CANDELS Multi-wavelength Catalogs: Source Identification and Photometry in the CANDELS ...

28 Pages·2016·8.23 MB·English

Accepted to the Astrophysical Journal Supplement Series
PreprinttypesetusingLATEXstyleemulateapjv.08/22/09
CANDELS MULTIWAVELENGTH CATALOGS: SOURCE IDENTIFICATION AND PHOTOMETRY IN THE
CANDELS COSMOS SURVEY FIELD
H. Nayyeri1,2, S. Hemmati2,3, B. Mobasher2, H. C. Ferguson4, A. Cooray1, G. Barro5,6, S. M. Faber5, M.
Dickinson7, A. M. Koekemoer4, M. Peth8, M. Salvato9, M. L. N. Ashby10, B. Darvish2,11, J. Donley12, M. Durbin4,
S. Finkelstein13, A. Fontana14, N. A. Grogin4, R. Gruetzbauch15, K. Huang16, A. A. Khostovan2, D. Kocevski17,
D. Kodra 18, B. Lee19, J. Newman18, C. Pacifici20,21, J. Pforr22,23, M. Stefanon24, T. Wiklind25, S. P. Willner10, S.
Wuyts26, M. Castellano14, C. Conselice27, T. Dolch28, J. S. Dunlop29, A. Galametz9, N. P. Hathi22, R. A.
Lucas4, H. Yan30
6
1 ABSTRACT
0
Wepresentamulti-wavelengthphotometriccatalogintheCOSMOSfieldaspartoftheobservations
2
by the Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey (CANDELS). The catalog

c is based on Hubble Space Telescope Wide Field Camera 3 (HST/WFC3) and Advanced Camera for
e
Surveys(ACS)observationsoftheCOSMOSfield(centeredatRA:10h00m28s,Dec:+02◦12(cid:48)21(cid:48)(cid:48)). The
D
final catalog has 38671 sources with photometric data in forty two bands from UV to the infrared (
1 0.3 8µm). Thisincludesbroad-bandphotometryfromtheHST,CFHT,Subaru,VISTAandSpitze∼r
2 Spa−ce Telescope in the visible, near infrared and infrared bands along with intermediate and narrow-
bandphotometryfromSubaruandmediumbanddatafromMayallNEWFIRM.Sourcedetectionwas

] conducted in the WFC3 F160W band (at 1.6µm) and photometry is generated using the Template
A FITting algorithm. We further present a catalog of the physical properties of sources as identified in
G theHSTF160Wbandandmeasuredfromthemulti-bandphotometrybyfittingtheobservedspectral
energy distributions of sources against templates.
.
h Subject headings: catalogs - galaxies: high-redshift – galaxies: photometry - methods: data analysis -
p techniques: image processing
-
o
r
t
s 1Department of Physics and Astronomy, University of Califor- 1. INTRODUCTION
a
niaIrvine,Irvine,CA92697
[ 2Department of Physics and Astronomy, University of Califor- TheCosmicAssemblyNear-infraredDeepExtragalac-

niaRiverside,Riverside,CA92521 tic Legacy Survey (CANDELS: PI. S. Faber and H. Fer-
1 3Infrared Processing and Analysis Center, California Institute guson;seeGroginetal.2011andKoekemoeretal.2011)
v ofTechnology,MS100-22,Pasadena,CA91125
4 4SpaceTelescopeScienceInstitute,3700SanMartinDrive,Bal- isthelargestMulti-CycleTreasury(MCT)programever
6 timore,MD21218,USA approved on the HST, with more than 900 orbits, and
5UCO/LickObservatory,DepartmentofAstronomyandAstro- it was designed to use deep observations by the Wide
3
7 ph6ysDicesp,aUrtnmiveenrtsitoyf AofstCraolnifoomrnyi,aU,SnaivnetrasiCtyruozf,CCaAlif9o5r0n6ia4,aUtSBAerke- Field Camera 3 (WFC3) and Advanced Camera for Sur-
0 ley,Berkeley,CA94720-3411,USA veys (ACS) instruments to study galaxy formation and
. 7National Optical Astronomy Observatory, Tucson, AZ 85719, evolution throughout cosmic time in five fields in many
2 USA different bands. The observations were done by the
1 8Department of Physics and Astronomy, The Johns Hopkins
HST/WFC3asthemainmodewithACSobservationsin
6 University,366BloombergCenter,Baltimore,MD21218,USA
1 9Max-Planck-Institut fur Extraterrestrische Physik, Giessen- parallel. The CANDELS images are publicly available,
bachstrasse1,D-85748GarchingbeiMunchen,Germany and multi wavelength photometric catalogs are made
:
v 10Harvard Smithsonian Center for Astrophysics, 60 Garden available by the CANDELS team following the release
i Street,Cambridge,MA02138,USA oftheimages. AllCANDELSphotometriccatalogswere
X 11Cahill Center for Astrophysics, California Institute of Tech-
nology,1216EastCaliforniaBoulevard,Pasadena,CA91125,USA selectedbasedontheWFC3F160Wbandandthisisthe
r 12Los Alamos National Laboratory, Los Alamos, NM 87544, referenceimageforalltheotherHSTandnon-HSTdata.
a
USA Thisprovidesadata-setwithconsistentphotometryand
13DepartmentofAstronomy,TheUniversityofTexasatAustin,
Austin,TX78712,USA
14INAF - Osservatorio Astronomico di Roma, via Frascati 33, 23ESA/ESTECSCI-S,Keplerlaan1,2201AZ,Noordwijk,The
I-00040MontePorzioCatone(RM),Italy Netherlands
15CenterforAstronomyandAstrophysics,ObservatorioAstro- 24HuygensLaboratoryNielsBohrweg22333CALeiden
nomicodeLisboa,TapadadaAjuda,1349-018Lisboa,Portugal 25DepartmentofPhysics,CatholicUniversityofAmerica,20064
16Department of Physics, University of California, Davis, CA WashingtonDC
95616,USA 26DepartmentofPhysics,UniversityofBath,ClavertonDown,
17ColbyCollege,4000MayflowerHill,Waterville,Maine04901 Bath,BA27AY,UK
18DepartmentofPhysicsandAstronomyandPITTPACC,Uni- 27SchoolofPhysicsandAstronomy,UniversityofNottingham,
versityofPittsburgh,Pittsburgh,PA15260,USA Nottingham,UK
19Department of Astronomy, University of Massachusetts, 710 28Department of Physics, Hillsdale College 33 E. College St.
NorthPlesantStreet,Amherst,MA01003,USA Hillsdale,MI49242
20NASAPostdoctoralProgramFellow 29InstituteforAstronomy, UniversityofEdinburgh, RoyalOb-
21GoddardSpaceFlightCenter,Code665,Greenbelt,MD,USA servatory,EdinburghEH93HJ,UK
22Aix Marseille Universite, CNRS, LAM (Laboratoire dAstro- 30Department of Physics and Astronomy, University of Mis-
physiquedeMarseille)UMR7326,13388,Marseille,France souri,Columbia,MO65211,USA
2
physical properties across all the fields targeted as part celestial equator. It was initially picked to maximize
of the survey. The CANDELS catalogs for the first two the visibility from observatories from both hemispheres
observed fields of UDS and GOODS-S (Galametz et al. and was specifically chosen to avoid any bright X-ray,
2013;Guoetal.2013)arealreadypubliclyavailable1 and UV or radio sources (Scoville et al. 2007b) and to be
the three fields of COSMOS (this work), EGS (Stefanon large enough for studies of large scale structure (e.g.
et al. 2016) and GOODS-N (Barro et al., in prep.) are Scoville et al. 2007a; Kovaˇc et al. 2010; Scoville et al.
in progress. The CANDELS observations are aimed at 2013; Darvish et al. 2014, 2015b).
achieving several major science goals that could only be The COSMOS field was targeted by CANDELS in
attained with data at the depth and resolution of CAN- a north-south strip, lying within the central ultra-deep
DELS. These include studying the most distant objects stripoftheUltraVISTAimaging(McCrackenetal.2012)
intheUniverseattheepochofreionizationinthecosmic and hence also the Spitzer SEDS imaging (Ashby et al.
dawn (e.g. Finkelstein et al. 2012a; Grazian et al. 2012; 2013)inordertoensurethebestpossiblesupportingdata
Yan et al. 2012; Lorenzoni et al. 2013; Oesch et al. 2013; at longer near-infrared wavelengths. The HST observa-
Duncanetal.2014;Bouwensetal.2015;Giallongoetal. tions cover an area of 216arcmin2 in the WFC3/IR
2015; Finkelstein et al. 2015; Song et al. 2015; Roberts- with parallel ACS obse(cid:39)rvations. The catalog was se-
Borsani et al. 2015; Caputi et al. 2015; Mitchell-Wynne lected in the HST/WFC3 F160W band and has multi-
et al. 2015), understanding galaxy formation and evolu- band data for 38671 objects from 0.3 to 8µm. These
tion during the peak epoch of star formation in the cos- fluxes are measured consistently a∼cross all these bands
mic high noon (e.g. Bruce et al. 2012; Bell et al. 2012; and in agreement with photometry measurement tech-
Kocevski et al. 2012; Wuyts et al. 2013; Lee et al. 2013; niques adopted by all CANDELS catalogs. We use the
Hemmati et al. 2014; Barro et al. 2014; Whitaker et al. unprecedenteddepthandresolutionprovidedbytheHST
2014; Williams et al. 2015; Hemmati et al. 2015) and for measuring the flux of the faintest targets across all
studying star-formation from deep UV observations and bands. By fitting this multi-waveband information with
cosmological studies from supernova observations (e.g. templatelibraries,wealsomeasuredphotometricredshift
Jonesetal.2013;Teplitzetal.2013;Rodneyetal.2014; and stellar mass for each object.
Strolgeretal.2015;Rodneyetal.2016). Thesemainsci- Thepaperisorganizedasfollow. Section2presentsthe
ence goals are described in more detail by Grogin et al. data used to compile the catalog. Section 3 describes
(2011) and Koekemoer et al. (2011). our photometry of the HST, Spitzer and ground-based
One of the major goals of modern observational cos- bands. Section 4 is devoted to data quality checks for
mologyistostudytheformationandevolutionofgalax- our TFIT photometry. Physical parameter estimation
ies with cosmic time. Recent advances in this frontier using the measured photometry is presented in Section
have been enabled by the availability of observations in 5. In Section 6 we investigate the applications of our
different wavelengths, targeting different populations of deep photometry on studies of high redshift star form-
galaxies (e.g. York et al. 2000; Giavalisco et al. 2004; ing and quiescent galaxies. We summarize our results in
Skrutskieetal.2006;Lawrenceetal.2007;Scovilleetal. Section 7. Throughout this paper we assume a cosmo-
2007b; Wolf et al. 2003; Bell et al. 2004; Faber et al. logical model with H =70kms−1Mpc−1, Ω =0.3 and
0 m
2007; Ilbert et al. 2013; Khostovan et al. 2015; Hemmati Ω = 0.7. All magnitudes are in the AB system where
Λ
et al. 2016; Vasei et al. 2016; Shivaei et al. 2015). The m =23.9 2.5 log(f /1µJy) (Oke & Gunn 1983).
AB ν
advent of the HST benefited many such studies by mak- The CAND−ELS×COSMOS photometry catalog will be
ingitpossibletohavethedeepestobservationsofthesky publiclyavailablethroughtheCANDELSwebsite2 along
in multiple bands. In particular the installation of the with the physical properties estimates and all the docu-
WideFieldCamera3(WFC3)onboardHSTinitiateda mentations. These will also be available on the Mikulski
newstageforstudyinggalaxyevolutionatnewextremes Archive for Space Telescopes (MAST)3, via the online
(McLeod et al. 2015; Oesch et al. 2016). version of the catalog and through Centre de Donnees
Galaxy populations at different look-back times range astronomiques de Strasbourg (CDS). We will also make
fromverybluestarforminggalaxiestoreddustyorvery these data available through the Rainbow Database4
oldsystems. Understandingtheevolutionofthesepopu- (P´erez-Gonz´alez et al. 2008; Barro et al. 2011).
lationsreliesontheavailabilityofmulti-wavelengthpho-
tometric data from the bluest to the reddest bands pos- 2. DATA
sible. The CANDELS multi-wavelength catalogs com-
The COSMOS field (Scoville et al. 2007b) has been
bine the best and deepest observations by the HST
observed in many different wavelengths from the X-
with the deepest ground-based observations and Spitzer
ray to the far infrared. There are observations from
Space Telescope data (Galametz et al. 2013; Guo et al. the CFHT/MegaPrime in the u∗, g∗, r∗, i∗ and z∗
2013). Thesecatalogsoftensofthousandsofextragalac-
(Gwyn 2012), from the Subaru/Suprime-Cam in the B,
tic sources, consistently measured across many bands g+, V, r+, i+ and z+ (Taniguchi et al. 2007), from
from 0.3 8µm, bring a unique opportunity to study
the VLT/VISTA in the Y, J, H and K bands (Mc-
galaxy∼evolu−tion. s
Crackenetal.2012),fromSpitzerinthefourIRACbands
The Cosmic Evolution Survey (COSMOS; Scov-
(Sanders et al. 2007; Ashby et al. 2013), VLA (Schin-
ille et al. 2007b) centered at RA:10h00m28s, nerer et al. 2007), XMM (Hasinger et al. 2007; Cap-
Dec:+02◦12(cid:48)21(cid:48)(cid:48)) is a 2deg2 field located near the pelluti et al. 2007), Chandra (Elvis et al. 2009; Civano
1http://vizier.cfa.harvard.edu/viz-bin/VizieR?-source=J/ApJS/ 2http://candels.ucolick.org/
206/10 and http://vizier.cfa.harvard.edu/viz-bin/VizieR? 3https://archive.stsci.edu/
-source=J/ApJS/207/24 4https://rainbowx.fis.ucm.es/Rainbow navigator public/
3
Figure 1. The transmission curves for the CFHT and ACS visible (top left), Subaru optical broad-band and narrow-band (top right),
WFC3andUltraVISTAnearinfraredandSpitzerinfrared(bottomleft)andtheSubaruintermediateandNEWFIRMmediumbandfilters
(bottomright)usedintheCOSMOSCANDELSTFITcatalog. Thesecoverobservationsat∼0.3−8µminfortytwofilters. Thefilters
areadoptedfromhttp://cosmos.astro.caltech.edu/page/filterset. TheeffectivewavelengthofthefiltersarereportedinTables1and2.
et al. 2016) and GALEX (Schiminovich et al. 2005). (J and H ) and in parallel by the Advanced Cam-
125 160
There are also numerous medium and narrow-band ob- era for Surveys (ACS) in the F606W and F814W fil-
servations available in the COSMOS from the Mayall ters (V and i ). The WFC3 observations covered a
606 814
NEWFIRM(Whitakeretal.2011)andSubaruSuprime- rectangular grid of 4 11 tiles ( 8(cid:48).6 23(cid:48).8) running
Cam (Taniguchi et al. 2015). The full COSMOS field north to south allowin×g for maxi∼mum c×ontiguous cover-
has been observed with HST in F814W (Koekemoer age in the near infrared. The field was observed over
et al. 2007) and contains more than 2 million galax- two epochs with each tile observed for one orbit in each
ies with multi-band data from the UV to the far-IR epoch. The one orbit observations were divided into two
(Mobasher et al. 2007; Capak et al. 2007; Ilbert et al. exposures in F125W ( 1/3 orbit depth) and F160W
2009, 2013). Furthermore, COSMOS has been followed ( 2/3 orbit depth) a∼long with parallel ACS observa-
upspectroscopicallybythelargerandmostsensitivetele- ti∼ons in the F606W and F814W (Koekemoer et al. 2011;
scopes/instruments like VLT/VIMOS (e.g. Lilly et al. Grogin et al. 2011). The exposures in each orbit were
2009; Le F`evre et al. 2015) and Keck/DEIMOS and dithered using a small scale pattern providing half-pixel
MOSFIRE among others5. This makes it possible to subsampling of the PSF and also ensuring that the hot
study different populations of galaxies from blue and pixels and persistences were moved around. The output
young systems at shorter wavelengths to red dusty or is a calibrated and astrometry-corrected mosaics of all
old objects at longer wavelengths. These ancillary data the exposures in the four individual HST bands. The
are accompanied by high resolution observations from astrometry is based on the CFHT/MegaCam i∗ imaging
CANDELSusingtheHubbleSpaceTelescopeinbothvis- supplementedbydeepSubaru/Suprime-Cami+ imaging
ible and near infrared. Figure 1 shows the transmission with absolute astrometry registered to the VLA 20cm
curvesforallbandsincludedintheCANDELSCOSMOS survey of the COSMOS field (Schinnerer et al. 2007; see
catalog. alsoKoekemoeretal.2007). Thisisalsotheadoptedref-
erence grid by the COSMOS team (Capak et al. 2007).
All the ground-based and Spitzer data (described in the
2.1. CANDELS HST Observations next Section) were aligned to the HST data astrometry
The CANDELS COSMOS field was observed by the using SWARP. For the present work we used the V0.5 re-
Wide Field Camera 3 (WFC3) in F125W and F160W lease of the HST/ACS and WFC3 data available from
5Foracompletelistofancillarydataontheentirefieldvisit: http:
//astro.caltech.edu/∼cosmos
4
Figure 2. SkycoverageofdataintheCOSMOSfield. TheWFC3F160Wmosaicsareshownasthegreyshadedregion. TheentireWFC3
footprintiscoveredbytheCFHT,SubaruandUltraVISTAobservations(whicharemuchlargerthanthescalesofthisFigure).
the CANDELS website6. The observation depth, effec- individual frames were combined, flat-fielded and pho-
tive wavelength and the PSF information for each of the tometry and astrometry calibrated. We refer the reader
HST filters are summarized in Table 1. to Capak et al. (2007) and Taniguchi et al. (2007, 2015)
for further description of the observations and data pro-
2.2. Ground-based Observations
cessing. There is also further observations of the COS-
The full COSMOS field was targeted by the MOSfieldbySubaru/Suprime-Camintwelveintermedi-
Canada-France-Hawaii Telescope (CFHT) 3.6m tele- atebandsalongwithnarrow-banddataacrosstwofilters
scopeMegaPrimeinstrumentinthe u∗, g∗, r∗, i∗ andz∗ (Capak et al. 2007; Taniguchi et al. 2007, 2015) cover-
optical bands as part of the CFHT Legacy Survey7 with ing the wavelength range of 4000 8500˚A. These
COSMOS being in the second Deep field. The Mega- observations were processed si∼milarly t−o the broad-band
Cam/MegaPrime camera used for the observations has
opticaldatadiscussedabove(Taniguchietal.2015). Al-
a pixel scale of 0.187arcsecpixel−1(Boulade et al. 2003). though slightly shallower than the optical broad-band
The final images are from MegaPipe8 (Gwyn 2008) with observations, these filters have higher resolving power
zeropointadjustments. Theimageprocessingandstack- than the former (with R = λ/∆λ 23; Taniguchi et al.
ing of the data is further described by Gwyn (2012). In 2015), equivalent to low resolution∼spectroscopy in the
this work we used the CFHTLS D2 mosaics 2009 release optical. The resolving power is even higher for the two
accessible from the CADC9. narrow-bandfilters(R 50 100;Taniguchietal.2015)
The COSMOS field was also observed by the although with smaller∼wave−length coverage. This pro-
Subaru/Suprime-Cam in the B, g+, V, r+, i+ and z+ vides a unique dataset for studying emission line galax-
broad-bandfilters. TheSuprime-Camhasafieldofview ies and high-redshift systems such as Lyman-α emitters
of34(cid:48) 27(cid:48) withapixelscaleof0.202arcsecpixel−1. The (Shimasaku et al. 2006; Iwata et al. 2009; Koyama et al.
data w×ere processed using the IMCAT package10. The 2014). Table 2 summarizes the Subaru intermediate and
narrow-band observations. We used the Subaru V2 mo-
6http://candels.ucolick.org/data access/Latest Release.html saicsforbroad-bandandNB816observationsandV1mo-
7http://www.cfht.hawaii.edu/Science/CFHTLS/ saicsforintermediateandNB711dataavailablefromthe
8http://www.cadc-ccda.hia-iha.nrc-cnrc.gc.ca/en/megapipe/ IRSA11 in this work.
cfhtls/index.html
9http://www.cadc-ccda.hia-iha.nrc-cnrc.gc.ca/en/search/
?collection=CFHTMEGAPIPE&noexec=true 11http://irsa.ipac.caltech.edu/data/COSMOS/images/subaru/
10http://www.ifa.hawaii.edu/∼kaiser/imcat/ mosaics/
5
Figure 3. Left: The1σ limitingmagnitudedistribution(perbinof0.01)intheWFC3F160Wdetectionbandineachpixelnormalized
toanareaof1arcsec2. Middle: Cumulativedistributionoftheareawithasensitivitygreaterthanagiven1σ limitingmagnitude. Right:
Distributionoftheexposuretime.
Figure 4. Left: The number of galaxies in the CANDELS COSMOS catalog in bins of F160W magnitude (black filled circles). The
number counts of the cold-mode selected galaxies (bright sample) and hot-mode only selected galaxies (faint sample) are shown in blue
andredrespectively. TheuncertaintiesassociatedwiththetotalcountsarePoissonerrors. Thecountsandtheassociateduncertaintiesare
reported in Table 3. Right: Number counts of the combined CANDELS COSMOS catalog compared to the CANDELS UDS (Galametz
etal.2013)andthe3D-HSTCOSMOS(Skeltonetal.2014).
The ground-based near infrared observations are from DR1mosaicsoftheUltraVISTAdatafortheCANDELS
the Visible and Infrared Survey Telescope for Astron- multi-wavelength catalog13.
omy (VISTA; Emerson & Sutherland 2010) 4.1m tele- The COSMOS field also has been observed by the
scopeVIRCAMlarge-formatarraycamera(Daltonetal. NOAO Extremely Wide-Field Infrared Imager (NEW-
2006) in the Y, J, H and K bands with mean pixel FIRM)ontheMayall4mtelescopeaspartoftheNEW-
s
scale of 0.34arcsecpixel−1. The complete contiguous FIRM Medium Band Survey14 (NMBS; Whitaker et al.
1.5deg2ofUltraVISTAobservationsweredoneusinga 2011). The NEWFRIM observations in the COSMOS
∼stripespatternwith 0.7deg2 ofthefieldobservedwith cover an area of 27(cid:48).6 27(cid:48).6 encompassing the CAN-
longer exposure (in f∼our stripes) separating the observa- DELS HST observation×s. The observations are over five
tions into deep and ultra-deep regions (McCracken et al. medium-band filters of J , J , J , H and H cover-
1 2 3 1 2
2012) with the CANDELS HST observations inside one ing the wavelength of 1 1.8µm and the K filter cen-
of the ultra-deep stripes. The data were pre-processed tered at 2.2µm. The thr−ee medium-band J , J and J
1 2 3
at CASU12, which includes dark subtraction, flat field- filters are a single broad-band J filter split into three
ing, gain normalization and initial sky subtraction (Ir- and the two medium-band H and H filters combine
1 2
win et al. 2004; McCracken et al. 2012). The data were into a single broad-band H (van Dokkum et al. 2009;
further processed at TERAPIX using an iterative sky- Whitaker et al. 2011). The final mosaic has a pixel scale
background removal technique and resampled to a pixel of 0.3arcsecpixel−1. Whitaker et al. (2011) further dis-
scaleof0.15arcsecpixel−1. McCrackenetal.(2012)give cuss the data processing. Here we used the first data
more details on the data processing. We used the final
13http://irsa.ipac.caltech.edu/data/COSMOS/images/
12http://casu.ast.cam.ac.uk/surveys-projects/vista/technical/ Ultra-Vista/
data-processing 14http://www.astro.yale.edu/nmbs/Overview.html
6
HST/F160W N Residual Subaru V
E
10"
0 0.05 0.1 0.15 0.2 0.25 0.3 -0.003 -0.002 -0.001 0 0.001 0.002 0.003
Residual UltraVISTA Y Residual IRAC Ch1
-0.015 -0.01 -0.005 0 0.005 0.01 0.015 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08
Figure 5. The WFC3 F160W source selection band (top left) along with TFIT residual maps in the optical (top right), near infrared
(bottomleft)andIRACinfrared(bottomright). ThemapsallshowthesameareaastheWFC3F160Wband,areinµJy/pixelunitsand
areallscaledlinearlyasshownbytherespectivecolorbars. ThebackgroundnoisefortheSubaruV,UltraVISTAY andIRAC3.6µmare
at the levels of 0.001µJy, 0.004µJy and 0.006µJy respectively (images are background subtracted as discussed in Section 3). The arrow
markedontheWFC3mapsshowsareferenceobjectatafluxdensityof74µJy.
release of the NMBS data (DR1) of the COSMOS field cfa.harvard.edu/SEDS/data.html for photometry mea-
available from the NOAO science archive15. surements in the 3.6µm and 4.5µm bands and the S-
COSMOS Spitzer/IRAC 5.8µm and 8.0µm GO2 (V2)
2.3. Spitzer InfraRed Observations data (Sanders et al. 2007) available from http://irsa.
ipac.caltech.edu/data/SPITZER/S-COSMOS/. Figure
The COSMOS field was observed by the Spitzer
2 shows the sky coverage of the different ground based
Space Telescope IRAC instrument (Fazio et al. 2004)
and space data in the COSMOS CANDELS.
at 3.6µm, 4.5µm, 5.8µm, 8.0µm as part of the S-
COSMOS (Sanders et al. 2007). The 3.6µm and 3. CATALOG PHOTOMETRY
4.5µm bands have much deeper data from the Spitzer
In generating the multi-wavelength catalog, the high
Extended Deep Survey16 (SEDS; Ashby et al. 2013).
resolutiondata(HST/ACSandWFC3)weretreateddif-
The SEDS observations cover a strip of 10(cid:48) 1◦ ori-
ferently from the low resolution data (ground-based and
entedNorth-SouthcoincidingwiththedeepVIS×TAdata
Spitzer IRAC).
mentioned above and incorporate previous 3.6µm and
4.5µm data from the S-COSMOS providing a uniform 3.1. HST Photometry
depth of 26mag (3σ) for all observations (Ashby et al.
Weperformedphotometryonthehighresolution(ACS
2013). The 5σ limiting magnitude and FWHM size + WFC3) data using SExtractor software version
of the PSF in each IRAC band are reported in Table
2.8.6 (Bertin & Arnouts 1996) in dual mode with the
1. In this work we used the SEDS first data release
WFC3 F160W as the detection band consistently with
(V1.2) (Ashby et al. 2013) available from https://www.
the multi-wavelength catalogs in the other four CAN-
DELS fields (Galametz et al. 2013; Guo et al. 2013).
15http://r2.sdm.noao.edu/nsa/nsa form.html SExtractor software was modified in several ways to
16The S-CANDELS data (Ashby et al. 2015) were not used here
enhancetheskymeasurement,addanewcleaningproce-
because they were not available at the time of catalog compila-
tion. dure and fix isophotal-corrected magnitude calculations
7
Table 1
SummaryoftheCANDELSCOSMOSBroad-BandData.
Instrument Filter(cid:63) Effective PSF 5σ limitingdepth‡ Version Reference
Wavelength† FWHM
(˚A) (arcsec) (ABMagnitude)
CFHT/MegaPrime u∗ 3817 0.93 27.31 July2009 Gwyn(2012)
g∗ 4860 1.08 27.69 − −
r∗ 6220 0.84 27.18 − −
i∗ 7606 0.85 27.23 − −
z∗ 8816 0.84 26.18 − −
Subaru/Suprime-Cam B 4448 0.95 27.98 V2 Taniguchietal.(2007)
g+ 4761 1.58 26.78 − −
V 5470 1.33 26.86 − −
r+ 6276 1.05 27.18 − −
i+ 7671 0.95 26.95 − −
z+ 9096 1.15 25.55 − −
HST/ACS F606W 5919 0.10 28.34 V0.5 Koekemoeretal.(2011)
F814W 8060 0.10 27.72 − −
HST/WFC3 F125W 12486 0.14 27.72 − −
F160W 15369 0.17 27.56 − −
VISTA/VIRCAM Y 10210 1.17 25.47 DR1 McCrackenetal.(2012)
J 12524 1.07 25.26 − −
H 16431 1.00 24.87 − −
Ks 21521 0.98 24.83 − −
Spitzer/IRAC 3.6µm 35569 1.80 24.41 V1.2 Ashbyetal.(2013)
4.5µm 45020 1.86 24.40 − −
5.8µm 57450 2.13 21.28 V2 Sandersetal.(2007)
8.0µm 79158 2.29 21.20 − −
Notes. (cid:63): Filtersadoptedfromhttp://cosmos.astro.caltech.edu/page/filterset. †: Calculatedas: λ =(cid:113)((cid:82) S(λ)λdλ)/((cid:82) S(λ)λ−1dλ)
eff
withS(λ)thefilterresponsefunction(Tokunaga&Vacca2005). ‡: The5σ limitingmagnitudecalculatedwithinacircularaperturewith
aradiusrap=FWHMofthePSFineachfilter.
2013; Guo et al. 2013), it is impossible to identify all
galaxies using a single set of parameters (signifying the
area, signal to noise, background, etc) for the extrac-
tion. The challenge is to detect the brightest targets
while avoiding blending and also detect the faintest ob-
jects without introducing spurious sources into the cata-
log. To this end we used two sets of SExtractor input
parameters. One set of parameters is aimed at bright
source detection with a focus on deblending extended
sources (cold mode), and a second set on faint galax-
ies (hot mode). The two catalogs generated by the hot
andcoldparameterswerethencombinedfollowingarou-
tineadoptedfromGALAPAGOS17 (Bardenetal.2012).
The combined catalog includes all the sources from the
cold mode catalog plus sources in the hot mode catalog
that do not exist in the cold mode as identified by the
Kron ellipse of a cold mode detected source as discussed
by Galametz et al. (2013). Figure 3 shows the 1σ de-
tection limit of the combined catalog computed over a
circularaperturewitharadiusof1arcsecalongwiththe
cumulativedistributionofthedetectionareaandtheex-
Figure 6. The5σlimitingmagnitudeofthedifferentobservations posure time distribution in the F160W band. Figure 4
in the CANDELS COSMOS (Tables 1 and 2) as a function of
shows the magnitude distribution of the sources in the
the wavelength. The symbol sizes are proportional to the filter
responsefunctionwidths. hot, cold and combined catalogs along with the compar-
isonoftheF160WcombinedcountswiththeCANDELS
UDS (Galametz et al. 2013) and 3D-HST (Skelton et al.
as discussed by Galametz et al. (2013). In order to 2014). Table 3 gives the number counts in magnitude
measure the photometry in the visible bands we PSF- bins for the combined catalog along with the associated
matched the ACS and WFC3 images and extracted the Poissonian uncertainties.
photometry from the matched images in dual mode.
AsshowninthephotometricstudiesoftheCANDELS 17http://astro-staff.uibk.ac.at/∼m.barden/galapagos/
UDS and CANDELS GOODS-S fields (Galametz et al.
8
Table 2
SummaryoftheCANDELSCOSMOSMedium-BandandNarrow-BandData.
Instrument Filter Effective PSF 5σ limitingdepth Version Reference
Wavelength FWHM
(˚A) (arcsec) (ABMagnitude)
Subaru/Suprime-Cam IA484 4849 1.14 26.34 V1 Taniguchietal.(2007,2015)
IA527 5261 1.60 26.04 − −
IA624 6232 1.05 26.21 − −
IA679 6780 1.58 25.65 − −
IA738 7361 1.08 25.94 − −
IA767 7684 1.65 25.33 − −
IB427 4263 1.64 25.91 − −
IB464 4635 1.89 25.66 − −
IB505 5062 1.44 25.82 − −
IB574 5764 1.71 25.66 − −
IB709 7073 1.58 25.79 − −
IB827 8244 1.74 25.44 − −
NB711 7120 0.79 25.56 − −
NB816 8149 1.00 26.10 V2 −
Mayall/NEWFIRM J1 10460 1.19 24.60 DR1 Whitakeretal.(2011)
J2 11946 1.17 24.32 − −
J3 12778 1.12 24.26 − −
H1 15601 1.03 23.86 − −
H2 17064 1.24 23.45 − −
K 21700 1.08 23.80 − −
Atthisstagewealsoassignedaphotometryflagtoev-
Table 3
eryobjectinthecatalog. Theflaggingsystemisthesame
HST/WFC3F160Wnumbercountsfromthecombinedhot+cold
asthatadoptedbytheotherCANDELSfields(Galametz catalog.
et al. 2013; Guo et al. 2013) and discussed in detail by
Galametzetal.(2013). Weuseazeroforagoodphotom- WFC3F160W† N(deg−2mag−1) PoissonUncertainty
etryintheflaggingandassignedavalueofoneforbright 15.25 70 50
stars and spikes associated with those stars as photome- 15.75 387 117
16.25 563 141
tryforobjectscontaminatedbythiswouldbeunreliable.
16.75 633 149
Aphotometricflagoftwoisassociatedwiththeedgesof
17.25 1196 205
the image as measured from the F160W RMS maps.
17.75 1653 241
18.25 2638 305
3.2. Ground-based and Spitzer Photometry 18.75 3166 334
19.25 4960 418
In order to measure the photometry in the ground-
19.75 6965 495
based and Spitzer bands, we used the Template FITting 20.25 9990 593
method(TFIT;Laidleretal.2007)similarlytotheother 20.75 13789 696
CANDELS multi-wavelength catalogs (Galametz et al. 21.25 17377 782
2013; Guo et al. 2013, Stefanon et al. 2016). TFIT is 21.75 27086 976
a robust algorithm for measuring photometry in mixed 22.25 33735 1089
resolution data sets. Sources that are well separated in 22.75 44956 1257
thehighresolutionimage(HST)couldbeblendedinthe 23.25 62720 1485
23.75 82841 1707
low resolution image (ground-based or Spitzer). TFIT
24.25 106200 1933
uses position and light profiles from the high resolution
24.75 133426 2166
image to calculate templates that are used to measure
25.25 156045 2343
thephotometryinthelowresolutionimage. Itdoesthat
25.75 178769 2508
by smoothing the high resolution image to match the 26.25 201810 2664
PSFofthelowresolutionimageusingaconvolutionker-
nel (Galametz et al. 2013; Guo et al. 2013). Fluxes in †: Bincentermagnitude.
the low resolution image are then measured using these
templates while fitting the sources simultaneously.
TFIT requires some pre-processing of low resolution smoothing and source masking that led to a noise
image in terms of orientation and pixel scale. The indi- map that was interpolated to determine the background
vidual steps taken are: (Galametz et al. 2013).
Background Subtraction: The low resolution im- Image Scale: TFITrequiresthelowresolutionimage
ages must be background-subtracted before running pixelscaletobeanintegermultipleofthehighresolution
TFIT. We used a background subtraction routine with detection image (0.06arcsecpixel−1 for the F160W) and
several iterations that included a first estimate through that both images have the same orientation. Further-
smoothing the image on large scales followed by PSF more the astrometry of the low resolution images must
9
be consistent with the high resolution observations. We along with the high resolution F160W detection band.
used SWARP to resample low resolution data sets to the The residual maps in the visible and near-infrared are
next larger pixel scale that is an integer multiple of the closetozeroandonlyshowresidualsinthecenterofvery
WFC3mosaicandalsouseditforastrometryandimage bright objects. However, as argued by Guo et al. (2013),
alignment. the residual maps are qualitative representations of the
Point Spread Function and Kernel: The point TFIT photometry measurements, and we later show in
spread function of both the high resolution and low res- our data quality checks that the photometry is properly
olution images are needed for the TFIT pre-processing. measured for these bright objects. Tables 1 and 2 sum-
We constructed the PSF by stacking isolated and unsat- marize the PSF FWHMs used for the high and low res-
urated stars in each band using custom IDL routines. olution images.
We also constructed a kernel to convolve the high reso- The final F160W SExtractor catalog has 38671
lution templates to the low resolution ones. The kernel sourcesoverthe216arcmin2areaoftheCANDELSCOS-
wasconstructedusingaFourierspaceanalysistechnique MOSfieldwithWFC3F160Wobservations. TFITkeeps
similartoGalametzetal.(2013),whichtakestheratioof theoriginalF160WSExtractorIDandcoordinatesfor
theFouriertransformofeachPSF.ThisgivestheFourier eachobjectinthecombinedhot+coldcatalogandthere-
transform of the kernel which is then transformed back fore we only need to combine corresponding entries from
intonormalspacegeneratingthekernel. Asdiscussedby the high resolution and low resolution catalogs. Figure
Galametzetal.(2013)alowpassbandfilterisappliedin 6 shows the 5σ limiting depth for all the filters in the
the Fourier space to cancel the high frequency fluctua- catalog as computed and tabulated in Tables 1 and 2.
tions and remove the effect of noise. For Spitzer/IRAC , We further derive and report a weight for each target in
which has the largest difference in resolution from HST, the catalog calculated from the F160W RMS maps at
one could use the PSFs directly as the convolution ker- the SExtractor positions for each object as described
nel (Galametz et al. 2013). We generated a model PSF in Guo et al. (2013).
by averaging a set of oversampled PSFs that measure
PSF variations across the detector. The final PSF is a 4. DATA QUALITY CHECKS
boxcarkernelsmoothedandfluxnormalizedmodelwhich
We checked the TFIT measured photometry by com-
alsoincorporatesallthePAsassociatedwithdifferentAs-
paring it with other independently measured photome-
tronomical Observation Requests (AORs). We refer the
try in the field. Additionally we checked the colors of
reader to Galametz et al. (2013) and Guo et al. (2013)
point sources in the catalog against model predictions
for more detail.
and color-color plots.
Dilation Correction for High Resolution Im-
ages: TFIT uses the area of the galaxy identified from
4.1. Stars Color Checks
the high resolution HST bands to measure the pho-
Thecolorofstarschangesasafunctionoftheirspectral
tometry. These are pixels defined by the high resolu-
typewhichinturndependsonthemass(andhencetem-
tion segmentation maps and are fed into TFIT in the
perature) among other parameters (Kurucz 1979; Van-
form of the isophotal areas of the high resolution image.
denberg 1985; Baraffe et al. 1998). Using this, we could
However, as demonstrated by Galametz et al. (2013),
compare the measured color of the point-like objects in
SExtractor usually underestimates the isophotal area
our catalog against predictions of the colors of stars de-
of faint or small galaxies, and this leads to an under-
rived from stellar physics. The predicted colors of stars
estimate of the flux for such systems. Galametz et al.
werecomputedfromthestellarlibraryofBaSeL(Lejeune
(2013) performed extensive tests to quantify and correct
et al. 1997; Westera et al. 2002) to measure the model
for this effect, the so-called “dilation correction”, by re-
finingandapplyingthepublicdilatecode. Thesesimu- stellarcolors. Wepresentthecomparisonsoncolor-color
plots using several observed filters. To measure the pre-
lations showed that the correction factor is negligible for
dictedphotometryfromthetemplates,weintegratedthe
objectswithlargeisophotalareasandthatitislargestfor
model stellar SED over the wavelength range of each fil-
objectswitharea<60pixels. Theoriginalisophotalarea
tertakingintoaccountthefilterresponsefunctions. This
size from SExtractor hence defines the dilation factor
provides the predicted colors of point sources. Figure 7
applied. We used the same criteria to correct our high
shows our TFIT measured colors for the point like ob-
resolution SExtractor segmentation maps as outlined
jects compared to the colors from the BaSeL stellar li-
by Galametz et al. (2013) and refer the reader to this
brary. The color trend of our point-like objects agrees
work for further details. Even after this correction, the
with the general distribution of colors predicted by the
total flux measurement is always bound by uncertainties
stellar models. This further confirms our measured pho-
incorporatedintotheaboveassumptionsandthisconsti-
tometry, specifically for the brighter sources, and shows
tutes one of the limitations of photometry estimation.
nosystematicbiasinthephotometry. Thescatteratthe
WeranTFITintwostages. DuringthefirststepTFIT
reddercolorismostlyassociatedwiththefaintersources
measured any remaining mis-alignment in the form of
inthecatalogandalsoduetotheintrinsicscatterofcol-
distortionormis-registrationbetweenthehighresolution
ors inherent to the library because of the degeneracies
and low resolution images in the form of shifted kernels
among the different populations of stars.
(Laidler et al. 2007; Galametz et al. 2013; Guo et al.
2013). In the second step, TFIT used the kernels mea-
4.2. Infrared Color Validation Check
suredfromthefirststeptocorrectthemisalignmentand
construct a difference residual map of the low resolution A validation check of the infrared colors of objects in
bands. Figure 5 shows the TFIT residual maps in the the catalog was done by using the Spitzer/IRAC TFIT
low resolution visible, near-infrared and infrared bands measured photometry of galaxies to identify luminous
10
4 3.0 5
FHT) 3 STA) 22..05 W 4
(C∗ 2 UVI 1.5 160 3
z ( F
− Ks 1.0 −2
T) 1 − W
(CFH 0 814W 00..05 F606 1
g∗ Model F 0.5 Model 0 Model
Catalog Catalog Catalog
1 0 2 4 1.0 0 2 4 11 0 1 2
u∗(CFHT) F814W F606W Y(UVISTA) F814W F125W
− − −
4 1.0 0.0
)
C
)
A 0.5 C
3 R A 0.5
W I R
(
14 m 0.0 (I
8 2 µ m 1.0
F 6 µ
− 3. 0.5 5
W − 4.
06 1 T) 1.0 W−1.5
6 H
F F 5
0 Model (C 1.5 Model F12 2.0 Model
∗
Catalog z Catalog Catalog
1 2.0 2.5
1 0 1 2 3 4 1 0 1 2 3 4 0 1 2
B(Subaru) r+(Subaru) g∗(CFHT) z∗(CFHT) r+(Subaru) Y(UVISTA)
− − −
Figure 7. Color-colordiagramsshowingtheTFITcolorofstars(determinedfromSExtractorasobjectswithCLASS STAR>0.9and
H160<22mag)inCANDELSCOSMOS(blue)andmodelstarsfromtheBaSeLlibrary(black)(Lejeuneetal.1997;Westeraetal.2002).
ThemodelcolorsofstarswerecomputedineachfilterbyintegratingthemodelSEDofstarsfromthelibraryoverthefiltertransmission
curves while the observed colors of the stars are directly from the TFIT catalog with no SED inferred zero-point corrections (Table 5)
applied. Thefiltersusedarefromhttp://cosmos.astro.caltech.edu/page/filterset.
and study AGN host galaxies (e.g. Laurent et al. 2000;
Farrahetal.2007;Petricetal.2011;Yanetal.2013;Lacy
et al. 2015). Several recent studies have used wide and
deep observations in the mid-infrared by Spitzer/IRAC
to successfully identify large samples of AGNs using flux
ratios (Lacy et al. 2004, 2007; Stern et al. 2005, 2012;
Donley et al. 2012; Messias et al. 2012). These color se-
lectionsarebasedonthepower-lawbehaviorofthemid-
infrared continuum of luminous AGNs caused by heated
dust producing a thermal continuum (Neugebauer et al.
1979;Ivezi´cetal.2002;Donleyetal.2012;Messiasetal.
2012).
Figure 8 shows the TFIT measured IRAC color distri-
butions of galaxies in our catalog that have a S/N > 5
in all four channels. According to these selections the
red IRAC colors are expected to be dominated by emis-
sion from AGN-heated dust, as also predicted by pre-
vious studies of SDSS and radio-selected quasars (Lacy
et al. 2004), whereas any stellar components usually
shifts the S /S ratio to bluer colors. The Donley
Figure 8. IRACcolor-colordiagramandtheAGNselectioncri- 5.8 3.6
et al. (2012) criteria are more conservative in selecting
teria from Lacy et al. (2004) (dashed magenta line) and Donley
et al. (2012) (solid magenta line). Black circles are galaxies from AGNs by removing star-forming and quiescent galaxies
the catalog with S/N > 5 in all four IRAC bands. Sources with identified through other selection methods from optical
CLASS STAR>0.95 and H160 <21 are shown by yellow circles. and near-infrared observations (such as the BzK and
X-raypoint-likedetectedsourcesfromXMM-Newtonwide-fieldob-
LBG selections; Madau et al. 1996; Giavalisco 2002).
servationsofCOSMOS(Brusaetal.2010)areshownbylargeblue
circles. Figure 8 also shows the IRAC color distribution of the
X-ray detected sources in COSMOS (Cappelluti et al.
2007) along with the IRAC color distribution of point
Active Galactic Nuclei (AGNs). Mid-infrared observa-
sources. The majority of the X-ray detected AGNs have
tions of galaxies have been used extensively to identify