CALIFA, the Calar Alto Legacy Integral Field Area survey. II. First public data release

2.2 MB·English

by  CALIFA collaboration

Astronomy&Astrophysicsmanuscriptno. CALIFA_DR1 (cid:13)cESO2012
November1,2012
CALIFA, the Calar Alto Legacy Integral Field Area survey:
II. First public data release(cid:63)
B.Husemann1,K.Jahnke2,S.F.Sánchez3,4,D.Barrado-Navascues4,S.Bekeraite˙1,D.J.Bomans5,6,
A.Castillo-Morales7,C.Catalán-Torrecilla7,R.CidFernandes8,J.Falcón-Barroso9,10,R.García-Benito3,
R.M.GonzálezDelgado3,J.Iglesias-Páramo3,4,B.D.Johnson11,D.Kupko1,R.López-Fernandez3,M.Lyubenova2,
R.A.Marino12,D.Mast4,A.Miskolczi5,A.Monreal-Ibero3,A.GildePaz7,E.Pérez3,I.Pérez13,14,
F.F.Rosales-Ortega16,T.Ruiz-Lara13,U.Schilling5,G.vandeVen2,J.Walcher1,J.Alves15,A.L.deAmorim8,
N.Backsmann1,J.K.Barrera-Ballesteros9,J.Bland-Hawthorn17,R.-J.Dettmar5,6,M.Demleitner18,A.I.Díaz16,
H.Enke1,E.Florido13,14,H.Flores19,L.Galbany20,A.Gallazzi21,B.García-Lorenzo9,10,J.M.Gomes22,N.Gruel23,
2
1 T.Haines24,L.Holmes25,B.Jungwiert26,V.Kalinova2,C.Kehrig3,R.C.KennicuttJr27,J.Klar1,M.D.Lehnert19,
0 Á.R.López-Sánchez28,29,A.deLorenzo-Cáceres10,E.Mármol-Queraltó9,10,I.Márquez3,J.Mendez-Abreu9,10,
2 M.Mollá12,A.delOlmo3,S.E.Meidt2,P.Papaderos22,J.Puschnig15,A.Quirrenbach30,M.M.Roth1,

t P.Sánchez-Blázquez16,K.Spekkens25,R.Singh2,V.Stanishev20,S.C.Trager31,J.M.Vilchez3,V.Wild32,
c
O L.Wisotzki1,S.Zibetti33,andB.Ziegler15

0
3 (Affiliationscanbefoundafterthereferences)

November1,2012
]
O
C ABSTRACT
.
h Wepresentthefirstpublicdatarelease(DR1)oftheCalarAltoLegacyIntegralFieldArea(CALIFA)survey. Itconsistsofscience-
p grade optical datacubes for the first 100 of eventually 600 nearby (0.005 < z < 0.03) galaxies, obtained with the integral-field
- spectrographPMAS/PPakmountedonthe3.5mtelescopeattheCalarAltoobservatory. ThegalaxiesinDR1alreadycoverawide
o
rangeofpropertiesincolor-magnitudespace,morphologicaltype,stellarmass,andgasionizationconditions.Thisoffersthepotential
r
t totackleavarietyofopenquestionsingalaxyevolutionusingspatiallyresolvedspectroscopy.Twodifferentspectralsetupsareavail-
s
ableforeachgalaxy,(i)alow-resolutionV500setupcoveringthenominalwavelengthrange3745–7500Åwithaspectralresolution
a
[ of6.0Å(FWHM),and(ii)amedium-resolutionV1200setupcoveringthenominalwavelengthrange3650–4840Åwithaspectral
resolutionof2.3Å(FWHM).WepresentthecharacteristicsanddatastructureoftheCALIFAdatasetsthatshouldbetakenintoac-
1 countforscientificexploitationofthedata,inparticulartheeffectsofvignetting,badpixelsandspatiallycorrelatednoise. Thedata
v qualitytestforall100galaxiesshowedthatwereachamedianlimitingcontinuumsensitivityof1.0×10−18ergs−1cm−2Å−1arcsec−2
0
at5635Åand2.2×10−18ergs−1cm−2Å−1arcsec−2at4500ÅfortheV500andV1200setuprespectively,whichcorrespondstolimiting
5
randgbandsurfacebrightnessesof23.6magarcsec−2and23.4magarcsec−2,oranunresolvedemission-linefluxdetectionlimitof
1
roughly1×10−17ergs−1cm−2arcsec−2and0.6×10−17ergs−1cm−2arcsec−2,respectively. Themedianspatialresolutionis3(cid:48).(cid:48)7,and
8
theabsolutespectrophotometriccalibrationisbetterthan15%(1σ). Wealsodescribetheavailableinterfacesandtoolsthatallow
.
0 easyaccesstothisfirstpublicCALIFAdataathttp://califa.caha.es/DR1.
1
Keywords. techniques:spectroscopic-galaxies:general-surveys
2
1
:
v
1. Introduction bundle (Kelz et al. 2006; Verheijen et al. 2004). A diameter-
i
X selected sample of 939 galaxies were drawn from the 7th data
The Calar Alto Legacy Integral Field Area (CALIFA) survey
releaseoftheSloanDigitalSkySurvey(SDSS,Abazajianetal.
r
a(Sánchezetal.2012a,hereafterS12)isanongoinglargeproject 2009)whichwillbedescribedinWalcheretal. (inpreparation),
of the Centro Astronómico Hispano-Alemán at the Calar Alto
hereafterW12. Fromthismothersamplethe600targetgalaxies
observatory to obtain spatially resolved spectra for 600 lo-
arerandomlyselected.
cal (0.005<z<0.03) galaxies by means of integral field spec-
Combining the techniques of imaging and spectroscopy
troscopy(IFS).CALIFAobservationsstartedinJune2010with
throughopticalIFSprovidesamorecomprehensiveviewofin-
the Potsdam Multi Aperture Spectrograph (PMAS, Roth et al.
2005),mountedtothe3.5mtelescope,utilizingthelarge(74(cid:48)(cid:48)× dividualgalaxypropertiesthananytraditionalsurvey. CALIFA-
64(cid:48)(cid:48)) hexagonal field-of-view (FoV) offered by the PPak fiber like observations were collected during the feasibility studies
(Mármol-Queraltóetal.2011;Viironenetal.2012)andthePPak
(cid:63) Based on observations collected at the Centro Astronómico His- IFSNearbyGalaxySurvey(PINGS,Rosales-Ortegaetal.2010),
pano Alemán (CAHA) at Calar Alto, operated jointly by the Max- apredecessorofthissurvey. Firstresultsbasedonthosedatasets
Planck-InstitutfürAstronomie(MPIA)andtheInstitutodeAstrofísica alreadyexploredtheirinformationcontent(e.g.Alonso-Herrero
deAndalucía(CSIC) et al. 2012; Rosales-Ortega et al. 2011, 2012; Sánchez et al.
Articlenumber,page1of27
E Sb
80
e] 4 S0 Sc
re 60 Sa Sd
g
e
d 3
[ 40

n
o 20
i 2
t
a
n
i 0
l
ec ur 1
D 20 o
l
o
c
z 00 00 00 00 72
s200 u 4 64e [%]
et 13 23 04 01 01 56mpl
targ 150 3 01 11 20 06 02 00 4408nal sa
er of 100 2 00 01 06 02 01 00 32n of fi
b 24o
um 50 00 01 03 03 00 00 16racti
N 1 F
8
00 50 100 150 200 250 300 350 24 23 22 21 20 19 18 17 16 0
Right Ascension [degree] absolute Magnitude (Mz)
Fig.1. DistributionofgalaxiesintheCALIFAmothersampleonthe Fig.2. Upperpanel:DistributionofCALIFAgalaxiesintheu−zvs.
sky (upper panel) and as a function of right ascension (lower panel). M color-magnitudediagram.BlackdotsindicategalaxiesintheCAL-
z
GalaxiesintheCALIFADR1samplearehighlightedbyfilledsymbols IFAmothersample(S12,W12)andcoloredsymbolsdenoteCALIFA
withredcolor. Thenumberdistributioninbinsof30◦ alongtheright DR1galaxies. Differentcolorsrepresentthemorphologicalclassifica-
ascensionisshowninthelowerpanelforthemothersample(grayarea) tionofgalaxieswhichrangefromellipticals(E)tolate-typespirals(Sd).
andtheDR1sample(redshadedarea). Lowerpanel: ThefractionofgalaxiesintheDR1samplewithrespect
to the expected final CALIFA sample distribution in bins of 1mag in
M and0.75maginu−z. Thetotalnumberofgalaxiesperbininthe
z
2011,2012b).CALIFAcanthereforebeexpectedtomakeasub- DR1sampleiswrittenineachbin. Thebinsforwhichthenumberof
stantialcontributiontoourunderstandingofgalaxyevolutionin galaxiesinthemothersampleislessthan5arepronetolow-number
variousaspectsincluding,(i)therelativeimportanceandconse- statisticsandenclosedbyamagentasquareforclarity.
quences of merging and secular processes, (ii) the evolution of
galaxiesacrossthecolor-magnitudediagram, (iii)theeffectsof
theenvironmentongalaxies,(iv)theAGN-hostgalaxyconnec- formationwastakenfromSIMBADforallgalaxieswhereSDSS
tion, (v) the internal dynamical processes in galaxies, and (vi) DR7spectrawereunavailable. ThereaderisreferredtoW12for
the global and spatially resolved star formation history of vari- adetailedcharacterizationoftheCALIFAmothersampleanda
ousgalaxytypes. thoroughevaluationoftheselectioneffectsimpliedbythecho-
In this article, we introduce the first data release (DR1) of sen selection criteria. From the CALIFA mother sample, 600
CALIFAwhichgrantspublicaccesstohigh-qualitydataforaset galaxies are randomly selected for observation purely based on
of100galaxies. ThepropertiesofthegalaxiesintheDR1sam- visibility,andwerefertothesegalaxiesasthevirtualfinalCAL-
plearesummarizedinSect.2. Wedescribethedatacharacteris- IFAsamplehereafter.
tics(Sect.3), datastructure(Sect.4), anddataquality(Sect.5)
The first 100 public DR1 galaxies were observed in both
ofthedistributedCALIFAdataasessentialinformationforany
spectral setups from the start of observations in June 2010 un-
scientificanalysis. Severalinterfacesareavailabletoaccessthe
til June 2012. We list these galaxies in Table 1 together with
CALIFADR1data,whichareintroducedinSect.6.
theirprimarycharacteristics. Thedistributionofgalaxiesinthe
sky follows the underlying SDSS footprint (Fig. 1). An excep-
tionistheadditionalcutinδ>7◦ thatwasappliedtotheregion
2. TheCALIFADR1sample
oftheNorthernGalacticCap,giventhesufficientlylargenumber
A sample of 939 potential CALIFA galaxies, also known as ofobjectswithbettervisibilityfromCalarAlto. Thenumberof
the “CALIFA mother sample”, was drawn from the photomet- galaxiesinDR1isnothomogeneousasafunctionofrightascen-
riccatalogSDSSDR7asoutlinedinS12andW12. Theprimary sion,α(J2000),andhasaclearpeakaroundα ∼ 255◦. Thiscan
CALIFA selection criterion was the angular isophotal diameter beexplainedbytheunexpecteddowntimeofthe3.5mtelescope
(45(cid:48)(cid:48) < D <80(cid:48)(cid:48))ofthegalaxies,whichwascomplementedby fromAugust2010untilApril2011duetooperationalreasonsat
25
limitingredshifts,0.005 < z < 0.03. Theseredshiftlimitswere theobservatory,whichdelayedthesurveyroughlyby8months.
imposedtoensurethatallinterestingspectralfeaturescanbeob- CALIFAobservations,therefore,sofarspanthreesummersea-
servedwithafixedspectralsetupoftheinstrumentandthatthe sons at Calar Alto, but only a single winter season. Neverthe-
samplewouldnotbedominatedbydwarfgalaxies. Redshiftin- less,thedistributionofphysicalpropertiesisnearlyrandom,as
Articlenumber,page2of27
Husemannetal.:TheCALIFAsurveyII.Firstpublicdatarelease
1
]
1 %]30
mag 2 e [
3 pl25
c 3 m
Mp sa20
ty [ 4 nal
si fi 15
n f
e o
d 5
r on 10
e i
b ct
m 6 a
u r 5
n F
14 08 28 17 27 06
7
17 18 19 20 21 22 23 0
E S0 SA SAB SB M
absolute r band magnitude
Galaxy type
Fig.3. Luminosityfunctionsinther bandoftheDR1sample(red Fig.4. ThefractionofgalaxiesintheDR1samplewithrespecttothe
points),theCALIFAmothersample(bluepoints),andthelow-redshift expectedfinalCALIFAsampledistribution,splitbyvisuallyclassified
sampleoftheNYUvalueaddedcatalog(blackline,Blantonetal.2005). morphology.Wedividethegalaxiesintoellipticals(E),lenticulars(S0),
Despite the small size of the DR1 sample the luminosity function is non-barredspirals(SA),weaklybarredspirals(SAB),stronglybarred
reproducedwell. spirals(SB),andongoingmergers(M)ofanytype.Themorphological
distributionoftheDR1sampleliesclosetothatofthemothersample.
ThetotalnumberofgalaxiesintheDR1foreachmorphologytypeis
writtenonthebar. ErrorbarsarecomputedfromthePoissonerrorsof
expected,andcoversgalaxieswithawiderangeofpropertiesas
theassociatedDR1numbercounts.
discussedbelow.
Thedistributionofgalaxiesinthecolor-magnitudediagram
(Fig. 2) shows that the DR1 sample covers almost homoge- late-type galaxies
neouslythefullrangeoftheCALIFAmothersample. Onaver- %]35 early-type galaxies
age, the DR1 targets comprises ∼20% per color-magnitude bin [

e30
ofthetotalexpectednumberwhenCALIFAiscompleted. How- l
p
ever, there is currently still a deficit of low luminosity galaxies m
a25
with intermediate colors. In other color-magnitude bins, espe- s

cieias,llyfluwcittuhaitniothnossceawnhbeereexthpelaminoetdhebrysatmhepleeffceocnttoaifnlsofwewnugmalbaxer- nal20
i
f
statistics. Figure 2 highlights the need to further increase the of 15
numbers to the full CALIFA sample to obtain enough galaxies
n
in each bin for a meaningful multi-dimensional statistical anal- o
i 10
t
ysis. In Fig. 3, we show the r band luminosity function (LF) c
a
of the DR1 sample as compared to the mother sample and the Fr 5
reference SDSS sample of Blanton et al. (2005). All technical
details on how we obtained the LFs are described in W12. We 0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
simplynoteherethatdespitethesmallsamplesizeofDR1,we
b/a axis ratios of galaxies
already reproduce the LF reasonably well. The turnover of the
LFatMr >−19.5isentirelyexpectedandunderstood. Fig. 5. The fraction of galaxies in the DR1 sample with respect to
Galaxymorphologieswereinferredbycombiningtheinde- the expected final CALIFA sample distribution, as a function of axis
pendent visual classifications of several collaboration members ratio (b/a). Galaxies were separated into early-type galaxies (E+S0)
as described in W12. In Fig. 4 we show the fraction of DR1 andlate-typegalaxies(Saandlater).TheCALIFAmothersampledoes
notincludeanyellipticalgalaxieswithb/a<0.3oranyspiralgalaxies
galaxieswithrespecttotheexpectedfinalsampledistributionfor
differentmorphologicaltypesgroupedintoelliptical,lenticular, withb/a>0.9.ErrorbarsarecomputedfromthePoissonerrorsofthe
associatedDR1numbercounts.
spiralgalaxies(separatedbydifferentbarstrength)aswellason-
going galaxy mergers. A more detailed classification of spirals
intoearly-andlate-typesisavailable,butwedonotdistinguish aKolmogorov-Smirnovtest,wefurtherquantifiedthatthemor-
betweenthemherebecauseofthestillmodestnumberofgalax- phologyandtheaxis-ratiodistributionofDR1isconsistentwith
ieswithinDR1. TheDR1coverageseemstobeconsistentwith being randomly drawn from the CALIFA mother sample with
a random selection because the fraction of DR1 galaxies with >95%confidence.
respect to the expected final sample is almost constant for all In Fig. 6, we present the distribution of stellar masses for
types. Axis ratios (b/a) were measured from the SDSS r band the DR1 galaxies. They were directly inferred from the CAL-
imagebycalculatinglightmomentsafterproperskysubtraction IFA data by applying spatially resolved spectral synthesis us-
andmaskingofforegroundstars(seeW12fordetails). Thefrac- ingtheStarlightcode(CidFernandesetal.2005)withacom-
tions of axis ratios, which can be used as an indication of the bination of single stellar populations (SSPs) from the libraries
inclinationofspiralgalaxies,coveredbytheDR1sampleisho- of González Delgado et al. (2005) and Vazdekis et al. (2010),
mogeneous with respect to the final sample (Fig. 5). Based on which both adopt a Salpeter initial mass function. Details of
Articlenumber,page3of27
1.5
Kewley et al. (2001)
Kauffmann et al. (2003) Seyfert (7)
20 Cid Fernandes et al. (2010)
s ) 1.0
e
i H
x
a 7/
al 0 0.5
g 15 50

1
R ]
D II 0.0
I
r of 10 ([O LINER (34)
be og 0.5
m l Inter. (14)
SF (28)
u
N 5
1.0
1.0 0.5 0.0 0.5
log([N II] 6583/H )
0
9.0 9.5 10.0 10.5 11.0 11.5 12.0
log(M /M ) Fig.7. Emission-linediagnosticdiagramforthemostcentralspaxelof
*
eachDR1datacube. Onlyobjectsforwhichallrequiredemissionlines
Fig.6. DistributionofstellarmassesforgalaxiesintheDR1sample. haveS/N>3areshown.ThedemarcationlinesofKewleyetal.(2001),
The stellar masses have been determined from the CALIFA data (see Kauffmann et al. (2003), and Cid Fernandes et al. (2010) are used to
textfordetails). classifythegalaxiesintostarforming(SF),intermediate,Seyfert,and
LINER-typegalaxies,whicharedenotedwithblack,magenta,red,and
greensymbols,respectively.
thesemeasurementswillbepresentedinGonzálezDelgadoetal. 3. Dataprocessinganderrorpropagation
(inpreparation). TheDR1galaxiescoverintermediatetohigh-
The instrument characteristics and observing strategy of the
mass galaxies, including more than 5 galaxies per 0.25dex bin
CALIFA survey define the requirements for the data reduction
between1010 and1012M andamedianvaluecloseto1011M .
(cid:12) (cid:12) scheme. These requirements are thoroughly described in S12
Theasymmetricdistributionisexpectedfromthedistributionin
and only briefly summarized here for completeness. The PPak
absolutemagnitudes(seeFig.2)andisinheritedfromtheCAL-
fiberbundleofthePMASinstrumentcomprises382fibers,each
IFAmothersampleduetoitsselectioncriteria(seeW12forde-
withadiameterof2.7(cid:48)(cid:48) projectedonthesky. TheprimaryFoV
tails).
hasahexagonalshapeof74(cid:48)(cid:48)×62(cid:48)(cid:48)insize,sampledby331fibers
withafillingfactorof∼60%. Asetof36fibersarededicatedto
Many different kinds of galaxies are already covered in the
sampletheskybackgroundandaredistributedin6smallerbun-
DR1sampleinsufficientnumberstoperformspatiallyresolved
dlesalongacircleof∼ 72(cid:48)(cid:48) radiusaroundtheFoVcenter. The
comparison studies. In addition to global galaxy parameters
remaining15fibersarecoupledtothecalibrationunitofthein-
presented previously, we show the [Oiii]/Hβ vs. [Nii]/Hα
strument.
emission-linediagnosticdiagram(Baldwinetal.1981;Veilleux
Each CALIFA galaxy is targeted twice with different spec-
& Osterbrock 1987) for the nucleus in Fig. 7. It is constructed
tralsettings,alow-resolution(V500)setupcoveringthenominal
fromthemostcentralspaxelatthecoordinatesofeachCALIFA
wavelengthrange3745–7500ÅatspectralresolutionofR∼850,
galaxy center (Table 1 excluding objects with a wrong astrom-
andamid-resolution(V1200)grismcoveringthenominalwave-
etryinthedataasdiscussedlaterinSect.5.2). Afterremoving
the stellar continuum using a spectral synthesis approach, we lengthrange3400–4840ÅataspectralresolutionofR ∼ 1650.
measured the emission line fluxes by fitting Gaussians to their The useful wavelength range is, however, reduced by internal
profiles in the residual spectra. A variety of different ioniza- vignetting within the spectrograph for several fibers to 4240–
tionmechanismscanbefoundintheDR1galaxysample,rang- 7140Å and 3650–4620Å in the worst case for the V500 and
ing from pure star formation to Seyfert-type active galactic nu- V1200,respectively. Threeditheredpointingsaretakenforeach
clei (AGN), with a significant number of galaxies in between. object to reach a filling factor of 100% across the entire FoV.
Also,Low-IonizationNuclearEmission-lineRegions(LINERs) The exposure time per pointing is fixed to 900s for the V500
(Heckman 1980), which are predominantly hosted by bulge- and 1800s for the V1200. The latter is further split into 2 or 3
dominated or elliptical galaxies, appear to be frequent. Robust individualexposures.
classificationofindividualgalaxiesaccordingtosuchascheme We are continuously upgrading the CALIFA pipeline (see
can be complicated, i.e., in border-line cases or due to system- S12 for details). The main improvements to the data reduction
aticuncertaintiesintheemission-linemeasurements,aswellas pipelineusedtoproducetheDR1dataarebrieflymentionedin
adoptingothertypesofdiagnosticdiagrams(e.g.CidFernandes the next section followed by a detailed characterization of the
et al. 2010), or a different classification scheme. This has been propagatednoise.
considered in the Kehrig et al. (2012) study of the ionization
sourceoftheinterstellarmediumintwoLINER-likeearly-type
3.1. ImprovementsontheCALIFAdatareductionscheme
CALIFA galaxies. Rather than to suggest or impose a certain
classificationofindividualgalaxynucleifortheDR1sample,we Atthe firststages ofthe project, we employedthe R3D data re-
present one particular representation here only to demonstrate duction package (Sánchez 2006) as the basis to develop a fully
thediversityoftheDR1galaxysample. automaticIFSdatareductionpipelinededicatedtotheCALIFA
Articlenumber,page4of27
Husemannetal.:TheCALIFAsurveyII.Firstpublicdatarelease
2.0
1.8 Observed (Ok) Error ( k)
3500000
1.6 Model (Mk) Residuals (Ok Mk)
1] 1.4 s3000000
A el
1.2 x
2 i2500000
m p
c 1.0 f
2s 0.8 er o2000000
erg 0.6 mb 1500000
16 0.06 u
0.04 N
0 1000000
1 0.02
[
0.00
f 500000
0.02
0.04
0.06 0
4500 5000 5500 6000 6500 4 2 0 2 4
wavelength[A] (O k M k)/ k
Fig.8. Exampleofastellarpopulationfitwithstarlighttothecentral Fig.9. Histogramoftheobserved(O )minusmodel(M )residu-
λk λk
V500 spectrum of NGC 5966, restricted to the unvignetted part. The alsateachwavelengthnormalizedbytheformalfluxerror((cid:15) ). The
λk
observed,modeledandresidualspectrumareshowntogetherwiththe modelspectraareobtainedfromstarlightfitsforthestellarcontinuum,
pipeline produced ±1σ error. Sky and galaxy emission line regions excludingthemainemissionlines,aswellasbadpixels.Thehistogram
maskedduringthefittingareindicatedbythegrayshadedareas. describesalmostaGaussiancenteredat0withadispersionof∼ 0.8.
ThebluelineindicatesaGaussiandistributioncenteredat0withadis-
persionof1,normalizedtothetotalnumberofpixelsforcomparison.
Ontheassumptionthatthespectralresidualsareduetonoise(i.e.,ne-
data. A description of that pipeline and the individual reduc-
glectingmodelimperfections),thisfiguredemonstratesthatthepipeline
tion steps was presented by S12. With the aim of improving
errorestimatesarereliable.
the flexibility, portability, maintenance and capabilities of the
pipeline,mostofthedatareductiontaskshavebeentransformed
toaPython-basedarchitecture. Neglecting systematic deviations of the model spectra with
respect to the real data in some wavelength regions, one would
Themaindriverforthepipelineimprovementswastoimple-
expect the residuals R to be typically of the order of (cid:15) . In-
mentthepropagationofthePoissonplusread-outnoiseaswell λk λk
deed, we find that this is the case. The histogram of U =
as bad pixels caused by cosmic ray hits, bad CCD columns, or λk
the effect of vignetting, from the raw data to the final CALIFA Rλk/(cid:15)λk(Fig.9)wasobtainedfromover108datapointsofnearly
105 spectra for the V500 spectra distributed within this DR1.
data product in a consistent way. This was not possible with
Emission lines and faulty pixels were excluded from the statis-
the reduction scheme adopted by the previous pipeline version
tics(shadedareasinFig.8),asthespectralfitsareonlymeantto
(V1.2),becausethenumerousinterpolationandresamplingsteps
reproducethestellarcomponent. Thehistogramfollowsnearly
did not allow a reliable error propagation. Details of the most
a Gaussian shape, centered at −0.04 with a standard deviation
importantimprovementsforthecurrentpipelineversion(V1.3c)
of 0.8. The expected standard deviation would be 1.0 if the er-
areprovidedinAppendixA.
rorspectrawouldstatisticallyagreewiththenoiseinthe resid-
The final output of the V1.3c pipeline are cubes of inten-
ual spectra and assuming that the stellar population model is a
sity, errors and bad pixels. Details on the data format are de- perfect fit. Selecting different wavelength ranges for this com-
scribedinSect.4. Belowweevaluatespecificallytheaccuracy
parison to avoid regions with systematic template mismatches
ofthepipeline-producednoiseestimatesandcharacterizeeffects
mainlychangesthecentroidofthedistributionwhereasthestan-
ofcorrelatednoisethatshouldbeproperlytakenintoaccountin
darddeviationisnearlyconstant.Thus,thepropagatederrorsare
anyscientificanalysis.
overestimatedby∼20%inthemean. Thiswasalsoverifiedfor
theV1200setupusingthesamemethod. Theoverallconclusion
ofthistestisthattheerrorsprovidedinDR1arerobust, witha
3.2. Accuracyofthepropagatednoise
slightsystematicoverestimationof∼20%thatcanbetakeninto
accountduringanyanalysisifneeded.
ThederivationoferrorspectrainIFSdataisacomplextask,for
which there are no well established recipes. It is therefore rel-
evanttoverifythereliabilityofthepipelineerrors((cid:15) ). Anap-
λ 3.3. Characterizationofspatiallycorrelatednoise
proximate assessment of the error spectra was carried out with
the aid of full spectral continuum fitting described by Cid Fer- Itisoftennecessarytospatiallyco-addspaxelsinthedatacubes
nandes et al. (in preparation), and Gonzalez Delgado et al. (in toachieveaminimumS/Ninthespectraforaspecificapplica-
preparation). The CALIFA spectra are decomposed in terms tion. ForCALIFAweadoptaninverse-distanceweightedimage
of combinations of SSP spectra from González Delgado et al. reconstructionschemesofar,whichaveragesthefluxamongall
(2005) and Vazdekis et al. (2010) including a dust extinction fibers within 5(cid:48)(cid:48) for a given spaxel in the final datacube by as-
termfollowingtheCardellietal.(1989)law.Anexampleofsuch suming a 2D Gaussian profile with a dispersion of 1(cid:48)(cid:48) for the
afitisshownforacentralV500spectruminFig.8forillustrative individual weighting factors (see S12 for details). Like many
purposes. For a certain spectrum k it presents the observation otherimageresamplingschemesitintroducessignificantcorre-
(O ),thebest-fitmodel(M ),theresiduals(R =O −M ), lationbetweenthespaxelsinthefinaldatacubes.Thiscanbeun-
λk λk λk λk λk
andtheprovidedpipelineerrorspectrum((cid:15) ). derstoodfromsimplearguments.EachCALIFAdatasetcontains
λk
Articlenumber,page5of27
5.0 Here, we characterize the effect of the correlated noise by
determiningtheratioofthe“real”error((cid:15) ),directlyestimated
V500 real
4.5 fromtheresidualsR ,totheanalyticallypropagatederror((cid:15) )
Target S/N = 20 λk bin
of binned spectra as a function of bin for a certain target S/N.
4.0
The results obtained for all DR1 datasets are shown in Fig. 10
3.5 for the two instrumental setups with a target S/N level of 20.
/albin3.0 Tlohgearoitbhsmerivcefduntrcetinodns1can be sufficiently described by a simple
re2.5 (cid:15) =(cid:15) (cid:2)1+αlogn(cid:3) , (1)
real bin
2.0 withnthenumberofspaxelsperbin.
The values for the slope α range from 1.35 to 1.45 with a
1.5 meanofα = 1.4,fortargetS/Nvaluesbetween10and60. The
best fit: =1.38 fits to the data result in less accurate exponents for target S/N
1.0
ratiosof 10and 60giventhe poorsampling towardslarge- and
5.0
small-size bins, but the overall shape is preserved. This means
V1200 thatαmainlydependsonthenumberofspaxelsperbinandnot
4.5
Target S/N = 20 onthetargetS/N.
4.0
3.5 4. CALIFAdataformatandcharacteristics
n
bi3.0 TheCALIFAdataarestoredanddeliveredasdatacubes(three-
/al dimensionaldata)inthestandardbinaryFITSformatandconsist
e
r2.5 offourFITSHeader/Dataunits(HDU).Thesedatacubesrepre-
sent (1) the measured flux densities, corrected for Galactic ex-
2.0
tinction as described in S12, in units of 10−16ergs−1cm−2Å−1
1.5 (primary datacube), (2) associated errors, (3) error weighting
factors, and (4) bad pixels flags (Table 2). They allow users to
best fit: =1.35
1.0 properlytakeintoaccountthecharacteristicsofeachdatasetcon-
0 20 40 60 80 100 cerningbadpixelsandnoisefortheirspecificanalysis. Thefirst
n
twoaxesofthecubescorrespondtothespatialdimensionalong
right ascension and declination with a 1(cid:48)(cid:48) ×1(cid:48)(cid:48) sampling. The
Fig.10. Ratiooftherealerror((cid:15)real)totheanalyticallypropagated third dimension represents the wavelength along a linear grid.
e(urrpopre(r(cid:15)bpina)neals)aanfudncVti1o2n00of(nlouwmebrepraonfeslp)adxaetlasopferDbRin1foatraalltatrhgeetVS50/N0 Thedimensionsofeachdatacube(Nα,Nδ,andNλ),aswellasthe
spectralsampling(d )andconstantresolution(δ )alongtheen-
of 20. Grey shades mark the 1σ, 2σ and 3σ levels. The blue lines λ λ
tirewavelengthrange,aresummarizedinTable3. Notethatthe
representthebestfitlogarithmicfunctionwithα=1.38andα=1.35,
spatialresolutionisworsethantheactualseeingduringtheob-
respectively.
servationbecausethelargeapertureofeachfiber(2(cid:48).(cid:48)7)strongly
undersamplesthepointspreadfunction. ThefinalCALIFAspa-
tial resolution is mainly set by the dither pattern and the image
993 physically independent spectra from the fibers of the three reconstruction scheme rather than the seeing. We evaluate this
dithered pointings, but the final datacube consists of more than aspartofourqualitycontroltestsdiscussedbelowinSect.5.2.
4000spectraatasamplingof1(cid:48)(cid:48)×1(cid:48)(cid:48)perspaxel. Thisautomat-
ically implies that the final spaxels cannot be completely inde-
4.1. Errordatacubes
pendentfromeachotherwithintheCALIFAdatacubesthrough
thecomplexcorrelationofsignalandnoisebetweenneighboring TheerrordatacubesinthefirstFITSextensioncorrespondtothe
spaxels. associated1σnoiselevelofeachpixelasformallypropagatedby
thepipeline.Asvalidationofoureffortstooptimizethepipeline,
Inthelimitcaseofco-addingthespectraoftheentirecube,
we verified that the measured noise is systematically lower by
thepipelineanalyticallycalculatesanerrorweightingfactorfor
∼20% than the formal error vector for individual spaxels (see
eachpixelsuchthattheformalerroroftheco-addedspectrumis
Sect.3.2). Inthecaseofbadpixels,weassignedanerrorvalue
identical to the one obtained by co-adding the individual fiber
that is roughly ten orders of magnitude higher than the typical
spectra. Of course, this is an unrealistic case, because only
value for the considered dataset. Any analysis based on the χ2
spectra within a small zone are typically co-added to preserve
statisticswillimplicitlytakeintoaccountbadpixelsiftheerror
some spatial resolution. A popular method for adaptive bin-
vectorisconsidered.Thecorrelationofnoisebecomesimportant
ningistheVoronoi-binningschemeimplementedforopticalIFS
onlyincaseswheretheCALIFAdataneedtobespatiallybinned
data by Cappellari & Copin (2003). It assumes that the spec-
asdiscussedinSect.3.3. ThesecondFITSextensionreportsthe
traarecompletelyindependentofeachothertocomputethere-
quired bin size for a given target S/N. Blindly adopting such a error scaling factor for each pixel in the limiting case that all
validspaxelsofthecubewouldbeco-added. Suggestionsabout
binningschemetoCALIFAdatacubeswillleadtoincorrectre-
how to deal with the correlated noise and proper usage of the
sults, because the assumption that the spectra are independent
errorscalingfactorsweredescribedaboveindetail.
is not valid. Either the bin sizes will be smaller than required
toachievethetargetS/N,oralternativelyexpressed,theerrorof 1 Anindependentestimateofthisempiricalrelation,proposedbyCid
theco-addedspectrawillbehigherthanformallyexpectedgiven Fernandes et al. (in preparation), is completely consistent within the
theerrorofindividualspaxels. estimateduncertainties.
Articlenumber,page6of27
Husemannetal.:TheCALIFAsurveyII.Firstpublicdatarelease
70 100 100
]
96 %
60 [
92 s 90
el
50 88 x
pi
el 40 84 d 80
pix 80 vali
y 30 76 f 70
o
20 72 n
o
68 ti 60
10 ac V500
V500 V1200 64 fr V1200
0 60 50
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 4000 4500 5000 5500 6000 6500 7000
x pixel x pixel wavelength [A]
Fig.11. FractionofvalidpixelsineachspectrumacrosstheCALIFAFoV(leftpanels)andasafunctionofwavelength(rightpanel)forthe
V500andV1200setup. Thespaxelsmostseverelyaffectedbythevignettingintheblueandredpartofthespectraleadtotheringlikestructure
aroundtheFoVcenter.ThebluepartoftheV1200datadoesnotshowavignettingeffectherebecausewecutoffthewavelengthrangeinthefinal
dataalreadyat3650Åduetotheverylowsensitivityintheblue.FourbadCCDcolumnsarevisiblethatleadtoasignificantlyreducedfractionof
validpixelsatnarrowwavelengthregions.
Table2.CALIFAFITSfilestructure
HDU Extensionname Format Content
0 Primary 32-bitfloat fluxdensityinunitsof10−16ergs−1cm−2Å−1
1 ERROR 32-bitfloat 1σerroronthefluxdensity
2 ERRWEIGHT 32-bitfloat errorweightingfactor
3 BADPIX 8-bitinteger badpixelflags(1=bad,0=good)
Table3.DimensionandsamplingofCALIFAdatacubes ThehexagonalPPakFoVisresampledtoarectangulargrid,
sothattheuncoveredcornersarefilledwithzeros.Thesearealso
Setup Nαa Nδa Nλa λstartb λendc dλd δλe flagged as bad pixels for consistency, whereas the residuals of
V500 78 73 1877 3749Å 7501Å 2.0Å 6.0Å brightnight-skyemissionlinesarenotflaggedasbadpixels.The
V1200 78 73 1701 3650Å 4840Å 0.7Å 2.3Å strengthoftheirresidualsisdifferentforeachdatasetandmight
behandleddifferentlydependingonthespecificdataanalysis.
Notes.(a)Numberofpixelsineachdimension.(b)Wavelengthofthefirst
pixelonthewavelengthdirection.(c)Wavelengthofthelastpixelonthe
wavelength direction. (d) Wavelength sampling per pixel. (e) Homoge- 4.3. FITSheaderinformation
nizedspectralresolution(FWHM)overtheentirewavelengthrange.
The FITS header contains the standard keywords that register
thespatialaxestothestandardWorldCoordinateSystem(WCS,
Greisen & Calabretta 2002) and the wavelength to the spectral
4.2. Badpixeldatacubes axis in a linear grid. Each CALIFA datacube contains the full
FITS header information of all raw frames from which it was
Bad pixeldatacubes are producedby the pipelineand stored in created,whereeachheaderentryisexpandedwithauniquepre-
thethirdFITSextension. Theyreportpixelsforwhichnosuffi- fix for a given pointing/frame. The prefixconsists of the string
cientinformationisavailableintherawdatabecauseofcosmic “PPAK” followed by the designation of the pointing PREF,
rays,badCCDcolumns,ortheeffectofvignetting. Thedataat whichcanbe“P1”,“P2”,or“P3”forthethreepointingsofthe
badpixelshavebeeninterpolated,anditisconsiderednotusable V500setup,or“P1F1”,“P1F2”,“P2F1”,etc...,fortheindividual
evenifthespectrumlooks“good”atthoselocations.TheV1200 framestakenforeachpointingoftheV1200setup.
dataislessaffectedbycosmicrayscomparedtotheV500data, The reduction pipeline also collects information regarding,
becauseseveralframesareobservedperpointing. Thedataata e.g.,skybrightness,flexureoffsets,Galacticextinction,approx-
badpixelarethereforerestoredfromtheotheravailableframe(s) imate limiting magnitude, etc., and adds it to the FITS header.
althoughthisresultsinalowerS/N. Header keywords that may be of general interest for the anal-
ysis and/or evaluation of the data are summarized in Table 4
The distribution of bad pixels is not homogeneous within
the datacube because of the vignetting effect as noted in S12. for convenience. Note that the systemic velocity of the galax-
ies (MED_VEL) automatically estimated by the pipeline is not
In Fig. 11, we present the typical fraction of valid pixels along
robustlymeasuredfromtheV1200data,becauseofthesmaller
the spatial and spectral axis for the V500 and V1200 setup, re-
wavelengthcoverageandlowerS/NcomparedtotheV500data.
spectively. Note that four bad CCD columns have been identi-
fied, which lead to four wavelength regions where the fraction
ofvalidpixelsissignificantlyreduced. Thevignettingeffectof
5. DataQuality
theinstrumentresultsinawavelengthcoveragereducedbyupto
∼25%forsomefibersontheblueand/orredsideofthespectral This first CALIFA data release provides science grade data of
range. 100 galaxies to the community. Hence, in order to define the
Articlenumber,page7of27
Table4.MainFITSheaderkeywordsandtheirmeaning
Keyword datatype Meaning
OBJECT string Nameofthetargetgalaxy
CALIFAID integer IDofthegalaxy
NAXIS1 integer Numberofpixelsalongrightascensionaxis(N )
α
NAXIS2 integer Numberofpixelsalongdeclinationaxis(N )
δ
NAXIS3 integer Numberofpixelsalongwavelengthaxis(N )
λ
CRPIX1 float Referencepixelofthegalaxycenterinrightascension
CRVAL1 float Rightascensionα(J2000)ofthegalaxycenterindegrees
CDELT1 float Samplingalongrightascensionaxisinarcsec
CRPIX2 float Referencepixelofthegalaxycenterindeclination
CRVAL2 float Declinationδ(J2000)ofthegalaxycenterindegrees
CDELT2 float Samplingalongdeclinationaxisinarcsec
CRVAL3 float WavelengthofthefirstpixelalongthewavelengthaxisinÅ
CDELT3 float SamplingofthewavelengthaxisinÅ
hierarchPIPEVERS string Versionofthereductionpipeline
hierarchPIPEUNITS string Unitsofthefluxdensity
hierarchPIPEGALEXTAV float GalacticV bandextinctionalongline-of-sight
hierarchPIPEREDUDATE string Date/timethedatawerereduced
hierarchPIPEVMAG3SIG float Estimated3σsurfacebrightnessdepthinmagarcsec−2
MED_VEL float Estimatedsystemicvelocityinkm/s
hierarchPPAKPREFDATE-OBS string Date/timeoftheobservation
hierarchPPAKPREFUT_START float StartoftheobservationinsecondsUTtime
hierarchPPAKPREFUT_END float EndoftheobservationinsecondsUTtime
hierarchPPAKPREFMJD-OBS float ModifiedJuliandateoftheobservation
hierarchPPAKPREFAIRMASS float Airmassduringtheobservation
hierarchPPAKPREFEXPTIM float Exposuretimeoftheobservation
hierarchPPAKPREFHVCOR float Line-of-sightdifferencetohelocentricvelocity
hierarchPPAKPREFEXT_V float V bandatmosphericextinctioninmagnitudes
hierarchPPAKPREFPIPEFLEXXOFF float Measuredflexureoffsetinx-directioninCCDpixels
hierarchPPAKPREFPIPEFLEXYOFF float Measuredflexureoffsetiny-directioninCCDpixels
hierarchPPAKPREFPIPESPECRES float Homogenizedspectralresolution(FWHM)inÅ
hierarchPPAKPREFPIPESKYMEAN float Meansurfacebrightness(µ )duringobservation
sky
set of galaxies to be released, a detailed characterization of the The QC parameters and QC flags can be found in Tables 5
dataqualityfortheexistingdataisneededtoselectanddefinea and6fortheV500andV1200data,respectively,whicharealso
minimalusefulsuiteofqualitycontrol(QC)parameters. availableinelectronicformontheDR1webpageandaccessible
Overall, the final data quality for CALIFA depends on a throughtheVirtualObservatory(seeSect.6.3fordetails).
numberofindependentfactorssuchas(i)generalinstrumentre-
liability and temperature stability, (ii) ambient conditions dur-
5.1. Initialreductionpipelineflags
ing observations, and (iii) the robustness of the data reduction
pipeline. WedefineasetofparametersandQCflagsintheareas Asaninitialdescriptiontoevaluatethedataandreductionqual-
of astrometric, spectroscopic and photometric characterization ity,weprovideeachofthefourflagspersetupthatcomedirectly
that are measured for all galaxies. We amend these parame- fromthereductionprocess. Someoftheextendedparametersin
ters with four quality flags for each of the two spectral setups the sections below will add to further validate and expand on
(V500/V1200; eightintotal)thatsummarizethereductionpro- theseinitialflags.
cessforeachgalaxyandindicatedataquality. Potential minor issues are noted with a value=1, while the
WestronglyurgeuserstoconsidertheseQCparameterswith standardcasereceivesvalue=0. Galaxieswithmajorissues,i.e.
regardtotheirscienceapplication.Forexample,ifasciencegoal flagvalues=2havebeenobservedaspartoftheCALIFAsurvey,
focusesprimarilyonkinematics,thenasubstantialoffsetinab- but were excluded from this data release. The reduction flags
soluteastrometricregistrationisnotalimitationfortheanalysis are:
of any one galaxy. However, if, under the same circumstance,
thesciencegoalrequiresderivingjointspatially-extendedinfor- – FLAG_RED_O:observingconditionsquality.
mationfrommatchingwiththeSDSS,thencautionisadvised. Set to 1 if the night sky is brighter than µ = 20.5
V
ThevaluesanduncertaintieslistedfortheindividualQCpa- magarcsec−2 (V500) or µ = 20.0 magarcsec−2 (V1200),
B
rametersareintendedtoprovidetheuserwithasummaryofthe respectively.Thereexistsacorrelationbetweenlimitingcon-
information content of the CALIFA data on a specific galaxy tinuumsensitivityofthedatacubesandthenight-skybright-
in order to allow for the initial assessment of whether or not a ness. For nights brighter than this limit the datacubes may
specificdatacubecanbeusedforaplannedanalysis. Inthefol- notbeasdeepastargeted.
lowing we display and comment on the distribution of the key Flagissetto2ifobservingconditionswerenon-photometric
flags and QC parameters while demonstrating the overall sam- orifbackgroundofunknownoriginisseeninthedata.
plepropertiesandthenatureofoutliers. – FLAG_RED_R:reduction/calibrationquality.
Articlenumber,page8of27
Husemannetal.:TheCALIFAsurveyII.Firstpublicdatarelease
Set to 1 if reduction deviates from the standard procedure,
e.g. due to a lack of directly associated continuum or arc- V500 observations
lamp exposures or a saturation in the available calibration V1200 observations
dFalatag.issetto2ifmanualinterventionwasrequiredtorecover es15
i
x
aminimumofpossibledatafromtherawdata. a
– FLAG_RED_W:wavelengthcalibrationquality. al
g
Set to 1 if the standard deviation of the pipeline-estimated
1
systemic velocity in different wavelength regions is larger R10
than25kms−1. D

f
Flagsetto2ifthedeviationismorethan2standarddevia- o

tions,or34kms−1. er
– FLAG_RED_V:visualinspectionquality. b
m
Summary of a visual inspection of the data. If there is an 5
u
obviousdefectaffectingasmallfractionoftheFoV(likethe N
presence of a strong line or a imperfectly traced fiber), flag
issetto1. Itissetto2whenitisevidentthatthedatahave
lowqualityforanyunknownreason.
0
Finally this flag is set to 1, even if all previous cases have 0.6 0.8 1.0 1.2 1.4 1.6 1.8
value=0,whenamastercontinuumormasterarcframe,in- seeing [arcsec]
steadofanindividualframe,hasbeenusedduringthereduc-
tionofanyoftheframesusedtocreatethefinaldatacube.
Fig.12. DistributionoftheseeingduringtheCALIFAobservationsas
measuredbytheautomaticDifferentialImageMotionMonitor(DIMM,
5.2. Astrometricaccuracyandspatialresolution Aceituno2004).
5.2.1. Astrometricregistrationaccuracy
– NGC4470: Galaxywithdifficulttodefinecenter. Offsetof
The first area of testing and evaluation of the pipeline output
∼9(cid:48)(cid:48)betweenthetwosetups.
is the absolute astrometric registration of the datacube coordi-
– NGC 4676A: As stated for V500 the center is not well de-
nate systems to the International Coordinate Reference System
finedforthisgalaxyandanoffsetof∼6(cid:48)(cid:48) existsbetweenthe
(ICRS). Astrometric registration is of central importance when
twosetups.
CALIFA data are to be combined with e.g. imaging data from
othersurveystoextractspatiallyresolvedinformationatcertain
For this DR1 we provide the pipeline registration as de-
locationsinagalaxy.ForthispurposetheCALIFApipelineuses
scribedaboveanddonotcorrectoffsetsfoundthroughthesesep-
asimpleschemewherebythetabulatedcoordinatesofthegalaxy
arateexternalchecks. Infuturedatareleasesitisplannedtoim-
V bandphotometriccenter, asgiveninTable1, areassignedto
plement more sophisticated registration methods directly in the
themeasuredbarycenterofthereconstructedimageintheCAL-
pipelinetherebyreducingthenumberofoutliersandimproving
IFAdatacubes.
theoverallastrometricregistrationaccuracy.Wedescribetheas-
Wetestedindependentlytherobustnessandaccuracyofthis
trometricoffsetsforDR1withtheflagFLAG_ASTR,givenfor
approach by both visual inspection of the assigned galaxy cen-
theV500andV1200inTables5and6, respectively. ForV500
tersinbothsetups,aswellasinV500matchingthegalaxyloca-
the values 0, 1, and 2 describe offsets <1(cid:48).(cid:48)4, 1(cid:48).(cid:48)4 to 3(cid:48).(cid:48)0, and
tionofareconstructedgbandfromCALIFAto,whereavailable,
>3(cid:48).(cid:48)0. IntheV1200tablevaluesforFLAG_ASTRof0and2re-
correspondingSDSSimages. ThedeviationsoftheCALIFAas-
fertorelativeoffsetsbetweenV500andV1200smallerorlarger
trometryfromtheICRSaretypicallysmall.
In the V500 setup the offset for most galaxies lies below than3(cid:48)(cid:48).
1(cid:48).(cid:48)4. However,threegalaxiesshowoffsetsbetween1(cid:48).(cid:48)4and3(cid:48)(cid:48),
andfurthertwosubstantiallylargeroffsetsof∼12(cid:48)(cid:48)and∼22(cid:48)(cid:48). In 5.2.2. Seeingandspatialresolution
thesecases,eitherthecentersofthegalaxiesarenotwelldefined
duetodustlanesinstronglyinclinedsystem,orthegalaxieshave Theachievablespatialresolutionofimagingdataisusuallyde-
abrightfieldstarneartheircenter: terminedbythetelescopeaperturetogetherwiththeatmospheric
and instrumental seeing during ground-based observations. In
– UGC10650: Edge-onspiralwithfuzzycenter. Registration
thecaseofCALIFA,thecoarsesamplingofthelargePPakfibers
shouldbebetterthan3(cid:48)(cid:48).
modifiedbytheadopted3-foldditheringpatternasdescribedin
– NGC 4676A: One galaxy of The Mice with unclear center
andapossibleoffsetof∼1(cid:48).(cid:48)5intheastrometricregistration. S12imposesadditionallimitationsonthefinalspatialresolution.
– NGC6032:Highlyinclinedgalaxywithprominentdustlane We measure the seeing in the CALIFA data from two in-
anddifficulttodefinecenter. ∼2(cid:48)(cid:48)offsetinthecoordinates. dependentsources. Thefirstone, thedifferentialimagemotion
– NGC 0477: Astrometric center set to nearby star; resulting monitor(DIMM,Aceituno2004),operatesfullyautomaticallyat
offsetis∼12(cid:48)(cid:48). theCalarAltoobservatoryduringthenight2. TheDIMMseeing
– NGC3991: Centerofthisinclinedspiralisnotthebrightest fortheDR1sampleisshowninFig.12andhasamedianvalue
component. Thecorrectcenterandhencecoordinatesystem of 0(cid:48).(cid:48)9 (FWHM). We confirm this value with measurements of
isoffsetby∼22(cid:48)(cid:48).
2 SincetheDIMMhastighteroperationalconstraints(humiditylower
IntheV1200setupwetestedtheastrometricco-registration than80%andwindspeedlessthan12ms−1)thanthe3.5mtelescope,
with respect to the V500 reference frame. Generally the regis- seeing information is not available for every CALIFA observation.
tration is accurate to better than 2(cid:48)(cid:48). Only for two galaxies we Hence DIMM seeing values can be missing from Tables 5 and 6, but
findamoresubstantialoffset: theoverallseeingdistributionisnotexpectedtobeverydifferent.
Articlenumber,page9of27
NGC 7321
30 7
s
20 xie6
a Fig. 13. Left panel: V band image of
ec] 10 gal 5 NGC 7321, an example of a CALIFA galaxy
[arcs 0 r of 4 wdaittah.aRstiagrhitnptahneeFl:oVDitsotrmibeuatsiounreothfethPeSFFWofHthMe
e
10 mb3 oftheCALIFAPSFasmeasuredfrom34cubes
u 2 withasufficientlybrightstarinthefield. The
N
20 median of the distribution is 3(cid:48).(cid:48)7; the spread
1 of values is due to the flux of the underlying
30 galaxystructureandundersamplingofthestel-
0
30 20 10 0 10 20 30 40 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 larimageandhencehigh-valueoutliersareup-
[arcsec] FWHM of PSF [arcsec] perlimitsinthesecases.
the width of the guide star on images taken by the Acquisition 25
andGuiding(AG)CameraofthePMASinstrument(Rothetal. V1200
2005),which,asexpected,hasaslightlylargermedianvalueof s V500
1(cid:48).(cid:48)15(FWHM).Thisconfirmsthatatmosphericandinstrumental ie20
x
seeingarenotthelimitingfactorsinthespatialresolutionofthe a
l
a
CALIFAcubes. g
1 15
Instead,thelargeapertureofeachfiber(2(cid:48).(cid:48)7diameter)inthe R
D
PPak fiber bundle does not permit to take full advantage of the
f
goodseeingconditions. ThefinalspatialresolutionoftheCAL- o 10
r
IFAdataissetbyfibersizeandtheditherschemetogetherwith e
b
the adopted image reconstruction algorithm. The width of the m
pointspreadfunction(PSF)isdirectlymeasuredfromtheV500 u 5
N
CALIFA datacubes in 35 cases where a bright field star is in
the FoV (see left panel of Fig. 13 for an example). The mea-
sured FWHM distribution of the PSF (Fig. 13 right panel) has 0 2 3 4 5 6 7
a well-defined median value of 3(cid:48).(cid:48)7 (FWHM). To verify these Spectral resolution (FWHM) [A]
measurements we simulated the expected CALIFA PSF for 1(cid:48)(cid:48)
Gaussian seeing while adopting the fiber sampling, dither pat- Fig.14. Distributionofthefinalspectralresolution(FWHM)inthe
tern and image reconstruction algorithm used for the DR1. We combinedcubesasmeasuredfromtheskylinesintheerrorcube. The
obtainedaFWHMforthePSFof≥ 3(cid:48)(cid:48) asalowerlimit, which verticaldashedlinesindicatethespectralresolutiontowhichthespec-
isinagreementwithourempiricalmeasurements. trawereaimedtobehomogenizedduringthedatareduction.Measure-
This spatial resolution is for a large part set by the adopted mentsforV1200(blue, left)areupperlimitssincethenightskylines
usedareintrinsicallyresolvedatthislevel. Thespreadisduetouncer-
kernel for the inverse-distance weighting scheme for the image
taintiesincalculatingthespectralresolutionfromonlyafewavailable
reconstruction (see S12 for details). It was chosen to produce
lines.
smooth images without obvious structure caused by the dither
pattern. Otherimagecombinationapproachessuchasdrizzling
(Fruchter & Hook 2002) can be used to reach closer to the in-
areavailableinthecorrespondingwavelengthrange.Sincethese
trinsicspatialresolutionofCALIFAdata,whichissignificantly
linesareresolvedattheresolutionofthissetupthederivedem-
betterthan3(cid:48)(cid:48) duetotheditheringschemeused. Thechoiceof
piricalresolutionisaconservativeupperlimit.
image combination procedure depends on the actual goal and
We tested the overall wavelength calibration in two ways.
willbeanareaofimprovementinfutureCALIFAdatareleases.
First, we measured the centroids of the same night sky lines
and determined their scatter across the FoV. In all cases, the
5.3. Wavelengthcalibrationaccuracy centroidsarefullyconsistentwithzerooffsetfromthenominal
wavelength, while the scatter is consistent with pure measure-
During the data reduction the spectral resolution was homoge- menterrorsandtheabsenceofanydetectablesystematicspatial
nizedtoreachatargetFWHMof6Å(V500)and2.3Å(V1200), variation.
respectively, over the whole wavelength range, as explained in Secondly,wetestedwhetherthereexistsanysystematicoff-
S12. To go beyond the simple flag-cut from the pipeline di- setsofthewavelengthcalibrationatdifferentwavelengthsofthe
agnostics we tested the spectral resolution and the wavelength
spectra. For this purpose we modeled the recession velocity in
calibration,anditsspectralandspatialconsistency. the spectra of every galaxy’s coadded central 10(cid:48)(cid:48) apertures in
The distribution of spectral resolutions for all galaxies is differentspectralregions. WedidthisusingthepenalizedPixel-
showninFig.14forbothinstrumentsetups. Itwasestimatedby Fitting method (pPXF, Cappellari & Emsellem 2004) to piece-
measuringthewidthofnightskylinesintheerrorcube, which wisefitstellarmodels, hencemeasuringcentralwavelengthsin
contains a reliable trace of the sky lines in the input data. For several (3–5) independent spectral bins showing strong absorp-
the V500 setup we find that it is centered on the target spectral tion features. The resulting velocity differences between these
resolution (6Å). This is not the case for the V1200 setup. For binswereagainalwaysconsistentwithpuremeasurementnoise
this setup only Hg pollution lines (Hgiλ4046Å and λ4358Å) andtheabsenceofsystematics.
Articlenumber,page10of27