.044
216.84
24, 70 and 160
2.5, 9.8, 15.9
6, 18, 40
60, 30, 6
0.5, 30, 275
19.3, 18.5, 16.1, 15.4
10.4, 3.5, -0.6
1.1, 1.1, 7.4, 4.0
6, 5.5, 3.0, 3.0
… , … , 0.5, 1
4.1, 23, 3
0.60, -3.7, -3.2
1, 5, 10
July 15 – 26, 2005
Oct. 26 – Nov. 2, 2005
July 27 – Aug. 3, 2005
Nov. 2-9, 2005
Taken from Meixner et al. (2006).
2
The catalogs and archives in this delivery, shown in Table 2, were produced by the IRAC
pipeline (Univ. of Wisconsin) and MIPS pipeline (Univ. of Arizona) and ingested into a
SAGE database for review, SAGE Team use and staging area for releases (STScI).
Table 2: Point Source lists for SAGE Epoch 1 delivery
Source List Title
Data
Volume
~4 million
sources,
~4
gigabytes
Wavelengths
(microns)
1.24, 1.66, 2.16
(2MASS)
3.6, 4.5, 5.8, 8
(IRAC)
SAGEarchiveIRACepoch1
~4.3
million
sources,
~4
gigabytes
1.24, 1.66, 2.16
(2MASS)
3.6, 4.5, 5.8, 8
(IRAC)
SAGEcatalogMIPS24epoch1
~60,000,
~0.2
gigbyte
24
SAGEcatalogIRACepoch1
Contents
IRAC epoch 1 catalog with
0.6 and 12 second frame
photometry bandmerged with
2MASS; high reliability
emphasized. Faint limits are
17, 16.5, 14 and 13.5 mags.
For IRAC 3.6, 4.5, 5.8, and 8
microns
IRAC epoch 1 archive with
0.6 and 12 second frame
photometry bandmerged with
2MASS; includes more
sources and more source
fluxes (fewer nulled
wavelengths), see section 3.2.
Faint limits are 17, 16.5, 14
and 13.5 mags. For IRAC 3.6,
4.5, 5.8, and 8 microns
MIPS 24 micron epoch 1
catalog high reliability
emphasized. Faint limit is 10
mag or ~0.7 mJy at 24
microns.
3. SAGEcatalogIRACepoch1 and
SAGEarchiveIRACepoch1
The point source lists presented in this document were extracted from IRAC epoch 1
single frame images processed with the SSC pipeline version S12.4.0 using a modified
version of DAOPHOT (Stetson 1987). The Wisconsin IRAC pipeline, which was
originally developed to process the GLIMPSE data (Benjamin et al. 2003), is described
by the GLIMPSE pipeline documents2. This pipeline was modified for the SAGE project
to handle the high dynamic range (HDR) data and longer exposure times. The details for
2
http://www.astro.wisc.edu/glimpse/docs.html
3
the SAGE IRAC epoch 1 processing can be found in Meixner et al. (2006). Here we
summarize the criteria for selection of the point sources to be in the catalog
(SAGEcatalogIRACepoch1) and the archive (SAGEarchiveIRACepoch1) after the source
lists are extracted from the images. Then we describe the quality of the resulting
SAGEcatalogIRACepoch1.
3.1
Bandmerging to Produce Sourcelists
The point source lists are merged at two stages using a modified version of the SSC
bandmerger3. Before the first stage, source detections with S/N less than 3 are culled.
During the first stage, or in-band merge, all detections at a single wavelength are
combined using position, signal -to-noise (S/N) and flux to match the sources. The 0.6
second flux is included if the signal-to-noise is greater than (5,5,5,7) and the magnitudes
are brighter than (12,11,9,9), for the four IRAC bands [3.6],[4.5],[5.8] and [8.0]
respectively. This prevents Malmquist bias from affecting the results. The 12 second flux
is included if the magnitude is fainter than the saturation limit of (9.5, 9.0, 6.5, 6.5) for
the four IRAC bands [3.6], [4.5], [5.8] and [8.0], respectively. When both criteria are
met, the 0.6 and 12 second fluxes are combined, weighted by the propagated errors.
Fluxes of sources within 1.6" in the IRAC frame are combined together or “lumped” into
one flux.
The second stage, or cross-band merge, combines all wavelengths for a given source
position using only position as a criterion in order to avoid source color effects. Crossband lumping is done with a 1.6" radius. Position migration can still occur in the
bandmerging process which results in a small number of sources that have sources within
1.6" of it. In the cross-band merge stage, we also merge with the 2MASS Point Source
Catalog only if it has a K detection. Note that the 2MASS information we include from
the 2MASS PSC is designation, cntr, fluxes, s/n, and a limited source quality flag. Users
should refer back to the 2MASS Point Source Catalog for the complete 2MASS
information about the source.
3.2
Catalog and Archive Criteria
The IRAC source lists were produced from photometry on individual BCD frames. The
12 second exposures suffer from cosmic rays. For this reason, a stringent selection
criteria was developed to limit false sources. To ensure high reliability of the final pointsource SAGEcatalogIRACepoch1 by minimizing the number of false sources, we adopt
the following selection criteria: Given M detections out of N possible observations, we
require that M/N >= 0.6 in one band, and M/N >= 0.4 in an adjacent band, with a S/N >
5, 5, 5, 7 for IRAC bands [3.6], [4.5], [5.8] and [8.0], respectively. As an example, a
source is typically observed twice at 0.6 second and twice at 12 second for a total of four
possible observations in each band. Such a source detected three times in band [3.6] with
3
http://ssc.spitzer.caltech.edu/postbcd/bandmerge.html
4
S/N > 5, and twice in band [4.5] with S/N > 5 would be included in the catalog and
archive. Sources are also culled when they are too close to another source because this
neighboring source could influence the flux for the source. We use the Archive list to
search for near neighbors, and cull from the Archive sources within 0.5", and from the
Catalog sources within 2".
A source may be reliably detected in two bands (usually [3.6] or [4.5]) but have
questionable flux in another (usually [5.8] or [8.0]). To ensure high quality fluxes for
each source, a flux/magnitude entry for a band in the catalog will be nulled, i.e. removed,
for any of the four following reasons. One, the source is brighter than the 0.6 sec.
saturation magnitude limits, 6.0, 5.5, 3.0, 3.0, for IRAC bands [3.6], [4.5], [5.8] and
[8.0], respectively. Two, the source location is flagged as near a frame edge, in a
saturated star wing or coincident with a bad pixel. Three, the S/N is less than [6, 6, 6,
10] for IRAC bands [3.6], [4.5], [5.8] and [8.0], respectively in order to mitigate
Malmquist bias. Four, for 12-second only data, if M<2 or M/N is less than 0.6 in order
to mitigate faint cosmic ray detections. Note that in HDR mode, data from one epoch
(N=4: 2 12 sec observations and 2 0.6 sec observations) having a result with N=2 is not
uncommon. Bright sources are only measurable in the 0.6 sec data and faint sources are
only measurable in the 12 sec data. Only stars in the intermediate range of brightness will
have useful detections in both the 0.6 and 12 sec images. If all fluxes for a source are
nulled, the source is removed from the catalog.
For the Archive, the nulling criteria are less stringent. The magnitude is nulled if the SN
is less than 5 in that band. For photometry with 12 second only data, if M is less than 2 or
if M/N < 0.3 the magnitude is nulled.
Future catalogs, combining both epochs and based on mosaic photometry (with cosmic
rays removed) will have different criteria. Characteristics of the epoch 1 source lists are
summarized in Table 2.
In the next few sections, we describe some preliminary quality checks performed on the
SAGEcatalogIRACepoch1 and SAGEarchiveIRACepoch1 including astrometry,
photometric precision, photometric accuracy and completeness.
5
Figure 1: Histogram of the position difference between the
SAGEcatalogIRACepoch1 sources and their corresponding 2MASS catalog
positions.
3.3
Astrometry
The IRAC catalog astrometry is referenced to the 2MASS astrometry through cross
matching of the IRAC and 2MASS frames. The absolute error in this 2MASS
astrometry is 0.3 1 sigma. We determine the astrometric quality of the
SAGEcatalogIRACepoch1 and SAGEarchiveIRACepoch1 by cross checking its
positions against the 2MASS catalogs. In Figure 1, we plot the histogram of the angular
separation between the SAGE IRAC catalog position and the corresponding 2MASS
position in 0.01 arcsecond bins. The peak of the histogram is at 0.15 and the majority of
sources have positional differences less than 0.3.
6
Figure 2: Uncertainty in Magnitude vs. magnitude for each IRAC band included in
the SAGEcatalogIRACepoch1 and SAGEarchiveIRACepoch1. Contours are used
to show the density of sources.
3.4
Precision and Accuracy of the Photometry
Figure 2 shows the photometric uncertainty due to the extraction. The jump in
uncertainties at the brighter magnitudes, e.g. 9.5 at 3.6 microns, shows at the photometry
boundary between the 0.6 and 12 second photometry. These brightest magnitudes have
larger errors because the exposure time is shorter.
To assure that our photometric calibration is uniform across the large area observed by
SAGE, and between different AORs, epochs, and wavelengths, we extract the
photometry for a network of absolute stellar calibrators custom-built for SAGE. These
are A0-A5V or K0-M2III stars selected from SIMBAD and their surface density within
the SAGE area is approximately 3 stars per square degree. The techniques used to
produce the complete UV to mid-IR absolute spectra are described by Cohen et al.
7
(2003). Any stars showing inconsistency between optical photometry, spectral type and
reddening, and 2MASS photometry were culled to produce the final list of 139 viable
calibrators for the epoch 1 data.
Figure 3 shows the excellent agreement between the SAGE magnitudes and the
predicted magnitudes of the calibration stars for the IRAC data. Errors in both the
extracted magnitudes and those predicted were added in quadrature to produce the
plotted error bars. Magnitude differences are much smaller than the one-sigma errors of
our photometry. The ensemble averaged differences and standard deviations in the four
IRAC bands between SAGE and predicted magnitudes are -0.013+/-0.051, 0.016+/0.059, -0.001+/-0.051, 0.007+/-0.046, respectively. The number of calibrators per
channel varies due to saturation and, for epoch 1 SAGE data, these were 75, 105, 135,
and 133 for IRAC [3.6], [4.5], [5.8], and [8.0] respectively.
The basis for calibration is identical for 2MASS (Cohen, Wheaton & Megeath 2003) and
IRAC (Cohen et al. 2003), with absolute zero points (expressed in Jy for zero
magnitude) for 2MASS J, H, Ks and the IRAC bands (3.6, 4.5, 5.8, and 8.0um) of
1594.0, 1024.0, 666.7, 277.1, 179.4, 113.9, 63.10 Jy, respectively. These zero points are
tied to the Midcourse Space eXperiment's absolute calibration which has an accuracy of
+/-1.1% (Price et al. 2004). The method employed to produce the SAGE network of
calibrators is identical to that used to create the suite of standards at the NEP (Cohen et
al. 2003) from which the IRAC primary calibrators were selected (Reach et al. 2005).
SAGE data in this release correspond to the SSC IRAC pipeline version S12.4.
However, the data plotted in Figure 6 were corrected to the SSC S13.2 pipeline
calibration by compensating for the updated values of FLUXCONV and magnitude zero
points of Reach et. al (2005). A preliminary examination of S13.2 reprocessed data
shows almost the same insignificant differences between expected and observed
calibrator magnitudes for a subsample of SAGE network stars reprocessed so far.
8
Figure 3: Plots demonstrating the quality of the SAGE flux and magnitude
measurements in the SAGEcatalogIRACepoch1 and SAGEarchiveIRACepoch1.
For each IRAC band, noted in the top left of the plots, the differences between
measured SAGE magnitudes and those predicted for the 139 calibration stars in the
SAGE field are plotted. Error bars are the RSS of the errors of both extracted and
predicted magnitudes for each star.
3.5
Completeness
The SAGEcatalogIRACepoch1 and SAGEarchiveIRACepoch1 has been designed for
reliability, not completeness. For this preliminary release of the data, we do not run a
thorough completeness test of the catalogs. Instead, as a gauge of its completeness, we
compare the number counts per magnitude bin of sources between the
SAGEarchiveIRACepoch1 with source counts based on the cleaner and deeper
9
combined epoch 1 & 2 mosaicked image for a particular region of the SAGE survey, tile
52 (Figures 4, 5, 6 & 7). These plots show that the epoch 1 lists are mostly complete
down to 12.5, 12.5, 12.5 and 9.5 in IRAC bands [3.6],[4.5],[5.8],[8.0] respectively, with
the big drop-offs at 16.5, 15.5, 14.0 and 13.0 for bands 1 through 4 respectively.
Divergence from the 'truth' list (obtained from mosaic photometry) occurs at the brighter
magnitudes, due to our stringent source-selection criteria that removes real sources in
addition to cosmic rays. Also completeness will be a function of background level which
will be more variable for IRAC [5.8] and [8.0].
Ch1 Number counts - tile 52
10000
1000000
Mosaic
SAGEcat
Mosaic
SAGEcat
100000
1000
100
1000
Cumulative Sum
Number
10000
100
10
10
1
1
5
10
15
Ch1 Mag
20
Figure 4: For IRAC Ch. 1 (3.6 micron): Comparison of the
SAGEarchiveIRACepoch1 source counts with the mosaicked photometry source
counts on tile 52 of the SAGE survey. The magenta and royal blue lines show the
differential number counts and the cyan and green lines show the cumulative total
number counts.
10
Ch2 Number counts - tile 52
10000
1000000
Mosaic
SAGEcat
Mosaic
SAGEcat
100000
1000
100
1000
100
10
10
1
1
5
10
Ch2 Mag
15
20
Figure 5: Same as Figure 7 except for IRAC Ch. 2 (4.5 microns).
Ch3 Number counts - tile 52
1000
100000
Mosaic
SAGEcat
Mosaic
SAGEcat
10000
Number
1000
100
10
10
1
1
5
10
15
Ch3 Mag
Figure 6: Same as Figure 7 except for IRAC Ch. 3 (5.8 microns).
11
Cumulative Sum
100
Cumulative Sum
Number
10000
Ch4 Number counts - tile 52
1000
100000
Mosaic
SAGEcat
Mosaic
SAGEcat
100
Number
1000
100
10
Cumulative Sum
10000
10
1
1
5
10
Ch4 Mag
15
Figure 7: Same as Figure 7 except for Ch. 4 (8.0 microns).
Catalog and Archive criteria have additional source completeness effects. Because source
reliability requires a two-band detection rule, source counts along frame overlap regions
can display apparent density enhancements and deficiencies when the source counts are
plotted near the limiting magnitudes. (see Figures 8 and 9). The effect is produced by a
'small numbers' complication inherent in the Catalog/Archive selection criteria. The
primary selection criteria is M/N>=0.6 in one band and M/N>=0.4 in an adjacent band.
Faint IRAC [3.6] sources use IRAC [4.5] as the primary confirming adjacent band. A
problem arises in the region of IRAC [4.5] overlap between AORs. In one direction of
the AOR overlap, the typical N is N=4, and in the other direction the typical N is N=3.
An apparent lack of IRAC [3.6] sources arises primarily on the N=3 boundary. This is
explained by the fact that for the M/N >= 0.4 condition to be satisfied when N=3, then M
must be >=2. For the vast majority of the survey area, as we approach the limiting
magnitude the typical N is N=2 (since the source will only be detectable by the longer 12
sec exposures) and then M only needs to be 1 to satisfy M/N >= 0.4. But in the boundary
case of N=3, M will need to be M>=2 in order to fulfill the condition of M/N >= 0.4.
Effectively this means M/N>=0.666 which is much more stringent than the actual M/N
>= 0.4 condition. Thus the issue in this case is a small numbers problem. This effect is
not seen on the N=4 boundary, where the M/N > 0.4 criteria is met by 2/4 which is
comparable to the typical survey region where N=2 and M=1 (1/2).
12
Figure 8: Source density image of IRAC band 1 (3.6 micron) sources between
magnitudes 17 and 18 showing some of the incompleteness effects in the
SAGEarchiveIRACepoch1.
Similarly, we discover on these boundaries an apparent increase in the number of IRAC
[4.5] sources (see Figure 9). This is a result of sensitivity and M/N selection effects. For
IRAC [4.5] overlap areas were N>=3, slightly fainter sources are slightly more reliably
exacted, and since IRAC [3.6] is more sensitive, these sources have a greater chance of
making the Archive, thus producing a slight apparent increase in IRAC [4.5] sources in
these overlap regions. This effect is seen in both the N=3 and N=4 overlap regions.
An additional source of incompleteness in this Catalog/Archive release was introduced in
the bandmerging process. The original algorthim was developed for the GLIMPSE
legacy project and designed for galactic coordinates. Bandmerging is done in subsections
(or tiles). Unlike GLIMPSE, SAGE is processed in RA/Dec coordinates, and
unfortunately the sub-tiles are slightly non-overlapping at some corners, producing
systematic small triangular gaps (Figure 9). Future releases will correct this.
13
Figure 9: Source density image of IRAC band 2 (4.5 microns) sources of
magnitudes between 16 and 17 showing some of the enhanced detections in the
overlap boundary regions and small triangular regions of missing sources.
4. SAGEcatalogMIPS24epoch1
The point source lists presented in this document were extracted from MIPS 24 micron
epoch 1 images. The Arizona MIPS team pipeline, MIPS Data Analysis Tool v3.02
(DAT; Gordon et al. 2005) was used to do the processing and mosaicing of the individual
images. The details for the SAGE MIPS 24 micron epoch 1 data processing can be found
in Meixner et al. (2006). Here we summarize the criteria for selection of the point
sources to be in the catalog after the sources are extracted from the images. Then we
describe the quality of the resulting SAGEcatalogMIPS24epoch1.
4.1
Catalog Criteria
The astrometry of the MIPS 24 micron frames was registered to the
SAGEcatalogIRACepoch1 using sources in common. A MIPS 24 micron source list was
extracted using StarFinder (Diolaiti et al. 2000) on the mosaicked MIPS 24 micron epoch
1 only images. To be included in the source list, the 24 micron detection must have a
signal-to-noise ratio >3 sigma and the shape of the source must be nearly point like with
a correlation value >0.80. The more highly redundant MIPS 24 micron observations
means that cosmic rays are not an issue with the source lists. However, background
14
emission fluxes may be and these background values are included in the catalog for
reference. The MIPS point source catalogs are created by merging from each AOR, the
point source lists produced from multiple AORs. For sources detected in multiple
AORs, the fluxes are averaged. A separate, independent point source list was derived
from the MIPS 24 micron epoch 2 only data using the same criteria. Because these
source lists were derived from independent data sets, we used the epoch 2 MIPS 24
micron source list as a way to check the validity of the epoch 1 MIPS 24 micron source
list. Our SAGEcatalogMIPS24epoch1 catalog includes those epoch 1 MIPS 24 micron
sources that are also found in the epoch 2 source list and that have a signal-to-noise ratio
>3 sigma. Our approach to the creation of the SAGEcatalogMIPS24epoch1 ensures it to
have a high reliability. Note, even though we use the epoch 2 source list as a means of
checking the validity of the sources, the fluxes for the sources are only derived from
epoch 1 because we want to maintain these catalogs as separate for variability studies.
In the next few sections, we describe some preliminary quality checks performed on the
SAGEcatalogMIPS24epoch1 including astrometry, photometric precision, photometric
accuracy and completeness.
4.2
Astrometry
The astrometry of the SAGEcatalogMIPS24epoch1 is referenced to the
SAGEcatalogIRACepoch1 by using a bright 8 micron source list for the astrometric
framework of the MIPS24 micron catalog. The histogram showing the differences
between the two catalogs for sources that match within a 1 radius reveals most of the
offsets sharply peaked at zero as expected (Figure 10).
15
Figure 10: Histogram of the offsets between sources found in both the
SAGEcatalogMIPS24epoch1 and the SAGEcatalogIRACepoch1.
4.3
Precision and Accuracy of the photometry
Figure 11 shows the photometric uncertainties due to the extraction process that limit the
precision of the photometry.
16
Figure 11: Uncertainty in magnitude vs. magnitude for MIPS 24 microns.
The accuracy of the MIPS 24 micron fluxes is based on the MIPS 24 micron absolute
calibration program. Analysis of this program by Engelbacht et al. (2006) demonstrates
an absolute calibration accuracy of 4% rms. The zero points for the MIPS 24 micron
magnitude is 7.17 Jy (Rieke, G. et al., in prep.). We also compare our extracted MIPS
24 micron magnitudes to the predicted SAGE magnitudes of calibration stars in the
SAGE fields in Figure 12. The solid line at zero shows where perfect agreement lies.
While there is an apparent average offset of 0.037 +/- 0.0034 magnitude, this offset is
well within the 0.072 magnitude scatter of the data points. This good agreement
provides confidence that the extraction of the MIPS 24 micron photometry does not
significantly degrade the accuracy of the photometry.
17
Figure 12 Difference between the measured SAGE magnitude and the predicted
magnitude for the calibration stars in the SAGE field. The results agree within 4%.
The predicted magnitudes are based on techniques by Cohen et al. (2003).
4.4
Completeness
We did not conduct a full completeness test for this release of the
SAGEcatalogMIPS24epoch1. Instead, we estimate the completeness by showing the
source properties of the catalog in comparison to the archive lists of both epoch 1 and 2
MIPS 24 micron data. The following histograms for the SAGEcatalogMIPS24epoch1
show that the catalog contains approximately half of the sources found in the archive lists
for the MIPS 24 micron epoch 1 & 2 point source lists. The
SAGEcatalogMIPS24epoch1 sources lack predominantly the sources fainter than 4 mJy.
18
Figure 13: Comparison of the flux histograms between the
SAGEcatalogMIPS24epoch1 and the total source lists derived from the source
extraction for epoch 1 & 2 for MIPS 24 microns, called the archives.
19
Figure 14: Comparison of the cumulative magnitude histograms between the
SAGEcatalogMIPS24epoch1 and the total source lists derived from the source
extraction for epoch 1 & 2 for MIPS 24 microns, called the archives.
20
5. Appendix of SAGEcatalogIRACepoch1 and
SAGEarchiveIRACepoch1 Formats
Notes:
1) The fields in the Catalog and Archive are the same. Data is delivered in
IPAC format
2) Where NULL values are not legal, NO is entered in the column. Otherwise,
the null value is given.
Column Name
Description
Units Data
Type
Format
Null
1 globalSourceID
Unique identifier
I*4
i10
NO
2 sourceCatalog
identifier for source catalog
ASCII
a8
NO
3
epoch
identifier for epoch of source
ASCII
a16
NO
4 designation
Source name
ASCII
a28
NO
Archive=> SSTISAGE1A JHHMMSS.SS+-DDMMSS.S
Catalog=> SSTISAGE1C JHHMMSS.SS+-DDMMSS.S
5 tmass_cntr
cntr from 2MASS Point
Source Catalog
I*4
i10
0
6 tmass_designation 2mass source name
ASCII
a32
7 ra
Right ascension (J2000)
deg
R*8
f11.6 NO
8 dec
Declination (J2000)
deg
R*8
f11.6 NO
9 dra
Error in right ascension
arcsec R*8
f4.1 NO
10 ddec
Error in declination
arcsec R*8
f4.1 NO
11 closeFlag
Close source flag
I*2
i3
-9
12-25 magi, dmag i Magnitude from
J,H,K (i=1-3)
mag
R*4 7(f7.3,f7.3) 99.999,99.999
and each IRAC band (i=4-7)
and 1 sigma error
26-39 fi,dfi
Fluxes and 1 sigma error in mJy
R*4 14(e11.3) -999.9,-999.9
each band i, i=1-7
40-43 rms_fi
rms dev. of individual
mJy R*4
4(e11.3) -999.9
detections from fi (i=4-7)
44-47 skyi
Local sky bkg. for band i
MJy/sr R*4
4(e11.3) -999.9
flux (i=4-7)
48-54 sni
Signal/Noise for band i
R*4
7(f6.2)
-9.99
flux (i=1-7)
55-58 srcdensi Local source density for band
#/sq' R*4
4(f5.1)
-9.9
i object (i=4-7)
59-62 mi
Number of detections for band
I*2
4(i3)
-9
i (i=4-7)
63-66 ni
Possible number of detections
I*2
4(i3)
-9
for band i (i=4-7)
21
Column Name
Description
70-73 sqfi
Source Quality Flag for band i flux (i=1-7)
Flux calc method flag for
band i flux (i=4-7)
74-77 mflagi
78
79
80
81
cx
cy
cz
htmID
Units Data Format
Type
x of unit vector to this source
y of unit vector to this source
z of unit vector to this source
20-deep HTM ID of this source
I*4
7(i10)
-9
I*2
4(i3)
-9
R*4
R*4
R*4
I*4
f30.20
f30.20
f30.20
i20
Designation is
SST = Spitzer Space Telescope
I = IRAC
SAGE= LMC Survey project
1 = epoch 1; 2=epoch 2; nothing if epoch 1+2
C = highly reliable Catalog
A = more complete Archive
J = 2000.0 epoch
HHMMSS.SS Right ascension (hr, min, sec) of source
sDDMMSS.S Declination (deg, min, sec) of source
Close source flag:
0 no sources in the Archive within 3" of the source
1 sources in the Archive between 2.5" and 3" of the source
2 sources in the Archive between 2.0" and 2.5" of the source
3 sources in the Archive between 1.5" and 2.0" of the source
4 sources in the Archive between 1.0" and 1.5" of the source
5 sources in the Archive between 0.5" and 1.0" of the source
6 sources in the Archive within 0.5" of the source.
Source Quality Flag (sqf)
bt = bit in sqf
value = 2(bit-1) i.e. bit#1 corresponds to 20=1
0 OK - no problems encountered
bt
Flag
1 poor pixels in dark current
2 flat field applied using questionable value
3 latent image
4 sat star correction
7 muxbleed correction applied
* 8 hot, dead or otherwise unacceptable pixel
22
Null
NO
NO
NO
NO
9 muxbleed > 3-sigma above the background
10 allstar tweak positive
11 allstar tweak negative
*13 confusion in in-band merge
*14 confusion in cross-band merge
*15 column pulldown correction
*16 banding correction
17 stray light
19 nonlinear correction not applied
*20 saturated star wing region
21 pre-lumping in in-band merge
22 post-lumping in cross-band merge
*30 edge of frame (within 3 pixels of edge)
31 truth list (for simulated data)
For 2MASS data:
2MASS Flag
=>
SQF
-----------"p" persistence
latent (bit 3)
"c" photom confusion corr sat star wing (bit 4)
"d" diffraction spike uncorrectable sat star (bit 20)
"s" electronic stripe muxbleed (bit 9)
bit values:
bt 1 => 1; 2 => 2; 3 => 4; 4 => 8; 5 => 16; 6 => 32; 7 => 64; 8 => 128
bt 9 => 256; 10 => 512; 11 => 1024; 12 => 2048; 13 => 4096; 14 => 8192; 15 => 16384;
bt 16 => 32768; 17 => 65536; 18 => 131072; 19 => 262144; 20 => 524288;
bt 21 => 1048576; 22 => 2097152;
bt 30 => 536870912; 31 => 1073741824
*The flags to possibly cull on. (Note that flag 8,20, and 30 have already been removed
from
Catalog or nulled.)
Flux calculation method flag:
mflag indicates by bit whether a given exposure time was present, and
whether that exposure time was used in the final flux.
present used
exp bit bit
0.6 1
2
2 3 4
12 5
6
30 7
8
100 9 10
23
i.e. if 0.6 and 12 were present, but only 12 was used, then
bits 1 and 5 will be set (fluxes present) and bit 6 will be set (12s
used), and the mflag will be 20+24+25=49
for SAGE HDR mode, the relevant numbers work out to be
3
19
48
49
51
short exp data used, long exp data absent
short used, long present but unused
long exp used, short absent
long exp used, short present but unused
both long and short present and used
===============================================================
Here is an example of a line of text for the Catalog (81 columns)
435834
64.799060
99.999
99.999
99.999
-999.9
0.089
-999.9
-999.9
-999.9
-9.99
2
4
1024
-9
9328356394387
iracc
-69.713677
99.999
16.236
99.999
-999.9
0.007
-999.9
-999.9
-9.99
-9.99
1
2
epoch 1
0.3
99.999
0.088
99.999
-999.9
-999.9
-999.9
0.010
-9.99
23.0
0
-9
-9
-9
0.14762783412474746000
SSTISAGE1C J041911.77-694249.2 0
0.3
0
99.999
99.999
99.999
99.999
99.999
-999.9
-999.9
-999.9
-999.9
-999.9
0.001
-999.9
0.015
-999.9
-9.99
12.37
-9.99
8.6
-9.9
-9.9
0
2
1
-9
-9
131072
48
48
-9
0.31371175741310424000
-0.93797172444184995000
For this source:
globalSourceID: 435834
sourceCatalog: iracc
epoch: epoch 1
designation: SSTISAGE1C J054225.00-701925.9
: Name
tmass_cntr: 1252742659
: 2MASS cntr from 2MASS PSC
tmass_designation: 04193115-6932567
ra,dec:
85.604200, -70.323868
: RA & Dec in degrees
dra,ddec:
0.3 0.3
: Error in RA and Dec (")
close:
0
: close source flag
[mag,dmag]: 14.307 13.485 13.201 13.142 13.279 13.071 12.916 : Magnitudes
[J,H,K,Bands 1-4]
0.033 0.030 0.040 0.055 0.074 0.074 0.064 : Uncertainties [J,H,K,Bands 1-4]
24
[f,df]: 3.018E+00 4.133E+00 3.496E+00 1.536E+00 8.761E-01 6.887E-01 4.303E-01 :
Fluxes (mJy)
9.172E-02 1.142E-01 1.288E-01 7.747E-02 5.981E-02 4.675E-02 2.528E-02 : df
(mJy)
[rms_f]: 2.050E-03 7.573E-03 1.679E-03 8.647E-03
: rms_flux (mJy) Bands 1-4
[sky]: 9.800E-02 5.500E-02 8.890E-01 8.680E-01
: Sky Bkg (MJy/sr) Bands 1-4
[sn]: 32.90 36.19 27.14 19.83 14.65 14.73 17.02
: Signal to Noise [J,H,K,Bands
1-4]
[srcdens]: 72.4 50.5 7.9 6.5
: Local Source Density Bands 1-4
[m]:
2 2 2 2
: Number of detections Bands 1-4
[n]:
2 2 2 2
: Number of poss detections Bands 1-4
[sqf]: 0 0 0 1024 0 0 512
: Source Quality Flag [J,H,K,Bands 1-4]
[mflag]: 48 48 48 48
: Flux Calculation Method Flag Bands 1-4
cx: 0.14762783412474746000
cy: 0.31371175741310424000
cz: -0.93797172444184995000
htmID: 9328356394387
6. Appendix: SAGEcatalogMIPS24epoch1 Formats
Notes:
1) Where NULL values are not legal, NO is entered in the column. Otherwise,
the null value is given.
Column Name
Description
1
2
3
4
5
6
7
8
9
globalSourceID
sourceCatalog
epoch
designation
ra
dec
dra
ddec
conf
10
11
12
13
14
flux24
dflux24
mag24
dmag24
sky
Unique identifier
I*4
identifier for source catalog
ASCII
identifier for epoch of source
ASCII
source name
ASCII
Right ascension (J2000)
deg
R*8
Declination (J2000)
deg
R*8
Error in right ascension
arc sec R*8
Error in declination
arc sec R*8
Confusion Flag for
I*2
band 24 or 70 or 160,
currently unused
24um flux
mJy
R*8
24um 1 sigma error
mJy
R*8
24um magnitude
mag
R*8
24um 1 sigma error
mag
R*8
Local sky bkg. for
band 24 or 70 or 160
MJy/sr
R*4
Signal/Noise for
15
SN
Units
25
Data
Format
Null
i10
a8
a16
a32
f11.6
f11.6
f4.1
f4.1
-9
NO
NO
NO
NO
NO
NO
NO
NO
e11.3
e11.3
f7.3
f7.3
-999.9
-999.9
-999.9
-999.9
e11.3
-999.9
16
17
18
19
20
21
flag
correlation
cx
cy
cz
htmID
band 24 or 70 or 160
The flag currently refers to how
many times a source was observed
by different AORs in a single
epoch. Is it was observed more
than once, the reported flux is
the average of the multiple
observations.
The correlation tells how like
the input PSF each poI*4 source
is. A value of 1 would be
Identical to the PSF. A value of
-1.0 means that the source was
found to be a close double and
decomposed I*4o two sources.
The current cut for the catalog
is a correlation of 0.8 or above,
thus there are no values between
0.0 and 0.8.
x of unit vector to this source
y of unit vector to this source
z of unit vector to this source
20-deep HTM ID of this source
R*4
f6.2
I*4
R*8
R*8
R*8
R*8
I*4
-9.99
-9
e11.3
f30.20
f30.20
f30.20
i20
-999.9
NO
NO
NO
NO
Designation is
SST = Spitzer Space Telescope
M1 = MIPS 24um
SAGE= LMC Survey project
1 = epoch 1+2 (in this case 1+2, indicates hi-reliability)
C = Catalog
J = 2000.0 epoch
HHMMSS.SS Right ascension (hr, min, sec) of source
sDDMMSS.S Declination (deg, min, sec) of source
===============================================================
Here is an example of a line of text for the Catalog (21 columns)
3226
mips24
epoch 1+2
SSTM1SAGE1C J050645.56-700556.4 76.690989
0.1
0.1
0
0.237
6.818
0.038
56.66
0
0.995
0.07836034479078390000
0.33125487624516487000
0.94028179995577383000
9314491318017
26
-70.098935
13.430
20.050
-
For this source:
globalSourceID: 435834
sourceCatalog: mips24
epoch: epoch 1+2
designation: SSTM1SAGE1C J050645.56-700556.4
: Name
ra,dec:
76.690989, -70.098935
: RA & Dec in degrees
dra,ddec:
0.1 0.1
: Error in RA and Dec (")
conf:
0
: confusion flag
flux24,dflux24: 13.430 0.237
: flux (mJy) and flux uncertainty
mag24, dmag24: 6.818 0.038
: Magnitude and magnitude
uncertainty
sky: 20.050
: Local sky background
SN: 56.66
: Signal to Noise
flag: 0
: flag
correlation: 0.995
: correlation
cx: 0.07836034479078390000
: x of unit vector to this source
cy: 0.33125487624516487000
: y of unit vector to this source
cz: -0.94028179995577383000
: z of unit vector to this source
htmID: 9314491318017
: 20-deep HTM ID of this source
7. Acknowledgements
The SAGE survey was proposed by a creative team of scientists listed in the SAGE
overview paper. At the core of the SAGE team are three data centers at which the
Spitzer data are processed and distributed.
 IRAC pipeline based at the University of Wisconsin: Barbara Whitney (lead/SSI),
Marilyn Meade (Univ. of Wisc.), Brian Babler (Univ. of Wisc.), Remy
Indebetouw (Univeristy of Virginia), Joe Hora (Harvard/CfA)
 MIPS pipeline based at the University of Arizona: Karl Gordon (Univ. of
Arizona/lead), Chad Engelbracht (Univ. of Arizona), Bi-Qing For (Univ. of
Texas), Miwa Block (Univ. of Arizona), Karl Misselt (Univ. of Arizona)
 SAGE Database based at STScI: Margaret Meixner (STScI/lead), Uma Vijh
(STScI), Claus Leitherer (STScI), Bernie Shiao (STScI)
Bernie Shiao and Alberto Conti of STScI were instrumental in setting up the SAGE
database. This work was funded by NASA Spitzer grant 1275598 and NASA grant
NAG5-12595 at STScI and Spitzer grant 1275471 with UC Berkeley and Spitzer grants
at the other SAGE institutions. Support from the SSC was critical to the success of the
SAGE project, in particular, Bill Mahoney, Nancy Silberman and Lisa Storrie-Lombardi.
27
8. References
Benjamin et al 2003, PASP, 115, 953
Cohen, M., Megeath, T.G., Hammersley, P.L., Martin-Luis, F., & Stauffer, J. 2003, AJ,
125, 2645
Cohen, M., Wheaton, W.A., & Megeath, T.G. 2003, AJ, 126, 1090
Diolaiti, E., Bendinelli, O., Bonaccini, D., Close, L., Currie, D., Parmeggiani, G. 2000
A&AS, 147, 335
Engelbracht, C. et al. 2006, MIPS 24 micron calibration, PASP, submitted
Gordon et al. 2005, PASP, 117, 503
Meixner, M. et al. 2006, AJ, 132, 2268. “Spitzer Survey of the Large Magellanic Cloud,
Surveying the Agents of a Galaxy’s Evolution (SAGE) I: Overview and Initial Results,”
Meixner, M.,; Gordon, K.; Indebetouw, R.; Hora, J.L.; Churchwell, E.B.; Whitney, B.;
Blum, R.; Reach, W.; Bernard, J-P.; Meade, M.; Babler, B.; Engelbracht, C.; For, B-Q.;
Misselt, K.; Leitherer, C.; Vijh, U.; Cohen, M.; Boulanger, F.; Frogel, J.A.; Fukui, Y.;
Gallagher, J.; Gorijian, V.; Harris; J.; Kelly, D.; Kemper, C.; Kawamura, A.; Kim, S.;
Latter, W.B.; Madden, S.; Mizuno, A.; Mizuno, N.; Mould, J.; Nota, A.; Oey, S.; Olsen,
K.; Onishi, T.; Paladini, R.; Panagia, N.; Perez-Gonzalez, P.; Shibai, H.; Shuji, S.; Smith,
L.; Staveley-Smith, L.; Tielens, A.G.G.M.; Ueta, T.; Van Dyk, S.; Volk, K.; Werner, M.;
and Zaritsky, D. 2006, AJ, 132, 2268
Price, S.D., Paxson, C., Engelke, C., & Murdock, T.L. 2004, AJ, 128, 889
Reach, W.T. et al. 2005 PASP, 117, 978
Rieke, G. et al. MIPS zero points, in prep.
Stetson, P. 1987, PASP, 99, 191
28