WISEPhotometry
WISECat
v02
2016-03-22
Michelle Cluver <michelle.cluver@gmail.com>

This table contains IR photometry of GAMA sources detected in the <a href="http://wise2.ipac.caltech.edu/docs/release/allwise/expsup/">WISE
AllWISE survey</a>. WISE surveyed the entire sky at 4 mid-infrared
wavelengths: 3.4 micron (W1), 4.5 micron (W2), 12 micron (W3) and 22
micron (W4), and performed source detection in each band
independently.

In v01 of this DMU, WISE sources from the <a href="http://wise2.ipac.caltech.edu/docs/release/allsky/expsup/">WISE All-Sky survey</a> were
matched to GAMA objects by associating each VISUAL_CLASS != (2 or 3)
object from <a href="/db/schema/table.php?id=256">TilingCatv41</a> (which covers the GAMA II equatorial survey
regions) with its nearest All-Sky source within a maximum matching
radius of 3 arcsec. In the present version of this DMU we have updated
the photometry and positions from the All-Sky survey with those from
the improved AllWISE survey by replacing the values of a given All-Sky
source with those of the nearest AllWISE source within a maximum
matching radius of 1 arcsec. If no ALLWISE source was found within 
1 arcsec of the All-Sky source, the All-Sky source was retained in the
catalogue (in order not to loose any sources from v01 to v02). We have
also added all new detections of GAMA sources in AllWISE (using a
maximum matching radius of 3 arcsec as before). Note that the same
WISE source may be associated with different GAMA objects. GAMA
objects without a matched WISE source are not included in this table.

Photometry is presented in the form of WISE profile-fit measurements
(w*mpro from the WISE ALLWISE Catalogue), WISE standard aperture
measurements (w*mag from the WISE ALLWISE Catalogue) and our own
isophotal photometry for resolved sources. Isophotal measurements (to
1 sigma of the background) are made using coadded (drizzled) images
created from the individual frames taken by WISE, using the software
package <a href="http://web.ipac.caltech.edu/staff/fmasci/home/icore.html">ICORE</a> by F. Masci (see also <a href="http://adsabs.harvard.edu/abs/2012AJ....144...68J">Jarrett et al. 2012</a>).

All magnitudes are on the AB system, and aperture measurements
(isophotal and standard aperture) have been aperture corrected,
including that given as "recommended" photometry.

"Recommended" photometry is provided for each source based on whether
the source is unresolved or resolved, flagged by the parameter
PHOTFLAG:
 1 = profile-fit photometry from the WISE ALLWISE Catalogue 
 2 = isophotal (W1 and W3) and "derived" (W2 and W4) photometry (see below)
 3 = standard aperture photometry from the WISE ALLWISE Catalogue (see below)
-1 = upper limits or null photometry 

Null values are represented as -999. Upper limits for the
profile-fitted measurements are provided (from the WISE AllWISE
Catalogue), where available, and are identifiable as having an
uncertainty of -999.


COVERAGE and SENSITIVITY:

WISE surveyed the entire sky, but due to the nature of the WISE
telescope orbit, the sky is <a href="http://wise2.ipac.caltech.edu/docs/release/allsky/expsup/sec6_2.html">not uniformly surveyed</a>. In addition, some
frames were lost due to contamination by the Moon. To indicate the
depth of coverage, we include the frame coverage and mean pixel
coverage. Low coverage affects the signal to noise of detections and
therefore positional accuracy. The coverage of the G09, G12 and G15
fields does not go below 8, which is the requirement to achieve the
nominal WISE sensitivity limits. 


RESOLVED SOURCES:

Details of the photometry measurements and performance are discussed
by <a href="http://adsabs.harvard.edu/abs/2014arXiv1401.0837C">Cluver et al. (2014)</a>, but the relevant details are summarised here.

A key feature of a catalogue containing WISE photometry is determining
which galaxies are resolved and ensuring their fluxes are measured
robustly. A broad indication is given by the reduced chi^2 (w*rchi2)
of the profile-fit solution, where high values of the parameter
indicate that the profile-fit measurement is underestimating the flux
of the source. Sources with w*rchi2 > 5 are usually well-resolved.

However, w*rchi2 by itself is not a robust measure of "resolvedness",
particularly where sources have a low S/N. For example, the
profile-fit measurement can be fooled for sources with relatively
bright cores that are just resolved by WISE - this happens with 2MASS
compact extended sources which have low w1rchi2, but the profile-fit
and isophotal photometry can differ radically; this is discussed and
illustrated in Cluver et al (2014). Therefore, all sources with
w1rchi2 > 2 and those with a match in the 2MASS Extended Source
Catalogue (XSC) should be tested to determine if they are resolved by
WISE.

The parameter RFUZZY is used as a measure of how resolved a source is
at the spatial resolution of the drizzled images (6.1 and 6.4 arcsec
at 3.4 and 4.6 micron, respectively). The RFUZZY parameter is measured
in the following way:

1) The source is rotated to determine the optimal 2nd moment
parameters (Rmajor, Rminor, position angle of major axis).

2) The source is divided along the major axis into two halves:
positive x, and negative x, where the central x position is the
nominal position of the source.
 
3) The 2nd order moment is computed for each half and derive the
Rmajor for each half: Rmajor(a), Rmajor(b).

4) RFUZZY = minimum(Rmajor(a), Rmajor(b)).

If RFUZZY < RFUZZY_limit, then the source is unresolved. The
RFUZZY_limit is determined by measuring sources that are expected to
be unresolved (see Cluver et al. 2014). The typical value for RFUZZY
is ~10.0" for high S/N point sources (stars), but the distribution can
be used to identify the empirical limit where "resolvedness" can be
identified. A power series fit to the 2 sigma mean of the distribution
yields the RFUZZY_limit curve - points that lie above this curve are
classified as resolved.


STANDARD APERTURE PHOTOMETRY:

The WISE AllWISE Catalogue also includes standard aperture photometry
(in W1 and W2) which is performed in a circular aperture of 8.25"
(corrected for point-source calibration) around detected sources. For
W1 and W2 these measurements do not suffer from the known bias at the
faint end of the profile-fit in the AllWISE Catalogue. Additionally,
these appear to be better suited to compact extragalactic sources and
more useful for combining with aperture measurements in other
wavelength bands.

These measurements from the WISE AllWISE Catalogue are reproduced in
this table as MAGSTDAP_W*. They are accompanied by a quality/confusion
flag (w*flg), presented as MAGSTDAPFLAG_W* in this table:

value    Condition
-----    ------------------------------------------------------
  0	 nominal: no contamination

  1 	 source confusion: another source falls within the measurement
  	 aperture

  2 	 bad or fatal pixels: presence of bad pixels in the measurement
  	 aperture (bit 2 or 18 set)

  4	 non-zero bit flag tripped (other than 2 or 18)

  8 	 corruption: all pixels are flagged as unusable, or the aperture
         flux is negative; in the former case, the aperture magnitude
         is NULL; in the latter case, the aperture magnitude is a 95%
         confidence upper limit

 16 	 saturation: there are one or more saturated pixels in the
 	 measurement aperture

 32	 upper limit -- the magnitude is a 95% confidence upper limit


RECOMMENDED PHOTOMETRY:

Details are discussed in Cluver et al. (2014), but are summarised
here.

1) Isophotal measurements are less accurate for low S/N sources: 
if W1iso S/N < 10, W2iso S/N < 10, W3iso S/N < 15, W4 S/N < 15 then
standard aperture (W1, W2) or profile-fit photometry (W3, W4) is used
where available.

2) If RFUZZY parameter (as measured in the W1 band) indicates an
unresolved source then the standard aperture photometry is used for
bands W1 and W2.

3) If W3rchi < 2, the source is classified as unresolved and we revert
to the profile-fit photometry for band W3 (if available).

4) If W4rchi < 2, the source is classified as unresolved and we revert
to the profile-fit photometry for band W4 (if available).


DERIVED PHOTOMETRY:

The W1 and W3 bands are more sensitive than the W2 and W4 bands,
respectively. We can retain this sensitivity by deriving the W2
measurement from the W1 isophotal measurement and the W1-W2 colour,
measured in a matched aperture within the smaller of the W1 or W2
isophotal aperture. This is reasonable since the emission is driven by
similar physics, and for SEDs matched photometry is preferable
(i.e. to the same sensitivity). Hence the smaller aperture (usually
W2) determined the W1-W2 colour and combined with the W1 isophotal
magnitude, the W2 "derived magnitude" is determined. A similar
procedure is used to derive W4 from W3, although in GAMA it is very
rare the sources are resolved in W3 and W4.


BLENDING FLAGS:

Point sources are actively deblended by the WISE pipeline and the WISE
AllWISE Catalogue provides the NA and NB blending flags included in
this catalogue. NA is 0 if no active deblending has been performed and
1 if sources have been deblended; this indicates an additional
uncertainty on the measurement provided. NB is the number of
components that were deblended.

In this catalogue we also provide contamination and confusion flags
(CC_FLAGS) consisting of a four character string, one character per
band [W1/W2/W3/W4], that indicates whether the photometry and/or
position measurements of a source are contaminated or biased due to
its proximity to an image artifact. The number 0 corresponds to the
source being unaffected by known artifacts. Additional codes are
explained in the <a href="http://wise2.ipac.caltech.edu/docs/release/allwise/expsup/sec2_1a.html#cc_flags">WISE Explanatory Supplement</a>.

In addition, we include the following deblend parameters from the
LAMDAR code (Wright et al. 2016):

DeblendedFraction:
  This is the deblended fraction, [Integral of Deblended Model] /
  [Integral of Undeblended Model], as determined by LAMDAR. It has
  values between 1 (no deblending was needed), and 0 (object has been
  deblended entirely. i.e. it contributes no flux).

SApFlag:
  This is a flag indicating if standard apertures overlap between GAMA
  objects and contaminants, based on a positional match. This assumes
  _everything_ is a standard aperture (both galaxies and
  contaminants), and just checks for matches (by RA/DEC) within
  2*StdApRad. The flag has values 0,1,2:
  0 - no, object does not intersect any other object
  1 - yes, object intersects one or more objects whose mag is less 
      than AB 18.5
  2 - yes, object intersects one or more objects of any magnitude

NnearSAp: 
  Gives the number of apertures that are overlapping with this object.

GApFlag:
  As above, but uses GAMA aperture major axis (defined in r-band;
  point sources have 0 area) + 1xPSF FWHM, instead of StdAps.
  0 - no, object does not intersect any other object
  1 - yes, object intersects one or more objects whose mag is less 
      than AB 18.5
  2 - yes, object intersects one or more objects of any magnitude

NnearGAp:
  Gives the number of apertures that are overlapping with this object.

Note: 
The DeblendedFraction value from LAMDAR incorporates the flux
information of every object, and uses full profile-weighted models
when deblending objects. Because of the way that these are defined,
the DeblendedFraction value intrinsically has source/contaminant flux
incorporated, while the flags SApFlag and GApFlag are just on-sky
intersections. For this reason (and also because it is physically
interesting) these two flags are separated into the two bins 1 and 2.

Cautions:
The *ApFlag values are based on positional matches within specific
radii. This means that: the SApFlag is correct (actually measures
overlap between standard apertures), because they are circular
apertures with radius equal to half my match radius. Therefore a match
implies overlap. The GApFlag is based on the intersection of each
object (GAMA aperture majax + 1xPSF FWHM) with every other
object. This means that the flag and the Nnear numbers assume:
  1) every object is a point source, except the one that is being
     focused on for the match
  2) objects are circular with radius = major axis + FWHM
Because of the above, the GApFlag is probably the most inaccurate. But
it is nonetheless informative if you are looking to compare how
(approximately) the StdAp overlap compares to raw overlap with GAMA
apertures at WISE resolution.

Comparison of WISE NA flag and LAMBDAR SApFlag:
* 49% have deblend=no in both WISE and LAMDAR
* 40% have deblend=no in WISE and deblend=yes (any) in LAMDAR
* 9%  have deblend=no in WISE and deblend=yes (rAB<18.5) in LAMDAR
* 88 sources have deblend=yes in WISE and deblend=no in LAMDAR
* 89% have deblend=no in WISE, corresponding to deblend=no or rAB>18.5
  in LAMDAR