A&A566,A108(2014) Astronomy DOI:10.1051/0004-6361/201321489 & (cid:13)c ESO2014 Astrophysics The VIMOS Public Extragalactic Redshift Survey (VIPERS)(cid:63) An unprecedented view of galaxies and large-scale structure at 0.5 < z < 1.2 L.Guzzo1,2,M.Scodeggio3,B.Garilli3,4,B.R.Granett1,A.Fritz3,U.Abbas5,C.Adami4,S.Arnouts4,6,J.Bel7,1, M.Bolzonella8,D.Bottini3,E.Branchini9,26,27,A.Cappi8,28,J.Coupon11,29,O.Cucciati8,16,I.Davidzon8,16, G.DeLucia12,S.delaTorre13,P.Franzetti3,M.Fumana3,P.Hudelot18,O.Ilbert4,A.Iovino1,J.Krywult14, V.LeBrun4,O.LeFèvre4,D.Maccagni3,K.Małek15,F.Marulli16,17,8,H.J.McCracken18,L.Paioro3,J.A.Peacock13, M.Polletta3,A.Pollo20,21,H.Schlagenhaufer22,19,L.A.M.Tasca4,R.Tojeiro10,D.Vergani23,G.Zamorani8, A.Zanichelli24,A.Burden10,C.DiPorto8,A.Marchetti1,25,C.Marinoni7,Y.Mellier18,L.Moscardini16,17,8, R.C.Nichol10,W.J.Percival10,S.Phleps19,andM.Wolk18 (Affiliationscanbefoundafterthereferences) Received16March2013/Accepted10March2014 ABSTRACT WedescribetheconstructionandgeneralfeaturesofVIPERS,theVIMOSPublicExtragalacticRedshiftSurvey.ThisESOLargeProgrammeis usingtheVeryLargeTelescopewiththeaimofbuildingaspectroscopicsampleof∼100000galaxieswithi < 22.5and0.5 < z < 1.5.The AB surveycoversatotalareaof∼24deg2withintheCFHTLS-WideW1andW4fields.VIPERSisdesignedtoaddressabroadrangeofproblemsin large-scalestructureandgalaxyevolution,thankstoauniquecombinationofvolume(∼5×107h−3Mpc3)andsamplingrate(∼40%),comparableto state-of-the-artsurveysofthelocalUniverse,togetherwithextensivemulti-bandopticalandnear-infraredphotometry.Herewepresentthesurvey design, the selection of the source catalogue and the development of the spectroscopic observations. We discuss in detail the overall selection functionthatresultsfromthecombinationofthedifferentconstituentsoftheproject.Thisincludesthemasksarisingfromtheparentphotometric sampleandthespectroscopicinstrumentalfootprint,togetherwiththeweightsneededtoaccountforthesamplingandthesuccessratesofthe observations.Usingthecatalogueof53608galaxyredshiftscomposingtheforthcomingVIPERSPublicDataRelease1(PDR-1),weprovidea firstassessmentofthequalityofthespectroscopicdata.Thestellarcontaminationisfoundtobeonly3.2%,endorsingthequalityofthestar–galaxy separationprocessandfullyconfirmingtheoriginalestimatesbasedontheVVDSdata,whichalsoindicateagalaxyincompletenessfromthis processofonly1.4%.Usingasetof1215repeatedobservations,weestimateanrmsredshifterrorσ/(1+z)=4.7×10−4andcalibratetheinternal z spectralqualitygrading.Benefitingfromthecombinationofsizeanddetailedsamplingofthisdataset,weconcludebypresentingamapshowing inunprecedenteddetailthelarge-scaledistributionofgalaxiesbetween5and8billionyearsago. Keywords.cosmology:observations–large-scalestructureofUniverse–galaxies:distancesandredshifts–galaxies:statistics 1. Introduction highlyeffectivedilutetracersoflargevolumes(Eisensteinetal. 2011;Ahnetal.2012). One of the major achievements of observational cosmology in In addition to changing our view of the galaxy distribution the 20th century has been the detailed reconstruction of the around us, the quantitative analysis of galaxy redshift surveys large-scalestructureofwhatisnowcalledthe“localUniverse” has consistently yielded important advances in our knowledge (z ≤ 0.2). Large redshift surveys such as the 2dFGRS (Colless of the cosmological model. Galaxy clustering on large scales etal.2001)andSDSS(Yorketal.2000;Abazajianetal.2009) is one of the most important relics of the initial conditions that have assembled samples of over a million objects, precisely shapedourUniverse,andtheobservedshapeofthepowerspec- characterising large-scale structure in the nearby Universe on scalesrangingfrom0.1to100h−1Mpc.TheSDSSinparticular trumP(k)ofdensityfluctuations(orofitsFouriertransform,the correlationfunctionξ(r))indicatesthatweliveinalow-density isstillextendingitsreach,usingluminousredgalaxies(LRG)as Universe in which only 25–30% of the mass-energy density is (cid:63) Based on observations collected at the European Southern provided by (mostly dark) matter. Combined with other obser- Observatory, Cerro Paranal, Chile, using the Very Large Telescope vations,particularlyanisotropiesinthecosmicmicrowaveback- under programmes 182.A-0886 and partly 070.A-9007. Also based ground (CMB), this observation has long argued for the rejec- on observations obtained with MegaPrime/MegaCam, a joint project tionofopenmodelsinfavourofaflatuniversedominatedbya ofCFHTandCEA/DAPNIA,attheCanada-France-HawaiiTelescope negative-pressurecosmologicalconstant(Efstathiouetal.1990). (CFHT),whichisoperatedbytheNationalResearchCouncil(NRC)of Thisconclusionpredatedthemoredirectdemonstrationviathe Canada, the Institut National des Sciences de l’Univers of the Centre HubblediagramofdistantTypeIaSupernovae(Riessetal.1998; National de la Recherche Scientifique (CNRS) of France, and the Perlmutteretal.1999)thattheUniverseiscurrentlyinaphase University of Hawaii. This work is based in part on data products of accelerated expansion. Subsequent large-scale structure and produced at TERAPIX and the Canadian Astronomy Data Centre as partoftheCanada-France-HawaiiTelescopeLegacySurvey,acollab- CMBdata(e.g.Coleetal.2005;Komatsuetal.2009;Hinshaw orative project of NRC and CNRS. The VIPERS website is http: etal.2013)haveonlyreinforcedtheconclusionthattheUniverse //www.vipers.inaf.it/ isdominatedbyarepulsive“darkenergy”.Currentobservations ArticlepublishedbyEDPSciences A108,page1of21 A&A566,A108(2014) areconsistentwiththelatterbeinginthesimplestformalready colour and stellar mass (e.g. Zehavi et al. 2004). At the same suggested by Einstein with his Cosmological Constant, i.e. a time,importantglobalgalaxypopulationtrendsinvolvingprop- fluidwithnon-evolvingequationofstatew=−1. ertiessuchasluminosities,stellarmasses,coloursandstructural Theoretical difficulties with the cosmological constant, parameters can be precisely measured when these parameters specifically the smallness and fine-tuning problems (e.g. are available for ∼106 objects, as in the case of the SDSS (e.g. Weinberg 1989) make scenarios with evolving dark energy an Kauffmannetal.2003). appealing alternative. This is the motivation for projects aim- In more recent years, deeper redshift surveys over areas of ing at detecting a possible evolution of w(z). Redshift surveys 1–2 deg2 have focused on exploring how this detailed picture are playing a crucial role in this endeavour, in particular after emergedfromthedistantpast.Thiswasthedirectconsequence the discovery of the signature of baryonic acoustic oscillations of the development during the 1990s of multi-object spectro- (BAO)fromthepre-recombinationplasmaintolarge-scalestruc- graphs on 8-m class telescopes. The most notable projects of ture. This “standard rod” on a comoving scale of ∼150 Mpc this kind have been the VIMOS VLT Deep Survey (VVDS; (Percival et al. 2001; Cole et al. 2005; Eisenstein et al. 2005) LeFèvreetal.2005),theDEEP2survey(Coiletal.2008)and providesuswithapowerfulmeantomeasuretheexpansionhis- the zCOSMOS survey (Lilly et al. 2009), which adopted vari- tory H(z) via the angular diameter distance (e.g. Percival et al. ous strategies aimed at covering an extended redshift range, up 2010;Blakeetal.2011a;Andersonetal.2012). to z ∼ 4.5. Such depths inevitably limit the angular size and An even more radical explanation of the observed accel- thus the volume explored in a given redshift interval, reflect- erated expansion could be a breakdown of General Relativity ing the desire of these projects to trace galaxy evolution back (GR) on cosmological scales (see e.g. Carroll et al. 2004; Jain to its earliest phases, while understanding its relationship with &Khoury2010).Suchascenarioisfullydegeneratewithdark environmentoveralimitedrangeofscales1.Evolutionarytrends energy in terms of H(z), a degeneracy that in principle can be inthedark-matter/galaxyconnectionwereexploredusingthese liftedbymeasuringthegrowthrateofstructure,whichdepends surveys (Zheng et al. 2007; Abbas et al. 2010), but none of onthespecifictheorydescribinggravity. these samples had sufficient volume to produce stable and re- There are in principle several experimental ways to mea- liablecomparisonsofe.g.theamplitudeandshapeofthecorre- sure the growth of structure. Galaxy peculiar motions, in par- lationfunction.OnlytheWideextensionofVVDS(Garillietal. ticular, directly reflect such growth. When the redshift is used 2008),startedtohavesufficientvolumeastoattemptcosmolog- asadistanceproxy,theyproduceameasurableeffectoncluster- ically meaningful computations at z ∼ 1 (Guzzo et al. 2008), ingmeasurements,whatwecallredshift-spacedistortions(RSD, albeitwithlargeerrorbars.Ingeneral,clusteringmeasurements Kaiser 1987). The anisotropy of statistical measurements like at z ∼ 1 from these samples remained dominated by field-to- thetwo-pointcorrelationfunctionisproportionaltothegrowth field fluctuations (cosmic variance), as dramatically shown by rateofcosmicstructure f(z),whichisatrademarkofthegrav- the discrepancy observed between the VVDS and zCOSMOS ity theory: if GR holds, we expect to measure a growth rate correlationfunctionestimatesatz(cid:39)0.8(delaTorreetal.2010). f(z)=[Ω (z)]0.55(Peebles1980;Lahavetal.1991).Ifgravityis Attheendofthepastdecadeitwasthereforeclearthatanew M modifiedonlargescales,differentformsarepredicted(e.g.Dvali step in deep redshift surveys was needed, if these were to pro- etal.2000;Linder&Cahn2007).Infact,althoughtheRSDef- ducestatisticalresultsthatcouldbecomparedonanequalfoot- fect has been well known since the late 1980s (Kaiser 1987), ingwiththosederivedfromsurveysofthelocalUniverse,such its potential in the context of dark energy and modified gravity as2dFGRSandSDSS.Followingthoseefforts,newgenerations hasbecomeclearonlyrecently(Guzzoetal.2008;Zhangetal. of cosmological surveys have focused on covering the largest 2007).TheRSDmethodisnowconsideredtobeoneofthemost possiblevolumesatintermediatedepths,utilizingrelativelylow- promising probes for future dark energy experiments, as testi- density tracers, with the main goal of measuring the BAO sig- fiedbytheexponentialgrowthinthenumberofworksonboth nalatredshifts0.4–0.8.ThisisthecasewiththeSDSS-3BOSS measurements(e.g.Beutleretal.2012;Blakeetal.2011a;Reid project (Eisenstein et al. 2011; Dawson et al. 2013), which ex- et al. 2012), and theoretical modelling (e.g. Song & Percival tendstheconceptpioneeredbytheSDSSselectionofLRG(e.g. 2009; Percival & White 2009; White et al. 2009; Scoccimarro Andersonetal.2012;Reidetal.2012).Similarly,theWiggleZ 2004;Taruyaetal.2010;Kwanetal.2012;Reid&White2011; survey further exploited the long-lived 2dF positioner on the delaTorre&Guzzo2012).Redshiftsurveysarethusexpected AAT 4-m telescope, to target emission-line galaxies selected tobeasimportantforcosmologyinthepresentCenturyasthey fromUVobservationsoftheGALEXsatellite(Drinkwateretal. were in the previous one, as suggested by their central role in 2010;Blakeetal.2011a,b).Boththesesurveysarecharacterised several planned experiments – especially the ESA dark-energy by a very large volume (1–2h−3Gpc3), and a relatively sparse mission,Euclid(Laureijsetal.2011). galaxy population (∼10−4h3Mpc−3). This is typical of surveys The scientific yield of a redshift survey, however, extends performed with fibre positioning spectrograph, which normally wellbeyondfundamentalcosmologicalaspects.Itisequallyim- can observe 500–1000 galaxies over areas of 1–2 square de- portanttoachieveanunderstandingoftherelationshipbetween grees. Higher galaxy densities can be achieved with such sys- the observed baryonic components in galaxies and the dark- tems via multiple visits, although this then limits the redshift matterhaloesthathostthem.Forthispurpose,weneedtobuild statistically complete samples of galaxies with measured posi- 1 ThePRIMUSsurvey(Coiletal.2011)isanotablerecentaddition, tions,luminosity,spectralpropertiesand(typically)coloursand with∼120000spectraforgalaxiesatz<1,collectedover7fieldsfora totalareaof9deg2.Redshiftsareobtainedwithalow-resolutionprism stellarmasses;inprovidingsuchdata,redshiftsurveysarethus (Cooletal.2013),yieldingtypicalerrorsoneorderofmagnitudelarger avitalprobeofgalaxyformationandevolution.Significantsta- thanthoseoftheVIMOSsurveys(seealsoSect.5.3).Assuch,analyses tistical progress has been made in relating the galaxy distribu- of these data have concentrated on galaxy evolution studies requiring tion to the underlying dark matter, via “halo occupation distri- lower precision on galaxy distances. Nevertheless, while we were re- bution”(HOD)modelling(Seljak2000;Peacock&Smith2000; visingthispaper,afirstdetailedstudyoftheclusteringofgalaxiesasa Cooray&Sheth2002),ofaccurateestimatesofthegalaxytwo- functionofluminosityandcolourwaspublishedinthearXiv(Skibba point correlation function, for samples selected in luminosity, etal.2014). A108,page2of21 L.Guzzoetal.:TheVIMOSPublicExtragalacticRedshiftSurvey(VIPERS) and/or volume surveyed. This approach has been taken by the i = 22.5, but observing all kinds of objects (stars and galax- AB GAMAsurvey(Driveretal.2011),whichaimstoachievesimi- ies),withlowsampling((cid:39)20%). larnumbersofredshiftstothe2dFGRS(∼200000),butworking Building upon this experience, VIPERS was designed to tor<19.8andouttoz(cid:39)0.5.Indeed,thehighsamplingdensity maximise the number of galaxies observed in the range of in- ofGAMAmakesitanimportantintermediatestepbetweenthe terest, i.e. at z > 0.5, while at the same time attempting to local surveys and the higher redshifts probed by the survey we select against stars, which represented a contamination up to arepresentinginthispaper,i.e.VIPERS. 30% in some of the VVDS-Wide fields. The latter criterion re- VIPERS stands for VIMOS Public Extragalactic Redshift quires multi-band photometric information and excellent see- Surveyandhasbeendesignedtomeasureredshiftsforapproxi- ing quality, but these qualities also benefit the galaxy sample, mately100000galaxiesatamedianredshiftz(cid:39)0.8.Thecentral whereawiderrangeofancillaryscienceisenabledifthegalaxy goal of this strategy is to build a data set capable of achieving surface-brightness profiles can be well resolved. The outstand- an order of magnitude improvement on the key statistical de- ing imaging dataset that was available for these purposes was scriptionsofthegalaxydistributionandinternalproperties,atan theCanada-France-HawaiiTelescopeLegacySurvey(CFHTLS) epochwhentheUniversewasabouthalfitscurrentage.Sucha Widephotometriccatalogue,asdescribedbelowinSect.3. datasetwouldallowcombinationwithlocalsamplesonacom- Thedesiredredshiftrangewasisolatedthroughasimpleand parable statistical footing. Despite being centred at z¯ ∼ 0.7, in robustcolour–colourselectiononthe(r−i)vs.(u−g)plane(as terms of volume and number density VIPERS is similar to lo- shown in Fig. 3). This is one of many ways in which we have calsurveyslike2dFGRSandSDSS.Allthesesurveysarechar- been able to benefit from the experience of previous VIMOS acterised by a high sampling density, compared to the sparser spectroscopicsurveys:wecouldbeconfidentinadvancethatthis samplesoftherecentgenerationofBAO-orientedsurveys. selection method would efficiently remove galaxies at z < 0.5, InthispaperweprovideanoverviewoftheVIPERSsurvey while yielding >98% completeness for z > 0.6, as verified in design and strategy, discussing in some detail the construction the results shown below. A precise calibration of this separa- of the target sample. The layout of the paper is as follows: in tion method was made possible by the location of the VVDS- Sect. 2, we discuss the survey design; in Sect. 3 we describe Wide(i < 22.5)andVVDS-Deep(i < 24)sampleswithin AB AB the properties of the VIPERS parent photometric data and the the W4 and W1 fields of CFHTLS, respectively. This was an build-upofahomogeneoussampleover24deg2;inSect.4we important reason for locating the VIPERS survey areas within discusshowfromthesedatathespecificVIPERStargetsample these two CFHTLS fields while partly overlapping the original atz>0.5hasbeenselected,usinggalaxycolours;inSect.5the VVDS areas, as shown in Fig. 1. The magnitude limit was set detailsoftheVIMOSobservationsandthegeneralpropertiesof as in VVDS-Wide, i.e. 17.5 ≤ i ≤ 22.5 (after correction for AB thespectroscopicsamplearepresented;inSect.6wediscussthe Galacticextinction). variousselectioneffectsandhowtheyhavebeenaccountedfor; The details of the star–galaxy separation are discussed in finally,inSect.7wepresenttheredshiftandlarge-scalespatial Appendix A, while the colour–colour selection is described in distributionofthecurrentsample,summarisingthescientificin- Sect.4. vestigationsthatarepartofseparatepaperscurrentlysubmitted orinpreparation. As a public survey, we hope and expect that the range of 3. Photometricsourcecatalogue science that will emerge from VIPERS will greatly exceed the core analyses from the VIPERS Team. This paper is therefore TheVIPERStargetselectionisderivedfromthe‘T0005’release alsotointroducethenewVIPERSdata,inviewofthefirstPublic of the CFHTLS Wide which was available for the first observ- DataRelease(PDR-1)2,whichisdescribedinmoredetailinthe ingseason2007/2008.Thisobjectselectionwascompletedand specificaccompanyingpaper(Garillietal.2014). improvedusingthesubsequentT0006release,aswedescribein thefollowing. The mean limiting AB magnitudes of CFHTLS Wide (cor- 2. Surveydesign responding to the 50% completeness for point sources) are ∼25.3,25.5,24.8,24.48,23.60inu∗,g(cid:48),r(cid:48),i(cid:48),z(cid:48),respectively.To VIPERS was conceived in2007 with a focus on clustering and RSD at z (cid:39) 0.5–1, but with a desire to enable broader goals construct the CFHTLS catalogues used here, objects in each tile were detected on a gri-χ2 image (Szalay et al. 1999) and involvinglarge-scalestructureandgalaxyevolution,similarlyto theachievementsof2dFGRSandSDSSatz (cid:39) 0.1.Thesurvey galaxies were selected using SEXtractor’s “mag_auto” mag- nitudes(Bertin&Arnouts1996),intheABsystem3.Theseare design was also strongly driven by the specific features of the themagnitudesusedthroughoutthiswork,aftertheyhavebeen VIMOSspectrograph,whichhasarelativelysmallfieldofview compared to fibre positioners ((cid:39)18 × 16arcmin2; see Sect. 6), correctedforforegroundGalacticextinctionusingthefollowing prescription: butalargeryieldintermsofredshiftsperunitarea. Given the luminosity function of galaxies and results from u = u∗ −4.716∗E(B−V) (1) previousVIMOSsurveysasVVDS(LeFèvreetal.2005;Garilli raw et al. 2008) and zCOSMOS (Lilly et al. 2009), we knew that a g = g(cid:48) −3.654∗E(B−V) (2) raw magnitude-limitedsamplewithiAB <22.5–23.0wouldcoverthe r = r(cid:48) −2.691∗E(B−V) (3) raw redshiftrangeouttoz∼1.2,andcouldbeassembledwithfairly i = i(cid:48) −1.998∗E(B−V) (4) shortVIMOSexposuretimes(<1h).Also,taking2dFGRSasa raw localreference,acomparablesurveyvolume∼5×107h−3Mpc3 z = z(cid:48)raw−1.530∗E(B−V), (5) could have been covered by mapping at this depth an area of ∼25 deg2. The first attempt towards this kind of survey was wheretheextinctionfactorE(B−V)isderivedateachgalaxy’s VVDS-Wide, which covered ∼8 deg2 down to a magnitude positionfromtheSchlegeldustmaps(Schlegeletal.1998). 2 Availableathttp://vipers.inaf.it 3 http://terapix.iap.fr/rubrique.php?id_rubrique=252 A108,page3of21 A&A566,A108(2014) -2 and T0006catalogs, limited to i < 23.0, were positionally CFHTLS W1 AB matched over the area of each hole, using a search radius of VIPERS W1 -3 VVDS deep 0.6arcsec.Allmatcheswithacompatiblei-bandmagnitude(de- finedashavingadifferencelessthan0.2mag)wereconsidered -4 27 26 25 24 23 22 21 20 19 as good identifications and used to verify the consistency be- -5 tweenthetworeleases. 18 17 16 15 14 13 12 11 10 For objects near the VIPERS faint limit, i.e. i ∼ 22.5, AB g) -6 9 8 7 6 5 4 3 2 1 thermsmagnitudeoffsetbetweenthetwocatalogueswasfound e C (d -7 tsomraallnegrethbaentwtheiesnfo0r.0b2ritgoht0e.r04obmjeacgts.(lGarigveernitnhitshreesuu-blta,nwde),caonnd- E D -8 cluded that the T0006 version of galaxy magnitudes could be used directly to replace the bad or missing magnitudes for the -9 originalT0005objectsintheholes.Thissolutionwasdefinitely preferabletoreplacingallmagnitudeswiththeirT0006values, -10 an operation that would have modified the target sample at the faintlimitsimplyduetostatisticalscatter. -11 Only a few of the T0005 holes arising from CCD failures -12 werenotfilledbytheT0006release.Tocompletetheseremain- 39 38 37 36 35 34 33 32 31 30 29 ing areas, Director’s Discretionary Time (DDT) was awarded RA (deg) at CFHT with MegaCam in summer 2009 (Arnouts & Guzzo, priv.comm.).Attheendof2010,thecombinationofnewT0006 observations and the DDTdata resulted in a virtually complete 5 CFHTLS W4 coverage in all five bands of the two VIPERS areas in W1 and VIPERS W4 W4. The last problem to be resolved was re-calibrating a few VVDS F22 4 small areas which were observed in T0006 with a new i-band filter,called“y”,astheoriginali-bandfilterbrokein2007.This procedureisdescribedinAppendixB. 3 g) 11 10 9 8 7 de 3.1. Tile-to-tilezero-pointhomogenisation ( 2 C E 6 5 4 3 2 1 The CFHTLS data are provided in single tiles of ∼1 deg D side, overlapping each other by ∼2 arcmin to allow for cross- 1 calibration.TheseareshowninFig.1fortheW1andW4fields, togetherwiththepositionofthetwoVIPERSareas.Tobuildthe VIPERSglobalcataloguewemergedadjacenttiles,eliminating 0 duplicatedobjects.Inthesecases,theobjectinapairhavingthe bestTerapixflagwaschosen;iftheflagswereidentical,theob- -1 jectatthegreaterdistancefromthetileborderwaschosen.Tiles were merged proceeding first in right ascension rows and then 336 335 334 333 332 331 330 mergingtherowsintoasinglecatalogue. RA (deg) Foranygalaxysurveyplanningtomeasurelarge-scaleclus- teringitiscrucialthatthephotometricorcolourselectionisas Fig.1.AreascoveredbyVIPERSwithintheCFHTLS-WideW1(top) homogeneous as possible over the full survey area in order to andW4(bottom)fields.Theinternalnumberingreportedoneachtileis avoidcreatingspuriousobjectdensityfluctuationsthatcouldbe linkedtotheCFHTLSnamingconventioninTablesC.1andC.2.Also mistakenasrealinhomogeneities.GiventhewaytheCFHTLS- shownarethepositionsoftheVVDS-Deep(LeFèvreetal.2005)and VVDS-Wide(Garillietal.2008)surveyfields. Wide catalogue has been assembled, verifying and correcting any tile-to-tile variation of this kind is therefore of utmost im- portance. In fact, it was known and directly verified that each When the first target catalogues were generated, the tileinT0005stillhadasmallbutnon-negligiblezero-pointoff- CFHTLSsurveyincludedsomephotometricallyincompletear- set in some of the photometric bands. These offsets are a con- eas (“holes” hereafter). In these areas one or more bands was sequenceofnon-photometricimagesbeingusedasphotometric eithercorruptedormissing.Inparticular,alloftheVIPERSW1 anchorfieldsintheglobalphotometricsolution. fieldatrightascensionslessthanRA(cid:39)02h09(cid:48)weremissingone Thesetile-to-tilecolourvariationsareevidentwhenstarsare band as CFHTLS Wide observations had not been completed. plottedinacolour–colourdiagram,asinFig.2.Inthisfigurewe Smaller survey holes were mostly due to the partial failure of showthe(u−g)vs.(r−i)coloursforstellarobjectsintwopartic- amplifier electronics (since all CCDs have two outputs, some ularlydiscrepanttiles(seeAppendixAfordetailsonhowstars imagesaremissingonlyhalf-detectorareas). andgalaxiesareseparated).Suchoffsetscanproducetwokinds In general, these missing bands meant that we were not ofsystematiceffectsinasurveylikeVIPERS.First,atile-to-tile able to select VIPERS targets in the affected areas and they difference in the selection magnitude (i band) would introduce were therefore excluded from our first two observing seasons a varying survey depth over the sky and thus a variation in the (2008 and 2009). The majority of these problems were fixed expectednumbercountsandredshiftdistribution.Secondly,the in Summer 2010 using the CFHTLS-T0006, which was care- colourswouldbeaffected,andthusanycolour–colourselection fully merged with the existing VIPERS target list. The T0005 (astheoneappliedtoselectgalaxiesatz > 0.5fortheVIPERS A108,page4of21 L.Guzzoetal.:TheVIMOSPublicExtragalacticRedshiftSurvey(VIPERS) Fig.2.Oneoflargesttile-to-tilemagnitudezero-pointvariationsinthe Fig.3.Distributioninthe(r−i)vs.(u−g)planeofi < 22.5galax- AB T0005 data. The position of the stellar sequence in the (g − r) vs. ieswithknownredshiftfromtheVVDS-Deepsurvey,showingthekind (u−g)planeiscomparedfortile#9andtile#11intheW4VIPERS ofselectionappliedtoconstructtheVIPERStargetsample.Thecolour area(seeAppendixC),showinganoffsetof∼0.15magin(g−r)and selectionofEq.(9)isdescribedbythecontinuousline,whichempiri- ∼0.06in(u−g)betweenthetwotiles. callysplitsthesampleintoz>0.5(redfilledcircles)andz<0.5(blue opencircles)byoptimisingthecompletenessandcontaminationofthe high-redshiftsample. targetcatalogue–seenextsection),wouldvaryfromonetileto another. AllfollowingstepsintheselectionofVIPERStargetgalax- Thewell-definedlocationofstarsincolour–colourspace,as ieswerethenoperatedoncolourscorrectedusingtheseoffsets, showninFig.2,suggestsatechniqueforapossiblecorrectionof i.e. thecolourvariations,i.e.usingtheobservedstellarsequenceas a colour calibrator (see High et al. 2009, for a similar more re- (u−g) = (u−g) −δ (6) uncorr ug centapplicationofthisregressiontechnique).Animportantas- (g−r) = (g−r) −δ (7) uncorr gr sumption of this correction procedure is that stars and galaxies (r−i) = (r−i) −δ . (8) areaffectedbysimilarzero-pointshifts,andthusthatstellarse- uncorr ri quencescanalsobeusedtoimprovethephotometriccalibration of extended objects. This assumption is quite reasonable and it 4. SelectionofVIPERSgalaxytargets isthesameadoptedatTerapixinthepasttocheckinternalcal- Aroundhalfofthegalaxiesinamagnitude-limitedsamplewith ibrationuntilthesecond-lastrelease,i.e.T0006.Withthelatest i < 22.5 are at z < 0.5. At the same time, the average num- release,T0007,thereareindicationsthatacontributiontothese AB ber of slits that can be accommodated within one of the four zero-point discrepancies could be also due to a dependence on seeingof mag_autowhenappliedtostellarobjects.Thiseffect VIMOSquadrants(seebelow)isapproximatelyfixed,forapar- ent sample with a given depth and clustering. This means that isnotfullyunderstoodyetanditsamplitudeissmallerthanthe inapuremagnitudelimitedsurveywithi <22.5,aroundhalf correctionsweoriginallyappliedtotheT0005data.Thepoten- AB of the slits would fall on z < 0.5 galaxies. Given the original tial systematic impact of this uncertainty, in particular on clus- goalofVIPERStobuildasamplecomplementarytolocalsur- tering analyses of the PDR-1 sample, is explicitly addressed in veys, a strategy was devised as to select a priori only galaxies thecorrespondingpapers(seee.g.delaTorreetal.2013). at higher redshifts, doubling in this way the sampling over the The colour corrections were carried out assuming (a) that high-redshift range. Using available magnitude-limited VVDS thei-bandmagnitudehadanegligiblevariationfromtiletotile, data,asimpleyeteffectiveandrobustcolourselectioncriterion and (b) taking the colours measured in tile W1-25 (see Fig. 1) wasdevisedthroughaseriesofexperiments.Themosteffective asthereferenceones.W1-25isthetileoverlappingtheVVDS- criteriaareshowninFig.3appliedtotheVVDSdata.Galaxies Deep survey, which was used to calibrate the colour selection areretainedinthesourcelistiftheircoloursobeythefollowing criteria as discussed in Sect. 4. By referring all colours to that relationship: tile,weassured(atleast)thatthecolour-redshiftcorrelationwe calibrated was applied self-consistently to all tiles. For all tiles (r−i)>0.5(u−g) OR (r−i)>0.7. (9) covered by VIPERS we measured therefore the (u − g) value oftheblue-endcut-offinthestellarsequence,clearlyvisiblein The resulting distribution of the true redshifts for the selected Fig.2,togetherwiththezeropointsderivedfromalinearregres- samplesisshowninFig.4,withthecorrespondinglevelofcom- siontothe(g−r)vs.(u−g)and(r−i)vs.(u−g)relationships pleteness shown in Fig. 5. To compute the latter quantity, we for stars. These two regressions give a consistent slope of 0.50 usedtheVVDSdata(bothDeepandWide),andplottheratioof and0.23,respectively,overalltiles.Thisallowedustocompute thenumbersofobjectsinaVIPERS-likeselectedsample,tothe threecolouroffsetsδ ,δ andδ foreachtile,correspondingto original total redshift sample. We call this quantity the Colour ug gr ri thevaluesrequiredtomatchthesamemeasurementsinW1-25. SamplingRate(CSR).Asindicatedbythecombinationofthese A108,page5of21 A&A566,A108(2014) Fig.4.Testofthecolour–colourredshiftselection,usinggalaxieswith knownredshiftfromtheVVDS-Deepsurvey,limitedtoi <22.5.The AB colourlocusinFig.3isusedtoseparateapriorigalaxieslyingbelow Fig.5. Direct verification of the completeness of the VIPERS colour (blue-dashedline)andabove(solidredline)z (cid:39) 0.5.Thedottedblack selectionasafunctionofredshift,usingbothVVDS-DeepandVVDS- line shows the global dN/dz of the sample. The VVDS-Deep sample Wide data, in W1 and W4 respectively. Note that the original colour hasbeenlimitedtoobjectsbelongingtotile#25(wherethebulkofthe criteriaweredefinedbasedonlyontheVVDS-Deepdata.Thecurves sampleisconcentrated),giventhatthishasbeenusedasthereference and points give the Colour Sampling Rate (CSR), i.e. the ratio of the fortheglobalcolourcalibrationdiscussedinthetext. numberofgalaxiessatisfyingtheVIPERScriteriawithinaredshiftbin andthetotalnumberofgalaxiesinthatsamebin.Bothfieldsprovide consistentselectionfunctions,indicatingthatthecolour–colourselec- tionfunctionisbasicallyunityabovez = 0.6andcanbeconsistently two figures, the VIPERS selection does not introduce any sig- modelledinthetransitionregion0.4<z<0.6. nificant colour bias (i.e. it selects virtually all galaxies) above z∼0.6,withanacceptablecontamination(∼5%)oflow-redshift interlopers. More quantitatively, some insight on the potential be noted also that, as visible in Fig. 4, a significant fraction incompleteness of the selection procedure – i.e. on how many of this incompleteness is concentrated in the transition region galaxiesthatshouldhavebeenincludedinoursampleatz>0.5 0.5 < z < 0.6andcanbemodelledasshowninFig.5anddis- arelost–canbederivedbylookingindetailatthefewoutliers cussedinsomedetailinGarillietal.(2014). in the blue histogram of Fig. 4. The tail of unselected objects An alternative technique to select a high-redshift sample at z > 0.5 in the VVDS calibration sample includes 46 cases. could have been to use photometric redshifts computed using 14oftheseareclassifiedasactivegalacticnuclei(AGN)bythe all five CFHTLS bands. We verified that this method provides VVDS, which explains why their colour–redshift relation does comparableperformanceintermsofcompletenessandcontam- not match the standard criteria defined for galaxies; 10 out of ination to the colour–colour selection. However we preferred a the14AGNshavez > 1.2andarethusoutofthetypicalrange simplecolour–colourcriterion,asitcanbereproducedprecisely usedbyVIPERSforstatisticalstudies;thusonlytheremaining4 at any time, while photometric redshifts depend inevitably also couldpotentiallybepartoftheVIPERStargetsample,although onthefeaturesofthespecificcodesandtemplateselectionused, onecannotdistinguishhowmuchtheactivenucleuscontributes whichwillevolvewithtime. to the overall magnitude (and thus, understand whether the ob- Finally, to further broaden the scientific yield of VIPERS, jectwouldbebrighterthanIAB =22.5basedonthesolemagni- thegalaxytargetcataloguewassupplementedwithtwosmallad- tudeofthehostgalaxy).7oftheremaingoutliershaveanerror ditionalsamplesofAGNcandidates.Theseincludeasampleof on the u-band magnitude which is larger than 0.1 mag, which X-rayselectedAGNsfromtheXMM-LSSsurveyintheW1field makes their u − g colour unreliable; another 3 have a redshift (Pierre et al. 2007), and a sample of colour–defined AGN can- flag = 1, i.e. their redshift has a ∼50% probability to be wrong didatesselectedamongobjectsclassifiedasstarsintheprevious (seeSect.5.3).Weareleftwith22furtheroutliers,amongwhich phase.Thesetwocataloguescontributedonaverage1–3objects 8 objects are at z > 1.2, while (after visual inspection) another per quadrant (against about 90 galaxy targets) with negligible 4 show noisy spectra suggesting an incorrect redshift, thus re- impactonthegalaxyselectionfunction.TheseAGNcandidates sulting in 10 further galaxies that apparently should have been areexcludedfromthecurrentPDR-1sample.Allthedetailson includedinthetargetsample. the selection criteria and the properties of the resulting objects Based on these figures, a conservative upper limit to the willbediscussedinafuturepaper. number of galaxies missed in the range 0.5 < z < 1.2 for the VIPERS-liketestsamplecanbeobtainedbysummingupthese numbers: 4 (AGN host galaxies), plus 3 (assuming conserva- 5. VIMOSobservations tively that all Flag = 1 redshifts are correct), 7 (wrong colour), 5.1. TheVIMOSspectrograph plus 10 (remaining galaxies with confirmed z > 0.5 redshifts). This corresponds to an estimated global incompleteness of the TheVIPERSprojectisdesignedaroundtheESOVIsibleMulti- z>0.5colour/redshiftselectionof24/1068,i.e.2.2%.Itshould Object Spectrograph (VIMOS), on “Melipal”, the ESO Very A108,page6of21 L.Guzzoetal.:TheVIMOSPublicExtragalacticRedshiftSurvey(VIPERS) Q1 Q2 N Q4 E Q3 2.4' 8.0' 2.0' 7.0' Fig.6. Example of the detailed footprint and disposition of the four quadrants in a full VIMOS pointing (W1P082 in this case). Note the re- constructedboundaries(solidredlines),whichhavebeentracedpointing-by-pointingthroughanautomaticdetectionalgorithmthatfollowsthe bordersoftheilluminatedarea.Thesecanvaryingeneralamongdifferentpointingsinthedatabase,inparticularduetotheCCDrefurbishment of2010andsometimestovignettingbythetelescopeguideprobearm. LargeTelescope(VLT)Unit3(LeFèvreetal.2003).VIMOSis spectralresolutionR (cid:39) 250overthewavelengthrange∼5500– a4-channelimagingspectrograph;eachchannel(a“quadrant”) 9500Å.Theinstrumenthasnoatmosphericdispersioncompen- covers∼7×8arcmin2 foratotalfieldofview(a“pointing”)of sator,giventhelargesizeofitsfield-of-viewattheVLTNasmyth ∼218arcmin2.Eachchannelisacompletespectrographwiththe focus((cid:39)1m).Forthisreason,observationshavetobelimitedto possibilitytoinsert∼30×30cm2 slitmasksattheentrancefo- airmasses below 1.7. For VIPERS observations we rarely went cal plane, as well as broad-band filters or grisms. The standard aboveanairmassof1.5. layoutofthefourquadrantsontheskyisreproducedinFig.6. Thefigureshowstheslitpositionsandtheresultinglocationof TopreparetheMOSmasks,directexposures(“pre-images”) thespectra,overlaidonthedirectpre-imageofpointingP082in need to be observed beforehand under the same instrumental theW1field. conditions. Object positions in these images are then cross- The pixel scale on the CCD detectors is 0.205 arcsec/pixel, correlated with the target catalogue in order match its astro- providing excellent sampling of the Paranal mean image qual- metric coordinates to the actual instrument coordinate system. ity and Nyquist sampling for a slit 0.5 arcsecond in width. This operation is performed during the mask preparation using For the VIPERS survey, we use slits of 1 arcsecond, together VMMPS,thestandardpackageforautomaticoptimisationofthe withthe“low-resolutionred”(LR-Red)grism,whichprovidesa positionsandtotalnumberofslits(Bottinietal.2005). A108,page7of21 A&A566,A108(2014) Fig.7.AfewrepresentativeexamplesofVIPERSspectraofearly-andlate-typegalaxies,chosenamongthedifferentqualityclasses(i.e.quality flags)andatdifferentredshifts.Thequotedfluxistheobservedone,withoutcorrectionsforfinite-slitlosses.Thetypicalabsorptionandemission featuresaremarked. In summer 2010, VIMOS was upgraded with new red- the standard CCD data reduction, spectral extraction and cali- sensitive CCDs in each of the 4 channels, as well as with a bration follow the usual recipes discussed in previous VIMOS new active flexure compensation system. The reliability of the papers (Le Fèvre et al. 2005; Lilly et al. 2009). The difference mask exchange system was also improved (Hammersley et al. in the case of VIPERS is that the only operation for which we 2010). The original thinned E2V detectors were replaced by stillrequirehumaninterventionistheverificationandvalidation twice-thicker E2V devices, considerably lowering the fringing ofthemeasuredredshift.Alldatareductionhasbeencentralised and increasing the global instrument efficiency by up to a fac- in our data reduction and management centre at INAF– IASF tor 2.5 (one magnitude) in the redder part of the wavelength Milano. When ready, the fully reduced data are made available range. This upgrade significantly improved the average quality to the team within a dedicated database. The full management of VIPERS spectra, resulting in a significantly higher redshift of these operations within the “EasyLife” environment is de- measurementsuccessrate. scribedinGarillietal.(2008).Figure7showsafewexamplesof VIPERS spectra, for galaxies with varying redshift and quality flag. In common with previous VIMOS surveys (e.g. Le Fèvre 5.2. Datareduction,redshiftmeasurementandvalidation et al. 2005; Lilly et al. 2009), all redshifts have been validated independently by two scientists but with some simplification VIPERS is the first VIMOS redshift survey for which the data to increase efficiency given the very large number of spectra. reduction is performed with a fully automated pipeline, start- Nevertheless,thisrequiredaverystrongteameffort.Twoteam ing from the raw data and down to the calibrated spectra and membersareassignedthesameVIMOSfieldtoreview,withone redshift measurements. The pipeline includes and updates al- of the two being the primary person responsible for that point- gorithms from the original VIPGI system (Scodeggio et al. ing.Attheendoftheprocessdiscrepantredshiftsresultingfrom 2005) within a complete purpose-built environment. Within it, thetworeviewersarediscussedandreconciled. A108,page8of21 L.Guzzoetal.:TheVIMOSPublicExtragalacticRedshiftSurvey(VIPERS) Thequalityofthemeasuredredshiftsisquantifiedatthetime confidence in a highly uncertain (flag = 1) spectroscopic red- ofvalidationthroughasimilargradingschemetothatdescribed shift,wouldbeincreasedincaseitscomparisontoz promotes phot inLeFèvreetal.(2005)andLillyetal.(2009).Thecorrespond- ittoflag=1.5. ingconfidencelevelsareestimatedfromrepeatedobservations, InallVIPERSpapers,redshiftscharacterisedbyaflagrang- asexplainedinSects.5.3and5.4: ing between 2.X and 9.X are referred to as reliable (or se- cure) redshifts and are the only ones normally used in the sci- – Flag4.X:ahigh-confidence,highlysecureredshift,basedon enceanalyses.Itmightsoundriskytoconsiderobjectswithflag ahighsignal-to-noiseratio(S/N)spectrumandsupportedby 9.2 as reliable. As explained above, these correspond to a red- obviousandconsistentspectralfeatures.Thecombinedcon- shiftmeasurementbasedononesingleemissionline(normally fidence level of Flag 4 + Flag 3 measurements is estimated [OII]3727Å),whichdoesnotagreewiththegalaxyphotometric tobe>99% redshiftestimate.Toconfimthis,weinspecteddirectlythespec- – Flag 3.X: also a very secure redshift, comparable in confi- traforall1027suchcasesinthePDR-1sample.For171ofthese dencewithFlag4,supportedbyclearspectralfeaturesinthe thesingle-linespectroscopicredshiftisclosetothephotometric spectrum,butnotnecessarilywithhighS/N. one,althoughnotsatisfyingthestatisticalcriteriatobedefinedin – Flag 2.X: a fairly secure, ∼95% confidence redshift mea- agreement.Thevastmajority(∼95%)ofthesecasespresentother surement, with sufficient spectral features in support of the featuresinthespectrumthatallowustopromotetheirflagto2. measurement. For the remaining 856, there is no way the observed emission – Flag 1.X: a tentative redshift measurement, based on weak linecouldbematchedtothephotometricredshifts,ifassociated spectralfeaturesand/orcontinuumshape,forwhichthereis tooneoftheotherstandardgalaxyemissionlines. ∼50%chancethattheredshiftisactuallywrong. – Flag 0.X: no reliable spectroscopic redshift measurement waspossible. 5.3. Erroronredshiftmeasurements – Flag9.X:redshiftisbasedononlyonesingleclearspectral emission feature, usually identified (in the VIPERS range) For783galaxiesintheVIPERSPDR-1samplearepeated,reli- ableredshiftmeasurementexists.Theseareobjectslyingatthe with[OII]3727Å. borderofthequadrants,wheretwoquadrantsoverlap,andwere – Flag–10:spectrumwithclearproblemsintheobservationor therefore observed by two independent pointings. In addition, dataprocessingphases.Inmostcasesthisisafailedextrac- during the re-commissioning of VIMOS after the CCD refur- tionbyVIPGI(Scodeggioetal.2005)orabadskysubtrac- bishmentinsummer2010,afewpointingswerere-observedto tionbecausetheobjectistooclosetotheedgeoftheslit. verifytheperformanceswiththenewset-up(Hammersleyetal. Serendipitous objects appearing by chance within the slit of 2010),targetinganother1357galaxies.Intotal,thisgivesasam- the main target are identified by adding a “2” in front of the ple of 1941 galaxies with double observations. 1215 of these main flag. Following human validation, a decimal fraction “.*” yield a reliable redshift (i.e. with a flag ≥2) in both measure- is added to the main flag, reflecting the agreement of the spec- mentsandcanbeconvenientlyusedtoobtainanestimateofthe troscopicmeasurement(z ),tothephotometricredshift(z ), internalrmsvalueoftheredshifterrorofVIPERSgalaxies. spec phot estimatedfromthefive-bandCFHTLSphotometry.Photometric The bottom panel of Fig. 8 shows the distribution of the redshifts have been derived using Le Phare (Ilbert et al. 2006; differences between these double measurements. The sign of Arnouts & Ilbert 2011), a code that provides us, on top of the thesedifferencesisclearlyarbitrary.Thesehavebeencomputed best redshift solution, z , with a specific 68% confidence in- as z − z , where “1” and “2” are chronologically ordered in phot 2 1 terval[zˆ ,zˆ ]foreachgalaxy.Toquantifythelevelof terms of observation date. Once normalised to the correspond- ph−min ph−max agreementbetweenz andz wealsoconsidertheoveraller- ing redshift expansion factor 1 + z, the overall distribution of spec phot rordistributionthatcanbeconstructedbyplottingspectroscopic these measurements is very well described by a Gaussian with and photometric redshifts against each other. Early in the sur- a dispersion of σ = 200 km s−1, corresponding to a single- 2 √ vey,weadoptedavalueofσz =0.025forthestandarddeviation object1σerrorσv =σ2/ 2=141kms−1.Intermsofredshift, (68%interval)ofthephotometricredshifts,slightlysmallerthan thisyieldsastandarddeviationontheredshiftmeasurementsof thecurrent morerobustestimateusing themedianabsolute de- 0.00047(1 + z). If we restrict ourselves to the highest quality viationintheVIPERSdata(σz = 0.03,Garillietal.2014).The spectra (i.e. flags 3 and 4), we are left with 655 double mea- decimalflagisdefinedasfollows. surements;theresultingrest-frame2-objectdispersionchanges very little, decreasing to σ = 193 km s−1. This indicates that 2 – Welookfirstatwhetherz isincludedinthe95.4%(2σ) spec flags2,3and4aresubstantiallyequivalentintermsofredshift interval definedby theoverall statisticsof photometric red- precision. shifts, i.e. the interval z ± 0.05 × (1 + z ). In this phot phot case,therearetwooptions: 1)ifzˆph−min < zspec < zˆph−max,i.e.thespectroscopicredshift 5.4. Confidencelevelofqualityflags alsofallswithinthe(stricter)68%intervaloftheindividual Repeatedobservationsallowustoquantifyinanobjectiveway PDF, this is defined as “full agreement” and a value 0.5 is thestatisticalmeaningofourqualityflags,whicharebynature addedtotheoriginal(integer)flag; subjective;theyareassignedbyindividualsinalarge,geograph- 2) if not, the two measurements are defined to be only in icallydispersedteam.Remarkably,thegradingsystemturnsout “marginalagreement”,andaflag0.4isadded. tobequitestableandwell-definedasweshowhereafter. – Whenneitherofthepreviouscriteriaissatisfied,avalue0.2 isadded. Let us define two redshifts as “in agreement” when – Whennozphotestimateisavailable,avalue0.1isadded. ∆z/(1+z)<3σz (cid:39)0.0025.Wecomparetheredshiftsofdouble measurements from the VIPERS sample only, considering the The rationale behind the decimal flag is to improve the confi- flag assigned to both measurements. Flags 3 and 4 are consid- denceinpoorlymeasuredspectroscopicestimates.Forexample, eredtogether,astheyshouldnotbedifferentinpracticeinterms A108,page9of21 A&A566,A108(2014) Table1.RedshiftconfidencelevelscorrespondingtotheVIPERSqual- ityflags,estimatedfrompairsofmeasurementsofthesamegalaxy. Flagclass zconfidencelevel 3+4 99.6% 2 95.1% 1 57.5% Twoofthesearebinarymasks(i.e.areasarefullyusedorfully lost). The first mask is related to defects in the parent photo- metric sample (mostly areas masked by bright stars) and the other to the specific footprint of VIMOS and how the differ- ent pointings are tailored together to mosaic the VIPERS area. Moreover,withineachofthefourVIMOSquadrantsonlyanav- erage40%oftheavailabletargetssatisfyingtheselectioncriteria areactuallyplacedbehindaslitandobserved,definingwhatwe call the Target Sampling Rate (TSR). Finally, varying observ- ing conditions and technical issues determine a variation from quadranttoquadrantoftheactualnumberofredshiftsmeasured Fig.8. Distribution of redshift differences between two independent withrespecttothenumberoftargetedgalaxies,whatwecallthe SpectroscopicSuccessRate(SSR). measurementsofthesameobject,obtainedfromasetof1215VIPERS galaxies with quality flag ≥ 2. In the bottom panel, the darker dots Detailedknowledgeofallthesecontributionsisacrucialin- correspond to top-quality redshifts (i.e. flags 3 and 4), which show gredientforanyquantitativemeasurementofgalaxyclustering. a dispersion substantially similar to the complete sample (see text). In principle, there will also be variations of the TSR and SSR Catastrophic failures (defined as being discrepant by more than ∆z = within a single quadrant, owing to the details of the response 6.6×10−3(1+z))haveobviouslybeenexcluded.Top:distributionofthe of slit assignment to small-scale clustering, and to internal dis- correspondingdifferences∆v = c∆z/(1+z).Thebest-fittingGaussian tortionsthatmaycausetheslitstobeslightlymisplacedonthe hasadispersionofσ = 200kms−1,correspondingtoasingle-object √2 sky.Theseeffectsarehardtorepresentsimply,sincetheycannot rmserrorσ =σ / 2=141kms−1.Intermsofredshift,thistranslates v 2 beviewedpurelyasaposition-dependentprobabilityofobtain- into a standard deviation of σ = 0.00047(1+z) for a single galaxy z ing a redshift. This is because the finite size of the slits means measurement. that close pairs of galaxies cannot be sampled, and there will alwaysbesomecomplexstructureinthestatisticsofpairsepa- ofstrictredshiftreliability.Wethereforeconsiderpairsofmea- rations owing to the survey selection. Once the main quadrant- surements,inthefollowingcases: based corrections are made, the only practical way of dealing 1. Both measurements have flag = 3 or 4: out of 655 pairs, 5 withtheseistousetheknownstatisticsofangularclusteringin havediscrepantredshift. the initial photometric catalogue in order to make a final small 2. Onemeasurementhasflag=2andtheother3/4:Inthiscase correctiontotheestimatedclusteringstatistics(delaTorreetal. we assume the measurement with flag 3/4 to be the correct 2013). one.Wehave10flag=2redshiftsthatarediscrepant,outof 345. 3. Both measurements have flag= 2: 22 out of 148 pairs have 6.1. RevisedCFHTLSphotometricmask discrepantredshift. The photometric quality across the CFHTLS images is tracked 4. One measurement has flag= 1 and the other has 2, 3 or 4: with a set of masks accounting for imaging artefacts and non- 121outof301arediscrepant. uniformcoverage.Weusethemaskstoexcluderegionsfromthe 5. Both measurements have flag= 1: 56 out of 74 are surveyareawithcorruptedsourceextractionordegradedphoto- discrepant. metric quality. The masks consist primarily of patches around With the reasonable assumption that when two redshifts are in bright stars (B < 17.5) owing to the broad diffraction pat- Vega agreementtheyarebothcorrect,usingthesedatawecanderivea tern and internal reflections in the telescope optics. At the core confidenceleveloftheredshiftmeasurementsforeachflagclass, of a saturated stellar halo there are no reliable detections, leav- which we report in Table 1. A final comment should be added ingaholeinthesourcecatalogue,whileinthehaloanddiffrac- concerningFlag9objects,i.e.thoseredshiftsbasedonasingle tionspikesspurioussourcesmayappearinthecataloguedueto emissionline(tipicallyinterpretedas[OII]λ3727),inparticular falsedetections.Wealsoaddtothemaskextendedextragalactic whentheydisagreewiththephotometricredshift(Flag9.2).We sourcesthatmaybefragmentedintomultipledetectionsorthat donothavesufficientstatisticsforthisclassinthesamplewith mayobscurepotentialVIPERSsources.Themasksarestoredin repeatedobservations.Theirreliabilityisdiscussedinmorede- DS9regionfileformatusingthepolygondatastructure. tailinthePDR-1datareleasepaper(Garillietal.2014). TerapixincludedabrightstarmaskaspartoftheT0006data releaseconsistingofstar-shapedpolygonscentredonthestellar halos. We found this mask to be too restrictive for VIPERS; in 6. Surveyselectionfunction particular,wefoundthatthearealostwasexcessiveneardiffrac- TheVIPERSangularselectionfunctionistheresultofthecom- tion spikes and within stellar halos. We follow the same strat- bination of several different angular completeness functions. egyinconstructingtheVIPERSmask,butinsteaduseacircular A108,page10of21