the detailed science case for the maunakea spectroscopic

TheDetailedScienceCasefortheMaunakeaSpectroscopicExplorer:

TheCompositionandDynamicsoftheFaintUniverse

01.01.00.003.DSNVersion:A

Status:Exposuredraft

2016-05-27

PreparedBy: Name(s)andSignature(s) Organization DateAlanMcConnachieMSEProjectScientist

MSEProjectOffice

2016-05-27

ApprovedBy: NameandSignature Organization DateRickMurowinskiMSEProjectManager

MSEProjectOffice

2016-05-27

Title: The Detailed Science Case Doc # : 01.01.00.003.DSN Date: 2016-05-27 Status: Exposure Draft Page: 2 of 210

ChangeRecordVersion Date Affected

Section(s)Reason/Initiation/Remarks

A 2016May27 All Exposuredraft(publicrelease)


TableofContents

Preface....................................................................................................................................6

Chapter1 Wide field spectroscopy for the 2020s and beyond: An overview of theMaunakeaSpectroscopicExplorer.........................................................................................11

1.1 ChapterSynopsis..........................................................................................................11

1.2 TheCompositionandDynamicsoftheFaintUniverse...................................................13

1.2.1 Theme1:TheOriginofStars,StellarSystemsandtheStellarPopulationsoftheGalaxy 14

1.2.2 Theme2:LinkingGalaxiestotheLargeScaleStructureoftheUniverse.....................16

1.2.3 Theme3:IlluminatingtheDarkUniverse...................................................................18

1.2.4 Fromsciencetorequirements:mappingthescienceflowdownforMSE....................20

1.3 MSEasauniqueanddiverseastronomicalfacility........................................................23

1.3.1 ThescientificcapabilitiesofMSE...............................................................................23

1.4 KeycapabilitiesforMSEscience...................................................................................25

1.4.1 Largeapertureandhighthroughput..........................................................................25

1.4.2 Operationatarangeofspectralresolutions..............................................................26

1.4.3 Dedicatedoperationsoveralonglifetime.................................................................26

1.5 CompetitionandsynergieswithfutureMOS................................................................27

1.5.1 4-mclassMOS...........................................................................................................28

1.5.2 8-mclassMOS...........................................................................................................29

1.6 Thefuture-scapeofMSE...............................................................................................30

1.6.1 WidefieldimagingoftheUniverse............................................................................32

1.6.2 TheeraofGaia..........................................................................................................35

1.6.3 TheSquareKilometerArray.......................................................................................37

1.6.4 Feedingthegiants.....................................................................................................40

1.7 MSEanditscapabilities:apriorityininternationalstrategicplanning..........................41

Chapter2 Theoriginofstars,stellarsystemsandthestellarpopulationsoftheGalaxy....44

2.1 Chaptersynopsis..........................................................................................................44

2.2 Thephysicsofstars:fromexoplanetarysystemstorarestellartypes...........................46

2.2.1 Exoplanetaryhostcharacterizationandtheirenvironmentsinthe2020s..................48

2.2.2 Stellarastrophysicswithtimedomainmulti-objectspectroscopy..............................54


2.2.3 SolarTwins:TheSunandSolarSysteminContext.....................................................55

2.2.4 WhiteDwarfsasProbesofGalacticFormationandEvolution....................................57

2.2.5 Fundamentalparametersforhighmassstars:Thechemicalevolutionofdiskgalaxies 59

2.3 ThechemodynamicalevolutionoftheMilkyWay.........................................................61

2.3.1 Gaiaandthenearfieldofgalaxyformation...............................................................62

2.3.2 ChemicaltaggingoftheouterGalaxy........................................................................71

2.3.3 ThemetalweaktailoftheGalaxy:firststarsandtheprogenitorsoftheMilkyWay..79

2.4 TheLocalGroupasatimemachineforgalaxyevolution...............................................82

2.4.1 Darkmatterandbaryonsatthelimitsofgalaxyformation........................................84

2.4.2 ThechemodynamicaldeconstructionofthenearestL*galaxy...................................90

Chapter3 LinkinggalaxiestothelargescalestructureoftheUniverse.............................95

3.1 Chaptersynopsis..........................................................................................................95

3.2 ExtragalacticsurveyswithMSE.....................................................................................96

3.2.1 Galaxyevolutionandthelinktolargescalestructure................................................96

3.2.2 Spectroscopicredshiftsurveys.................................................................................103

3.2.3 Bringingittogether:theroleofMSE.......................................................................106

3.3 Galaxiesacrossthehierarchy......................................................................................107

3.3.1 ThelowestmassstellarsystemsintheLocalUniverse.............................................108

3.3.2 Theevolutionoflowmassgalaxies..........................................................................111

3.3.3 Spatiallyresolvedstudiesofgalaxies.......................................................................119

3.3.4 Thegrowthofgalaxiesthroughmergers..................................................................127

3.4 Steppingthroughthehierarchy:thenearestgalaxygroupsandclusters.....................131

3.4.1 Clustersaslaboratoriesofgalaxyevolution.............................................................131

3.4.2 TheVirgoandComaClusters...................................................................................135

3.5 Galaxiesacrossredshiftspace....................................................................................138

3.5.1 Redshiftevolutionofclusteringandhalomodels....................................................138

3.5.2 ThechemicalevolutionofgalaxiesacrossCosmictime............................................145

Chapter4 IlluminatingtheDarkUniverse.......................................................................149

4.1 ChapterSynopsis........................................................................................................149

4.2 TheUniverseinabsorption.........................................................................................151

4.2.1 TheInterstellarMedium..........................................................................................151


4.2.2 TheIntergalacticMedium........................................................................................162

4.3 Timevariableextragalacticastrophysics.....................................................................173

4.3.1 Reverberationmappingofblackholesandtheinnerparsecofquasars...................173

4.3.2 SupernovaeIaandmassivetransientdatasets........................................................177

4.4 ThedarksectoroftheUniverse..................................................................................180

4.4.1 Thedynamicsofdarkmatterhalosfromsub-luminoustoclustermassscales.........182

4.4.2 Peculiarvelocitiesandthegrowthrateofstructures...............................................192

4.4.3 Probinghomogeneity,isotropyandtheback-reactionconjecture...........................196

4.4.4 Cosmicwebdynamicsandgalaxyformation............................................................197

4.4.5 Kinematiclensing....................................................................................................198

References...........................................................................................................................201


Preface

P.1 PurposeandScope

This is the MSE Detailed Science Case (DSC). It is the highest level document in the MSEdocumenthierarchy,andformsthefoundationfortheLevel0ScienceRequirementsDocument(SRD). TheMSEDSCprovidesthesciencenarrativedescribingtheprincipalenvisionedsciencegoalsofMSE and describes its impact on a broad range of science topics. Science ReferenceObservations (SROs) describe in detail specific transformational observing programs for MSEthatspantherangeofsciencedescribed in theDSC.ScienceRequirementsaredefinedas thesciencecapabilitiesrequiredtoconducttheSROs.TheSROsareincludedasAppendicestotheDSC.

P.2 AnoteonthestructureoftheDSC

The DSC is an extensive document covering a range of science topics. All chapters and allappendicescanbereadasstand-alonedocuments.Allchaptersbeginwitha1–2pagesynopsisthatsummarizesthecontentofthechapter.Thefirstchapterprovidesasummaryoftheentiredocument.A 10page summaryof theentireproject aimedat the international astronomy community ispresentedinanaccompanyingdocument,“AconciseoverviewoftheMaunakeaSpectroscopicExplorer”.

P.3 CreditandAcknowledgements

This document represents contributions from over 150 scientists across the internationalastronomical community from the past 5 years. It has been compiled and edited by theMSEProjectScientistinclosecollaborationwithmembersoftheMSEScienceTeam.ItisbasedonalargenumberofdocumentsdevelopedbymembersoftheMSEScienceTeamanditsprecursorproject,theNextGenerationCanada-France-HawaiiTelescope(ngCFHT).The MSE Project Office was established in 2014 after precursor studies led to the ngCFHTFeasibilityStudies.TheMSEScienceTeamthenundertookamulti-stageprocesstobetterdefinethe science capabilities ofMSE and develop the associated science case.White Papers weredrafted highlighting important science areas not necessarily previously considered by thengCFHTstudies.Phase1studies,underthecoordinationof three leads,werethenconducted.The purpose of these studies was to identify the most compelling science areas for furtherdevelopment in three broad areas (Stars and theMilkyWay, The LowRedshiftUniverse, TheHighRedshiftUniverse)andtopresentthesciencecasesforeach.Cooperationbetweenthesegroups was encouraged for areas of overlap. This led to the first set of proposed ScienceReference Observations. From this initial set, a shortlist of 12 SROs was selected for furtherdevelopment.TheseSROsformthebasisforthederivationoftheScienceRequirements.


Themainchaptersof theDetailedScienceCasearebasedonthecompilationofmaterial thathas been developed forMSE during thismulti-stage process, including the ngCFHT FeasibilityStudy.ThefinalsetofSROsispresentedinasAppendicestotheDSC.All original science documents on which the DSC is based are available athttp://mse.cfht.hawaii.edu/docs/sciencedocs.php

P.3.1 Scienceteamleadsandcoordinators

CarineBabusiaux(StarsandtheMilkyWay)MichaelBalogh(ThelowredshiftUniverse)SimonDriver(ThehighredshiftUniverse)PatCôté(Leadauthor,thengCFHTFeasibilityStudy2012)

P.3.2 ScienceReferenceObservationsleadsandcontributors

(FullauthorlistsforeachSROaregivenintheAppendices)LeadsCarineBabusiaux(SRO-3,MilkyWayarchaeologyandthe insituchemical taggingoftheouterGalaxy)MichaelBalogh(SRO-6,Nearbygalaxiesandtheirenvironments)HeleneCourtois(SRO-12,Dynamicsofthedarkandluminouscosmicwebduringthelastthreebillionyears)Luke Davies (SRO-11, Connecting high redshift galaxies to their local environment: 3Dtomographic mapping of the structure and composition of the IGM, and galaxies embeddedwithinit)SimonDriver(SRO-9,ThechemicalevolutionofgalaxiesandAGNoverthepast10billionyears(z<2)LauraFerrarese(SRO-7,BaryonicstructuresandthedarkmatterdistributioninVirgoandComa)SarahGallagher(SRO-10,MappingtheinnerparsecofquasarswithMSE)RodrigoIbata(SRO-4,StreamkinematicsasprobesofthedarkmattermassfunctionaroundtheMilkyWay)NicolasMartin(SRO-5,DynamicsandchemistryofLocalGroupgalaxies)AaronRobotham(SRO-8,Evolutionofgalaxies,halosandstructureover12Gyrs)KimVenn(SRO-2,Rarestellartypesandthemulti-objecttimedomain)EvaVillaver(SRO-1Exoplanets)ContributorsJo Bovy, Alessandro Boselli,Matthew Colless, Johan Comparat, Pat Côté, Kelly Denny, Pierre-Alain Duc, Sara Ellison, Richard de Grijs, Mirian Fernandez-Lorenzo, Laura Ferrarese, KenFreeman,RajaGuhathakurta,PatrickHall,AndrewHopkins,MikeHudson,AndrewJohnson,NickKaiser, Jun Koda, Iraklis Konstantopoulos, George Koshy, AndrewHopkins, Khee-Gan Lee, Adi


Nusser, Anna Pancoast, Eric Peng, Celine Peroux, Patrick Petitjean, Christophe Pichon, BiancaPoggianti,CarloSchmid,PrajvalShastri,YueShen,ChrisWillot

P.3.3 DSCEditorsandprincipalcontributorstonewmaterialintheDSC

MichaelBalogh(Lowmassgalaxies)JoBovy(CDMhalosubstructures)PatCôté(Lowmassstellarsystems)ScottCroom(IFUstudiesofgalaxyevolution)SaraEllison(Galaxymergers)LauraFerrarese(TheVirgoCluster)SarahGallagher(Reverberationmapping)RosineLallement(TheInterstellarMedium)NicolasMartin(TheLocalGroup)PatrickPetitjean(TheInterclusterMedium)CarloSchimd(TheDarkUniverse)DanSmith(SKAsynergies)MatthewWalker(Nearbydwarfgalaxiesanddarkmatter)JonWillis(Galaxyclusters)

P.3.4 Phase1studyreportco-authors

Carine Babusiaux, Michael Balogh, Alessandro Bosselli Matthew Colless, Johan Comparat,Helene Courtois, Luke Davies, Richard de Grijs, Simon Driver, Sara Ellison, Laura Ferrarese,RodrigoIbata,SarahGallagher,ArunaGoswami,MikeHudson,AndrewJohnson,MattJarvis,EricJullo,NickKaiser,Jean-PaulKneib,JunKoda,IraklisKonstantopoloulous,RosineLallement,Khee-Gan Lee,NicolasMartin, JeffNewman, Eric Peng, Celine Peroux, Patrick Petitjean, ChristophePichon, Bianca Poggianti, Johan Richard, Aaron Robotham, Carlo Schmid, Yue Shen, FirozaSutaria,EdwarTaylor,KimVenn,LudovicvanWaerbeke,ChrisWillot

P.3.5 ScienceWhitePapers

GiuseppinaBattaglia(TheaccretionhistoryoftheMilkyWayhalothroughchemicaltagging)PatCote(UpgradestotheMSEPointDesignSpecifications)ArunaGoswami(Studyofoldmetal-poorstars)RicharddeGrijs(Spatiallyresolvedstellarpopulationsandstarformationhistoriesofthelargestgalaxies)Pierre-AlainDuc(Probingthelargescalestructuresofmassivegalaxieswithdwarfs/GC/UCDsinavarietyofenvironments)PatHall,CharlingTao,YueShen,SarahGallagher(OpportunisticTransientTargeting)MishaHaywood,PaoladiMatteo(BulgesciencewithMSE)MishaHaywood,PaoladiMatteo(Galacticarcheology:TheoutskirtsoftheMilkyWaydisk)MattJarvis(SKASynergies)IraklisKonstantopoulos(TheRedshiftRejectRubric)IraklisKonstantopoulos(Dense,Low-MassGalaxyGroups:APhase-SpaceConundrum)IraklisKonstantopoulos(StarClustersasChronometersforGalaxyEvolution)


RosineLallement(GalacticInterStellarMedium:the3DerawithMSE)PatrickPetitjean(ThreeDimensionalReconstructionoftheInterGalacticMedium)CharliSakari,KimVennetal.(IntegratedLightSpectroscopyatHighResolutionwithMSE)CarloSchmidetal.;updatedOctober2014(MSE:avelocitymachineforcosmology)ArnaudSeibert(MilkyWaystructureandevolution,ISM)YueShen(Multi-ObjectAGNReverberationMappingwithMSE)SivaraniThirupathi(TracersofPre-galacticandextragalacticLithiumabundancesinMilkyWay:towardssolvingthelithiumproblem)YutingWang(Optimaltestsongravityorcosmologicalquantitiesonlargescales)YipingWang(SFHandscale-lengthevolutionofmassivediskgalaxiestowardshigh-z)

P.3.6 OtherContributionsandFeedback

(notpreviouslymentioned)FerdinandBabas,SteveBauman,ElisabettaCaffau,MaryBethLaychak,DavidCrampton,DanielDevost, Nicolas Flagey, Zhanwen Han, Clare Higgs, Vanessa Hill, Kevin Ho, Sidik Isani, ShanMignot,RickMurowinski,GajendraPandey,DerrickSalmon,ArnaudSiebert,DougSimons,ElseStarkenburg,KeiSzeto,BrentTully,TomVermeulen,KanoaWithington

P.3.7 MSEScienceTeammembership

NobuoArimotoMartinAsplundHerveAusselCarineBabusiauxMichaelBaloghMicheleBannisterGiuseppinaBattagliaHarishBhattSSBhargaviJohnBlakesleeJossBland-HawthornAlessandroBoselliJoBovyJamesBullockDenisBurgarellaElisabettaCaffauTzu-ChingChangAndrewColeJohanComparatJeffCookeAndrewCooperPatCoteHeleneCourtoisScottCroom

RicharddeGrijsPaolaDiMatteoSimonDriverPierre-AlainDucSaraEllisonGinevraFavoleLauraFerrareseHectorFloresKenFreemanBryanGaenslerSarahGallagherPeterGarnavichKarolineGilbertRosaGonzalez-DelgadoArunaGoswamiPuragraGuhathakurtaPatHallGuentherHasingerMishaHaywoodFalkHerwigVanessaHillAndrewHopkinsMikeHudsonNaraeHwang

RodrigoIbataPascaleJablonkaMatthewJarvisUmanathKamathLisaKewleyIraklisKonstantopoulosGeorgeKoshyRosineLallementDamienLeBorgneKhee-GanLeeGeraintLewisRobertLuptonSarahMartellNicolasMartinMarioMateoOlgaMenaDavidNatafJeffreyNewmanGajendraPandeyEricPengEnriquePérezCelinePerouxPatrickPetitjeanBiancaPoggianti


FranciscoPradaMathieuPuechAlejandraRecio-BlancoAnnieRobinAaronRobothamWillSaundersCarloSchimdArnaudSeibertPrajvalShastriYueShen

DanielSmithC.S.StalinElseStarkenburgFirozaSutariaCharlingTaoKarunThanjuvurSivaraniThirupathiLaurenceTresseBrentTullyLudovanWaerbeke

KimVennEvaVillaverMatthewWalkerJian-MinWangYipingWangYutingWangJonWillisDavidYongGongboZhao

P.3.8 ngCFHTScienceFeasibilityStudymembership

(notpreviouslymentioned)

PatrickBoisse,JamesBolton,PiercarloBonifacio,FrancoisBouchy,LenCowie,DavidCrampton,KatiaCunha,MagaliDeleuil,ErnstdeMooij,PatrickDufour,SebastienFoucaud,KarlGlazebrook,John Hutchings, Jean-Paul Kneib, Chiaki Kobayashi, Rolf-Peter Kudritzki, Damien Le Borgne,Yang-Shyang Li, Lihwai Lin, Yen-Ting Lin, Martin Makler, Norio Narita, Changbom Park, RyanRansom,SwaraRavindranath,BachamEswarReddy,MarcinSawicki,LucSimard,RaghunathanSrianand, Thaisa Storchi-Bergmann, Keiichi Umetsu, Ting-Gui Wang, Jong-Hak Woo, Xue-BingWu


Chapter1 Wide field spectroscopy for the 2020s and beyond: An overview of theMaunakeaSpectroscopicExplorer

1.1 ChapterSynopsis

• Afullydedicated,11.25maperturewidefieldhighlymultiplexedspectroscopicfacility• Thedefinitivedeconstructionofallcomponentsandsub-componentsoftheGalaxythrough

insituchemo-dynamicalanalysesofmillionsofstarsovertheentireluminosityrangeofGaiatargets

• Linkinggalaxiestotheirsurroundinglargescalestructurethroughemissionandabsorptionlinestudies

• Measuringthedynamicsofdarkmatteracrossallspatialscales• AnessentialfeederfacilityfortheVeryLargeOpticalTelescopes(VLOTs)• The“go-to”follow-upfacilityforthenewgenerationofmulti-wavelengthimagingsurveys

The Maunakea Spectroscopic Explorer (MSE) has a unique and critical role in the emergingnetworkof international astronomical facilities.As the largest groundbasedoptical andnear-infrared telescope aside from the TMT, the E-ELT and GMT (we refer to these facilitiescollectively as the Very Large Optical Telescopes, VLOTs), its wide field will be dedicated tohighlymultiplexed spectroscopy at a broad range of spectral resolutions.MSE is designed toenabletransformationalscience inareasasdiverseastomographicmappingofthe interstellarand intergalactic media; the in-situ chemical tagging of thick disk and halo stars; connectinggalaxies to their large scale structure;measuring themass functions of cold darkmatter sub-halos in galaxy and cluster-scale hosts; reverberationmapping of supermassive black holes inquasars; next generation cosmological surveys using redshift space distortions and peculiarvelocities. MSE is an essential follow-up facility to current and next generations of multi-wavelength imaging surveys, including LSST,Gaia, Euclid,WFIRST, PLATO, and the SKA, and isdesigned to complement and go beyond the science goals of other planned spectroscopiccapabilities like VISTA/4MOST,WHT/WEAVE, AAT/HERMES and Subaru/PFS. It is an essentialfeeder facility for the VLOTs, namely E-ELT, TMT and GMT, and provides the missing linkbetweenwidefieldimagingandsmallfieldprecisionastronomy.

MSEisan11.25mapertureobservatorywitha1.5squaredegreefieldofviewthatwillbefullydedicatedtomulti-objectspectroscopy.3200ormorefiberswill feedspectrographsoperatingatlow(R ~2000–3500)andmoderate(R~ 6000)spectralresolution,andapproximately1000fiberswill feed spectrographsoperating at high (R ~40000) resolution.At low resolution, theentire optical window from 360nm – 950nm and the near infrared J and H bands will beaccessible, and at moderate and high resolutions windows with the optical range will beaccessible.Theentiresystemisoptimizedforhighthroughput,highsignal-to-noiseobservationsofthefaintestsourcesintheUniversewithhighqualitycalibrationandstabilitybeingensuredthroughthededicatedoperationalmodeoftheobservatory.ThediscoveryefficiencyofMSEisan order ofmagnitude higher than any other spectroscopic capability currently realized or indevelopment.Figure 1 isa renderingofa cut-awayofMSEshowing thesystemarchitectureandmajorsubsystems.


Figure1:Cut-awayofMSErevealingthesystemarchitectureandmajorsubsystems.

MSE is a rebirth of the 3.6m Canada-France-Hawaii Telescope as a dedicated, 11.25m,spectroscopic facility within an expanded partnership extending beyond the originalmembership. The Mauna Kea Comprehensive Management Plan anticipates CFHTredevelopment.MSEwill reuse thecurrentbuildingandpier (minimizingworkat thesummit)and the ground footprint will remain the same. MSE will benefit from CFHT's 40 years ofexperience onMaunakea and a support staff deeply rooted in the Hawaii Island community.CFHT/MSE is actively engaging theHawaii community throughout theplanningprocess and iscommitted tobalancing cultural andenvironmental interests in thedesignof theobservatoryand theorganizationof thenewpartnership.With their support,MSEwill begin constructiononceithaspassedtheConstructionProposalReviewandtheMSEpartnershipapprovesfunding,


andonceallnecessarypermittingandlegalrequirementshavebeenfulfilled.This includestheapprovalofanewmasterleasefortheMaunaKeaScienceReserve.

MSE isa response to the internationally recognizedneed forhighlymultiplexed,deep,opticalandnear-infrared spectroscopy. Thediverse andexciting scientific opportunities enabledby awide-field, large-aperture, spectroscopic survey telescope have long been recognized by theinternationalastronomicalcommunity(theinitialcallsforsuchafacilityappearedmorethan15years ago). The suite of powerful photometric imaging and astrometric telescopes that havenowstartedoperations,orareduetobeginimminently,havesharpenedtheneedforwide-fieldspectroscopic follow up and have resulted in such a facility becoming the most glaring andimportant “missing capability” in the portfolio of international astronomical projects. Variousnationalastronomicalstrategicplansnowplacethedevelopmentofafacilitythatmeetstheseneedsathighpriority.

In summary, MSE is the realization of the long-held ambition for wide field multi-objectspectroscopy at optical and near-infraredwavelengths. It is the result of diverse needs acrossdiversefieldsofastronomyallconvergingonasinglecapability.MSEoccupiesacriticalnexusinthe international network of astronomical facilities in the 2020s, and will enable stand-alonetransformative science from our understanding and characterization of exoplanet hosts, tochemicaltaggingoftheMilkyWay,tothephysicsofsupermassiveblackholesandthedriversofgalaxyformationatcosmicnoon.Thisdocumentpresentsthecontextandthesciencethathasdriven the development and detailed design of MSE, and illustrates the very broad range ofastrophysicsthatwillbeimpactedbythisuniqueandessentialObservatory.

1.2 TheCompositionandDynamicsoftheFaintUniverse

MSE will unveil the composition and dynamics of the faint Universe through chemicalabundancestudiesofstarsintheouterreachesoftheGalaxy,throughstellarpopulationstudiesofdistant and lowmassgalaxies, and throughdynamical studiesacrossall spatial scales fromdwarf galaxies to the largest scale structures in the Universe. MSE is designed to excel atprecision studies of faint astrophysical phenomena that are beyond the reach of 4-m classspectroscopicinstruments.Abroadrangeofspectralresolutions–includingR=2000,R=6000andR=40000–ensurethatthespecializedtechnicalcapabilitiesofMSEenableadiverserangeof transformational astrophysics from stellar through to cosmic scales. Specifically, theseinclude:

• (Chapter2)TheOriginofStars,StellarSystemsandtheStellarPopulationsoftheGalaxy• (Chapter3)LinkingGalaxiestotheLargeScaleStructureoftheUniverse• (Chapter4)IlluminatingtheDarkUniverse

Each chapter of the Detailed Science Case can be read as a stand-alone section, and eachchapter starts with a 1 – 2 page synopsis of the content of that chapter. Reference ismadethroughout to a suite a “Science Reference Observations” (SROs), that are included asAppendices,andwhichcanalsobereadasstand-alonesections.TheSROsspanthefullrangeofsciencethatMSEisexpectedtoundertake,andhighlightarangeofpossibleoperationalmodes.Importantly, they are compelling, high impact, transformational science programs that are


uniquely possible with MSE, and it is from these that the science requirements – i.e., thefundamentaldesignofthefacility–arederived.Theyincludethefirstlargescalereverberationmappingexperimenttoprobetheinnerparsecofthousandsofquasars;ameasurementofthecold darkmatter sub-halomass function of the Galaxy through comprehensive and precisionradialvelocitymappingoftidalstreams;derivationofthehalooccupationfunctionthroughtheconstructionofeightphoto-zselectedsurveycubesthatprobeacosmologicallyrelevantvolumeto low stellar mass and which cover the peak of the star formation history of the Universe;tomographic reconstruction of the interstellar medium in the Galaxy and the intergalacticmediumtohighredshift;chemicaltaggingofstars inthehaloandthickdiskoftheMilkyWaythroughin-situanalysesofstarsthatcovertheentiremagnituderangeofGaiatargets.The reader is referred to the relevant chapters or appendices for detailed discussion of thediversesciencecaseforMSE.Here,webrieflyhighlightsomeofthekeysciencethemesineacharea.

1.2.1 Theme1:TheOriginofStars,StellarSystemsandtheStellarPopulationsoftheGalaxy

Science Reference Observations for The Origins of Stars, Stellar systems and the StellarPopulationsoftheGalaxy• SRO-01:Exoplanetsandstellarvariability• SRO-02:Thephysicsofrarestellartypes• SRO-03:TheformationandchemicalevolutionoftheGalaxy• SRO-04:Unveilingcolddarkmattersubstructurewithprecisionstellarkinematics• SRO-05:ThechemodynamicaldeconstructionofLocalGroupgalaxies

MSE will produce vast spectroscopic datasets for in situ stars across all components of theGalaxyatallgalactocentricradii,acrossthefullmetallicitydistributionoftheGalaxy,andacrossthe full luminosity rangeofGaia targets. ExplorationofourGalaxyprovidesuswith themostdetailedperspectivepossibleontheformationofplanetarysystems,theirhoststars,andtheirparent stellar populations. At the largest scale, the properties of the stellar populations –including their spatial distributions, energy, angular momenta, and chemical properties –provide uswith the fossil information necessary to decompose theMilkyWay galaxy into itsprimordialbuildingblocks, and in sodoing reveal thehistoryof its formationand the sitesofchemicalnucleosynthesis.

The key measurables are well-defined spectral line features for accurate stellar parameters(temperature,gravity,metallicity,rotationalvelocities,magneticfieldsstrengths,etc.),chemicalabundances,andradialvelocities.Suchadetailedperspectivenecessitatesthehighestsignal-to-noisedataobservedathighspectralresolution:atthehighestresolutionofMSE(R=40000),syntheticspectrashowthatlinesaredeblendedat~0.1Å,rotationalandradialvelocitiescanbemeasured to better than~0.1 kms-1 through cross-correlation of many lines (>100), andchemical abundances can bemeasuredwith Δ[X/Fe] <0.1 dex, when the signal-to-noise ratio(SNR) > 30. Here, the capabilities of MSE are unmatched amongst any existing or plannedfacilities. Further, an increasing fraction of astronomical resources are being devoted to


characterizing the transient,moving and variable universe.MSE, thanks to its large aperture,wide field and high multiplexing, is well positioned to pioneer multi-object time domainoptical/IRspectroscopicstudiesoffaintstellarsources.

Specifically,MSEwillprovidecompletecharacterizationathighresolutionandhighSNRofthefaintendofthePLATOtargetdistribution(g~16)visiblefromthenorthernhemispheretoallowfor statistical analysis of the properties of planet-hosting stars as a function of stellar andchemical parameters. With high velocity accuracy and stability, time domain spectroscopicprograms will allow for highly complete, statistical studies of the prevalence of stellarmultiplicityintotheregimeofhotJupitersforthisandothersamplesandalsodirectlymeasurebinary fractions away from theenvironmentof the SolarNeighbourhood.Rare stellar types–including amenagerie of pulsating stars, but also halowhite dwarfs and solar twins –will beeasilyidentifiedwithinthedatasetofmillionsofstarsobservedperyearbyMSE,andwillallowforpopulationstudies that incorporateaccuratecompletenesscorrectionsandwhichexaminetheirdistributionwithintheGalaxy.

Figure 2: Main panel shows the relative flux of a synthetic spectrum of a metal poor red giant star at theintermediateMSEspectralresolutionofR~6000,alongwithsomeofthestronglinestellardiagnosticsaccessibleatthisresolution.HighlightedregionsshowthenormalizedfluxinthreewindowsobservablewiththehighresolutionmodeofMSE.AmagnifiedregionoftheUVwindowshowsexamplesofthespeciesthatwillbeidentifiedathighresolution.MSEchemicaltaggingsurveyswill identifyspeciessamplingalargeanddiversesetofnucleosyntheticpathwaysandprocesses.SeeChapter2formoredetails.

MSEwillhaveunmatchedcapabilityforchemicaltaggingexperiments.Recentworkinthisfieldhas started to reveal the dimensionality of chemical space and has shown the potential forchemistrytobeusedinadditionto,orinsteadof,phasespace,torevealthestellarassociationsthat represent the remnants of the building blocks of the Galaxy. In particular, MSE willdecompose the outer regions of the Galaxy – where dynamical times are long and whosechemistryisinaccessiblefrom4-mclassfacilities–intoitsconstituentstarformationeventsbyaccessingarangeofchemicaltracersthatsamplealargenumberofnucleosyntheticpathways.In so doing, MSE will provide a high resolution, homogenous sampling of the metallicitydistributionfunctionoftheGalaxy.Itwilltracevariationinthemetallicitydistributionacrossall


componentsandasa functionofradius.Atthemetal-weakend, [Fe/H]<-3,suchananalysisrevealsthedistributionoftheoldeststars intheUniverse,and isavitalcomponenttotracingthebuild-upoftheGalaxyattheearliesttimes.

MSEwillextendanalysisoftheresolvedstellarpopulationsoftheGalaxytotheLocalGroupandbeyond.VariablestarssuchasCepheidscanbespectroscopicallystudiedthroughouttheLocalGrouputilizing the time-domainstrengthsofMSE,andmassivestarswillbeused to trace themetallicity gradients of disks in galaxies in the next nearest galaxy groups. Such studies willprobetheinside-outformationofgalaxydisksthroughindividualstellarmetallicitiesobservedatmoderatespectralresolution.M31andM33areparticularlyprominentextragalactictargetsforresolved stellar population analysis, and MSE will provide the definitive chemodynamicaldeconstruction of our nearest L* galaxy companion by a focused program targeting all giantbranch candidates observed inM31 out to approximately half the virial radius of the galaxy.SuchaspectroscopicpanoramaisunprecedentedandanaturalcomplementtothecoreMilkyWaysciencethatformsthedriversforMSEinthisfield.

MSEisuniquelysuitedtoprecisionstudiesofthestellarpopulationsoftheMilkyWay,andwillprovideatransformativeimpactinourunderstandingofthelinksbetweenstellarandplanetaryformation,thesitesofchemicalnucleosynthesis,stellarvariability,andthechemicalstructureoftheGalaxythroughinsituanalysisofstarsinallGalacticcomponents.ItwillpushthesestudiesouttothenearestgalaxiestoprovidepanoramiccomplementstothedrivingMilkyWayscience.MSEwill leveragethedatasetcreatedbyPLATOforexoplanetaryandstellarastrophysics,andwill provide theessential complement to theGaia space satellitebyprovidinghigh resolutionchemicalabundancesacrosstheentireluminosityrangeofGaiatargets.

1.2.2 Theme2:LinkingGalaxiestotheLargeScaleStructureoftheUniverse

ScienceReferenceObservationsforLinkingGalaxiestotheLargeScaleStructureoftheUniverse• SRO-6:GalaxiesandtheirenvironmentsinthenearbyUniverse• SRO-7:Thebaryoniccontentanddarkmatterdistributionofthenearestmassiveclusters• SRO-8:Multi-scaleclusteringandthehalooccupationfunction• SRO-9:ThechemicalevolutionofgalaxiesandAGN

MSE will provide a breakthrough in extragalactic astronomy by linking the formation andevolutionofgalaxies to thesurrounding largescalestructure,across the full rangeof relevantspatialscales,fromkiloparsecstomegaparsecs.WithintheΛCDMparadigm,itisfundamentaltounderstandhowgalaxiesevolveandgrowrelativetothedarkmatterstructureinwhichtheyareembedded.ThisnecessitatesmappingthedistributionofstellarpopulationsandsupermassiveblackholestothedarkmatterhaloesandfilamentarystructurethatdominatethemassdensityoftheUniverse,andtodosooverallmassandspatialscales.

MSE is optimized for studying the evolution of galaxies, AGN and the environmental factorslikely to influence their evolution.MSEwill be the premier facility for unscrambling the non-linearregime(smallscaleclustering,mergers,groups,tendrilsandfilaments),thebaryonregime(i.e.,metallicity/chemicalevolution),andtheevolutionand interplayofgalaxies,AGN,andtheIGM, incorporating environmental factors and the key energy production pathways.MSEwill


builduponthe impressive legacyofcurrentspectroscopicsurveys inthenearandfarfieldsbygoing several magnitudes deeper, covering an order of magnitude larger sky area, andconducting both stellar population studies of the nearest galaxies and precision dynamicalanalysesacrosstheluminosityfunctionandacrosscosmictime.

InthelocalUniverse,MSEwillmeasurethedynamicsandstellarpopulationsofeverybaryonicstructureintheVirgoandComaclustersdowntoextremelylowstellarmass.Thiswillallowthephysical association of dwarf galaxies, nucleated dwarfs, globular clusters and ultra-compactdwarfstobeprobedthroughchemodynamicalanalysisofunprecedenteddatasets(bothinsizeandcompleteness).Quenchingprocessesoperatingintheclusterenvironmentcanbeexaminedforall systemsatall radii, and inparticularanalysisof the stellarpopulationsof thousandsoflow mass galaxies all in the same cluster will give direct access to the relative roles ofenvironmentally-drivenandinternally-drivenprocessesacrosstheluminosityfunctionandtoitsfaintestlimits.

MSEwillbeabletoconductthedefinitivespectroscopicsurveyofthelowredshiftUniversefordecades to come by obtaining spectra for every z < 0.2 galaxy in a cosmologically-relevantvolume spanning 3200 deg2, allowing every halo with mass Mhalo>1012M¤

and four or moregalaxies to be detected. Follow up in select regionswill result in a sample covering 100 deg2spanning four decades in halo mass, accounting for the majority of stellar mass in the low-redshiftUniverse.ThiswillallowthemeasurementoftheshapeofthestellarmassfunctiontothescaleoftheLocalGroupdwarfgalaxies,overalargevolumeandarangeofenvironments.

Figure 3: Left panel: Cone plot showing an exampleMSE survey compared to other notable benchmark galaxysurveys.ConesaretruncatedattheredshiftatwhichL*galaxiesarenolongervisible.Rightpanels:Lookbacktimeversuscosmicstarformationrate(rightaxis)usingtheparameterizationofHopkins&Beacom(2006).Grey linesindicatethewavelengthofOII3727Å.Alsoindicatedonthiswavelengthscaleisthewavelengthrangeofallhighlymultiplexedspectroscopicfacilities indevelopment.Theseareoffsetverticallyaccordingtotheirsystemetendue(left axis). The light conedemonstrates the reachofMSE for extragalactic surveysusing ahomogeneous setoftracersatallredshifts.SeeChapter3formoredetails.

Uniquely, thebroadwavelength rangeandextreme sensitivityofMSEallowsextrapolationofthis “low redshift” survey concept out to beyond cosmic noon. An ambitious set of photo-zselected survey cubeswill allowMSE tomeasure the build up of large scale structure, stellar


mass, halo occupation and star formation out to redshiftsmuch greater than 2. By targeting(300 Mpc/h)3

boxes, each cube will measure “Universal” values for an array of potential

experiments. These survey volumeswill trace the transition frommerger-dominated spheroidformationtothegrowthofdisks,coveringthepeakinstar-formationandmergeractivity.Thiscombination of depth, area and photo-z selection is not possible without a combination offuture imagingsurveysandMSE.Assuch,MSEwillbeabletoproducethedefinitivesurveyofstructure,halosandgalaxyevolutionover12billionyears.

MSEwillbeunparalleledinitsabilitytoconnectthepropertiesofgalaxiestotheirsurroundinglarge scale structure. Itwill link the properties and diversity of galaxies observed in the localuniversetotheevolutionaryprocessesthathaveoccurredacrosscosmictime.Insodoing,MSEwill fulfill threedistinctroles inthe2025+eraforextragalacticastrophysics:dedicatedscienceprograms,coordinatedsurveyprogramsandfeederprogramsfortheVLOTs.Allthreerolesarecritical, transformational, and will lead to major advancements in the area of structureformation, galaxy evolution, AGN physics, our understanding of the IGM, and the underlyingdarkmatterdistribution.

1.2.3 Theme3:IlluminatingtheDarkUniverse

ScienceReferenceObservationsforIlluminatingtheDarkUniverse• (SRO-03:TheformationandchemicalevolutionoftheGalaxy)• (SRO-04:Unveilingcolddarkmattersubstructurewithprecisionstellarkinematics)• SRO-06:Thebaryoniccontentanddarkmatterdistributionofthenearestmassiveclusters• SRO-10:Mappingtheinnerparsecofquasarsthroughreverberationmapping• SRO-11:LinkinggalaxyevolutionwiththeIGMthroughtomographicmapping• SRO-12:Apeculiarvelocitysurveyoutto1GpcandthenatureoftheCMBdipole

MSEwill cast lighton someof thedarkest componentsof theUniverse. Thebackboneof thelarge scale structures in which galaxies are embedded is dark matter, that is invisible inelectromagnetic radiation. Embedded alongside galaxies in these large scale structures is theintergalactic medium, a major reservoir of baryons in the Universe but one which is bestobservedintheopticalwavelengthrangethroughitsabsorptioneffectsonbackgroundgalaxies.Within galaxies, the interstellarmedium is similarly probed through absorption studies in theoptical, and the central supermassive black holes are generally only probed through indirecttechniques.MSEwillprovideunprecedentedperspectivesoneachofthesecomponentsofthedarkUniverse,andprovideasuiteofcapabilitiesideallysuitedtoprovidingcriticaldatafornextgenerationcosmologicalsurveys.

Specifically,MSEwillprovidetomographicreconstructionofthethreedimensionalstructuresofthe Galactic interstellar medium and the intergalactic medium through high signal-to-noiseabsorption studies alongmillions of sight lines. These programswill be orders ofmagnitudeslarger than the currently state-of-the-art.Whether focused on stellar or galactic science, therelationsbetweenstarsandgalaxiesandtheirsurroundingmediaisakeyissueinunderstandingthecyclingofbaryons. In theextragalactic regime,MSEcan target the IGMat z~2–2.5, toreconstruct the dark matter distribution and associated galaxies at high-z, allow detailedmodeling of galactic scale outflows via emission lines, investigate the complex interplay


between metals in galaxies and the IGM, and provide the first comprehensive, moderateresolutionanalysisofalargesampleofgalaxiesatthisepoch.

Figure4:Recoveryofline-of-sightvelocitydispersion(left)anddarkmatterdensity(right)profilesasafunctionofstellar spectroscopic sample size for dwarf galaxies, demonstrating the need for extensive datasets with MSE.Shaded regions represent 95% credible intervals froma standard analysis ofmock data sets consisting of radialvelocities for N = 102, 103 and 104 stars (median velocity errors of 2 kms−1), generated from an equilibriumdynamicalmodelforwhichtrueprofilesareknown(thickblacklines).SeeChapter4fordetails.

MSEwill allow for anunprecedented, extragalactic timedomainprogram tomeasuredirectlythe accretion rates and masses of a large sample of supermassive black holes throughreverberationmapping. This information is essential for understanding accretion physics andtracing black hole growth over cosmic time. Reverberation mapping is the only distance-independentmethodofmeasuringblackholemassesapplicableatcosmologicaldistances,andcurrentlyonly~60local,relatively low-luminosityAGNhavemeasurementsoftheirblackholemasses based on this technique. A ground-breaking MSE campaign of~100 observations of~5000 quasars over a period of several years (totaling~600 hours on-sky) tomap the innerparsecofthesequasarsfromtheinnermostbroad-lineregiontothedust-sublimationradiuswillmap the structure and kinematicsof the innerparsec arounda large sampleof supermassiveblack holes actively accreting during the peak quasar era and provide a quantum leap incomparisontocurrentcapabilities.Further,theresultingwell-calibratedreverberationrelationforquasarsofferspromiseforconstructingahigh-zHubblediagramtoconstraintheexpansionhistoryoftheUniverse.

MSE is theultimate facility forprobing thedynamicsofdarkmatterovereveryandall spatialscalesfromthesmallestdwarfgalaxiestothelargestclusters.Fordwarfgalaxiesinthevicinityof theMilkyWay,MSEwillobtaincompletesamplesof tensof thousandsofmemberstars toverylargeradiusandwithmultipleepochstoremovebinarystars.Suchanalyseswillallowtheinternaldarkmatterprofiletobederivedwithhighaccuracyandwillprobetheoutskirtsofthedark matter halos accounting for external tidal perturbations as the dwarf halos orbit theGalaxy.Anewregimeofprecisionwillbereachedthroughthecombinationoflargescaleradialvelocity surveys usingMSEwithprecision astrometric studies of the central regionsusing the


VLOTs. In theGalactic halo, large scale radial velocitymapping of every known stellar streamaccessiblefromMaunakeawithhighvelocityprecisionwillrevealtheextenttowhichthesecoldstructures have been heated through interactions with dark sub-halos and for the first timeplace strong limitson themass functionof all subhalos (luminousandnon-luminous) in an L*galaxy. On cluster scales, MSE will use galaxies and globular clusters as dynamical tracers inexactly the sameway as stars in dwarf galaxies, to provide a fully consistent portrait of darkmatterhalosacrossthehalomassfunction.

Finally,MSEprovidesaplatformforenablingnextgenerationcosmologicalsurveys,particularlythroughlargescalesurveysoftheinternaldynamicsofgalaxiesaspartoffuturemulti-objectIFUdevelopments. Such surveys allow for dynamical tests probing the scale of homogeneity, cantestthebackreactionconjectureinGeneralRelativity,andprovidesanovelmeansofmeasuringgravitational lensing via thevelocity field that complements standardmeasurementsbybeingsensitivetoverydifferentsystematics.

MSEwill transformthestudyof someof thedarkestcomponentsof theUniverse.Absorptionline studies, reverberation mapping programs, and precision velocity studies of stars andgalaxiesastracerparticlesutilizeaspectsoftheperformanceofMSE–sensitivity,stabilityandspecialization–thatareuniquetothefacilityandwhichconsequentlyallowuniqueinsightsintokeyaspectsof thebaryoncycle,blackholephysics,andcosmology.MSErepresentsasuiteofcapabilitiesthatisanessentialplatformforthedevelopmentofcosmologicalprogramsthatgobeyondthecurrentgenerationandwhicharerelevantonthe2025+timescale.

1.2.4 Fromsciencetorequirements:mappingthescienceflowdownforMSE

Chapters2–4discuss indetail theanticipated scienceobjectivesofMSE,with the importantcaveat that it is impossible to predict how astrophysics will evolve over the next decade inresponsetodiscoveriesandrealizationsthathavenotyetbeenmade.Nevertheless, it isclearthatdiversefieldsofresearchhaveconvergedinrequiringdedicatedlargeaperturemulti-objectspectroscopy,and it isclear that thecapabilitiesofMSEwillbeacriticalcomponentof futurelinesofastronomicalenquiry.Reference is made throughout this document to a suite of Science Reference Observationspresented in the Appendices. These have been defined by the international science team asspecific,detailed,scienceprogramsforMSEthataretransformativeintheirfieldsandwhichareuniquelypossiblewithMSE.This lastpoint isparticularly importantgiventhe largenumberofspectroscopicresourcesthatarebecomingavailableandwhicharediscussed inmoredetail inSection1.5.TheSROshavebeenselectedtospantherangeofanticipatedfieldsinwhichMSEisexpected to contribute, and they have been developed in considerable detail. The ScienceRequirements forMSE – i.e., the highest level design requirements for the facility – are thendefinedasthesuiteofcapabilitiesnecessaryforMSEtocarryouttheseobservations.TheMSEarchitectureandtechnicalrequirementsflowdirectlyfrom,anddeliverthecapabilitydescribedwithin,theScienceRequirements.Figure5showsthesciencedevelopmentprocedureforMSE,anditsrelationtothekeydocumentsthatdefinethedesignofMSE.


Figure5:Overviewof the sciencedevelopmentprocess todate forMSE, including the relationships to themaindocumentsthatdefinethedesignofMSE.

Weexpect thatmanyof the SROsdiscussed in this documentwill beprecursors toobservingprograms thatMSEwill carryout.However, it isnotnecessarily the case that this is so, sincecomefirstlightofthefacility,anewsetofforefrontsciencetopicsmighthaveemergedthattheusersofMSEwillwanttoaddress.Nevertheless,thesuiteofcapabilitiesthatMSEwillhave,inresponsetotheSROsasdescribedintheAppendices,willensurethatMSEisatthecuttingedgeofastronomicalcapabilitiesin2025andbeyond.Table 1 cross references each SRO to each of the high levelMSE science requirements. Thescience requirements are described and discussed in detail in theMSE Science RequirementsDocument.

Science Requirements Document

Science Reference Observations

ngCFHT Feasibility StudyDetailed Science Case

White Papers

Phase 1 SRO studies

Operations Concept

Document

Observatory Requirements

Document

Observatory Architecture Document

MSE Science Development Phase

Level 0

Level 1

MSE Foundational Documents


Table 1: Science Reference Observations cross referenced to Science Requirements. Dark green boxes indicateimport impacts of the SRO on the requirement; light green indicates moderate impact of the SRO on therequirement;greyindicatesminimalornoimpactoftheSROontherequirement.

DSC-SRO-01

DSC-SRO-02

DSC-SRO-03

DSC-SRO-04

DSC-SRO-05

DSC-SRO-06

DSC-SRO-07

DSC-SRO-08

DSC-SRO-09

DSC-SRO-10

DSC-SRO-11

DSC-SRO-12

Exoplanethosts

Timedom

ainstellarastrophysics

Chemicaltagging

intheouterGalaxy

CDMsubhalosand

stellarstreams

LocalGroupgalaxies

Nearbygalaxies

VirgoandComa

Halooccupation

GalaxiesandAGN

TheInterGalacticMedium

Reverberationmapping

Peculiarvelocities

REQ-SRD-011 Lowspectralresolution

R~2000(whitedwarfs)

R~3000 R~3000 R~2000-3000 R~3000 R~3000 R~1000-

2000

REQ-SRD-012

Intermediatespectralresolution

Anyrepeatobservatio-ns

Essential,R~6500

Essential,R~6500

Essential,R~6500

Velocitiesoflowmassgalaxies

Velocitiesoflowmassgalaxies

R~5000

REQ-SRD-013 HighspectralresolutionR>=40000 R>=20000 Essential,

R~20-40K Essential Youngstars

Brightglobularclusters

REQ-SRD-021 Etendue~2000sq.deg@g=16 all-sky

1000ssq.deg;g~20.5@

1000ssq.deg,g~22withhighvaccuracy

100ssq.deg,g~24

3200(100)sq.deg,i<23(24.5)

~100sq.deg,i<24.5

~1000sq.deg,i<24.5

~300sq.deg,i<25

40sq.deg,r<24.0

7sq.deg,5000targetstoi=23.25

all-sky

REQ-SRD-022

Multiplexingatlowerresolution

~5000galaxies/sq.deg

100stargets(galaxies/GCs)/sq.deg

>5000galaxies/sq.deg

770galaxies/sq.deg

600AGN/deg

1000sgalaxies/sq.deg

REQ-SRD-023

Multiplexingatmoderateresolution

1000sstars/sq.degtog~23


few-thousandsstars/sq.deg

500galaxies/sq.deg

REQ-SRD-024

Multiplexingathighresolution

~100stars/sq.deg@g=16

~1000stars/sq.degtog=20.5

~1000stars/sq.degtog=20.5


REQ-SRD-025 SpatiallyresolvedspectraGoal Goal Yes

REQ-SRD-031

Spectralcoverageatlowresolution

0.37-1.5um

0.37-1.5um

0.36-1.8um

0.36-1.8um

0.36-1.8um(0.36-

0.36-1.8um

Opticalemissionlines

REQ-SRD-032

Spectralcoverageatmoderateresolution

Stronglinediagnosticsinoptical

CaTessential

CaTessential

Goal:complete

Goal:complete

Goal:complete

REQ-SRD-033

Spectralcoverageathighresolution

Stronglinesforvelocities;tagging

Stronglinesforvelocities

Chemicaltagging

Stronglinesforvelocities

REQ-SRD-034

Sensitivityatlowresolution

i=24.5 i=24.5 i=25.3 i=25/H=24 i=23.25 i=24.5

REQ-SRD-035

Sensitivityatmoderateresolution

g>20.5 g~23 i=24 i=24.5 i=24.5 r~24

REQ-SRD-036

Sensitivityathighresolution

g=16@highSNR g=20.5 g=20.5 g~22

REQ-SRD-041 Velocitiesatlowresolutionv~20km/s v~20km/s v~100km/s v~20km/s v~20km/s v~20km/s

REQ-SRD-042

Velocitiesatmoderateresolution

v~1km/s v~1km/s v~5km/s v~9km/s v~9km/sv~20km/s(10km/sgoal)

REQ-SRD-043

Velocitiesathighresolution

v<100m/s v~100m/s v~100m/s v<1km/s

REQ-SRD-044

Relativespectrophotometry

~4% Critical:3%

REQ-SRD-045

Skysubtraction,continuum

few% few% few% few% <1% <1% <1% <0.5% <0.5% <1% <1% <1%

REQ-SRD-046

Skysubtraction,emissionlines

important(CaTregion)

important(CaTregion)

critical critical critical critical critical critical

REQ-SRD-051 Accessiblesky

Platofootprint(ecliptic)

Gaiafootprint(allsky)

Gaiafootprint(allsky)

Gaia,PS1,HSCfootprint

Northernhemisphere(M31,M33,dec~+40)

LSSToverlapuseful(10000sq.deg)

NGVSfootprint;dec+12

LSSToverlapuseful(10000sq.deg);Euclid(allsky)



allskytargetdistribution

allskytargetdistribution

REQ-SRD-052 Observingefficiencymaximise maximise maximise maximise maximise maximise maximise maximise maximise maximise maximise maximise

REQ-SRD-053 ObservatorylifetimeMonitoring>=years

Monitoring>=years

Survey>=5years

Survey>=5years

Survey~100snights

Survey~fewyears

Survey~100snights

Survey~7years

Survey~100snights

Survey~100snights

Monitoring~5years

Survey~years

Ope

ratio

ns

Resolvedstellarsources Extragalacticsources

Spectral

resolutio

nFocalplane

inpu

tSensitivity

Calibratio

n


1.3 MSEasauniqueanddiverseastronomicalfacility1.3.1 ThescientificcapabilitiesofMSE

Table2:SummaryofthescientificcapabilitiesofMSE

ThediversescienceenabledbyMSEanddiscussedinthisdocumentspansallastronomy,fromplanetformation,themicrophysicsofstarsandtheinterstellarmediumthroughtothedynamicsofdarkmatterandthephysicsofblackholes.WhatarethekeycapabilitiesofMSEthatfacilitatethisimpactanddiversity,andhowdoesitcomparetootherspectroscopicresources?Table2showstherelevantscientificcapabilitiesofMSE.Thetelescopewillhavean11.25msegmentedprimary (effective aperture of >10m) and a 1.5 square degree field of view. Three differentspectral resolution settings are possible. At the lowest resolution, over 3000 spectra can beobtainedinasinglepointingthatspanstheentireopticalspectrumintothenearinfrared(goal:H band). At moderate resolution, the same number of spectra can be obtained forapproximatelyhalftheopticalwaveband.Atthehighestresolution,approximately1000spectracan be obtained per pointing for windows in the optical region of the spectrum. At allresolutions,MSEcanpushtothefaintesttargets,andwillbewellcalibratedandstableoveritslifetime. Located at the equatorial site ofMaunakea (ℓ = 19.9°),MSEwill access the entirenorthern hemisphere and half of the southern sky, making it an ideal follow-up and feederfacilityforalargenumberofexistingandplannedground-andspace-basedfacilities.Itisnotoriouslydifficulttocapturethescienceperformanceofaninstrument/facilityinasinglemetric.Nevertheless,forillustrativepurposesweattempttodothisbydefiningaparameter,η,thatlooselyrepresentsthe``discoveryefficiency"ofaMOScapability.Here,

𝜂 = !!!! ! !!"# !

!"!,

where

• 𝐷!!isthediameteroftheprimarymirror,• Ωistheinstantaneousfieldofview,• 𝑁!"# isthenumberofsimultaneousspectra,• 𝑓 isthefractionaltimeavailableforscheduling,and• 𝐼𝑄 isthetypicalimagequality(FWHM)deliveredbythesiteandfacility.

AccessibleskyAperture(M1inm)Fieldofview(squaredegrees)Etendue=FoVxπ(M1/2)2

Modes Moderate IFU

0.36-0.95μm J,Hbands 0.36-0.45μm 0.45-0.60μm 0.60-0.95μmSpectralresolutions 2500(3000) 3000(5000) 6000 40000 40000 20000Multiplexing >3200

Spectralwindows ≈Half λc/30 λc/30 λc/15

Sensitivity m=23.5*

Velocityprecision 9km/s�Spectrophotometicaccuracy <3%relative♯Dichoricpositionsareapproximate

*SNR/resolutionelement=2 � SNR/resolutionelement=5

♮ SNR/resolutionelement=10 ★SNR/resolutionelement=30

Wavelengthrange 0.36-0.95μm0.36-0.95μm�0.36-1.8μm

Low High

IFUcapable;anticipatedsecond

generationcapability

30000squaredegrees(airmass<1.55)11.25m1.5149

>1000

m=20.0�m=24*20km/s�

<3%relative

Full>3200

<100m/s★N/A


Figure 6: The discovery efficiency of MSE relative to other existing and planned spectroscopic instruments, asdefined in the text. The top panel shows instruments that operate at low and intermediate resolutions (heredefinedasR<10000)andthebottompanelshowsinstrumentsthatoperateathighresolution(R>10000).Judgedbythismetric,MSEishighlycompetitiveacrossallofitsmodesofoperation.

Figure 6 compares this parameter for a number of facilities. Here, the facilities have beendivided into ``low/intermediate" (1000 < R < 10000) and ``high" (R > 20000) resolutioncategories,althoughtheexactresolutionatwhichthissplit ismadeisarbitrary. Judgedbythismetric,MSEoutperformsallother facilitiesbyawidemarginforbothhighand lowresolutionscience.However,giventheuncertaintiesinherentinattemptingtocomparealargenumberofdisparate facilities throughasinglemetric, it is illustrative toconsider theperformanceof theinstrumentslistedinFigure6asitrelatestospecificsciencegoals:

• Atlowresolution,considertheSROdescribedinAppendixJthatwillmaptheaccretionprocesses occurring within the inner parsec of quasars through a large-scalereverberation mapping campaign. This program requires good SNR observations of alargenumberof relatively faintquasarsat low/moderate resolutionoveranextendedperiodof time. 10mclass telescopes are thereforeessential.At low redshift,manyofthediagnosticstobemonitoredareintheopticalregionpriortobeingredshiftedintotheNIR,andsoa largewavelengthrange isnecessary. OnlyMSEandSubaru/PFSaretherefore capable of this science, although the narrower wavelength range ofSubaru/PFS means such studies are more limited in redshift space. Critical to thisscience,however,isaccuratespectrophotometryandcalibration,tobeabletocompareobservations separated by up to 5 years in duration. Such precise observations are astrength of a dedicated, specialized facilitywhere the entire system is optimized and


stable,andismuchmoredifficulttoachieveforafacilityinstrumentsuchasSubaru/PFSthatmovesonandoffthetelescopeatregularintervals;

• Athighresolution,considertheSROdescribedinAppendixDthatwillmeasurethecolddark matter sub-halo mass function of the Galaxy through precision velocitymeasurementsofstarsintidalstreamsinthehalo.SuchobservationsrequirehighSNRobservations of very faint stars (g > 20), in order to be able to have the statisticsnecessary to probe the kinematic substructure of very diffuse and distant structures.Moderateorhighresolutionobservationson10mclassfacilitiesarerequired,andtheseadditionallyprovideimportantmetallicitydiagnosticsinordertoisolatethepopulationsof interest. In addition, however, data spanning hundreds or thousands of squaredegreesarerequiredduetotheveryextendednatureofthestreamsbeingprobed,andin order to probemultiple streams at different locations in theGalactic halo. Field ofview andmultiplexing capabilities are essential.MSE is therefore the only instrumentthatcanconductthisscienceprogram.

1.4 KeycapabilitiesforMSEscience

MSEisuniquelypowerfulforspectroscopicsurveyscienceduetoarangeoffactors,includingitslargeaperture, rangeof spectral resolution, excellent sitequality, long lifetimeanddedicatedoperations. Given the intense focus on wide field spectroscopy and the large number ofinstrumentsshowninFigure 6, it isworthdiscussingthekeyscience-enablingcapabilitiesforMSEinmoredetail.

1.4.1 Largeapertureandhighthroughput

MSE is the fourth largestoptical andnear-infrared telescope tobebuilt, after theE-ELT,TMTandGMT(werefertothesefacilitiescollectivelyastheVeryLargeOpticalTelescopes,VLOTs).ThelargeapertureofMSEisfundamentaltoitsimportancewithinthenetworkofinternationalastronomicalfacilities.Firstly,smalleraperturetelescopescannotprovideessentialfollow-upoffaint sources identified by imaging surveys on large aperture telescopes like Subaru or LSST.Secondly,asafeeder facilityfortheVLOTs,MSEmustbeabletoprovidereasonablesignal-to-noisemeasurementsoffaintsourcesthatwillbefollowedupwithhigherspatialresolution(e.g.,IFUs),higherspectralresolution,and/orhighersignal-to-noiseonthesegiantfacilities.MSEwillutilizethefulllight-gatheringpowerofitslargeapertureinpursuitofitsdrivingsciencegoals. There are numerous excellent projects underway or under development that provideMOS capabilities for 4 m class apertures. MSE science, however, is focused towards anunderstanding of the faintUniverse - such as intrinsically faint stars, the distant Galaxy, lowmassgalaxies,andthehighredshiftUniverse-thatisnotaccessiblewithsmallerapertures.Thisopensupextensivenewareasofresearch,suchasin-situchemodynamicalstudiesofthedistantMilkyWaystellarhalo,enablingallGalacticcomponentstobestudiedinanunbiasedmanner.Indeed,MSEistheonlyfacilitythatwillbeabletoconductdetailedchemicalabundancestudiesacross the full luminosity range of targets identified by the ESA/Gaia mission. Further, forextragalacticscience,thedynamicsandstellarpopulationsofdwarfgalaxiesthatfallbelowthedetection threshold of smaller-aperture facilities at low andmoderate redshiftswill be easily


accessible,andMSEwillallowtheanalysisofsub-L*galaxiesouttohighredshift.Again,thelightgatheringpowerofthelargeapertureisessentialinallowingunbiasedanalysesofthesegalaxypopulationsthatcouldnotbeachievedwithsmallerapertures.

1.4.2 Operationatarangeofspectralresolutions

A core requirement of MSE is that it operates at a range of spectral resolutions and this isessential to itssuccessfuloperation.Firstly,asalreadydescribed, itenablesadiverserangeofscientificinvestigations,andMSEisexpectedtocontributeequallyto“LocalUniverse”science–where the brighter targets are generally more well-suited to high and intermediate spectralresolutions – and “DistantUniverse” science –where themoredistant faint targets generallyrequire lower spectral resolutions. This diversity is essential to ensure success as a facility, incontrasttoinstruments(whereamorelimitedrangeofcapabilitiesisoftenacceptableduetoafocus on a narrower range of science cases). Good examples of this latter approach areMayall/DESI (that focusesuponthedeterminationofspectroscopicredshiftsof threedifferentsub-classesofgalaxiesforprecisedeterminationofcosmologicalparameters)andAAT/HERMES(thatisfocusseduponthederivationofthechemicalabundancesofstarsinspecificcomponentsoftheGalaxy).

As well as promoting scientific diversity, a range of spectral resolutions is essential to theoperationalefficiencyofMSE. Inparticular,MSEwillobserve throughout the lunarcycle,withhigher resolutionobservationsofbright targetsexpected todominateduringbright time,andfainter extragalactic targets expected to dominate the observing during dark time. Operatingwith high efficiency is important for all facilities, but it is especially true forMSE where theaccumulationoflargedatasetsisadrivingrequirementfornearlyallobservingprograms.

1.4.3 Dedicatedoperationsoveralonglifetime

MSEprovidesaspecializedtechnicalcapability,namelylarge-aperturewidefieldextremeMOS.MSE will conduct spectroscopy of a vast number of astronomical sources to producehomogeneously-calibrated,well-characterized,datasets.For instrumentsthatmoveonandofftelescopesat regular intervals,data issuessuchascalibration,stabilityandreproducibilitycanbeproblematicsincetheinstrumentisnotbeingleftinastableconfiguration.ThisisincontrasttothesituationforMSE,wherethebasicoperationalphilosophyofspecializationenableshighquality and stable data (e.g., fibre coupling will be left in place all the time, no prime focuschanges,WFCorADCremovals,etc.).Indeed,thisallowsforsciencecasesthatwouldotherwisebe very difficult to address on other instruments (for example, time resolved high-resolutionspectroscopyandquasarreverberationmapping).

Theimpressivepowerofferedbyalarge,homogeneousandwell-characteriseddatasetsuchasthatofferedbyMSEhasbeenmostsuccessfullydemonstratedby theSloanDigitalSkySurvey(SDSS). It shouldperhapscomeasno surprise that the science thatultimatelyemerged fromSDSS–andwhichcontinuestoemerge–wasfarmorediversethananticipatedatitsoutset.Theimportance of the spectroscopic component of SDSS can scarcely be overestimated: it isuniversally recognized that the science impact of SDSS would have been greatly diminishedwithout its spectroscopic element, for the simple reason that broadband photometry alone


providesonlyzeroth-orderinformationonthephysicalpropertiesofastrophysicalsources.SDSShad a profound impact on topics as far ranging as small bodies in the solar system to thereionizationoftheUniverse–abroadappealthatunderliesitsrepeatedrankingasthehighest-impacttelescopeofthelastdecade(e.g.Madrid&Macchetto2006,2009;Chenetal.2009)1.The overwhelming success of SDSS is despite the fact that the telescope is a relatively small-aperture facility by modern standards, and it is located at a site that, while good, does notcompete in terms of median image quality with Maunakea. A large part of this success canthereforebetracedtotheextremelywell-calibratedandwell-characterizednatureofthedata.AsthefourthlargestOIRtelescopeinoperationinthe2020s,andtheonlylargeaperturefacilityfully dedicated to wide field multi-object spectroscopy at equatorial latitudes, MSE can beviewedasanevolutionoftheSDSSconcept.ThemajordifferencesarethatMSEisrealizedonafacility that has~25 times larger aperture than SDSS and that is located at arguably the besttelescopesiteontheplanet.ThesciencepotentialforMSEisthereforesignificant.

AnadditionalanalogybetweenSDSSandMSE lies in the fact that, in contrast tomanyof theinstrumentsorfacilitieslistedinTable2,MSEisnotbuilttoaddressasinglesciencecaseortoconduct a single survey. Rather, MSE is built to provide a single, missing, important, set ofcapabilities(largeaperturededicatedwidefieldMOSatarangeofspectralresolutions).MSEisdesignedandoptimizedtoexcelatprovidingthesecapabilities,and insodoingenablesavastrange of science and strategic surveys. Similarly, SDSS has demonstrated flexibility andresponsiveness to science over its lifetime though successive generations of the telescopeaddressingnewandstrategic sciencegoals: theLegacySurvey;SEGUE Iand II,Marvels,BOSS,eBOSS,APOGEEIandII.Overitslonglifetime,MSEisexpectedtoaddressaverylargerangeofsciencecases,andinstrument/telescopeupgradeswilloccur.ButliketheSDSSoverthelastfewdecades,MSEwillremaintheworld’spremierresourceforthespectroscopicexplorationoftheUniverse.

1.5 CompetitionandsynergieswithfutureMOS

Figure 6 shows a large number of MOS instruments and facilities, many of which will beoperatingontimescalesthatoverlapwithMSE.Here,weprovideabriefoverviewofthemajornewprojectsinthisarea.

1A search on the Astrophysical Data System for publications that mention “SDSS” in theirabstract–presumablythosepapersthateitherrelyonSDSSdataforanalysis,interpretationorcomparison–returnsover11000paperswithoveraquarterofamillioncitations.


1.5.1 4-mclassMOS

Table3:Summaryofnewopticalandinfraredmulti-objectspectroscopicinstrumentsandfacilities.

LAMOST (2011): The Large Sky AreaMulti-Object Fiber Spectroscopic Telescope (LAMOST) –also known as the Guo Shoujing Telescope – is a 4m reflecting Schmidt telescope with a 5degreediameter fieldof viewwithadedicated suiteof spectrographscapableof carryingoutR ~1500 spectroscopy for~ 4000 objects per field. It is undertaking two main surveys, theLAMOSTExtraGAlacticSurvey(LEGAS)andtheLAMOSTExperimentforGalacticUnderstandingandExploration(LEGUE).

AAT/HERMES(2015):TheHighEfficiencyandResolutionMulti-ElementSpectrograph(HERMES)isafourchannelfiberfedspectrographonthe3.9mAnglo-AustralianTelescope(AAT)thatusesthe 392 fiber 2 Degree Field (2dF) fiber positioner system. HERMES operates at R = 28 000,although a higher resolution of R~ 50 000 is possible using a slit mask at the cost ofconsiderable light loss. The main program being undertaken by HERMES is the ``GalacticArchaeologywithHERMES"(GALAH)survey,thattargetsatotalof1millionstarsbrighterthanV~14mag,measuringaradialvelocityforeachaswellaschemicalabundancesfor~15differentelements.

WHT/WEAVE (2017): The WHT Enhanced Area Velocity Explorer (WEAVE) is based on thesuccessful 2dFpositioner systemof theAAT,modified to allow~1000 fibers tobepositionedover a 2 degree field for the 4.2m William Herschel Telescope. The spectrographs provideR ~ 5000over the fullopticalwavelengthrange (370–1000nm),andR ~20000overamorerestrictedrange.WEAVEalsopossessesnotableIFUcapabilities,specifically20deployableIFUseachwiththirtysevenfiberswithadiameterof1.3”each,aswellasalargeformatIFUwith540fiberswithdiametersof2.6”each.

VISTA/4MOST(2017):The4-meterMulti-ObjectSpectroscopicTelescope(4MOST)isa4squaredegreefieldofviewfacility.2400targetscanbeobservedperfield,feedingbothhighresolution(R~20000) and intermediate resolution (R~5000) spectrographs. 4MOST will operate as adedicated facility enabling a wide range of spectroscopic surveys, and in many operationalrespectsisa4manalogytoMSE.

Class Facility/InstrumentFirstlight

(anticipated)Aperture(M1inm)

FieldofView(sq.deg) Etendue Multiplexing

Wavelengthcoverage(um)

Spectralresolution(approx) IFU

Dedicatedfacility

Comparison SDSSI-IV Existing 2.5 1.54 7.6 640 0.38-0.92 1800 Yes Yes

GuoShoujing/LAMOST Existing 4 19.6 246 4000 0.37-0.90 1000-10000 No YesAAT/HERMES 2015 3.9 3.14 37.5 392 windows 28000,50000 No No

0.37-1.00 5000windows 200000.39-0.95 5000windows 18000

Mayall/DESI 2018 4 7.1 89.2 5000 0.36-0.98 4000 No Yes0.8-1.8 4000windows 200000.38-1.26 20000.71-0.89 50000.36-1.8 30000.36-0.95

50%coverage6500

windows 40000

Secondgeneration

Yes10-m

8-m

MSE 2024 11.25 1.5 149 3468

7.4 1000 No No

8.2 1.25 66 2400 No No

39.5 1000 Yes Yes

YesNo240031.42017

2017

2018

2019

4 3.14

2.54

8.2 0.14

WHT/WEAVE

VISTA/4MOST

VLT/MOONS

Subaru/PFS

4-m


Mayall/DESI(2018):TheDarkEnergySpectroscopicInstrument(DESI)willgoonthe4-mMayalltelescope.ItisfocusedonperformingaStageIVDarkEnergymeasurementofbaryonicacousticoscillationsby targetinga totalof25million luminous redgalaxies, emission linegalaxiesandquasarsoveraperiodof5years.5000objectscanbeobtainedper8squaredegreefieldofview,ataspectral resolutionofR=4000by feeding10 identical three-armspectrographs.DESIwillalso operate additional surveys during bright time, including targeting some 10 million starsbrighterthanV ~ 18.

1.5.2 8-mclassMOS

Table4:MSEincomparisontootherplannedMOSinstrumentson8-mclasstelescopes.

VLT/MOONS(2018):TheMulti-ObjectOpticalandNearInfraredSpectrograph(MOONS)isa500square arcminute field of view spectrograph for the 8.2m VLT operating at R~5000 and R~ 20000andcapableoftaking1000spectrasimultaneously(nominally,500objectand500skyspectra). In addition to its field of view, it differs from most other MOS spectrographs bytargetingthered/NIRwavelengthrange,0.8–1.8μm,makingitideallysuitedforstudiesofthebulgeandotherGalacticfieldswithhighreddening,aswellasthehighredshiftUniverse.Subaru/PFS (2019):ThePrimeFocusSpectrograph(PFS)forthe8.2mSubarutelescope(Vivesetal.2012;Sugaietal.2012) isa2400objectspectrographwitha1.25squaredegreefieldofview that covers the optical/NIR wavelength range (0.38 – 1.26μm). It primarily operates atR ~ 2000,althoughthereareplansforanadditionalR ~ 5000modeintheregionoftheCalciumTriplet.A“StrategicSurveyProgram”willusethisinstrumentfor~300nightsdistributedovera5-yearperiodtoconductaprogramtoexplorethenatureofdarkenergythroughobservationsofemissionlinegalaxies,andaGalacticArchaeologyprogramforthederivationofvelocitiesandstellarparametersfor~1millionstars(Takadaetal.2014).

DedicatedfacilityAperture(M1inm)FieldofView(sq.deg)EtendueMultiplexingEtenduexMultiplexing

Observingfraction

Spectralresolution(approx)

4000 20000 2000 5000 3000 6500 40000

Wavelengthcoverage(um)

0.8-1.8 windows 0.38-1.26 0.71-0.89 0.36-1.8 0.36-0.9550%coverage

windows

IFU

10.2(first5years)0.2-0.5afterwards?

<1?

No No Yes

7400(=0.014) 158400(=0.31) 517100(=1.00)

8-12mclassfacilitiesVLT/MOONS Subaru/PFS MSE

8.2 8.2 11.250.14 1.25 1.57.4 66 1491000 2400 3468

No No Secondgeneration


1.6 Thefuture-scapeofMSE

Figure 7: Some of the most notable next generation facilities that, together with MSE, will help define theinternationalnetworkofastronomicalfacilitiesoperationalbeyondthe2020s.


Astronomyisenteringanew,multi-wavelengthrealmofbigfacilities.Figure7showssomeofthe most notable of these new observatories. Collectively, they represent many billions ofdollarsofinvestmentandmanydecadesofdevelopmentbylargeteamsfromtheinternationalcommunity. They probe questions as diverse and fundamental as the origin of life and thecompositionofourUniverse,andtheirscientificexploitationhasalreadybegun.Atthetimeofwriting, ALMA is now in routine operation, and the first data release fromGaia is imminent.These areno longer “next generation” ambitions, but insteadare the cutting edgeofwhat ispossiblenow.Perhapsmostexcitingly,ourviewoftheUniverseisnolongerlimitedonlytotheelectromagnetic spectrum, such that the term “multiwavelength” is no longer sufficient tocompletelydescribe thewindows into theUniverse thatwehaveatourdisposal. Indeed, thedetection of the first gravity waves in February 2016 by LIGO promises to provide afundamentallynewwayofexploringtheUniverse.Eachof the facilitiesshown inFigure 7hasawell-definedsciencecaseandcanoperateasastand-alone facility.However, therecenthistoryofastronomydemonstrates that it is throughthecombinationofdatafromthesenewfacilitiesandextanttelescopesthatmanyofthemajoradvanceswill bemade. The science thatwill ultimately emerge from this collaboration is farmorediversethanwecancurrentlyanticipate.The development of MSE explicitly recognizes the emerging existence of an internationalnetwork of astronomical facilities thatwill define the future of frontline astrophysics.MSE ispositioned to be a critical hub in this network, with scientific capabilities that naturallycomplement and extend the capabilities of the facilities shown in Figure 7. Many of thesynergiesareobvious,manycannotyetbeanticipated,butallareimportanttothefuturehealthofastronomy.Here,wediscussMSEinthisglobalcontextofastronomicalcapabilities.


1.6.1 WidefieldimagingoftheUniverse

Figure8:Calendarcomparingvarious imagingsurveycapabilitiesacross theGammaraytosub-mmspectrum(asindicatedbycolor).TheFigureofMeritforthesesurveysiscalculatedusingAreaxFoV/(IQ)2andisindicatedbygrayscale.FigurefromMeyeretal.(2015).

Figure 8showsthetimelineforalargenumberofopticalandNIRimagingsurveys(green)aswellasat shorter (blue/purple)and longerwavelengths (orange/red).Complete,deep,opticalimagingof the skywill shortlybeavailable througha combinationof surveyprograms. In thenorthernhemisphere,theentire3πsteradiansofskyvisiblefromHawaiihasbeenmappedusingthePanSTARRS(PS1)telescopeinthegrizyfiltersto5-σABlimitingpoint-sourcemagnitudesof(21.9, 21.8, 21.5, 20.7, 19.7) mags, respectively. In the south, SkyMapper is undertaking itsSouthernSkySurvey,anall-hemispheresurvey(inuvgriz)thatwillbeapproximately0.5–1magdeeperthantheSDSS.Hyper-Suprime-Cam(HSC)ontheSubarutelescopeonMaunakeaisthelargestmosaiccameraevercommissionedonan8m-classtelescope,andiscurrentlyconductinga 5 year strategic survey to map~ 1500 square degrees along the celestial equator tounprecedenteddepth.Fromthesouth,theDarkEnergySurveyisusingtheDarkEnergyCamera(DEC)onthe4mBlancotelescopetomapanareaof5000squaredegrees.Manyotherextensivewidefieldimagingsurveyson4–8mtelescopesalreadyexist,bothatopticalwavelengths(e.g.,the CFHT Legacy Survey and other CFHT programs; the Subaru SuprimeCam imaging archive)


andatNIRwavelengths (e.g. theUKIDSSnorthern sky surveyusingUKIRT; various surveysonVISTA).Figure7demonstratesthatthemostpowerfulground-basedimagingcapabilityonthehorizonis the Large Synoptic Survey Telescope (LSST). This facility is the top-ranked, ground-basedfacilityinthe2010decadalsurveyforUSastronomy.TobesituatedatCerrroPachon,LSSTisa6.2m(effectiveaperture)telescopewithafieldofviewof3.5squaredegreesthatwillseefirstlightin2021.Itsprimarymissionistoundertakea10-yearprogramtomonitor~25000squaredegrees, building up a deep image of the sky through the co-addition of~1000 exposures ateachposition.Eachindividualexposurelasts~15sandhasasingle-passdepthofr~24.5.Thefinal,co-added,LSSTsurveydepthisr~27.5.

The discovery space of LSST is enormous. The design of the surveys has been guided by fourmain science themes– “Taking an inventoryof the Solar System”, “Mapping theMilkyWay”,“ExploringtheTransientOpticalSky”,and“ProbingDarkEnergyandDarkMatter”. Inpractice,however,LSSTcanbeexpectedto impacteveryareaofastronomy.Overthe10yearbaseline,LSSTwilltake~5.5millionimagesandwillhavemeasured~7trillionsingleepochsources,foratotalof~37billiondistinctobjects.Itisarequirementthatphotometriczeropointsarestableto~10 millimags; with nearly 1000 visits per patch of sky, LSST will probe the variable andtransientopticalskywithunprecedentedaccuracyandcadence.

ThescientificsynergiesbetweenMSEanddeepimagingsurveyssuchasLSST,Subaru/HSC,PS1and others, that togethermap the entire sky, are extensive. It isworth emphasizing that themajorityofthebillionsofsourcesthatwillbephotometricallyidentifiedwillbefaint,faroutsidethe capabilities of 4-m class spectroscopy. As an 11.25m aperture facility, MSE obtainspectroscopicdataforsourcesacrossthefullmagnituderangeofPS1,andwillbeabletoobtainspectra for sources identified ina singlepassof LSST.Smalleraperture spectroscopic facilitiesareunabletoexploitthediscoverypotentialofthesesurveystothesamedegreeasMSE,andemphasizesthestrongneedforlargeaperture,dedicatedMOS.Whilethereisalongheritageoftheobvioussynergiesbetweenopticalandinfraredimagingandspectroscopic surveys,newmodesof follow-uparise from thededicated spectroscopic surveycapabilitiesofMSEthathighlightthediversityofsciencethat itenables.Duetotheircadence,bothPS1andLSSTarepowerful transientdiscoverymachines; LSSTexpects to issue105–106transient alerts per night. Similarly, SKA and othermulti-wavelength survey facilitieswill alsoidentify very large number of transients, and follow-up of photometrically identifiedoptical/near-IRcounterpartsofradio,X-rayandgamma-raytransientswillbeessential.MSE isuniquelyplacedtoundertakeasystematic, longtermopportunisticscienceprogram insupportofLSST,SKAandothertransientsurveys(seeChapter4).Givenitsequatoriallocation,MSEprovidesgoodaccess tobothhemispheres, and therewill be~1500 squaredegrees

per

nightofLSSTimagingaccessibletoMSE,inwhichweestimatetherewillbe~300newSNeand>75000variablestars(Ridgwayetal.2014,ApJ,796,53),aswellasmanymoreexotictransientphenomena.ForanygivenpointingofMSE,oneor two fibersonaveragecanbeallocated tosources recently identified as transients by LSST (or other surveys). Given MSE is alwaysobtainingspectraformanythousandsoftargetsperhour,thenthisquicklyamortizestoensure


thatMSEwilldominateanewregimeoffainttransients(probingveryearlyorverylatestages,aswellaslargedistancesandlowluminosities),Indeed,MSEwilllikelydedicatemoreofitstimetotransientspectroscopythananyother10mclasstelescope.InspectionofFigure8revealsthat,whileLSSTdominateswidefieldimagingfromtheground,imaging from space is dominated by two missions: Euclid andWFIRST. Euclid, scheduled forlaunchin2020,isanESA-ledmissiontounderstandthenatureofdarkenergyanddarkmatterthroughweaklensingandgalaxyclustering.It isa1.2mNIR/opticaltelescopethatwill imageaminimum area of 15 000 square degrees at |b| > 30 degrees during its anticipated 5-yearlifetime.LimitingABmagnitudesareexpectedtobe24inYJH,and24.5inasingle,broadopticalfilter (RIZ).Overall,more thanabilliongalaxieswillbeobserved.WFIRST isacomplementary,NASA-led mission that has a considerably larger aperture than Euclid (at 2.4m) and that isequipped with a wide field imager, an integral field unit, and a coronographic imager forexoplanet imaging. In terms of imaging, one of its primary missions will be to map~2000squaredegreesofskytodepthsofJ~26.7,inadditiontoasetofultra-deepdegree-scalefields.Precisionphotometryandastrometrywillbeobtainedforallsources.Importantly,upto25%ofWFIRSTtimeislikelygoingtobeP.I.-time,ensuringthatdeeppointedsurveysofvariousfieldswill occur. Euclid andWFIRSTwill together provide an unprecedentedwide-field NIR datasetthat,whilefocusedprimarilyoncosmologicalsciencequestions,willhavealastinglegacyforallofastronomy.Examples of the sort of scientific opportunities produced by Euclid andWFIRST forMSE arenumerous and discussed throughout this document. A driving theme of extragalactic sciencewithMSEislinkinggalaxiestotheirsurroundinglargescalestructure(Chapter3).Assuch,theemergenceofstructureonkiloparsecscalesisofgreatinterest.Driveretal(2013)proposedthatgalaxies form via two stages, firstly bulge formation via some dynamical hot process (i.e.,collapse,rapidmerging,disc instabilitiesand/orclumpmigration),andsecondlydiscformationvia a more quiescent dynamically cool process (i.e., gas in-fall and minor-merger accretionevents).Thepivotal redshift for theseprocesses isat z~1.5.At later times,duringperiodsoflow interaction rates (see Robotham et al 2014), bars emerge along with disc-buckling andpseudo-bulgeformation.InthepresentdayUniverse,systemsadheretotheHubblesequence,at earlier times galaxies appear more lie train-wrecks, as unveiled by the Hubble SpaceTelescope. While the Hubble Space Telescope (and its successor the James Webb SpaceTelescope)provide(andwillprovide)highqualityimaging,thefieldsofviewareverysmallandonlytheCOSMOSsurveyextendstomorethanasquaredegree.Whilethisprovidesinsightintothe morphologies of galaxies at very high redshifts, the epochs and process by which eachstructure(bulge,bar,disc)emergeisempiricallyunconstrained.

EuclidandWFIRSTwillprovideextremelydeephigh-spatialresolutionimaging(~0.2”)andwilldiscernstructuretosub-kpcscalesouttoz>2.5atrest-opticalwavebandsoverextremelylargeareas. Thiswill allow a directmeasure of the epochs atwhich the various structures emerge(e.g.,bulge,disc formation)andhowtheyevolve (i.e,growthof spheroids,bulges, anddiscs).However, tofill inthevoidbetweentheverynearbysurveyssuchasSDSSandGAMAandtheverydistantsurveyswithHSTandJWSTrequiresthecombinationofEuclid/WFIRSTimagingwithphotometric-redshifts (LSST)andMSEspectra.Throughthiscollaboration,wecanextractstar-formationrates,metallicities,andcanconfirmpairmembershiptoestablishmergerrates.


The combination of data from Euclid/WFIRST/LSST (and SKA; see Section 1.6.3) with MSEspectralanalysiswillprovideacompleteblueprintofgalaxyevolutionfromthepresentepochtothepeakofthecosmicstar-formationera(i.e.,z=0toz~2.5).Samplessizesneednotbelarge;the crucial elementwill be the ability toobtain reasonable SNR spectra (~30) for high-z (z~2.5),faint(i~25mag)systemswithahigh-levelofcompleteness.MSEistheonlyspectroscopicfacilitythatprovidesthesecapabilities.

1.6.2 TheeraofGaia

Figure 9: Top left: Expected radial velocity accuracy for the RVS instrument aboard Gaia for stars of differentspectral types. The lowerblack curve shows the radial velocity accuracyexpected for aG2V starobserved for atotalof1hourwithMSEinhigh-resolutionmode.Bottomleft:Typicalerrorsintransversevelocityasafunctionofdistance for starsofdifferent spectral typeobservedwithGaia. The shadedgrey region shows the approximateerrorscalculatedforG2Vstarswithg=20.5withMSE.Topright:ApproximateexposuretimesrequiredforMSEtoreachtargetsignal-to-noiseratios,plottedasafunctionofg-bandmagnitude.Thehorizontaldottedlineindicatesthenominalsurveyexposuretimeof1hour.Middleright:Depthreachedforold,metal-poorhalopopulationsasafunctionofg-bandmagnitudeforthreedifferenttracerpopulations.Bottomright:CumulativenumbercountsforMilkyWaystarsina100deg2regionatthenorthGalacticcap(blackcurve).StarcountsforthreeseparateGalacticcomponents(youngdisk,thinkdiskandhalo)areshownincolour.

Gaia is a landmark astrometric space mission that has as its primary focus a detailedunderstandingofthestructureandcompositionofourGalaxy.Launchedin2014, ithasbegunits 5 year mission to produce a three-dimensional stellar map of the Milky Way withunprecedented precision. Gaia is conducting an all-sky survey to measure the positions ofroughly1billionstars,approximately1%oftheentirestellarcontentoftheMilkyWay.Infact,allastronomicalpointsourcesbrighterthanV~20magwillbecatalogued,andobjectsbrighterthan V~ 15 mag will have their position measured to better than 20 microarcseconds (theapproximatewidthofahumanhairviewedatadistanceof1000km).Beyondastrometry,Gaiawillobtainmulti-bandphotometryforallsources,and it isadditionallyequippedwithaRadialVelocity Spectrometer (RVS) thatwillmeasure velocities forobjectsbrighter thanV~17mag(roughly 150 million stars) to an accuracy of 1 – 10kms-1. Basic astrophysical information –includinginterstellarreddeningandatmosphericparameters–willbeacquiredforthebrightest


~5millionstars.Chemicalabundanceinformationwillalsobeprovidedforafewelements(i.e.,Mg,Si,Ca,TiandFeforstarsofspectraltypeF-G-K)forstarsbrighterthan12thmagnitude.

The Gaia dataset is a unique and comprehensive resource for astronomy, given theunprecedentedaccuracywithwhichitwillmeasurethe3-DpositionsofGalacticstars.Thereisanear perfect synergy betweenMSE and Gaia onmultiple fronts, and these synergies featureheavilyinChapter2inparticular.

Theimportanceofground-basedspectroscopytosupplementtheGaiadataiswellrecognizedintheinternationalcommunity,particularlyasitrelatestothechemicaltaggingofindividualstars(Chapter 2). For chemical evolution studies, Gaia will identify hundreds of millions of starsfainterthanG~17forwhichtheRVScannotprovideanyspectral information. Inresponsetotheimmediateneedfordeepspectroscopicobservationsoffaintstarsforchemicalabundanceanalysis, some300co-investigatorsare involved in theESO-Gaia survey (Gilmoreetal. 2012)since2011.ThisisadedicatedefforttousetheVLT'shigh-resolution,multiplexedspectrographsto observe~ 105 stars with FLAMES/GIRAFFE (R~ 20000, V < 20) and~ 104 stars withFLAMES/UVES(V<15)overaperiodof5years.Resultsfromthissurveyarealreadyinformingcritical aspects of next generation spectroscopic surveys, includingMSE, specifically informingtrade-offs between resolving power, wavelength range and multiplexing. As discussed inChapter2thesetradesareparticularlyimportantforchemicaltaggingstudiessinceweaimtomeasuretheabundancesofenoughchemicalelements,withsufficientlydiversenucleosyntheticorigins,withasufficientlyhighprecision.

Thegeneralgoalofchemically taggingstars in theGalaxy iscommonbetweenMSEandotherhigh resolution spectroscopic surveys, such as VISTA/4MOST and WHT/WEAVE. These othersurveys will provide essential insight into the nearby Galaxy and the main disk in particular.However, only MSE has the resolution and throughput to conduct these experimentsthroughout theGalaxy, via in-situ chemical analysis of every component and sub-component,andacrosstheentireluminosityrangeofGaiatargets.

Intermsofdataonthedynamicsofstars,theastrometricaccuracyofGaiaisonlymatchedbysimilarlyaccurateradialvelocitiesfortheverybrightestsubsetofitstargets.Forrelativelyfaintstars,MSE is required in order to provide precise radial velocities to complement the precisetransverse velocities measured by Gaia. Figure 9 illustrates this with a comparison of theanticipated velocity accuracy ofMSE compared with the RVS on Gaia, including showing thedistanceout towhich certain stellar tracers canbeobservedandwhat this corresponds to interms of contributions from different Galactic components. We return to discussion of GaiasynergiesinChapter2.


1.6.3 TheSquareKilometerArray

Figure10:Calendercomparingvariousradiosurveyfacilities inoperationnowandthenearfuture.TheFigureofMerit is calculated using (Area/T)2 x FoV / IQ2 and is shown as grayscale for each survey. Note that “H-I”correspondstofacilitiesthatareabletoobserveHIinthelocalUniversei.e.,thatcanobservefrequenciesashighas1.4GHz.FigurefromMeyeretal.(2015).

Figure 10 shows the major radio surveys and facilities currently operational or planned to beoperational prior to 2025. Inspection of this figure shows that the SKA will have a profoundimpact. Among the many science goals of this transformational telescope, the SKA has thecapability of detecting Milky Way-type galaxies via synchrotron radiation into the epoch ofreionization,findingAGNofalltypesandluminosity,andtracingtheneutralhydrogencontentofgalaxiestoz~2.


Figure11:Star formationratesensitivityofselectedfar-infraredobservatoriesandradiocontinuumsurveysasafunctionofredshift.Theredlinesshowtheconfusion-limitedSFRsensitivityoftheHerschelSpaceObservatoryat250μmand theSCUBA-2 instrumenton the JCMTat850μm (solidanddashed, respectively). TheblackdashedcurveshowstheSFRsensitivityofradiosurveysfromtheJ-VLAoverStripe82(Heywoodetal.inprep;solidline),andthecoloredlinesshowthethreetiersoftheLOFARSurveysKSP;Tier1(dottedline),Tier2(dot-dashedline),and Tier 3 (dot-dot-dot-dashed). Spectroscopic redshifts are necessary to provide the x-axis for real data in thisplot,andwillbeprovidedforLOFARbyWHT/WEAVE.MSEandSKAcanexpectsimilarsynergiesinthisfield.FigurefromSmith(2015).

Optical spectroscopy is crucial to maximise the scientific output of the SKA, and to gain thebiggestleapinourunderstandingofgalaxyformationandcosmology.Forexample,theextremesensitivity of the SKA means that it will be able to detect star-forming galaxies in radiocontinuumemissiontohigh(z>1)redshifts.However,thelackofspectralfeaturesintheradiocontinuum means that additional optical spectra are required. While photometric redshiftsprovide reasonable precision, they are insufficient for most other science, for examplequantification of the galaxy environment, or the measurement of ages and metallicity etc.Discrimination of AGN from star formation activity likewise requires optical emission linediagnostics.

AgoodexampleofthetypeofsynergywecanexpectbetweenSKAandMSEinthisarenacanbeseenwiththeproposedWEAVE–LOFARsurvey(P.I.DanielSmith),tobeconductedusingthe4-mclassWHT/WEAVEspectroscopicfacility.Figure11showsthestarformationratesensitivityof the three teirs of the LOFAR radio continuum survey compared to other surveys. ThesensitivityofTier3oftheLOFARsurveyisextremeandwillallowtheroutinedetectionofsub-millimeter-like galaxies at z > 5. However, the critical redshift information on the x-axis ofFigure11islackingwithoutopticalspectroscopyandhereWEAVEwillprovideessentialdata.Fulldetailsof theWEAVE–LOFARsurveyareavailableelsewhere (Smith2015).MSEandSKAcanexpectequallystrongsynergiesastheyrelatetotheSKAradiocontinuumsurveys.


Figure12: FractionofHISKAsurveysources(leftpanel,SKA1;rightpanel, SKA2)thataredetectedinanr-bandapparentmagnitude-limited sample. For SKA1, four surveys are considered (including thenowdefunct SKA-SURmode): 3 deg2 using SKA1-MID (black), 300 deg2 using SKA1-SUR (red), 3000 deg2 using SKA1-SUR (green), and30000deg2usingSKA1-SUR(blue).ForSKA2,threesurveysareconsidered:60deg2(black),600deg2(red),and6000deg2(green).Dottedlinesindicatether-magnitudelimitneededtoachievematchesfor90%ofsourcesineachSKAsurveyarea(given2000hrsofintegrationforSKA1-MID&SKA2surveys,and2yearsoftelescopetimeforSKA1-SURsurveys).Dashedlinesindicate90%completionthresholds.FigurefromMeyeretal.(2015).

ThechallengeposedbythesensitivityofSKAforopticalastronomyissummarizedinFigure12.This shows the results of some simulations presented in Meyer et al. (2015), in which theycalculatethefractionofHIsourcesthatwillbeidentifiedinSKAthatwillalsobeidentifiedintheoptical, as a function of r-band limitingmagnitude. Several different configurations for SKA 1(leftpanel)andSKA2(rightpanel)areconsidered(includingthenowdefunctSKA-SURmode).Clearly,eventomatchtoSKA1detectionthresholdsrequiresopticaldataextendingtor~24.While easy to achieve with imaging, this is at the limit of 4-m spectroscopic capabilities; forexample, the detection limits for one of the most ambitious extragalactic surveys to beconductedonVISTA/4MOST,calledWAVES,ismarkedonFigure12,andeventhisisbelowthe90% completeness threshold for several of the possible configurations. In addition, theconstruction of large statistical datasets to capitalize on the SKA 1 data requires the surveyspeedofthespectroscopicfacilitiestobehigh,andthisscaleswiththeapertureofthefacility.While the sensitivity limits of SKA 2 are uncertain at this stage due to the range of possibleoptionsopentoSKAafterSKA1iscomplete,itisobviousthat4-mclassspectroscopicfacilitieswill be unable to provide the spectroscopy necessary to best match these data. Very largeaperturefacilitiessuchasMSEarethereforeessential.

NumerousothersynergiesexistbetweenMSEandSKA.Forexample,eventheSKAwillnotbesensitiveenoughtodirectlydetectHI inallbut thegasrichgalaxiesout toz ~ 2.Therefore,akeyfocusisonstackinggalaxieswithknownredshifts.Here,spectroscopicredshiftsarefarmorevaluablethanphotometricredshifts.MSEcouldprovidetheidealspectroscopicsampletocarryout HI stacking analyses of galaxies, subdivided by various galaxy properties such as age,


morphology,redshiftetc.

MSEandSKAnaturallyprovidemeasurementsofdistincttracerstotracetheunderlyingdensitydistributionoftheUniverseviagalaxyredshiftscatalogues(e.g.,preciseredshifts for luminousredgalaxiesfromMSEandpreciseredshiftsforlowermass,gas-richgalaxiesdetectedinHIwiththeSKA).Thesegalaxiestracetheunderlyingdark-matterdistributionwithvastlydifferentbias,thus allowing the so-called “multi-tracer” technique (Seljak 2009) to be used in order toovercomecosmicvarianceeffects(e.g.,Ferramachoetal.2014).Suchanapproachisnecessarysinceitisincreasinglyapparentthatcosmologyisbecominglimitedbysystematicsnotstatistics(seeChapter4).

A further areaof direct synergybetweenMSEand SKA is toobtain spectroscopic redshifts ofdistant galaxies that areused forweak lensing. Inmuch the sameway as for optical surveys,redshift information adds significant power to weak lensing analyses, allowing the growth ofstructuretobetraced.Radioweaklensingisinitselfextremelycomplementarytoopticalweaklensing surveys, as the systematic uncertainties are different, e.g., thewavelength-dependentPSFisknownanalyticallyinradiointerferometrybutthesourcedensityisgenerallylower(untilSKA2).ThusthesameMSEspectroscopicsurveysthatwillbeusedtosupplementopticalweaklensingsurveyswillplaythesameroleinradioweaklensingwiththeSKA.

1.6.4 Feedingthegiants

The Thirty Meter Telescope (TMT), the European Extremely Large Telescope (E-ELT) and theGiant Magellan Telescope (GMT) will become the premier astronomical OIR facilities fordetailed,high spatial resolutionviewsof the faintest astronomical targetswhen they see firstlightatthestartofthe2020s(TMT&E-ELT–2024;GMT–2025,althoughscienceoperationsare planned earlier with a reduced number of segments). Together, each of these ≥USD1Bfacilitiescanaccesstheentireskywithunprecedentedcollectingareasandwithfieldsofviewoforderafewarcminutes.Essential to the efficient scientific exploitation of these forefront observatories is targetidentification. Ideally, thiswill use a coordinated suite of supporting facilities to ensure theseforefrontfacilitiesmaximizetheirscienceimpactbytargetingthemostscientificallycompellingphenomena.MSE will occupy an important role in the era of the VLOTs through its ability to providestatisticallysignificantsamplesofOIRspectraofrelativelyfaintsources identified inwidefieldimagingsurveysandwidefieldsurveysatotherwavelengths.Targetscanbeselectedfromthissample based on criteria specific to the individual science case, using spectral informationderivedover the samewavelength towhich theVLOTsare sensitive.Observationswith thesegiant facilities can then focus on higher SNR, higher spectral resolution and or higher spatialresolution(inthecaseofspatiallyresolvedsources).Giventheplethoraofsourcesidentifiedbycurrentand futurewide field surveysat faintmagnitudes, this typeof filtering isessential fornearlyallscienceprograms.


ThewidefieldperspectiveofMSEandthehighprecisionsmall fieldperspectivesoftheVLOTswillnaturallyencouragethedevelopmentofcombinedscienceprogramsbetweenthefacilities.Forexample,Chapter4describesprecisionmeasurementsofthedarkmattermassprofilesofMilkyWay satellites through spatially complete radial velocity surveys extending out to verylargeradius.Here,itisessentialtosamplestarsthatcoverthefullrangeoforbitalparameters.TheVLOTs,withfieldsofviewsthatarerelativelysmallincomparisontomanyMilkyWaydwarfgalaxies, can nevertheless provide critical astrometric data in the central regions of the darkmatterhalo.Thisallowsthederivationoftangentialvelocitiesforasubsetofstarsthatprovidessignificantleverageindistinguishingcoredprofilesfromcuspedprofiles.Onlargerastrophysicalscales, mass measurements of galaxy clusters using complete kinematic data for clustermemberscanbeobtainedusingMSE,andcompared toprecision lensingmassmeasurementsforthesameclustersusingtheVLOTs(Chapter4).

1.7 MSEanditscapabilities:apriorityininternationalstrategicplanning

The plethora of deep imaging and astrometric surveys at optical wavelengths, and of surveymissions at other wavelengths, has resulted in significant focus turning to how to obtaincomplementary OIR spectral data for the (literally) billions of sources that these surveys willidentify. Wide field spectroscopy is now seen as a critical “missing link” in the internationalportfolioofastronomicalfacilities,especiallyatlargeapertureswhereMSEistheonlyfacilityinactivedevelopment.Forexample:

• [Australia]TheAustralianAstronomyDecadalPlan2016–2025(AustralianAcademyofScience 2015) explicitly recognizes the importance of large aperture wide fieldMOS:“Building on Australia’s world-leading expertise in optical multi-object spectroscopy,development of an 8-metre class optical/IR wide-field spectroscopic survey telescopetowards the end of the decade would complement Australian optical astronomers’leadingpriorityofaccess toamulti-purpose8-metre classopticalobservatory. Suchafacility,most likelyconstructedbyan internationalpartnership,wouldaddress severalofthefundamentalastronomicalquestionsofthecomingdecade,includingthenatureofdarkmatteranddarkenergy,theformationandevolutionoftheMilkyWayandhowstarsandgalaxiesprocesschemicalelements.Itwouldalsoprovidefollow-upspectraofobjects identified by the SKA and imaging telescopes like the US-led Large SynopticSurveyTelescope(LSST).”

• [Canada] The Long Range Plan 2010 (Pritchet et al. 2010) notes that the 10m classtelescope equipped with an extremely multiplexed spectrograph would “…have atransformativeimpactinawiderangeoffields,includingstellarstructureandevolution,large-scalestructure,galaxyformationandevolution,darkmatter,AGNphysics,andtheepochofreionization.Furthermore,…[it]…wouldbeauniqueresourcefor follow-upspectroscopy, both for the European Gaia satellite mission, and also for LSST andEuclid/WFIRST”. The report goes on to describe the science case for such a facility as“unassailable”.Subsequently,theMid-TermReviewofLRP2010(Thackeretal.2016)commentsdirectlyonMSEandnotesthat“ThescientificandcollaborativeopportunitiesavailabletoMSE


partnersareadirectindicatorofthestrategicrelevanceofMSEtothefutureastronomylandscape.” They make the following recommendation “The MTRP stronglyrecommends that Canada develop the MSE project, and supports the efforts of theproject office to seek financial commitments from Canadian and partner institutesources.”

• [ESA, ESO] The Report by the ESA-ESO Working Group on Galactic Populations,ChemistryandDynamics(Turonetal.2008)madesomeprescientrecommendationsonthe need to improve wide-field spectroscopic capabilities worldwide, with an eyetowardsmaximizing the scientific legacyof theGaia dataset. These recommendationsinclude the development of dedicated, highly multiplexed, 4m- and 8m-classspectroscopictelescopes,notingthat``Ourtermsofreferenceweretoproposeasetofrecommendations toESAandESO foroptimizing theexploitationof their current andplannedmissions.However,theGalaxyisanall-skyobject;infact,fromtheground,theouter parts of the Galaxy are best observed from the Northern hemisphere, as theextinctionison-averagelowerthere.Inparallelwith[earlierrecommendations]thereisa real need for dedicated highly multiplexed spectrographs in the northernhemisphere".

Morerecently,theprioritisationoftheESOscienceprogrammeforthe2020shasbeencompleted. Highly multiplexed spectroscopy has received strong community backingdue to its broad importance for a vast range of science2. AWorking Group has beenestablished to investigate the science case and synergistic opportunities and practicalrequirementsforadedicatedground-basedwidefieldspectroscopicsurveytelescopeinthe2020s,thatisexpectedtoreportinmid-2016.

• [USA] The “Astro 2010 Decadal Review” (Blandford et al 2010) noted that ``Thepropertiesofdarkenergywouldbe inferredfromthemeasurementofboth itseffectsontheexpansionrateanditseffectsonthegrowthofstructure(thepatternofgalaxiesand galaxy clusters in the universe). In doing so it should be possible to measuredeviations from a cosmological constant larger than about a percent. Massivelymultiplexed spectrographs in intermediate-class and largeaperture ground-basedtelescopeswouldalsoplayanimportantrole."

Subsequently, the US National Research Council commissioned a study entitled“OptimizingtheUSGroundBasedOpticalandInfraredAstronomySystem”(Elmegreenet al. 2015) in which an official recommendation includes “The National ScienceFoundation should support the development of a wide-field, highly multiplexedspectroscopic capability on a medium- or large-aperture telescope in the SouthernHemisphere to enable a wide variety of science, including follow-up spectroscopy ofLarge Synoptic Survey Telescope targets. Examples of enabled science are studies ofcosmology,galaxyevolution,quasars,andtheMilkyWay”

2http://www.eso.org/public/about-eso/committees/stc/stc-85th/public/STC-551_Science_Priorities_at_ESO_85th_STC_Mtg_Public.pdf


MSEislocatedatarguablythepremiersiteontheplanetforground-basedopticalandinfraredastronomy,with access to the entire northern sky andmore thanhalf of the southern sky atreasonable airmass. The science presented herein reflects several years of effort by theMSEScience Team in the development of transformational science cases that demonstrate theunique, high impact and exceptionally diverse science enabled by large aperture, wide fieldMOS.MSEisdesignedtoprovideanaturalnextstepbeyondwhatisenvisionedforthecurrentand imminent generation ofMOS instruments. The “stand-alone” science potential ofMSE isawesome,butmoreover the strategic importanceofMSEwithin the international networkofastronomical facilities cannot beoverstated. This is reflectedby the strongbacking that largeaperture,widefieldMOShasontheinternationalscene,andforwhichMSEistherealizationofthatambition.


44

Chapter2 Theoriginofstars,stellarsystemsandthestellarpopulationsoftheGalaxy2.1 Chaptersynopsis

• ExoplanetaryhostcharacterizationintheeraofPLATO• Widefieldstellarspectroscopyinthetimedomain• PopulationstudiesofrarestellartypesacrossallGalacticcomponents• ThedetailedchemicalevolutionofstarsacrosstheentireluminosityrangeofGaiatargets• In-situchemicaltaggingofthedistantstellarhaloandtheinterfacesbetweencomponents• ThespatialdistributionofthefirststarsthroughanalysisofthemetalweakGalaxy• ThechemodynamicaldeconstructionofthenearestL*galaxy

Science Reference Observations for The Origins of Stars, Stellar systems and the StellarPopulationsoftheGalaxy• SRO-01:Exoplanetsandstellarvariability• SRO-02:Rarestellartypesandthemulti-objecttime-domain• SRO-03:MilkyWayarchaeologyandinsituchemicaltaggingoftheouterGalaxy• SRO-04:Unveilingcolddarkmattersubstructurewithprecisionstellarkinematics• SRO-05:ThechemodynamicaldeconstructionofLocalGroupgalaxies

MSE will produce vast spectroscopic datasets for in situ stars across all components of theGalaxyatallGalactocentricradii,acrossthefullmetallicitydistributionoftheGalaxy,andacrossthefullrangeofluminosityofGaiatargets.ExplorationofourGalaxyprovidesuswiththemostdetailedperspectivepossibleontheformationofplanetarysystems,theirhoststars,andtheirparent stellar populations. At the largest scale, the properties of the stellar populations –including their spatial distributions, energy, angular momenta, and chemical properties –provide uswith the fossil information necessary to decompose theMilkyWay galaxy into itsprimordialbuildingblocks, and in sodoing reveal thehistoryof its formationand the sitesofchemicalnucleosynthesis.

The key measurables are well-defined spectral line features for accurate stellar parameters(temperature,gravity,metallicity,rotationalvelocities,magneticfieldsstrengths,etc.)chemicalabundances,andradialvelocities.Suchadetailedperspectivenecessitatesthehighestsignal-to-noisedataobservedatgenerallyhighspectralresolution:atR=40000,syntheticspectrashowthat lines are deblended at~0.1 Å), rotational and radial velocities to better than~0.1 km/sthroughcross-correlationofmanylines(>100),andchemicalabundanceswithΔ[X/Fe]<0.15dex(0.1dex),whenSNR>30.Here,thecapabilitiesofMSEmakeitunmatchedamongstanyexistingorplannedfacilities.Further,anincreasingfractionofastronomicalresourcesarebeingdevotedto characterizing the variable universe.MSE, thanks to its large aperture,wide field and highmultiplexing, iswellpositioned topioneer the fieldofwide-field, largeaperture, timedomainoptical/IRspectroscopy.

Specifically,MSEwillprovidecompletecharacterizationathighresolutionandhighSNRofthe


45

faintendofthePLATOtargetdistribution(g~16)visiblefromthenorthernhemispheretoallowfor statistical analysis of the properties of planet-hosting stars as a function of stellar andchemical parameters. With high velocity accuracy and stability, time domain spectroscopicprograms will allow for highly complete, statistical studies of the prevalence of stellarmultiplicityintotheregimeofhotJupitersforthisandothersamplesandalsodirectlymeasurebinary fractions away from theenvironmentof the SolarNeighbourhood.Rare stellar types–including amenagerie of pulsating stars, but also halowhite dwarfs and solar twins –will beeasilyidentifiedwithinthedatasetofmillionsofstarsobservedperyearbyMSE,andwillallowforpopulationstudies that incorporateaccuratecompletenesscorrectionsandwhichexaminetheirdistributionwithintheGalaxy.

MSEwillbeunmatchedasachemicaltaggingexperiment.Recentworkinthisfieldhasstartedtorevealthedimensionalityofchemicalspaceandhasshownthepotentialforchemistrytobeusedinadditionto,orinsteadof,phasespace,torevealthestellarassociationsthatrepresentthe remnantsof thebuildingblocksof theGalaxy.MSEwillpush these techniques forward tohelprealizethe“NewGalaxy”,asoriginallyenvisionedbyFreeman&Bland-Hawthorn(2002).Inparticular,MSEwill decompose the outer regions of theGalaxy–where dynamical times arelong and whose chemistry is unaccessible from 4-m class facilities – into its constituent starformation events by accessing a range of chemical elements that sample a large number ofdifferent nucleosynthetic pathways. In so doing, MSE will provide a high resolution,homogenoussamplingofthemetallicitydistributionfunctionoftheGalaxy.Itwilltracevariationin themetallicitydistributionacrossall componentsandasa functionof radius.At themetal-weak end, [Fe/H] < -3, such an analysis reveals the distribution of the oldest stars in theUniverse,andisavitalcomponenttotracingthebuild-upoftheGalaxyattheearliesttimes.

MSEwillextendanalysisoftheresolvedstellarpopulationsoftheGalaxytotheLocalGroupandbeyond.VariablestarssuchasCepheidscanbespectroscopicallystudies throughout theLocalGrouputilizing the time-domainstrengthsofMSE,andmassivestarswillbeused to trace themetallicity gradients of disks in galaxies in the next nearest galaxy groups. Such studies willprobetheinside-outformationofgalaxydisksthroughindividualstellarmetallicitiesobservedatmoderatespectralresolution.M31andM33areparticularlyprominentextragalactictargetsforresolved stellar population analysis, and MSE will provide the definitive chemodynamicaldeconstruction of our nearest L* galaxy companion by a focused program targeting all giantbranch candidates observed inM31 out to approximately half the virial radius of the galaxy.SuchaspectroscopicpanoramaisunprecedentedandanaturalcomplementtothecoreMilkyWaysciencethatformsthedriversforMSEinthisfield.

MSEisuniquelysuitedtoprecisionstudiesofthestellarpopulationsoftheMilkyWay,andwillprovideatransformativeimpactinourunderstandingofthelinksbetweenstellarandplanetaryformation,thesitesofchemicalnucleosynthesis,stellarvariability,andthechemicalstructureoftheGalaxythroughinsituanalysisofstarsinallGalacticcomponents.ItwillpushthesestudiesouttothenearestgalaxiestoprovidepanoramiccomplementstothedrivingMilkyWayscience.MSEwill leveragethedatasetcreatedbyPLATOforexoplanetaryandstellarastrophysics,andwill provide the essential complement to theGaia space satellite by providing high resolutionchemicalabundancesacrosstheentireluminosityrangeofGaiatargets.


46

2.2 Thephysicsofstars:fromexoplanetarysystemstorarestellartypes

MSEisacriticalcomponentforstellarastrophysicsinthe2020sthatleveragesthedatasetsthatwill be obtained by the multitude of facilities described in Chapter 1. Amortised over itslifetimes,MSEwillcoverlargefractionsoftheGalacticvolumeandsurveymanymillionsofstarsper year. The sheer increase in sample sizewill create countless opportunities for – andwillmotivate original science questions in – stellar structure, stellar processes, and localenvironment.Thediscoveryofsignificantsamplesofrareobjectsthroughblindsurveys,e.g.,Li-rich stars, C-rich stars, solar twins and solar analogues, peculiar AGB stars, metal-line whitedwarfs,orevenpotentiallyfirststarremnants(e.g.,Ramerizetal.2010,Liuetal.2014,Carolloetal.2014,Datsonetal.2014,Vennetal.2014,Koesteretal.2014,Aokietal.2014)willhaveatransformativeimpactinmodelingtheirphenomena.Thisextendstoincreasingthesamplesizesofobjects innearbydwarfgalaxies,wheredifferences in theirenvironmentsandhistoriescanprovide importantconstraints for stellarnucleosynthesis,e.g.,progenitormassandmetallicitydependentyields fromAsymptoticGiantBranch (AGB)starsandsupernova (SNe)events (e.g.,Shetrone et al. 2003, Venn et al. 2004, Herwig 2005, Heger &Woosley 2010, Nomoto et al.2013,Frebeletal.2014,Schneideretal.2014).

In many ways, stellar astrophysics forms the backbone of modern astronomy. Stellarevolutionarytheory,guidedbymeasurementsforfundamentalparametersfromhigh-resolutionspectroscopy,underpinsourmodelsofgalaxyevolutionandcosmology.Furthermore,itisnowrecognized that the formationof planetary systems is closely connected to that of the stellarhosts,asevidencedbytheapparenthighefficiencyofplanetformationamongmetal-richstars.In the coming decade, a number of the key questions in stellar astrophysics will need to beaddressed using very large, systematic spectroscopic surveys conducted at high spectralresolution:i.e.,roughlyanorderofmagnitudehigherthanthatoftheSDSS(R~1800).MSE,asan ambitious,wide-area stellar survey,will uncover rare and important stellar types, such assolartwinsandfaint,metal-poorwhitedwarfsassociatedwiththeMilkyWaythickdiskorhalo.The detection and characterization of such objects using R ≥ 20000 optical spectroscopy isrequired to deepen out understanding of, e.g., planetary formation processes and theirdependenceonenvironment,thechronologyofdiskandhaloformation,andtheendstagesofstellarevolutionforlow-massstars.Targetedprogramswillexploretheevolutionandchemistryofmassivestarsinthelocalvolume,andMSEcanpioneerthefieldoflargeaperture,widefieldtime-domain stellar spectroscopy. The latter has the potential to improve dramatically ourunderstandingofstellarmultiplicity(includingtheinterconnectionsbetweencompanionsofallsorts,fromlow-massstars,tobrowndwarfsandexoplanets)andpulsating,eclipsingoreruptivestars.

In the fieldof stellarastrophysics, the transformational roleofMSE isbasedon threedistinctatributes:first,inthecollectionofspectraforstellarsamplesofunprecedentedsize,depthandresolution;second,inthespectralmonitoringoftimevariablesources;third,inaccessingstellarpopulations belonging to Galactic components that are difficult to access with currentspectroscopicinstrumentsduetolimitedapertureandmultiplexing.

The stellar populations that are most difficult to access using current spectroscopic surveysincludethoseintheouterhalo,outerdisk,andfaintstarsinthebulge.TheMSEhighresolution


47

mode(R≥20,000)canprovidehighprecisioninstellarparameters,radialvelocities,rotationalvelocities,chemicalabundances,andisotopicratios,whichareoftenneededfordetailedstellarmodelingandobservationaltesting.Forexample:

• Intheouterhalo,highresolutionspectraofuniquestars(e.g.,chemicallypeculiarstars,remnantsoffirststars,RRLyraestars,andhalodwarfs)canprovidenewinformationforstellar nucleosynthetic yields, stellar pulsation theory, and ages for halowhite dwarfs(e.g.,Ivansetal.2003,Aokietal.2013,2014,Yongetal.2013,Cohenetal.2013,Drakeetal.2013,Kilicetal.2005,Kalirai2012).

• Intheouterdisk,youngmetal-poorstarsfoundinstarformingregionscanbecomparedtothetypicalouterdiskfieldstars,providingvaluableinformationontheouterdiskstarformation environment (e.g., Friel et al. 2010, Frinchaboy et al. 2013, Carrero et al.2015).

• IntheGalacticbulge,starstendtobemoremetal-richthaninthesolarneighbourhoodand likely to have formed early, thus high resolution spectroscopy of the faint stars(dwarfs, and those in high extinction regions) provide unique constraints for stellarnucleosynthesis, stellar interiors, and stellar atmospheres at higher metallicities, inhigherdensityenvironments,andwithuniquechemicalcompositions(e.g.,Fulbrightetal.2006,2007,Bensbyetal.2010,2011,GarciaPerezetal.2013,Johnsonetal.2014;seealsothediscoveryofancientmetal-poorstarsinthebulgebyHowesetal.2015).

ThenatureandstructureoftheGalacticcomponentsreferredtoaboveisaprimaryfocusofthecomplementary Galactic Archaeological surveys discussed in Section 2.3, but the underlyingquestionsareuniquetostellarastrophysics:

1. IstheSununusualbecauseithostsaplanetarysystem?Detailedspectroscopicanalysesforahandfulof“solartwins”haveshownthattheirchemistrydiffers fromthatoftheSun, i.e., they contain higher fractions of refractory elements. In the case of the Sun,such elements may have been consumed by its surrounding planetary system (e.g.,Ramerizetal.2010).However, the interpretationof this result isdifficultbecause theeffects are subtle and there are alternative explanations for the observed abundancepatterns (e.g., chromospheric activity, stellar seismology, or even first ionizationpotentialeffects).Because fewer thana fewdozensolar twinsarepresentlyknown,alargersampleisessentialifwearetounderstandtheoriginofthisdifference;

2. Timedomainstellarspectroscopy.Alargenumberofimagingsurveysareexploringandcharacterizing the variable sky down to very faint magnitudes Eminent among theseendeavors isPLATO, thatwill surveynearlyhalf the skyandmonitor sourcesdown tog~16(withthecoresamplebeingbrighterthanthislimit).Spectralmonitoringofmanyofthesevariablesources,andmanyofthePLATOfields,willberequiredtounderstandtheirnatureandcanbeexpected to reveala richdiversityof interestingastrophysicalphenomena. Multi-epoch spectral monitoring over wide areas will in turn require alarge-aperture, highly-multiplexed spectrograph. Obvious applications in the stellarregime include spectroscopic follow-up of transit-selected planetary host candidates,binary stars, pulsating and/or eclipsing variables (see Chapter 4 for discussion ofextragalactictransientandvariablephenomena);

3. Howold is theGalactic halo?Whitedwarfs canbeused to age-date their host stellar


48

populationusingcoolingcurves.Unfortunately,theyaresofaintthatmostknownwhitedwarfsarerestrictedtothesolarneighbourhood.Ahandful(2–4)ofwhitedwarfshasbeen found with high proper motions, suggesting they are members of the Galactichalo,yetonly~30pcaway.Theagesof11–12Gyrmeasuredforthesestars(Kilicetal.2012,Bergeronetal.2005)areingoodagreementwithwhitedwarfcoolingcurveagesin theglobular clustersNGC6397andM4 (Hansenetal. 2004,2007).Nevertheless, amuch larger sample of more distant halo white dwarfs (with medium-resolutionspectroscopy) could be combined with detailed white dwarf model atmospherescalibrated using nearby objects (having high-resolution spectroscopy) to provide anindependentmeasurementoftheagedistributionfortheGalactichalo;

4. Where are theminority populations in the Galactic Bulge? Stars in the Galactic bulgetend tobemetal-rich,aswouldbeexpectedwhenstar formationoccurs rapidly (e.g.,Lecureur et al. 2007, Fulbright et al. 2007, McWilliam et al. 2008). However,microlensing studies toward the bulge (Bensby et al. 2010, 2011) and some recentspectroscopic surveys (Gonzalez et al. 2011) find some stars with lowermetallicities.Veryrecently,asampleofextremelymetalpoorstarshavebeenidentifiedinthebulge(Howes et al. 2015). Clearly, very large sample sizes are needed to identify andcharacterizethemostmetal-poorstellarpopulationsinthebulge;

5. How does environment impact the chemical evolution of disk galaxies?Precise stellarparametersformassivestarscanbedeterminedusinglow-resolutionspectroscopyandnarrowbandphotometricindices(e.g.,Kudritzkietal.2012;Firnstein&Przybilla2012).Because these stars are bright, they can be studied individually in galaxies out to~4Mpc: i.e.,at thedistanceof theSculptorandM81groups. Inprinciple,metallicityandreddeningmapscanbedeterminedforthedisksofstar-forminggalaxiesinordertotesttheeffectsofenvironmentondiskevolution.Whileindividualgalaxiescanbeexaminedwith present-day instruments, a comprehensive study will require homogeneousobservationsformanystarsbelongingtoalarge,statisticallyrobustsampleofgalaxies.

Inwhat follows,wedescribe a number ofMSE legacy science programs in the field of stellarastrophysics. Some of these programswill require datasets that cover a large fraction of theGalacticvolumeandthatsampleallmajorGalacticcomponents(includingthehalo,thethinandthick disk, and the bulge). Others require observations with specific cadence, and others aresmaller,targetedprogramsthatcouldcomplementlargerscalesurveys.

2.2.1 Exoplanetaryhostcharacterizationandtheirenvironmentsinthe2020s

ScienceReferenceObservation1(AppendixA)

Thecharacterizationandenvironmentsofexoplanethosts

MSE will provide spectroscopic characterization at high resolution (R~40000) and high SNR(~100)ofthefaintendofthePLATOtargetdistribution(mV~16),forstatisticalanalysisofthepropertiesofplanet-hostingstarsasafunctionofstellarandchemicalparameters.MSEisableto cover the full ~three-quarters of the PLATO sky (~15000 square degrees) accessible fromMaunakea in less than 800 hours. Such a survey will provide homogeneous spectral typing,metallicity estimates and chemical abundances for all PLATO targets at the faint end. Suchinformation will otherwise remain unavailable for most of this important sample. With high


49

velocity accuracy (~100ms-1) and stability, MSE will conduct a monitoring campaign of~100squaredegreesinthePLATOfootprint,potentiallyinthenorthernlong-durationfield.Thistimedomain spectroscopic program will directly measure binary fractions away from the SolarNeighbourhood andwill allow for a well characterized statistical studies of the prevalence ofstellarmultiplicity intotheregimeofhot Jupiters.Such information is importanttounderstandthephysicalmechanismsgoverning stellarmultiplicity, planetmigrationand their dependenceonhoststarproperties.

At the timeofwriting,3411planetshavebeenconfirmed in2550planetary systems3,usingavariety of methods (1) radial velocity searches; (2) photometric transits; (3) astrometricsearches; (4) direct imaging; (5) transit timing variations; (6) pulsar timing; (6) microlensingsurveys.Transitsurveysarenowresponsibleformorethanthreequartersofdetections,thanksin large part to the overwhelming success of the Kepler spacemission (Borucki et al. 2010).Differentsearchmethodsprobedifferentplanetaryparameters,andeachapproachhasitsownadvantagesandbiases.However,bycombiningresultsfromthevariousmethods,it ispossibletobuildupasurprisinglycompletepictureofplanetarysystems.For instance, radialvelocitiesprovideimportantconstraintsonplanetarymasses,whileplanetaryradiicanbemeasuredusingphotometrictransits.Thus,ithasbeenpossibleinrecentyearstolayimportantgroundworkintheburgeoningfieldofcomparativeplanetology.

Eventherelativelywell-characterizedpopulationofgiantplanetsshowsasurprisingdiversityinproperties,withorbitalconfigurationsofallkindsobserved:i.e.,veryshortorbitalperiods,largeeccentricities, large inclinations, etc. Thanks to Kepler, we now able to detect planets withmasses and radii notmuch larger than those of the Earth (and even identify some intriguingmultiple-planet systems containing rocky planets). In general, a planet’s fundamentalparameters (e.g., radius, mass, density and orbit) provide the basic data needed for acharacterization of extrasolar planets. These allow an assessment of the overall nature of agivenplanet (i.e.,whether it is agasgiant, aNeptune-like system,ora terrestrialplanet)andprovide important insights into its interior structure. A measured orbit also allows us toinvestigatethedynamicalevolutionoftheplanet,andtosetconstraintsonthecurrenttheoriesofplanetaryformation.

3Datafromexoplanets.eu,May25,2016


50

Figure 13: Top panel: schematic diagram showing the possible layout of the PLATO observing fields, relative toCoRoTandKepler,incelestialcoordinates,fromRaueretal.(2014).ThePLATOfieldofviewis48.5x48.5squaredegrees,anditwillobserveroughlyhalftheskytomV~16.TheareaabovethedashedlineiseasilyaccessibletoMSE(roughly¾ofthePLATOfields,or~15000squaredegrees).MSEcanobservethefaintendofthePLATOtargetdistributionathigh spectral resolution (R~40000) andveryhigh SNR (~100) in short exposures (<300s),withavelocity precision of order~100m/s. Possible observing strategies for MSE in the field of exoplanetary hostsinclude(a)coveringallofthePLATOtargetswithsingleexposures;(b)targetingtheKeplersurveyfields,whichwillstillhave limitedspectralcoverageat the faintend in the2020s. In the lower leftpanel,weshowthatthetotal∼115 square degree survey area can be covered, with high spatial completeness, in 84 MSE pointings; (c) amonitoringcampaignofasubsetofPLATOtargets.Inthelowerrightpanels,weshowapossiblelayoutof91MSEfields covering the central 136 square degrees of the northern PLATO long duration field (thatwould takeMSEabout8hourstoobserve).(FigureadaptedfromRaueretal.2014).

Majorthrustsofexoplanetresearchgivenrecentdiscoveriesinclude(butarenotlimitedto):

1. Surveyssensitivetoplanetsinthelowerpartofthemass–radiusdiagram,coveringanextendedrangeinorbitalperiod.AmajorgoalcontinuestobethediscoveryofanEarth-likeplanet inthehabitablezoneofaG-typestar.ForDopplersearches,thisrequiresaradial velocity precision of about 10 cm s−1 over a period of several years. For aphotometric detection, it requires a measurement of the transit signal of a few x0.001%.

2. Thedetectionofgiantplanetsbeyondthesnowline(3–5AU),includingthediscoveryofJoviananalogs.

3. Athoroughassessmentofhowthebulkpropertiesofplanetsdependonthepropertiesoftheirstellarhosts.Thisrequiresdetectionofplanets(andothercompanions)around


51

alargesampleofdiversestars,fromF-toM-typedwarfs,aswellasmoderatelyevolvedstars.

4. A determination of the uniqueness of our Solar System, and a comparison of itsproperties to those of planetary systems in different environments and at differentevolutionarystages.

Nosinglefacilityiscapableofaddressingalloftheissues.Forinstance,anumberofongoingorplannedsurveysare focusingon improveddetection limitsviaextendedtimecoverageand/orincreased radial velocity precision. At present, there are about fifteen instrumentsworldwidethatcandelivera radialvelocityaccuracyof~10ms−1orbetter.Somehavebeenmonitoringbright stellar samples formore thanadecade,and somecanachieveaprecisionof~1ms−1:i.e.,HARPS(ESO3.6m),HIRES(Keck10m),SOPHIE(OHP1.9m)andHARPS-N(TNG3.6m).Thus,thedomainofmassiveplanetswillberelativelywellexploredinadecade’stime.Forlower-massplanets (or those having long orbital periods), ESPRESSO is in active development. This is anultra-high-precisionradialvelocityspectrographfortheVLTsthataimstoachieveaprecisionofafewx10cms−1(Pepeetal.2010).

ForMSE,theradialvelocityprecisionattainableinhigh-resolutionmodeisoforder~100ms−1.While thiswill certainly not competewithdedicatedDoppler-search instruments,MSE’s largeaperture,widefieldandextrememultiplexingwillallowittomakeauniquecontributiontothequestionofhowexoplanetpropertiesdependonenvironment,arecurringthemeintheabovelist,andtoexaminethestatisticsofmassiveplanetarycompanions.

For illustration, we consider the roleMSEmight play in the era of PLATO. PLATO (PLanetaryTransitsandOscillation instars) isamedium-classESAmissionwitha targeted launchdateof2024. PLATO has an enormous field of view (2232 square degrees), a large dynamic range(observingstarsfrom4–16magnitude)anditwillobserveroughlyhalfthesky(seeaschematicdiagramof its possible sky coverage inFigure 13, adapted fromRauer et al. 2014). For thebrightest subsample, (4 – 11magnitude, estimated to be around 85000 stars), itwill provideaccurate planetary parameters, such as density, mass, radius and age, down to Earth-sizeobjects.Twolongpointings(highlightedinFigure13)areincludedintheobservingstrategytodetectandcharacterizeplanetsinthehabitablezonesofsolar-likestars.Roughly1millionstarswillbeobservedwithPLATOto16magnitude,andmassiveplanetswillstillbedetectedatthisfaintlimit,providingstatisticallyrobustsamplesforinvestigation.


52

Figure14: Illustrationof thepossible roleofMSE in characterizing the low-masscompanionsof solar-typestars.The panel on the left shows velocity induced by a companion plotted against (minimum) companionmass. Thedashedredcurvesshowthereflexvelocitiesexpectedforcompanionswithorbitalperiodsbetween1and105days,asindicatedinthefigure.Allcalculationsassumea1M¤primaryandanorbitaleccentricityofe=0.2.ThedashedhorizontallineshowstheexpectedtypicalprecisionofMSEradialvelocitymeasurementsforstarsbrighterthang~16.5.ThepanelontherightshowsthesensitivityofMSEintermsofcompanionmassandorbitalperiod.FigureadaptedfromGrether&Lineweaver(2006).

There are numerous, unique, observational strategies for MSE to complement PLATO, anddetailed analysis will be required to determine the optimal survey parameters. Figure 13showsaschematicdiagramofthePLATOskycoverage,ofwhichroughly15000squaredegreesisaccessibletoMSE(thenon-shadedareainFigure13).Whilethisisalargearea,thesetargetsareamongthebrightestthatMSEwillstudy,andonlyrelativelyshortexposures(<300seconds)are required inorder toobtainhighSNR (~100)athigh resolution (R~40000). Indeed, suchaprogramcouldevenbereservedforpoorconditions(e.g.,poorseeing,brighttime),suchistherelative easewithwhich it canbe conducted given the capabilities ofMSE. The entire PLATOsurveyareacouldbecoveredbyMSE in less than800hours, and thiswouldprovidedetailedstellarparametersand chemical abundance information forall PLATO targetsat the faintendwheresuchinformationwillotherwiseremainunavailableformostofthesample.

Alternatively,orinaddition,monitoringcampaignsofselectregionsofskywillprovideessentialtimedomainspectroscopicinformationtocomplementtheextensivephotometricmonitoringofthese stars. In this respect, it is worth highlighting that, even by the mid-2020s, thespectroscopicdataavailableforthefaintendoftheKeplersamplewillstillbelimitedduetotherelativelyfaintlimitofthissurvey(~17magnitude).ExcellentsynergiescurrentlyexistbetweenKepler and MOS programs on smaller facilities, with perhaps the best example being the


53

APOGEE – Kepler campaigns (see Fleming et al. 2015, Rodrigues et al. 2014, and referencestherein) and the same can be expectedwithMSE for fainter targets. The lower left panel ofFigure 13 showshowMSE could survey the entire Kepler region in~84pointings. Similarly,regions of the PLATO sky can be identified formonitoring byMSE: the lower right panels ofFigure 13 shows a zoom-in of the northern long duration PLATO field, and a possibleMSEsurvey configuration in its central parts is highlighted. The 91 MSE fields shown coverapproximately132 squaredegrees, and canbeobservedbyMSE in less than8hours.Repeatvisitstothesefields,perhapsmonthlyorwithavariablecadence,couldbeusedtodevelopanextensivecharacterizationofthetimevariablespectralcharacteristicsofthefaintPLATOstellarpopulations.

A survey of this sort would have a two-fold purpose, and the deep, time domainMOS dataprovided by MSE would be a unique contribution to ground-based support of theseexoplanetaryspacemissions.First,itwouldprovideforalargesub-sampleofPLATO(orKepler)targets,homogeneousspectraltypingandmetallicityestimatesforallstarsusingacommonsetof spectral features. There is emerging evidence that the correlation between planetaryfrequencyandhostmetallicity(seeSection2.2.3)exhibitedbymassiveplanetsmaynotholdforthe lowest-massplanets,sohomogeneousmetallicityestimatesforthehoststarsofplanetarysystems of all sorts, based on high-SNR, homogeneous, optical/IR spectroscopy,would be animportantdatasettocomplementtheextensivetransitobservations.Secondly,atthefaintendPLATOwillbesensitivetomassiveplanets,especially“hotJupiters”,whoseformationisthoughtto involve scatteringand/ormigrationprocesses.Thephysical interpretationof these systemswouldbenefit froma thorough characterizationof stellar companionsof all sorts— from theregimeof low-massstarsdowntobrowndwarfs.AlthoughMSE’svelocityprecisionwouldnotbe sufficient to detect exoplanets themselves, itwould bemore than adequate to perform acompletecensusoflow-massstellarcompanionsandbrowndwarfs.MSEshouldtakeadvantageofitsstabilityandhighqualitycalibrationstoconductlong-termmonitoringprogramsthatotherMOSinstrumentstomeasurestellarmultiplicityandaugmentthePLATOsciencegoals.Figure14 illustrates graphically the detection efficiency of MSE for these studies as a function ofcompanionmassandperiod.

Finally,theseexampleshaveconsideredthespecificcaseofthePLATOmission.However,thereare several other opportunities for MSE to provide valuable contributions to exoplanetaryresearchbyusingitsuniquesetofcapabilities.Forexample,amajorcomponentofexoplanetaryresearch will continue to be exoplanet detection via direct imaging studies. In such surveys,panoramic spectroscopy for millions of Galactic stars would be invaluable for optimizing thetargetselectionbyprovidingaccuratestellarparameters,suchaseffectivetemperature,surfacegravity, and also [Fe/H] or some elementary abundances of key elements; continued radialvelocitymonitoringwouldalsoallowapartofthecompanionpopulationsinthetargetstarstobeidentifiedandcharacterized,asdescribedabove.Aparticularlyexcitingapproachwouldbetoselectimagingtargetsusingacombinationofmulti-bandimagingandwide-fieldspectroscopy—toidentifystarsofvariousagesusingCaIIHandKstrength,Liabundances,rotationalvelocities,and/orUVWspacemotions.Combinedwithdirectimagingdataandradialvelocitymonitoring,thiswouldallowamorecompletecharacterizationofthetimeevolutionofplanetarysystems.


54

2.2.2 Stellarastrophysicswithtimedomainmulti-objectspectroscopy

Figure 15: Left panel: Composite colour-magnitude diagram for 11 124 RR Lyr variables of type RRab from theOGLE-IIIsurvey,whichcarriedoutan8yearmonitoringcampaignofa∼100squaredegreeregioninthedirectionoftheGalacticbulge.Aroughlyfive-foldincreaseinthenumberofbulgeRRLyrvariablesisexpectedfromtheVVVsurvey(Minnitietal.2010,Hempeletal.2014).FigurefromPietrukowiczetal.(2012).Rightpanel:ThedistributionofCepheidsinM31fromthePOMMEsurvey(PixelObservationsofM31withMEgacam).TheredhexagonsshowpossibleMSEfieldsforspectroscopicmonitoringcampaigns.NotethattheCepheiddistributionissimilartothatofthe brightest stars in M31, although there are small differences due to reddening variations, gas content andchangesinunderlyingstarformationhistory.FigureadaptedfromFliri&Vals-Gabaud(2012).

For thestudyofvariablestars— includingeclipsingandpulsatingvariables—theyields fromthenextgenerationof imagingtelescopes(Pan-STARRS,Skymapper,ATLAS,LSST,etc)promisetobeimmense.Asanindicationofwhatthefuturemayhold,theleftpanelofFigure15showsresults froma search for RR Lyrae variables basedon theOGLE-IIImicrolensing surveyof theGalacticbulge.Thissurvey,coveringatotalareaof~100deg2,discoveredmorethan16000RRLyrae variables of all types, producing phased light and colour curves for each object(Pietrukowiczetal.2012).TheVISTAVariablesintheViaLacteasurvey(VVV;Minnitietal.2010,Hempeletal.2014)isexpectedtoincreasethebulgeRRLyraesampleto40000–70000.Thegainsexpectedforothertypesofvariableswillbeequallydramatic;forexample,therightpanelofFigure15showresultsfromthePOMMEsurvey(Fliri&Vals-Gabaud2012)whichusedCFHTtomonitorthediskofM31overaperiodofabout2.5months.Thisfigureshowsthedistributionofmore than 2500 Cepheidswith periods in the range 2 to 80 days, highlighting the familiarring-likedistributionofmassive,brightstars inM31. Inshort, thecomingdecadeshouldseearevolutioninthestudyofvariablestars.

Cepheids,RRLyraesandothertypesofpulsatingstarsarenotonlycornerstonesofthecosmicdistance ladder, but also important testbeds for stellar evolutionary models. While there ismuch to be learned fromphotometry alone, the addition ofmulti-epoch spectroscopywouldmakepossibleanumberof investigations thathaveheretoforebeen impossible.For instance,RR Lyrae variables in the bulge regionhavemagnitudes in the range 15 ≤ I ≤ 18 and velocitysemi-amplitudesofA~20kms−1ormore,whileCepheidsinM31withperiodslongerthan3–4


55

dayshave19≤r≤23andA≥15kms−1.Thesemagnitudesandpulsationvelocitiesdemonstratethe importanceofspectroscopicmonitoringcampaignswithMSE: i.e., repeatedradialvelocitymeasurements, at a resolutionofR~ 6500 (i.e., εv~3 kms−1), forhundredsor thousandsofvariables in selected fields. Such datawill be combinedwith contemporaneous,multi-colour,photometric data to carry out Baade-Wesselink studies for stellar samples of unprecedentedsize. Because the Baade-Wesselinkmethod represents a direct route to themeasurement offundamental parameters including stellar radius and distance, such programs would haveobvious implications forourunderstandingof themechanisms thatdrivestellarpulsationandthe evolution of stars through the instability strip. In the case of M31, a number of otherattractivetargetsforspectroscopicmonitoringwouldalsobeavailableincludingbright,eclipsingbinaries(whichwouldprovidedirectmassmeasurementsforstarsattheupperendofthemainsequence)andnovae,wherepioneeringefforts inspectralclassificationbasedonahandfulofobjects in Local Group galaxies has revealed an unexpected diversity in spectral properties(Shafteretal.2012).

For the roleMSEmightplay in the studyofexplosive transients, includinghigh-z supernovae,seeChapter4.FormorediscussionofMSEandLocalGroupgalaxies,seeSection2.4.

ScienceReferenceObservation2(AppendixB)

Rarestellartypesandthemulti-objecttime-domain

TheroleofMSEforstellarastrophysicsisbasedonthreedistinctattributes:first,inthecollectionofspectraforstellarsamplesofunprecedentedsize,depthandresolution;second,inthespectralmonitoringoftimevariablesources;third,inaccessingstellarpopulationsbelongingtoGalacticcomponents that are difficult to access with current spectroscopic instruments due to limitedapertureandmultiplexing.Thesheerincreaseinsamplesizewillcreatecountlessopportunitiesfor – andwill motivate original science questions in – stellar structure, stellar processes, andlocal environment. The discovery of significant samples of rare objects through blind surveys,e.g., Li-rich stars, C-rich stars, solar twins and solar analogues, peculiar AGB stars,metal-linewhite dwarfs, or even potentially first star remnants will have a transformative impact inmodeling their phenomena, specifically by allowing for population studies that incorporateaccuratecompletenesscorrectionsandwhichexaminetheirdistributionwithintheGalaxy.Thisalsoextendstoincreasingthesamplesizesofobjectsinnearbydwarfgalaxies,wheredifferencesintheirenvironmentsandhistoriescanprovideimportantconstraintsforstellarnucleosynthesis,e.g. progenitor mass and metallicity dependent yields from AGB and SNe events. Here, weoutlineabaselinesetofprogramsandanalysesthatwillprovideatransformativelegacyforMSEinthebroadarenaofstellarastrophysics.

2.2.3 SolarTwins:TheSunandSolarSysteminContext

Solar-type stars are obvious targets in the search for exoplanetary systems. In recent years,possible connections between the “architecture” of planetary systems and the fundamentalparameters of their host stars, including their chemical makeup, have emerged as keyconstraintsonproposedmechanismsofplanetformation.


56

Figure16:MetallicitydistributionsfortheCORALIEsample(toppanels),theHARPSsample(middlepanels),andtheunion of the two samples (bottom panels), as presented in Sousa et al. (2011). The left panels shows the fullsamplesdistributionof [Fe/H],while thecenterplots foreachsampleshowstheplanethost [Fe/H]distribution.Therightplotsshowsthefrequencyofplanetsforeach[Fe/H]bin.FigurefromSousaetal.(2011).

It has been recognized for some time that the stars found to host planets in radial velocitysurveys–usuallyhighmass,giantplanets–tendtohaveratherhighmetallicities(e.g.,Gonzalez1997).Withtheever-increasingnumberofDoppler-selectedexoplanetarysystems,ithasbeenshownmany times that the shift inmetallicity—usually characterized by [Fe/H]—betweenplanet-hostingFGKdwarfstars,andthoseknownnottohostcloselyorbitinggiantplanets,isofhighstatisticalsignificance.Moststudiesagree(e.g.,Fischer&Valenti2005)that,inthemajorityofcases,themetal-richnatureofthegiant-planet-hostingstarsisintrinsictothestaritself.Thissuggests thatgiantplanetsarepreferentially formed inmetal-richenvironments. Indeed,dataon582 low-activityFGKstars intheHARPSvolume-limitedsampleconfirmthatplanet-hostingstars tendtobeenhanced inmetals (Figure 16). Intriguingly, this trend isstrongest forstarswithgiantplanets;starswhichhostJovian-massplanetstendtobemoremetal-richthanthosestarswhichhaveonlyNeptunian-massplanetsorsmaller(e.g.,Ghezzietal.2010;Buchhaveetal.2012).

Recent results measuring the metallicities of a large number of stars hosting small planetssuggestthattheplanetscanbedividedintothreeregimes,namelyterrestrialplanets,gas-dwarfplanets(smallplanetswithlowermeandensities),andgas-giants(Buchhaveetal.2014)wheretheplanetsintheterrestrialplanetregimehaveametallicityconsistentwithsolar.ThishasbeenconfirmedbyananalysisbyBuchhave&Latham(2015),whoanalyseasampleofstarshostingtransiting terrestrial planets and compare them to an equivalent sample without transitingplanets.Theyfindnostatisticaldifferencebetweentheaveragemetallicityofthetwosamples.Theseresultssuggeststhatmetallicityplaysarolenotjustintheformationofgiantplanets,butalsointhedistributionofplanetarymasseswithinexosolarsystems.


57

Additional,intriguing,chemicalabundancetrendshavebeennotedinstarsthatmayberelatedtothepresenceofplanets.Usingatechniqueinwhichspectrallinesin“solartwins”(i.e.,starshaving stellar parameters similar to those of the Sun) are compared in a careful differentialabundanceanalysis,Melendezetal. (2009) foundthat thechemicalcompositionof theSun isanomalous with respect to most (85%) of the solar twins. Compared with its twins, the Sunexhibitsadeficiencyofrefractoryelements(i.e.,thosewithhighcondensationtemperatures,Tc)relativetovolatileelements(thosewithlowTc).Thisfindingmaybeasignthatplanetformationoccurred more efficiently around the sun compared with the majority of solar twins.Furthermore,within thecontextof this scenario, it seems likely that theobservedabundancepatterns are specifically related to the formation of terrestrial planets: i.e., while refractorymaterialismissinginthesun,thosedust-formingelementsareoverabundantinmeteoritesandterrestrialplanets.

Ifthissuppositioniscorrect,theconsequencesareexciting:itpointsthewaytousingchemicalabundance distributions as diagnostics of inner, terrestrial-type planetary architectures. If thechemicalsignaturefoundintheSunis indeedduetotheformationofterrestrialplanets,thenthestudyofmetal-richsolaranalogsandF-typedwarfscanprovide furtherclues toplanetaryformationmechanisms(Ramirezetal.2010).Inshort,exploringthepossibilitythatplanetmassmay be correlated with the chemical composition of the stellar host, using a large andhomogeneousstellarsample,isofutmostimportance.

Thesestudieshavedemonstratedthattherearesubtle,butfascinating,patternsinthechemicalabundancedistributionsof stars hosting exoplanets. Suchpatternshavepotential significancebeyond a simple empirical shift towards higher metallicities. Probing the trends in chemicalabundance distributions requires an ability to derive stellar parameters and chemicalabundances for many different elements in a very large sample of solar-type stars. With asample of~100000 or more FGK main-sequence stars, the intrinsic dispersions of particularchemical familiescouldbemeasured.Thiswouldmake itpossible toestablishwhetherornotthepeculiarsignaturesseeninthesmallnumbersofplanet-hostingandnon-planet-hostingstarsto-date are found in a general census of solar-type main-sequence stars. In particular, theavailabilityoftargetsfromPLATO(aswellastheKeplerfields) isaresourcetobeexploitedbyMSE.AtaresolutionofR≥20000,awell-calibrated,uniformspectraldatasetcouldbeusedtoextract accurate chemical abundance distributions of awide variety of elements, covering allnucleosyntheticoriginsandchemicalbehaviours.

2.2.4 WhiteDwarfsasProbesofGalacticFormationandEvolution


58

Figure17:Toppanel:BescanonGalacticmodelsimulationofthenumberofwhitedwarfsdetectableper100squaredegreesintheCFHTLegacyfortheu-bandall-skyuniverse(Luau)survey,comparedtoSDSS.Inthenearfuture,wecanexpect thephotometric sampleofwhitedwarfcandidates to increasesignificantly,particularly for themoredistant Galaxy. Bottom panel: BescanonGalacticmodel of the cumulative luminosity function of Galacticwhitedwarfs.Thereareapproximately170000whitedwarfsbrighterthang~20.4andatdeclinationsgreaterthan-30

degreesaccessiblebyMSE.

Prior to the SDSS, only~2500 white dwarfs were known. The SDSS DR4 catalogue of whitedwarfs (Eisenstein et al. 2006) more than doubled this number, all with spectroscopicconfirmations.ThegrowthoftheSDSSspectroscopicwhitedwarfsamplehascontinued(DR7–Kleinmanetal.2013;DR10–Kepleretal.2015a),andincludingthemostrecentsamplebasedonSDSSDR12(Kepleretal.2015b)nowhasmorethan30000members.TheSDSScatalogueshavebeenusedforstatisticalstudiesofwhitedwarfstellarphysicsandparameters(e.g.,Kepleretal.2007;Tremblayetal.2011)aswellasforstudiesoftheagesoftheGalacticthinandthickdisks(e.g.,Harrisetal.2006).Thecatalogueincludesraresub-classesofmagneticandpulsatingwhite dwarfs, and white dwarfs in binary systems (which have also been identified by theUKIDSS survey, Steele et al. 2011). These latter systems can be further studied using time-resolved spectroscopy in order to identify compact, post common-envelope binaries: theseobjects—whicharetheprecursorsofcataclysmicvariables,X-raybinariesandpossiblySNeIa— allow the various formation mechanisms proposed for these systems to be carefullyexamined.

Atpresent, SDSScatalogues contain fewwhitedwarfsbelonging to theGalactichalo (i.e., theSDSS “20 pc” sample is estimated to be only 80% complete, Holberg et al. 2008). However,analysesof four, inner-halowhitedwarfsmoving throughthesolarneighbourhood (Kilicetal.2012; Kalirai 2012) yield ages in good agreement with those of old stellar clusters havingmeasuredwhitedwarfages(e.g.,NGC6397andM4;Hansonetal.2007,2004).Thisisanareawhere MSE can be expected to have a profound impact on white dwarf studies: MSE canidentifynewwhitedwarfsintheGalactichalo,whicharecriticalforimprovedageestimates,aswell as the discovery of additional rare objects, e.g., white dwarfs polluted by accretion ofterrestrialobjects,magneticstars,etc.

Themeasurementofprecisestellarparametersforwhitedwarfs isonlypossiblebycombiningphotometry, from surveys like SDSS, Gaia, Pan-STARRS or LSST, with follow-up spectroscopy.Recently, a new survey initiated at CFHT, called the “Legacy for the u-band all-sky Universe”(Luau)issurveyingalargefractionofthenorthernhemispheretoaubanddepthconsiderablydeeperthanSDSS(u~24.5),thatwillbeanidealdatasetforidentifyingwhitedwarfcandidates


59

outtomuchlargerradiusthanSDSS,includingahigherfractionofGalactichalodwarfs(seethetoppanelofFigure17).

MSEisthedefactoproviderofspectroscopyforthisandanyotherwhitedwarfsample,givenitsmuchdeeper limitingmagnitudethanSDSS,Gaia,or2–4mclasstelescopes ingeneral. Itwillplay an invaluable role by providing good SNR spectra (15 – 50was shown tobe ideal in theanalysis of SDSS spectra by Tremblay et al. 2011) for a volume-limited sample: e.g., within adistance of ~100 pc. Indeed, the bottom panel of Figure 17 illustrates the possibility fortransformationalgainswithMSErelativetoSDSSforwhitedwarfstudies,potentiallythroughaprogramsuchas thatdescribed inSRO-02.Thedashedcurveshowsthepredicted,cumulativenumber of white dwarfs down to g~21 for declinations δ ≥ −30◦. Down to this limitingmagnitude,weexpectroughly~8whitedwarfsperMSEfield,averagedoverthis~3×104deg2region.Forasubsetofthenewlydiscoveredwhitedwarfs,propermotionsanddistancesfromGaiawould further improve the precision of themeasured stellar parameters. Inmost cases,intermediate-resolutionspectroscopy,R=6500,issufficient.Indeed,spectraatR=2000wouldbeadequate formostwhitedwarfs (i.e.,DAandDB)where log g~8 and theHandHe linesused for stellar parameter determinations are very broad. Higher resolution would beparticularlyinterestingforothersubtypes.Forexample,thosewithmetallines,whichmayhavebeenpollutedbytheaccretionofplanets,couldbeusedtodeterminethechemicalcompositionoftheformerrockyplanets.Needlesstosay,itwouldalsobepossibletoanalysethosesystemswithmagneticfieldstoalevelofprecisionthatisimpossiblewithSDSS.

2.2.5 Fundamentalparametersforhighmassstars:Thechemicalevolutionofdiskgalaxies

Figure18: Leftpanel:HST/ACScolour-magnitudediagramused toselectbluesupergiant starcandidates inM81.Rightpanel:LocationofthebluesupergiantstartargetsinM81.FigurefromKudritzkietal.(2012).

Bothhigh-andlow-resolutionspectroscopicsurveyscanplayanimportantroleinthederivationof fundamental parameters for massive stars. Detailed stellar parameters and chemical


60

abundances for massive, hot stars can be determined with exquisite precision from high-resolutionspectroscopy(e.g.,Firnstein&Przybilla2012),whilemedium-resolutionspectroscopycanbeusedtomeasurebasicparametersandmetallicities.Forexample,Kudritzkietal.(2012)have shown that Keck/LRIS spectroscopy (R ~ 2000) can be used to measure reliabletemperaturesandluminositiesforhot,massivestarsbyfittingtheBalmerjumpandBalmerlineprofiles. Because these stars are intrinsically bright (i.e.,MV > −6),medium-resolution spectracanbeusedtoexaminestars indistantgalaxies(e.g.,starforminggalaxies intheSculptorandM81groups).Spectral templatescanthenbeusedtoestimatemetallicities in individual stars,andthusexaminedirectlythemetallicitygradientsofdiskgalaxies.

Figure 19: Left panel: Comparison of metallicities in M81 deduced from blue supergiant stars with planetarynebulaeoxygenabundances (blue and red, respectively). Rightpanel: Comparisonofmetallicitiesdeduced frombluesupergiantstarswithoxygenabundancesderivedfromHII regions (blueandred, respectively).Therandomuncertaintiesinthesemeasurementsare∼0.1–0.2dex.FigurefromKudritzkietal.(2012).

Inthesimplestpictureofdiskformation,inwhichadiskformsthrough“insideout”gravitationalcollapse, one expects the metallicity to decrease as a function of Galactocentric radius (seeChengetal.2012).Ontheotherhand, theexistenceof radial flows (e.g.,Spitoni&Matteucci2011)orthepresenceofaturbulentdiskatearlytimes(Brooksetal.2009;Dekeletal.2009)couldalterthispicturesignificantly.Strongradialmigrationmechanismsmightalsowashoutanexistinggradientinthestars(Roskaretal.2008;Schoenrich&Binney2009),whileradialbreaksmay indicate either dynamical interactionswithin disks, ormerger events in outer disks (e.g.,Pohlen & Trujillo 2006; Younger et al. 2007; Bigiel et al. 2010). The measurement of diskmetallicitygradientsfromindividualstarsmeansnotonlythatcoverageismorehomogeneousthan would be possible with H II regions, but it ensures that metallicity and reddening aresampled together, rather than relyingon justoxygenabundances.Needless to say,metallicitymeasurements in a large and diverse sample of galaxies beyond theMilkyWay will providecrucial environmental information that can be used to disentangle competing disk formationscenarios.

ToillustratethetransformationalrolethatMSEcouldplayinthisfield,weconsiderthecaseofM81,thenearbySabgalaxyatadistanceof~3.5Mpc.Inthisgalaxy,starsasfaintasV~21.5havebeenobservedwithKeck/LRISatSNR~50,andthemetallicitygradientswerefoundtobeconsistent with the oxygen abundance gradients from HII regions, with comparable


61

uncertainties(seeFigure18).ItisinterestingthatthegradientinM81seemstohaveevolvedoverthelast~5Gyr,astheoxygengradientderivedfromPNeissteeper(seeFigure19).WithMSE,adedicatedsurveyofnearbydiskgalaxieswouldbepossible,includinggalaxiesspanningarange ingroupand fieldenvironments.Previously, thiskindofworkwas restricted toUVandhigh-resolutionspectroscopicanalyses;newtechniques,combinedwithawealthofhigh-qualityspectra fromMSE,wouldallowanambitious,comprehensive,surveytobemountedfromtheground.

WithMSE,high-resolution spectroscopy couldbeused toprovidea large statistical sampleofchemistriesandrotationratesofmassivestars.Thesedatawouldbeusefulinaddressingissuesrelated to rotational mixing (e.g., Maeder & Meynet 2010; Ekstrom et al. 2012). Since theinclinationangle forany individualstar is rarelyknown,a largestatisticalsample isoneof thebest ways to look for correlations between stellar line strengths/chemical abundances andstellar rotation rate (e.g., Hunter et al. 2009; Brott et al. 2011). Rotation is a fundamentalparameter in the analysis of massive stars since it is thought to affect core masses, andtherefore, ages, of these stars. In addition, rotation drives the mixing of nuclear processedmaterial from a stellar interior to its surface, and it is thus one of the key observationalconstraints instellarastrophysics (see,e.g.,Vennetal.2002).Forverymassivestars, rotationrates are thought to affect Eddington luminosities, and thereforemass loss rates aswell, andultimatelytoleadtothedevelopmentofWolf-Rayetstarsandluminousbluevariables.Thereisalso somespeculation that rotation is affectedby stellarmetallicity (Penny&Gies2009), andthattheseparameterscombinetoinfluencestellarevolutionpredictions,includingevolutionarymasses. A large statistical sample of measured chemical abundances and rotation rates formassive, bright stars inM31,M33, and other LocalGroup, northern hemisphere star-forminggalaxieswouldbetheobviouswayinwhichtotestthisprediction.ThesestarstypicallyhaveV≤18andarethuseasilywithinreachofMSEinitshighresolutionmode.

Ofcourse,rotationisalsoimportantforthe(mainsequence)evolutionoflowermassstars,asitaffects directly convective mixing efficiencies and, ultimately, the interpretation of surfacechemistry measurements (including the lithium isotopic abundances that are compared toStandardBigBangpredictions;Asplundet al. 2006; Sudaet al. 2011).With a large sampleofaccurate,homogeneousrotationratesmeasuredformanythousandsofbrightstarswithMSEinhigh-resolutionmode,itwouldbepossibletostudytheseeffectsstatistically.

2.3 ThechemodynamicalevolutionoftheMilkyWay

ScienceReferenceObservation3(AppendixC)

MilkyWayarchaeologyandtheinsituchemicaltaggingoftheouterGalaxy

MSEwillcarryouttheultimatespectroscopicfollowupoftheGaiamission.NootherplannedorproposedsurveywillbeabletochemicaltagstarsattheoutskirtsoftheMilkyWayanddowntothefaintestGaiastars.OurviewoftheMilkyWaywillbecompletelyrevisitedimminentlywiththefirstreleaseoftheGaiadata.MSEisdesignedtohaveallthekeyelementstobetheessentialspectroscopic tool for Galactic Archaeology in the post-Gaia area, able to probe the GalaxythroughchemicalanddynamicalstudiesofindividualstarsacrossallGalacticcomponentsatallradii in theGalaxy,andthroughabsorption linestudiesof the intervening interstellarmedium.


62

We highlight here the main questions that we currently know MSE will uniquely be able toanswer as well as the essential elements for its implementation. This includes extensivediscussion of trades between spectral resolution andwavelength coverage tomaximizeMSE’sutilitytoefficientlyprobechemicalspace.

2.3.1 Gaiaandthenearfieldofgalaxyformation

Figure 20: Projected darkmatter distribution of a suite of high resolution simulations ofMilkyWaymass darkmatterhalosat z=0 from theCaterpillarProject. Theunderlying cosmology isbasedon results fromPlanck (ThePlanck Collaboration et al. 2014), and candidate halos were originally selected for resimulation based on asimulationofamuchlargervolume.Boxwidthineachpanelis1Mpc,andthewhitecirclesmarkthevirialradius.FigurefromGriffenetal.(2016).

Themeasurementof the fluctuations in the cosmicmicrowavebackground (CMB)allowus todeterminekeycosmologicalparameters,includingthetotaldensity,matterdensityandbaryonicdensityoftheuniverse.ThepicturethatemergesisofaUniversedominatedbyacosmologicalconstant,withamassbudgetlargelydominatedbynon-baryonicmatter(Komatsuetal.2011).Theoverall success of this LambdaColdDarkMatter (ΛCDM) cosmological paradigmon largescales is impressive. By following cosmological simulations of the growth of the primordialfluctuations that gave rise to the overdensities where structures form, ΛCDM is now able tomake predictions on smaller (i.e., galactic) scales. Figure 20 shows a suite of high resolutionsimulationsofthestructureofdarkmatteronthescaleofgalaxiesthemassoftheMilkyWay,wheretheunderlyingcosmologyisbasedonresultsfromThePlanckCollaborationetal.(2014).It isonthesescalesthatobservationsandtheoryshowimportantdiscrepancies.Thedetailsofgalaxy formation and evolution within this framework are not well understood due to thecomplexities of baryonic physics, and so it is on these small scales that the prevailing


63

cosmologicalparadigmfacesitstoughestchallenges.

Advancing our understanding of the galaxy formation process in general requires detailedobservations of individual systems, and, naturally, our Galaxy provides the most importantlaboratory for exploring chemodynamical evolution on these scales. Indeed, theMilkyWay isnowrecognisedasbeingacomplexensembleofstellarpopulations,eachhavingcharacteristicageandmetallicitydistributions,anduniquedynamicalproperties.

Eggenetal.(1962)werethefirsttoshowthatstellarabundancesandkinematicscouldbeusedto understand the formation and evolution of our Galaxy, and thus guide ideas of galaxyformation in general. Indeed, their analysis of 221 very nearby stars remains, arguably, thesinglemostinfluentialobservationalpaperongalaxyformation.Init,theyproposedthatmetal-poorstarsinthehaloofthegalaxywereformedduringtherapidcollapseoftheprotocloudthateventuallybecametheMilkyWay.AnalternativeproposalwasofferedbySeale&Zinn(1978),whoseanalysisof the stellarpopulationsofanumberofGalacticglobular cluster led themtoinferthattheywereformedinindependent“protogalacticfragments”thatlaterassembledtheouter parts of the Galaxy. The current paradigm — whereby gas collapses in dark matterstructuresthat formdisksandeventuallymergetogetherorareaccretedby largersystems—containskeyelementsofbothoftheseoriginalscenarios.

Figure21:SimulationofthedistributionoffieldredgiantbranchstarsontheskyatvariousdistancesfromtheSunfor the stellar halo of the simulation Aq-A. The different colours correspond to stars originating in differentprogenitors.Differentprogenitorscannotbedistinguishedbetween10and30kpc inphotometricsurveysalone,and velocities and chemical abundances are essential to reveal the individual events. Figure from Helmi et al.(2011).

These seminal papers demonstrated that information relating to the chemical and dynamicalcharacteristics of the early protogalaxy, its constituent building blocks, and the subsequentevolutionary processes that have produced the Milky Way (and its subcomponents) can beexplored through chemodynamical studies of modern-day stellar populations. Contemporarystudiesof thedynamicsofstars in theGalaxyshowthis isapproach ispowerful:Helmi (2001)and others showed that stars that were accreted into theMilkyWay from the samemergereventcanstillbeidentifiedthroughtheirenergiesandangularmomenta,eveniftheynolongerformanobviousstellarstreamacrossthesky(incontrasttocoherent,present-dayfeaturessuchastheSagittariusandOrphanstreams,andotherprominentsubstructures).Figure21showsarealizationof theexpectedstellardistributionofa“MilkyWay”galaxyhalofromtheAquariusconsortium.Dependingon thedistance regimebeingexamined, it is clear thatunraveling theindividual events that contributed to the formation of this component requires informationbeyondthethreespatialcoordinates.Extendingtheseconceptsintochemicalspace,Freeman&Bland-Hawthorne(2002)proposedidentifyingstarsthatwerebornfromthesamegascloudbythemeasurementof abundances formany chemical species, and “tagging” thosewith similar


64

abundancepatternsi.e.,identificationofGalactic“buildingblocks”bychemistryalone.

The chemical signatures, energies and angularmomenta of the individual stars in the Galaxythusencode its formationandevolutionaryhistory,andcanprovideuswith themostpreciseunderstandingoftheprocessesthroughwhichtheMilkyWaywasassembled.Incontrasttotheearly study of Eggen et al. (1962), MSE will provide these parameters for millions of starsthroughouteverycomponentandsubcomponentoftheGalaxy,andcontributetothecreationofthemostdetaileddataseteverassembledforasinglegalaxy.

TheanalysisofEggenetal. (1962)wasbasedontheSolarneighbourhood; thehalostars thatwereobservedarethatverysmallsubsetthathappentobeatapointintheirorbitwheretheyare passing close to the Sun. More than 50 years later, the high resolution large scalespectroscopicsurveysthatexistorareimminent–inparticularAAT/HERMESandVISTA/4MOST–aresimilarlyprobingtheverynearbyUniversewherethethindiskcomponentisdominant.Acritical component of Galactic science with MSE in comparison to every other existing orproposedspectroscopic facility is theabilitytoaccessthedetailedchemodynamicalsignaturesofeveryGalacticcomponentandsub-componentthroughin-situanalysisofindividualstars.

ThecapabilitiesofMSEinthisfieldareperfectlymatchedtothecapabilitiesofGaia,alandmarkastrometricspacemissionthathasasitsprimaryfocusadetailedunderstandingofthestructureandcompositionofourGalaxy.Launchedin2014,ithasbegunits5yearmissiontoproduceathree-dimensional stellar map of the Milky Way with unprecedented precision. Gaia isconductinganall-skysurveytomeasurethepositionsof roughly1billionstars,approximately1%oftheentirestellarcontentoftheMilkyWay.Infact,allastronomicalpointsourcesbrighterthan V � 20 mag will be catalogued, and objects brighter than V~15 mag will have theirposition measured to better than 20 microarcseconds. Beyond astrometry, Gaia will obtainmulti-band photometry for all sources, and it is additionally equipped with a Radial VelocitySpectrometer (RVS) thatwillmeasurevelocities forobjectsbrighter thanV~17mag (roughly150 million stars) to an accuracy of 1 – 10kms-1. Basic astrophysical information – includinginterstellar reddening and atmospheric parameters – will be acquired for the brightest~5millionstars.Chemicalabundanceinformationwillalsobeprovidedforafewelements(i.e.,Mg,Si,Ca,TiandFeforstarsofspectraltypeFGK)forstarsbrighterthanG=12.


65

Figure22: (a)ExpectedradialvelocityaccuracyfortheRVS instrumentaboardGaiaforstarsofdifferentspectraltypes.ThelowerblackcurveshowstheradialvelocityaccuracyexpectedforaG2Vstarobservedforatotalof1hourwithMSEinhigh-resolutionmode.(b)Typicalerrorsintransversevelocityasafunctionofdistanceforstarsofdifferent spectral typeobservedwithGaia.The shadedgrey region shows theapproximateerrors calculated forG2Vstarswithg=20.5observedwithMSE.

Theimportanceofground-basedspectroscopytosupplementtheGaiadataiswellrecognizedintheinternationalcommunity,particularlyasitrelatestothechemicaltaggingofindividualstars(Section 2.3.2). For chemical evolution studies,Gaiawill identify hundredsofmillions of starsfainter than G~17 for which the RVS cannot provide any spectral information. Indeed, theastrometricaccuracyofGaiaisonlymatchedbysimilarlyaccurateradialvelocitiesfortheverybrightest subset of its targets. For relatively faint stars, MSE is required in order to provideprecise radial velocities to complement the precise transverse velocities measured by Gaia.Figure 22 illustrates this with a comparison of the anticipated velocity accuracy of MSEcomparedwith the RVS on Gaia. Clearly, the kinematic data provided byMSE is an essentialcomplement to the astrometric data from Gaia across the entire magnitude range of Gaiasources.


66

Figure23:DifferentialstarcountsasafunctionofmagnitudeforthethreemainGalacticcomponents,basedontheBescanconmodeloftheGalaxy(Robinetal.2003,A&A,409,523)fora100squaredegreeregioninthevicinityofthenorthGalacticcap.Theshadedregionindicatesthemagnituderangeaccessibleathighresolutionto4mclassspectrographs(typicallyoperatingatR~20000–40000).MSEistheonlyfacilityabletoaccessthethickdiskandspheroidathighresolutionintheregionsoftheGalaxyinwhichtheyarethedominantcomponents.MSEtargetsathighresolutionspanthefullluminosityrangeoftargetsthatwillbeidentifiedwithGaia.

ThegraspofMSEforexplorationoftheGalaxy isawesome.Figure 23 showsthedifferentialstellarnumberdensityofeachof themain componentsof theGalaxy, as calculated from theBesanconmodel(Robinetal.2003,A&A,409,523)fora100squaredegreeregionaroundthenorthGalacticcap. In themagnituderegimethat isbestaccessibleathighresolutionwith4mclassspectrographs(shadedareaofFigure23),thethindiskisthedominantcomponent:evenattheGalacticpole,thindiskstarsout-numberhalostarsintherange14≤g≤16byafactorof15:1,andthisratioincreasesdramaticallywithdecreasing|b|.Forexample,AAT/HERMES(thathasabrightermagnitude limitofV=14)estimate thatonly0.2%of their targetswillbehalostars4.Fainterthang≥16,thethickdiskbecomesthedominantcomponentathigherlatitudesandiseasilyaccessiblewithMSEevenwithminimalpre-selectionoftargets.Asan11maperturefacility,MSEwillobtaingoodSNRathighresolution inreasonableexposuretimeseven in themagnituderangewherethestellarhaloisdominant(i.e.,g≥19.5athigherlatitudes).

4http://www.mso.anu.edu.au/galah/survey_design.html#selection_criteria


67

Figure 24: Distance range probed as a function of magnitude by using two different stellar tracers: horizontalBranchstars(inred)probetogreaterdistancesatafixedmagnitudethanturn-offstars(inblue).Notethattipoftheredgiantbranchstars(notshown)probebeyondtheMilkyWayvirialradius.Theshadedgreyareacorrespondsto themagnitude range covered by the 4m-class spectroscopic surveys, and is therefore not a priority forMSE.HorizontaldashedlinesindicatethedistancethresholdrequiredtoprobesomeprominentGalacticstructuresandsatellites.ExtinctionshiftstheredandbluelinestotherightasillustratedwiththeAV=2magpresentedhere.Foreachcoloredline,thedottedsectionrepresentsthedistance–magnituderangewhereGaiaparallaxaccuracieswillbebetter than20%.Thesolid linethereforerepresents thedistance–magnituderangewheredistancesderivedthroughspectroscopywillbeessentialtocomplementGaia.

Figure24showstheeffectivedistancerangeprobedbyMSEasitextendsdownthemagnituderange of Gaia sources. Here, distance as a function of (Gaia) magnitude is shown for twopossible stellar tracerpopulations, amain sequence turn-off star (blue lines) andahorizontalbranchgiant star (red lines). The twodifferent linesper tracer indicate theeffectofdifferentassumptions regardingGalacticextinction.Horizontaldashed lines correspond toapproximatedistance thresholds for probing various stellar components and/or satellites. The dottedsegmentsofthecoloredlinesindicatetheregimesinwhichGaiaparallaxeswillhavebetterthan20%accuracy.Beyondthis,spectroscopywillprovideimportantconstraintsonthedistancesofthetargets.

ThepotentialroleforMSEinfurtheringourunderstandingoftheGalaxyinthepost-Gaiaeraisextensiveandcoversallcomponentsandsub-componentsoftheGalaxy,anditisnottheintentofthisdocumenttodetailallpossiblescienceprograms.Representativesubjectsinclude:

• TheInterstellarMediumandcarriersofDiffuseInterstellarBands


68

MSEwill be theonly instrumentable toobserveathigh resolutionmillionsofobjectswith known distance, allowing an outstanding step forward in the knowledge of themulti-phasegalacticISM.

• TheGalacticdisko Theouterdisk

Extragalacticobservationshaverevealedallthecomplexityoftheouterregionsof galaxy disks: truncated/anti-truncated surface brightness profiles, breaks inmetallicityprofiles,U-shapedageprofiles,complexstarformationhistories,etc.FortheMilkyWay,itsouterregionsarestill largelyunknown.MSEwillbeableto reach starsbeyond theedgeof theouterdisk. Itwill quantify itsboundaryandsubstructures,provideadetailedchemicaldescriptionoftheouterdisk,andtraceitsformationhistoryanditslinkwiththeinnerdisk

o DiscdynamicsMSEiscrucialtomapthe3dimensionalvelocityfieldbeyondtheextendedsolarneighbourhood.Adetailed3dimensionalmapofthevelocityfieldcontainsnotonlyinformationthemassdistributionoftheGalaxy,italsoholdsthesignaturesoftheongoingperturbationsofthedisk,suchasthespiralarmsorthecentralbar and their back reaction. It allows measurement of the various patternspeedsatplayintheGalaxyandthelocationoftheirresonances.

• TheGalacticbulgeo Bulgedwarfs

TheinnerregionsoftheMilkyWaykeeptraceoftheearlyphasesofformationoftheGalaxy,andofitssubsequentevolution.Nohighresolutionsurveyisyetable to reach thebulgedwarfs,but thediscoverypotential is enormousgiventhatthebulgeislikelyhometosomeoftheoldeststarsintheGalaxy.ThemaincompetitorofMSEforthebulgesciencecasewillbeVLT/MOONS(near-infraredoverarelativelysmallfield,andthereforeparticularlyefficientforthebulgeinthemostcrowdedandextinctedregions).

o ThebulgestarformationhistoryThebulgeistoofartouseGaiaastrometrytoderiveagesinthebulge.Forthis,one relies entirely on spectroscopy. MSE is the only high-resolutionspectrograph that can reach the bulge turn-off. This is at V ~20 in Baade’sWindowandH ~17.5,thereforetoofaintforVLT/MOONS.

o LinkingtheinnergalacticsubstructurestotheouterGalaxyThe detailed abundance distribution of the bulge and its outskirts, and itscomparisonwiththoseobtainedforthehaloandthediskstudies,willrevealthecontinuity/ discontinuity between these populations. The bulge is denselypopulatedanddifferentpopulationsdominateatdifferent longitude/latitudes.Reconstructing the distribution of its chemistry, kinematics and age profilesthereforerequiresgoodstatisticsoveralargearea,upto|b|>10−20degrees).

o SearchingforprimordialpopulationsIt ispossible that thebulgecontains the remainsof themostancientmergersthatshapedthecoreofourGalaxy(e.g.,Howesetal.2015).Stars inthe innerMilky Way are kinematically well mixed, and so chemical abundances are


69

requiredtorevealtheirorigins.Thestructureandkinematicsofthebulgeisthatof a barred system presenting a boxy/peanut shape. However, there are alsoindications of the presence of a primordial structurewithin the inner galacticregions, but its relative mass and connection with the local old structures isunknown.Wearedependentupondetailedchemicalstudiesformanyelementsinordertoidentifythepresenceofaseparateprimordialbulgecomponentanddistinguish it, or link it, with the structures we identify at our solarneighbourhood.

• TheStellarHaloo Structureandsubstructure

Observationally, there is evidence that the Milky Ways stellar halo containsplenty of substructure. There are also evidences for different formationmechanismstobeatplayinthestellarhalooftheMilkyWay,withaninnerandouter component having significantly different structure, metallicity andkinematicproperties(e.g.Carolloetal.2007).KinematicsandchemistryoftheGalactichalouptoitsvirialradiuswillprovideuswithstrongconstraintsontheformationandevolutionofourGalaxy.

o In-situversusaccretedstarsintheinnerhaloMSEwill definitively quantify the portion of halo in-situ stars versus accretedones, measure potential gradients in the in-situ population and record thechemicalimprintoftheaccretedprogenitorformationhistory.

o ThefirststarsTheunprecedentedsizeandqualityofthestellarspectroscopicdatasetofMSEwill enable a detailed analysis of the metal-weak tail of the halo metallicitydistributionfunction(MDF).

ExaminationofFigure23andFigure24emphasizesthecrucialroleofMSEforunderstandingtheotherwise leastaccessiblepartsofourGalaxy i.e.,thefaintanddistantregimeswheretheouterdisk,thickdiskandespeciallystellarhaloaredominant.ThisisnottosaythatMSEwillnotprovidecriticaldataforexaminationofthethindiskorbulge;itisclearfromthelistabovethatMSEwillhavesignificantimpactonthestudyofthesecomponents.However,itdoesrecognizetheneedtooptimallydeploythespectroscopicresourcesavailableinthe2020stobestadvanceourunderstandingoftheevolutionoftheMilkyWay,andtooptimizeMSEsoitisofbestutilitytothosesciencecasesforwhich itcanbeuniquely,andpowerfully,suited.Further,onlyMSEcancomplementGaia in the faint regimebyproviding thedetailedchemistryanddynamicsoftargetstarsovertheentiremagnituderangeofsourcesobservedbythissatellite;apredictionoftheexpectedsourcedensityofGaiatargetsaccessibletoMSEisshownasanAitoffprojectioninFigure2.

Aslistedabove,amajorsourceofGaia–MSEsynergyisintheunderstandingoftheinterstellarmedium through absorption line studies of intervening gas clouds along sight-lines to brightstars.MSEisliterallyunmatchedinthisfield,andextensivediscussionofthisimportantaspectofthestructureofourGalaxycanbefoundinChapter4. Intheremainderofthissection,wefocusdiscussionontheresolvedstellarpopulationsoftheGalaxy.


70

Figure25:Aitoffprojectionofthepredictedskydensityoftargets(perdeg2)toG=20observedbyGaiathatwillbeaccessibletoMSE(δ>−30◦).Inputdatacredit:GaiaDPACCU2.

ScienceReferenceObservation4(AppendixD)

Unveilingcolddarkmattersubstructureswithprecisionstellarkinematics

AspartofanextensiveGalacticarchaeologymission,MSEwillmeasuretheaccuratevelocitiesofmillionsofstars instellarstreamsdistributedthroughouttheGalactichalo.Stellarstreamsaredynamically cold features that have been stripped from globular clusters and dwarf galaxies.However, according to ΛCDM, theMilkyWay halo should host many thousands of dark sub-halos.Astheseorbit intheGalacticpotential,theywillproduceaheatingeffectoncoldstellarstructures such as stellar streams. Bymapping the kinematics of a large number of streams,extendingoverlargeareasofskyandatarangeofradiifromthecenteroftheGalaxy,MSEwillbeabletomeasuretheheatingcausedbydarksub-halosandconstraintheirmassdistribution.Thediscoveryofevenasingleextendedcoldstellarstreamplacesstrong limitsontheallowedmassdistributionoftheseotherwiseinvisiblestructures.


71

2.3.2 ChemicaltaggingoftheouterGalaxy

Thehalocontainssomeoftheoldeststars intheGalaxy—includingthosethatformedonlyafew Myr after the Big Bang (Frebel & Norris 2015). Both the kinematics and chemicalcompositions of these old stars provide us with strong constraints on the formation andevolutionofourGalaxy.Manyof the“proto-galactic fragments”whoseexistence in theouterhalo was implied in the seminal work of Searle & Zinn (1978) have since been identifiedphotometrically intheSDSS,mostnotably inthe“fieldofstreams”(Belokurovetal.2006;seeFigure 26 for a recent visualization of the data from Bonaca et al. 2012). Such streams aregenerallyonlyobservableifthetimethathaspassedsincetheymergedwiththeGalaxyisshortcompared to thedynamical timescalesat their location. Streams in theouterhalo tend tobelongerlivedandthuseasiertorecognizeascoherentstructures(Johnstonetal.1999).

Figure26:StellardensitymapsoftheMilkyWayhaloinCelestialcoordinates(rightascensionincreasestotheleft,declination increases to the top) in the footprintof theSDSS imaging survey.Amatched filteranalysis revealsalargenumberof stellar streams andother substructures, themost prominent ofwhich are labeled.Figure fromBonaca,Giguere&Geha(2012).


72

AcriticalweaponinthearsenalavailabletoMSEliesnotjust indynamicalanalyses,butinthederivationof detailed chemical abundances. Recognising the suggestionby Freeman&Bland-Hawthorne (2002) that “themajor goal of near-field cosmology is to tag individual starswithelementsof theprotocloud”,MSEwillprovideabundanceratios forelements formedthroughmultiplenucleosyntheticchannelsformanymillionsoftargets.Ofcourse,thisappliesnotjusttothestellarhalobutalsotootherMilkyWaycomponents,includingthedisk,thickdiskandbulge.As discussed in the preceding section, the study of the chemical properties of the halo hasrelied,withafewexceptions,onlocalsamplesofhalostarsthatpassnearenoughtotheSuntobeobservableathighspectralresolution,whereasMSEwillmoveshifttheparadigmtowardsinsituanalyses.

Figure 27: Alpha elements as a function of [Fe/H] in a variety of nearby dwarf galaxies. Left panel: Sagittarius(red/gray filled squares: Sbordone et al. 2007;McWilliam& Smecker-Hane 2005), Fornax (blue/dark-gray opentriangles: Letarte et al. 2010) and Sculptor (green/light-gray filled circles: Hill et al. 2012). Right panel: Sextans(orange/light-gray filledcircles:Kirbyetal.2010;Aokietal.2009;Shetroneetal.2003). Inallpanels, solid linesshow the average trend and dotted lines show the 1-σ dispersion for each dwarf galaxy. The underlying pointsshowstarsintheMilkyWay(Cayreletal.2004;Vennetal.2004).(FigurefromHilletal.2012).

Figure 27providesa tasterof thepowerof largechemicalabundancedatasets inhelpingtounravel the formation ofMilkyWay and its satellites. Here, Hill et al. (2012) show the alphaelementdistributionasafunctionof[Fe/H]forMilkyWaystars(blackpoints;Cayreletal.2004,Vennetal.2004)aswellasseveralprominentMilkyWaydwarfspheroidalsatellites(Sagittarius– red/gray filled squares – Sbordone et al. 2007,McWilliam& Smecker-Hane 2005, Fornax –blue/dark-gray–opentriangles–Letarteetal.2010,Sculptor–green/light-grayfilledcircles–Hill et al. 2012, Sextans – orange/light-gray filled circles – Kirby et al. 2010, Aoki et al. 2009,Shetroneetal.2003).It isclearthatthedwarfspheroidalsshowadifferentdistributioninthisparameterspacefromtheMilkyWay(andpotentiallyfromeachother).Specifically,the“knee”in thedistributionofMilkyWaystars,atwhichpoint thealphaelementabundance isbroadlyconstantwithdecreasing[Fe/H],informsusontheearlieststarformationepochsthatmadethestars thatwe see in this Galactic component. Type II supernovae are alpha-rich, and explode


73

only a few-to-tens ofMyrs after a star formation event. Type Ia supernovae are iron rich incomparison, and take at least a few hundred Myrs prior to contributing to the chemicalabundances.Thepositionofthekneeinthedistributionisthereforeameasureoftheintensityofearlystarformation,sinceakneeathighermetallicityindicatesmoreintensestarformation(i.e.,moreTypeIIsupernovaecontributedtoboththealphaandironabundancepriortoTypeIasupernovaecontributingtothe ironabundanceonly).Thedwarfspheroidals, incomparisontotheMilkyWayhalo, appear tohaveevolvedmorepassively since their knee (if indeedone isevenpresent!)isatconsiderablylowermetallicity.

Whilethe insightsavailablethroughchemicalabundancestudiesarepowerful, largestatisticaldatasets of high resolution chemical abundances are only now emerging, prominent amongwhich isSDSS/APOGEEandAAT/HERMES.Assuch,therehasbeenawealthofrecentanalysesdiscussing the practical application of chemical tagging to these datasets. Key questions thathavebeenaddressed,andwhichareessentialforMSEdevelopmenttoacknowledge,include:

Figure28:Cumulativeposteriordistributionfunctionsfortheintrinsicabundancescatterin15elementsobtainedfor three clusters (M67 in red, NGC 6819 in yellow, NGC 2420 in blue) based on analysis of SDSS/APOGEEspectroscopy.Thecumulativedistributionfunctionforallthreeclustersisshowninblack,andthecombined95%upper limit isgiven ineachpanel.The intrinsic scatterofeachelement in theseclusters (star formationsites) isextremelysmall,asdemandedforchemicaltaggingtobesuccessful.(FigurefromBovy2016).


74

i. Thechemicalhomogeneityofstarformationevents

A fundamental premise of chemical tagging is that stars born in an individual cluster (whichhenceforthweadoptasshorthandfor“astarformationevent”,regardlessofwhethertheeventformsa recognizable stellar cluster) share a single chemical “fingerprint”, such that there is achemical signature that canbeused to relate the starseven if theydonot cluster strongly inphasespace.Until recently,however, thishadnotbeentestedtohighaccuracy (althoughseework by De Silva et al. 2007, 2009; Ting et al. 2012; Friel et al. 2014; Önehag et al. 2014).Traditional techniques to determine stellar abundances have an accuracy that, empirically,appears tobe limited to~0.1dex forwhatever abundance is beingmeasured.Generally, thisaccuracyreflectssystematicuncertaintiesintheunderlyingstellarmodelsusedtointerpretthedataandinunderstandingtheinstrument-dependentsignaturespresentinthedata.

Bovy (2016) undertakes a novel analysis of three different star clusters observed in theSDSS/APOGEE dataset that attempts tomeasure the chemical homogeneity, or otherwise, ofeachcluster,withminimalrecoursetomodels.Rather,theworkinghypothesisofthisanalysisisthat a homogeneous stellar cluster is one in which all the stars form a one dimensionalsequence, with any variations between members due solely to stellar mass (as well asobservationaluncertainties).

Bovy (2016) test their hypothesis at the level of the stellar spectra, and results are shown inFigure 28. Here, the solid black line is the overall posterior probability of the spread inabundanceforeachofthe15elementsindicated,andcoloredlinescorrespondtotheindividualclusters (M67 in red,NGC6819 inyellow,NGC2420 inblue).This figure shows that, foreachelementineachcluster,theintrinsicspreadbetweenthestarsintheclusterisatthelevelofafewhundredthsofadex,atmost.Thisprecisionisnearlyanorderofmagnitudebetterthancanbe achieved using standard techniques, and implies that the fundamental assumption ofchemicaltaggingissound.


75

Figure 29: The four principal components of a 17 element abundance space for ametal poor halo sample fromBarklemetal(2005).FigurefromTingetal.(2012).

ii. Thedimensionalityofchemicalabundancespace

Theapplicationof chemical tagging requires themeasurementof a largenumberof chemicalspeciesperstar.However,exactlyhowmanymeasurementsarerequired,andwhatspeciesaremostvaluable,isacomplexquestion.Forexample,twodifferentspeciesthatarefoundtovaryin lock-stepwith eachother as a functionof anyother variablewill provide considerably lessdiscriminating power than two species whose behavior over a large sample of stars is lesscorrelated.Clearly,thedimensionalityofchemicalabundancespaceislinkedtothenumberofuniquepathwaysbywhichchemicalenrichmentcantakeplace.

Tingetal. (2012)presenta landmarkstudyofthestructureofchemicalabundancespaceasa


76

pathfinder study for the AAT/HERMES spectrograph. These authors conduct a principalcomponentanalysisoftheabundancespaceofstarsobservedinseveraldifferentenvironments,includingthesolarneighbourhood,halometalpoorstars,etc.Forthemetalpoorstars (-3.5<[Fe/H]<-1.5dex),theyuseadatasetofnearly300starsfromBarklemetal.(2005)andtheFirstStarsSurvey (Cayreletal.2004,Francoisetal.2007,Bonifacioetal.2009) inwhich therearemeasurementsfor17differentchemicalelements.

Figure29showsthemain4principalcomponentsofthemetal-poorstarsanalysedbyTingetal. (2012). Here, neutron capture elements are different shades of red (dark red representlighters-processelements,redrepresentsheaviers-processelementsandorangearemostlyr-process elements), light odd-Z elements are dark brown, alpha elements are blue, iron-peakelements are green, and Cr andMn are black. In full, they find that there are typically 7 – 9dimensionsinchemicalabundancespaceforthesolarneighbourhoodandmetalpoorstars,andthat dwarf galaxies may have extra dimensions, perhaps associated with their longer starformationtimescales.

Aswell as leading to a better understandingof the structure of chemical space, the principalcomponents shown in Figure 29 are a fascinating way to reveal the nature of the sites ofnucleosynthesis.Forexample:

• ThefirstprincipalcomponentinFigure 29revealsthepresenceofasitethatproducesboth r process elements andalphaprocess elements (perhaps r-process core collapseSNe?)

• Thesecondprincipalcomponentrevealsthepresenceofaprocessthatproducesananti-correlation between alpha elements with iron peak and neutron capture elements(perhapsnormalcorecollapseSNe?)

• Thethirdprincipalcomponentrevealsaprocessproducingananti-correlationbetweenalpha elements and iron peak elements with neutron capture elements (perhapshypernovae?)

• The fourth principal component shows a strong contribution to Cr and Mn (both ofwhicharesynthesizedintheincompleteSi-burningregion).

In addition to providing insight into the chemical space thatMSEwill explore, the analysis ofTingetal.(2012)alsoshowsthepowerofprincipalcomponentanalysistochemicalabundancestudies.Ultimately,thedimensionalityofchemicalabundancespacewillbeabletobeexploredmorefullybyMSEusingdataformillionsofstarsacrossallcomponentsandsub-components,incontrasttothe~300starsinFigure29.


77

Figure30:Greyscalepointsshowprojectionsofapproximately105individualstarsfromtheSDSS/APOGEEsurvey.Thetopsetofpanelsshowprojectionsinabundancespaceonly.Clusteringanalysisofthis(15-dimensional)spacerevealssubstructures.Oneofthesemostprominentsubstructuresisshownasthelargepoints.Thelowerpanelsshowthestructureofthisfeatureinphasespacecoordinatesthatwerenotusedforitsidentification.Infact,thisfeatureisduetotheSagittariusdwarfspheroidalgalaxy.FigurefromHoggetal.(2016).

iii. Samplingrateandpracticalclustermasslimitsforchemicaltagging

Ting et al. (2015) examine someof the survey considerations for successful chemical tagging.Keyparameterstheyconsiderinclude:

• Sampling rate, that is the number of targets in the survey compared to the possiblenumberoftargets(Ntarget/Nsample).Highersamplingisrequiredinordertobesensitivetolowermassclusters;


78

• Thenumberof“distinct”regionsofchemicalspaceinwhichtheclustersaredistributed,

Ncells.Therelevantregionofchemicalspacetobeexploredwillhaveafinitevolume,inwhich each of the clusters will occupy a unique point. However, once measurementerrorsareconsidered,itisclearthatsomeclustersmaybecomeindistinguishablefromother clusters in chemical space, particularly if the number of clusters is high, thevolume of chemical space is small, and/or measurement errors are large, i.e., thedimensionality of chemical space, and typical measurement uncertainities, are bothcriticalparametersindefiningNcells;

• Thenumberandmassdistributionof individual star clusters (“star formationevents”)contributing to the signal. Clearly,measurement of these quantities is a fundamentalgoalofchemicaltaggingexperiments.

TheanalysisofTingetal.(2015)isfocusedtowardsstudiesofthestellardisk.Here,dynamicaltimesare relatively short andadditionaldynamicalprocesses can complicate theanalysis. Forexample, radialmigrationcanact tomovestarsoutof the surveyvolume inwhich theywereborn(or introducenewstarsthatdidnotforminthisregion intothesurveyvolume), therebyacting to make identification of the relevant, individual clusters, more difficult. Shorterdynamicaltimesalsomeansthatphasespacecoordinatesbecomeincreasinglydifficulttoapplyasadditionalconstraints.

Despite these additional difficulties, Ting et al. (2015) conclude that, through strategicobservationalprograms,itispossibletostatisticallyreconstructtheslope,highmasscut-offandevolutionoftheclustermassfunctionforstudiesthatfocusonthestellardiskoftheMilkyWayforsurveysthatsampleoforder1millionstars(presumably,themostchallengingenvironmentasdescribedabove).

ArecentanalysisbyHoggetal.(2016)isanicedemonstrationofchemicaltagginginaction:thetoppanelsofFigure 30 showsvariousprojectionsof105 starsobservedwithSDSS/APOGEE.Thedarkestpointsshowaprominentsub-groupinginthisspace,identifiedviaanalgorithmthatidentifiesclusteringinN-dimensionalspace.Thelowerpanelsshowthesesamestarsinvariousprojections of phase space. Even although the phase space coordinates were not used toidentify these stars, it is clear that the starsdo forma coherent structure. In fact, these starsbelongtotheSagittariusdwarfgalaxy.

These recent analyses of different aspects of chemical tagging emphasise the importance ofcarefullyplannedobservingprogramsto(i)bestsamplethestructuresbeingprobed(ii)tobestprobe the numerous dimensions of chemical abundance space (iii) to best use phase-spaceinformation in addition to chemical abundance information, and (iv) to best enable precisiondeterminationofindividualabundances.SRO-03describesindetailtheseconsiderationsastheyapply to MSE, with specific emphasis on important trades between multiplexing, spectralresolutionandwavelengthcoverage.BaselineGalacticArchaeologyprogramsaredescribed inSRO-03andSRO-4, thatwillproducehighquality,precisionabundances fora largenumberofchemicalelementsinmillionsofstarsthatspanallthemaincomponentsoftheGalaxy,focusingontheinterfaceofthethickdiskandhalo,andthatreachestotheedgeoftheMilkyWay.


79

2.3.3 ThemetalweaktailoftheGalaxy:firststarsandtheprogenitorsoftheMilkyWay

TheGalacticArchaeologyprogramsdescribedinSRO-03andSRO-04willenableavastrangeofscienceon the formationof theGalaxyand stellarnucleosynthesis, andwill provideaholisticperspectiveontheformationoftheGalaxythroughmulti-parameteranalyses.Wehavealreadyemphasised that amultitude of science cases spanning all aspects of theGalaxy are possiblewith such a program. In keeping with the focus of this document on science cases that areuniquelypossiblewithMSE,wenowdiscussinmoredetailaparticularlycompellingcomponentofSRO-03thatisfocusedonthemetal-poorGalactichalo.

Theunprecedented sizeof the stellar spectroscopicdataset forMSEwill enable thedefinitiveanalysis of the metal-weak tail of the halo metallicity distribution function (MDF). This keyobservablehasadirectbearingonmodelsfortheformationofthefirststars,andonthedarkbaryoniccontentofgalaxies.ThefirststarstobeformedaftertheBigBangwereformedwiththe “primordial” chemical composition: i.e., hydrogen and helium, plus traces of lithium. Aprotogalactic cloud consisting of such a gas may have had difficulty in providing coolingmechanismsefficientenoughtoallowtheformationoflow-massstars.Severaltheoriesonstarformationpostulatetheexistenceofa“criticalmetallicity”belowwhichonlyextremelymassivestars can form (Bromm & Loeb 2003; Schneider et al. 2011 and references therein). Othertheories invoke fragmentation to produce low-mass stars at any metallicity (Nakamura &Umemura2001;Clarketal.2011;Greifetal.2011).

Theimplicationoftheseconsiderationsforthebaryoniccontentofgalaxiesisobvious:ifthefirstgeneration of massive stars that reionized the universe formed along with low-mass stars, alarge fractionof thesewouldnowbepresentasold, coolwhitedwarfs.On theotherhand,asmallfractionofthese(essentiallythoseofmasslessthan0.8M⊙)wouldstillbeshiningtodayandcan inprincipalbeobserved.Further, if there isacriticalmetallicity, thenthemetal-weaktailoftheMDFoughttoshowasharpdropatthisvalue.MSEcanexamineboththechemicalabundancedistributionsofoldstarstosearchfornucleosyntheticsignaturesofthesefirststars,butalsoconstructapreciseMDFtosearchforevidenceforacriticalmetallicity.


80

Figure 31: Stellarmass density profiles for two different realizations of a cosmologically-consistentMilkyWaymasshalo(topandbottompanels),forallstars(leftpanel)andthreemetallicitycuts(fromlefttoright),[Fe/H]<−2,−3,and−4,andthreeredshiftsofre-ionizationineachpanel(asdescribedbytheinsettext).Thedashedlinesmarkpowerlawswithslopes−2,−3,and−4.FigurefromTumlinson(2010)

Detailedmodeling of the formation of aMilkyWaymass galaxy in which the distribution ofthesefirststarsaremonitoredrevealfascinatinginsightsintotheexpectedgrowthoftheMilkyWaywithtime.Tumlinson(2010)simulateasuiteofsixMilkyWayanaloguesand incorporatebaryonicprocessesontopofthedarkmatter“scaffolding”toallowanalysisoftheevolutionoftheimpliedstellarpopulations.Theyfindthatthefractionofstarsdatingfromtheoldestepochsisastrong functionof radiuswithin theGalaxyandwithmetallicity.The former trendreflectsthe insideoutgrowthofdarkmatterhalos,and impliestheoldeststarsaresomeofthemosttightly bound to the Galaxy. The latter trend reflects the trend of increasingmetallicity withtime.

Figure 31 shows the radial distributionof stars in a coupleof thedifferenthalo realisationsfromTumlinson(2010)asafunctionofmetallicity.Starsthatformedatredshiftsz=6–10likelyhavemetallicitiesmoremetalpoorthan[Fe/H]=-3dex.Totestthevalidityofthesemodels,andtoprobetheformationoftheGalaxyanditsstarsattheearliesttimes,itisthereforeimportantnotjusttofindmetalpoorstars,buttodevelopadetailedunderstandingoftheshapeofthefullMDFand its spatialvariation in theGalaxy, includingout tovery large radius.Suchaprogramrequiresinsituanalysisoflargenumbersofmetalpoorstars,andisacoresciencegoalofMSE.


81

Figure32:Themetallicitydistributionfunction,[Fe/H],basedonthehigh-resolutionabundanceanalysisofYongetal. (2013a) at the metal-weak end, [Fe/H]<-3 dex. Left panels show a linear scaling, and right panels show alogarithmicscaling.Theupperandlowerpanelsrefertotherawdataandthosefollowingcompletenesscorrectionson the range −4.0 < [Fe/H] < −3.0, as described by Yong et al. (2013b). Green and grey color-coding is used topresent the contribution of C-rich and C-normal stars for which measurement was possible, respectively. ThedashedlineshowstheHamburg-ESOMDFbasedonthedataofSchorcketal.(2009).FigurefromFrebel&Norris(2015).

ToputthepotentialofMSEstudiesofthemetal-weaktailoftheGalaxyincontext,Figure32showsthestate-of-the-artMDFforstarsmoremetal-poorthan[Fe/H=-3dex.Forhomogeneity,thisisbasedonahighresolutionanalysisbyYongetal.(2013a).AlsoshownasadashedlineistheMDFderivedfromtheHamburg-ESOSurvey(Schorcketal.2009).TheHamburg-ESOsurveyisrecognizedasalandmarkstudyofthemetallicityoftheGalaxyhalo,andisbasedonatotalof1638stars.TheYongetal.(2013a)samplehas86starswith[Fe/H]<-3dex,ofwhich32have[Fe/H] < -3.5 dex. Current surveys will increase the known sample of metal-poor starssignificantlyoverthenextfewyears,andwillproduceimportantlistsofcandidatemetalpoorstarsforspectroscopicfollow-up(e.g.,theSkyMapperproject–whichhasalreadyfoundastarwith [Fe/H]~-7, Keller et al. (2014) – and the Pristine project on CFHT(P.I. E. Starkenburg).However, full characterisation of theMDF through in-situ spectroscopic analyes requires thelargeaperture,highresolution,highlymultiplexedcapabilitiesofMSE.ThroughaprogramsuchasthatdescribedinSRO-03,MSEwillprovidethedefinitivestudyofthemetal-weakstructureoftheGalaxy,byprovidingmetallicities fora sampleof severalmillionstarsat resolutionR>20000,allowingacomplete,homogeneous,characterizationofthehaloMDFdownto[Fe/H]∼–7.


82

2.4 TheLocalGroupasatimemachineforgalaxyevolution

Figure33:DensityslicesoftheIllustrissimulationatdifferentredshifts,eachwithasidelengthof106.5co-movingMpc.Thesquareineachpanelindicatestheco-movingsizeofaLocalGroup-typestructureateachredshift.Thesearecomparable insizetotheHSTUltra-DeepField.Thezoom-inpanelsshowsanexpectedimageofthegalaxiesvisibleatz=7±0.02usingJWST(leftpanel),andthroughthestellarfossilrecordintheLocalGroup(rightpanel;whitesymbolsindicatestructuresstillpresentintheLocalGroupatz=0,whereasgoldsymbolsrepresentobjectsthatarenolongerpresentintheLocalGroupatz=0).(FigurefromBoylan-Kolchinetal.2015).

ThecollectionofgalaxiesintheimmediatevicinityoftheMilkyWay—theMagellanicClouds,the Andromeda Galaxy, Triangulum, and the~ 100 currently known dwarf galaxies thatconstitutethe“LocalGroup”—arethenearestexamplesofgalaxiesspanningawiderangeofmorphological types. Generally speaking, they are the only galaxies in the Universe— asidefromtheMilkyWayitself—forwhichlargenumbersofindividualstarscanbespectroscopicallyobserved from the ground. As such, these nearby systems offer unique insights into galaxyformationandevolution

The size of the Local Group is of order a couple ofMpc in diameter, as defined by the zero-velocity surface where the gravitational attraction of the galaxies of the Local Group exactlybalances the Hubble flow (McConnachie 2012). A recent analysis of the formation of LocalGroup-likestructuresincosmologicalsimulationsshowsthat,atz=7,theco-movinglinearsizeof the Local group is 7Mpc (i.e., a volume of~350Mpc). At early epochs, the Local Grouptherefore probes a cosmologically representative volume of the Universe. At z ≤ 3, the co-movingsizeoftheLocalGroupexceedsthatoftheHubbleUltra-DeepField,andissimilartothe


83

volumesthatwillbeprobedbyJWST(seeFigure33).

Figure34:Thesolidblacklinerepresentsthecumulativeluminosityfunctionofgalaxiesatz=7(Finkelsteinetal.2015),extrapolatedbeyondMUV=-18tofaintmagnitudes.ThethinkgreylinesrepresentthecumulativeluminosityfunctionsinsimulationsofLocalGroupenvironments(theELVISsimulations,seeGarrison-Kimmeletal.2014),andthedashedblacklinerepresentsthemean.TheapproximatedetectionlimitsofHST,JWSTandtheLocalGroupareshown, highlighting the complementarity of the near and far-field regimes for a full understanding of galaxyevolutionacrosstheluminosityfunction.(FigurefromBoylan-Kolchinetal.2016).

The complementarity of studies of the Local Group population of galaxies to high redshiftstudiesofgalaxyevolutionwithJWST,WFIRSTandotherfuturefacilities,isclear.Atearlytimes,the Local Group and JWST explore similar volumes. The Local Group contains representativenumbersof faintgalaxies thatwillneverbedetectableathigh redshiftwith JWST,andwhoseformation canbe explored through their stellar fossil record.Meanwhile, JWST is sensitive togalaxiesthatareintrinsicallybrighterand,therefore,rarer.ThisisshownexplicitlyinFigure34.

LocalGroupgalaxiessubtendlargeanglesontheskyandconsistofamultitudeofstarsthatare,individually,quitefaint.AcompleteexplorationofthedeepfieldoftheLocalGroupthereforerequireswidefields,extremesensitivities,andhighmultiplexing.MSEisthereforeessentialforthisfield,beitforthelowestmassorhighestmassgalacticsystemsthatarepresent,aswenowdiscuss.ScienceReferenceObservation5(AppendixE)

ThedynamicsandchemistryoftheLocalGroupgalaxies

WithMSE,wewill be able to conduct a systematic survey of the kinematics and chemistry of


84

LocalGroupgalaxieswithin1Mpcwithalevelofdetailthathasneverbeenachievedbeforeandcannotbeachievedanywhereelse.Gaiaisgatheringunprecedentedastrometricmeasurementsthatprovidetheabilitytomapthesix-dimensionalphase-spacestructureofagalaxywill likelyusher in a “Golden Age” of Milky Way research. However, without essential complementaryinformation from the nearest galaxies, conclusionswemay draw about the process of galaxyformation from a single galaxy may be premature. It is in this context that we propose toconducttheultimatechemodynamicaldecompositionofLocalGroupgalaxiesacrossthegalacticluminosity function, including a large population of faint dwarfs as well as the sub-L* galaxyM33,andtheL*galaxyM31.Thesegalaxiesrepresenttheobvioussteppingstonesbetweenourhighly detailed description of the Milky Way, and the low-resolution studies of more distantgalaxies, in the realm of a fewMpc and beyond,wherewe can obtain significant samples ofgalaxies(asafunctionofgalaxytype,environment,mass,etc).

2.4.1 Darkmatterandbaryonsatthelimitsofgalaxyformation

The Local Group’s dwarf galaxies include the nearest, smallest, oldest, darkest and leastchemically-enriched galaxies known. Their extreme properties have made these objects thefocusof investigations ranging fromgalaxy formation to cosmologyandevenparticlephysics.Severalgenerationsof imagingsurveys,combinedwithtargetedspectroscopic follow-up,havedramatically improved knowledge of these galaxies’ luminosity functions and spatialdistributions, internal structure and kinematics, chemical abundance patterns and star-formation histories (Mateo 1998, Tolstoy et al. 2009, McConnachie 2012, and referencestherein).

Over the last decade, large-scale imaging surveys (e.g., SDSS, Pan-Starrs, Dark Energy Survey,PAndAS,etc.)havenearlyquadrupledthenumberofknowndwarfgalaxiesintheLocalGroup,revealing tens ofM31 satellites aswell as a vast population of “ultra-faint”Galactic satellitesthatreachesastonishinglylowluminositiesofLV∼102−3L⊙.(e.g.,Willmanetal.2005,Zuckeretal. 2006, Belokurov et al. 2007,McConnachie et al. 2009, Koposov et al. 2015, Laevens et al.2015, Bechtol et al. 2015, Drlica-Wagner et al. 2015). The ability of large imaging surveys todeliver data sets that are simultaneously deep, wide and homogeneously processed has alsoenabled discoveries of low surface brightness features such asM31’smetal-poor stellar halo(Chapman et al. 2006, Kalirai et al. 2006), numerous stellar streams in the Milky Way (e.g.,Belokurov et al. 2006), distorted morphologies in the outermost regions of several dwarfgalaxies (e.g., Muñoz et al. 2010), and distant tracers of the Galactic potential (e.g., bluehorizontalbranchstars;Brownetal.2010,Deasonetal.2012).

On the spectroscopic front, over the same timeframemulti-object spectrographs at 6 − 10mclasstelescopeshaverevealedintriguingdetailsaboutthechemicalanddynamicalpropertiesofthesestellarpopulations.Velocitydispersionprofilesmeasuredfromhundredstothousandsofstarspergalaxyimplydominationbydarkmatteratallradii(Walkeretal.2007).Inmanycases,even theoldest stellarpopulations in these systemscanbedivided intomultiple componentsthat provide clues about formation and evolution: typically a relatively metal-rich, spatiallycompactandkinematicallycoldsub-populationisembeddedwithin–andtracesthesamedarkmatterpotentialas–arelativelymetal-poor,kinematicallyhottersub-population(Tolstoyetal.2004,Battagliaetal.2006,Battagliaetal.2011).ThefaintestGalacticsatellitesnotonlyextend


85

thegalacticluminosity–metallicityrelationbyatleastthreeordersofmagnitudeinluminosity(Kirbyetal.2013),theyalsohostsignificantnumbersofextremelymetal-poorstars([Fe/H]<-3)that encode information about early supernovaeevents (Kirby et al. 2008, Frebel et al. 2010,Simon et al. 2015). Furthermore, chemical abundance patterns are uncovering the extent towhich theGalactichalogrewbydevouring the siblingsof its surviving satellites (Tolstoyetal.2009,andreferencestherein),andmostrecentlyareevenhelpingtosettledecades-longdebateregarding the astrophysical mechanisms that lead to “rapid” neutron-capture production ofheavyelements(Jietal.2015,Roedereretal.2016).

Yetduetothelackofadedicatedlarge-aperture,wide-field,multiplexingspectroscopicfacility,imaging surveys have dramatically outpaced their spectroscopic counterparts thus far. Whilerecent and/or future large-scale spectroscopic surveys target primarily extragalactic sources(e.g.,SDSS,BOSS,DESI)and/orGalacticstars(e.g.,SEGUE,APOGEE,Gaia,Gaia-ESO),nonehasasignificant componentdevoted tobuilding large, uniformand spatially completedata sets fortheresolvedstellarpopulationscomprisingtheLocalGroup’sdwarfgalaxies.

Figure35:Possiblestellar-spectroscopicsamplesizesforknownLocalGroupdwarfgalaxieswithMV≥-16.5.Asafunctionofgalacticluminosity,blackpointsindicatenumberofstarsbrighterthanafiducialmagnitudelimitofV<23.0. Connected red points indicate sizes of samples available in the current literature. Marker types specifywhetherthedwarfgalaxyisisolated(triangles;seetheSolosurveyofHiggsetal.2016)orasatelliteoftheMilkyWay(circles)orM31(squares).DatafromMcConnachie(2012);Muñozetal.(2005);Gehaetal.(2006);Kochetal.(2007);Lewisetal. (2007);Martinetal. (2007);Simon&Geha(2007);Mateoetal. (2008);Walkeretal. (2009a);Fraternalietal.(2009);Carlinetal.(2009);Kochetal.(2009);Belokurovetal.(2009);Gehaetal.(2010);Willmanetal.(2011);Koposovetal.(2011);Hoetal.(2012);Tollerudetal.(2012);Frinchaboyetal.(2012);Kirbyetal.(2013);Collinsetal.(2013);Walkeretal.(2015a,b);Kirbyetal.(2015);Walkeretal.(2016);Torrealbaetal.(2016).

ThistechnologicalshortcomingofLocalGroupastronomyisauniqueopportunityforMSE,asits


86

unprecedented combinationof depth,wide fieldof view, spectral resolutionandmultiplexingwillbetransformativeinthisfield.Inordertohelpquantifytheimprovementsonecanexpect,weconsiderthespectroscopicsamplesizesthatwouldbeachievableinanMSEsurveyofLocalGroup galaxies. For all known systems less luminous thanM32 (MV ≥ 16.5), we estimate thenumberofstarsbrighterthanafiducialmagnitudelimitofV≤23.Figure 35comparesthesenumberstothe largestspectroscopicsamplesizesthatarecurrentlyavailable inthe literatureforeachobservedgalaxy.InmostcasesanMSEsurveywouldgrowtheavailablespectroscopicsamplebymorethananorderofmagnitude.Forthe“ultrafaints”(MV>-5),thismeanssamplesreaching hundreds to one thousand member stars. For the Milky Way’s “classical” dwarfspheroidals(−7≤MV≤−13)andmoreluminousbutfartherobjectslikeNGC185,NGC205andNGC6822,itmeanssamplesreachingintothetensofthousandsofmemberstars.

Figure36:Examplesofobservationalselectionbiasincurrentlyavailablespectroscopicsamplesforindividualdwarfgalaxies.Left:Spatialdistributionofcandidateredgiantbranch(RGB)starswithV≤23mag(blackpoints) intheFornaxdwarf spheroidal (Bateetal.2015), comparedwith thedistributionof starswithpublishedspectroscopicmeasurements fromsurveyswithVLT/FLAMES(Battagliaetal.2006),Magellan/MMFS(Walkeretal.2009b)andKeck/Deimos(Kirbyetal.2010).Thedashedellipserepresentsthe“tidal”radiusfromthebest-fittingKingmodeltoRGBcandidates (Bateetal.2015).Right:Ratioofnumberof spectroscopicallyobservedstars toRGBcandidateswithinanangularseparationRfromthecenterofFornax.

ItisnotjustthesizeofthedatasetsthatMSEwillimpact.Giventhemodestfieldsofview(≤0.5degree) that are available for conducting multi-object spectroscopy at the largest-aperturetelescopes available today, nearly all published spectroscopic data sets for Galactic satellitessuffer fromspatial incompleteness;morespecifically, selectionbiases that leaveouter regionseitherundersampledorneglectedaltogether.Eventhoughouterregionsareofkeeninterestfortracingstellarpopulationgradients(Harbecketal.2001,McConnachieetal.2007),investigatingtheouterstructureofdarkmatterhalos(Walkeretal.2007),identifyingkinematicsignaturesoftidaldisruption(Muñozetal.2006)aswellasevolutionofinternalhalostructure(El-Badryetal.2015)andevenmeasuring systemicpropermotions (Kaplinghat&Strigari 2008;Walkeret al.2008), the relatively low fractions of bona fidemembers at large radius has made thoroughobservationsoftheseregionsprohibitivelyexpensive.


87

Figure 36demonstrates thedegree towhich thisproblemafflictsevenFornax, forwhich in-dependent surveys with VLT/FLAMES, Magellan/MMFS and Keck/Deimos have delivered therichestavailablespectroscopicdatasetsforanyLocalGroupdwarfgalaxy(N∼3000members;Battagliaetal.2006;Walkeretal.2009b;Kirbyetal.2010).Theleft-handpaneldisplaysthespa-tialdistributionofredgiantbranch(RGB)candidateswithV≤23(Bateetal.2015),comparedtodistributionsofspectroscopicallyobservedstarsfromeachsurvey.Theright-handpanelshowshow the number of spectroscopically observed stars within a given separation from Fornax’scenter compares to the number of RGB candidates. Figure 36 makes clear that even the“richest”existingdatasetsarewoefullydeficient:notonly is the fractionofspectroscopically-observedRGBcandidateslessthan3%withinFornax’sestimatedtidalradius,but<10%ofthespectroscopically observed stars lie beyond one half-light radius (∼ 0.25 degrees) from thecenter of Fornax. This selection bias propagates into biased estimates of chemodynamicalquantitieslikevelocitydispersionandmeanmetallicityandchemicalabundancepatterns,allofwhichexhibit radialgradientsout tothe limitsofcurrentsamples.Furthermore, thecentrally-biasedspatialsamplingcanentirelymissimportantfeaturesthatbecomeapparentonlyatlargeradius(e.g.,diffusetidaldebristhatcanbedetectedoutto∼10half-lightradiifromthecenter;Muñozetal.2006).

In contrast to the current situation, MSE’s wide field will enable the first spatially completesurveys of the Milky Way’s relatively large (0.1 ≤ Rhalf ≤ 1 degree) satellites. MSE’sunprecedentedcapabilitieswillbeparticularlywell-suitedtoobservingeventhemostenigmaticGalactic satellites like Crater 2which, albeit relatively luminous atMV∼ -8, remainedhiddenuntil this year due to its large size (Rhalf ∼ 0.5 degree) and correspondingly low surfacebrightness(31mag/arcsec2;Torrealbaetal.2016).

Howdoesthisimprovementinthesizeandcompletenessofdatasetshelpstudiesofthenearestdwarfgalaxies?Onesourceofkeen interest in thesespectroscopicsamples is the informationthey convey about the dark matter content of dwarf galaxies (see also Chapter 4). Suchinformation is crucial for investigations ranging from testsof cosmologicalmodels (e.g., thosebuilt on “cold” vs “warm” vs “self-interacting” dark matter) as well as interpreting (non-)detectionsofdarkmatterdecayand/orannihilation.Inordertogaugetheeffectsofsamplesizeon inferencesaboutdarkmatter,we compare results fromdynamical analysesof threemockdatasetsconsistingofN=102,N=103andN=104 stars. Ineachcase theartificial sample isdrawn from a phase-space distribution function describing a stellar population that follows aPlummer (1911)surfacebrightnessprofileand tracesagravitationalpotentialdominatedbyaNavarro, Frenk & White (1997) dark matter halo. The analysis uses standard Bayesianprocedurestofitsimultaneouslyforthevelocityanisotropy,surfacebrightnessanddarkmatterdensity profiles (a by-product is specification of the line-of-sight velocity dispersion profile),similar to the procedures described by Geringer-Sameth et al. (2015) and Bonnivard et al.(2015).


88

Figure37:Recoveryof intrinsic line-of-sightvelocitydispersion (left)anddarkmatterdensity (right)profilesasafunctionof spectroscopic sample size. Shaded regions represent 95% credible intervals froma standard analysis(basedontheJeansequation)ofmockdatasetsconsistingoflineofsight(LOS)velocitiesforN=102,103and104stars(medianvelocityerror2kms−1),generatedfromanequilibriumdynamicalmodelforwhichtrueprofilesareknown(thickblacklines).

Figure 37 compares the resulting inferences for velocity dispersion and darkmatter densityprofiles, displayingbands that enclose 95% credible intervals in each case.Given the samplesthatMSEcanprovide, inferencesaboutkinematicsanddarkmatter contentofdwarfgalaxieswill become dramatically more precise. We will infer the dark matter densities of ultrafaintsatelliteswithprecision similar towhat isachieved todayonly for themost luminousclassicaldwarfs, for whichwewill infer density profileswith unprecedented precision from pc to kpcscales.Thisimprovementwillrenderdynamicalanalyseslimitedbysystematics(e.g.,triaxiality,non-equilibriumkinematics)insteadofstatistics.

As a practical illustration of the sheer power of MSE for chemodynamical studies of nearbydwarfgalaxies,weconsiderNGC6822,oneofthenearestdwarfirregulargalaxiesatadistanceof 500 kpc. It is one of the more intriguing targets for detailed study because of ongoingdisturbances in itsHI velocity fieldandveryactive star formation.There is someevidence foryoung stellar populations associatedwith infalling HI clouds, and for deviations from circulardisk rotation. However, the large angular scale of the system (~1 degree across) and thelikelihoodthatthesubstructuresarerepresentedbyonlyasmallfractionofthestarsmeansthatthe system remains poorly understood, even with a small (but steadily growing) sample ofstellarspectrafromVLT/FLAMESandKeck/DEIMOS.

Figure 38showsthespatialdistributionofredgiantbranchcandidatesinNGC6822basedonCFHT/MegaCamimaging,comparedtotheMSEfieldofview.ThenumberofpotentialtargetsinNGC6822 available for spectroscopywithMSE dwarfs the capabilities of existingmulti-objectspectrographson8-10m-classtelescopes.Forinstance:

• thenumberofpotential“redstar”targets(i.e.,starswithages>500Myr,mostofwhicharelikelyolderthan1Gyr)is:


89

o ~2500starsabovetheTRGB,including~103carbonstars(18<I<19.5).o ~7500starswithin0.5magoftheTRGB(19.5<I<20)o ~30000starsbetween0.5and1.5magoftheTRGB(20<I<21).

• thenumberofpotential“bluestar”targetsis:o ~1000starswithages<50Myr(15.5<g<18.5)o ~1000starswithages50–200Myr(18.5<g<20.5)o ~2000starswithages200–400Myr(20.5<g<21.5)

Figure 38: Spatial distribution of red giant branch stars (red dots) in the barred irregular galaxy NGC6822 fromCFHT/MegaCamobservations(seeHiggsetal.2016).ThecontoursshowtheboundaryoftheHIdiskfromdeBlok&Walter (2006). TheMSE fieldof view is shown for comparison; clearly,MSE is ideal forwide field spectroscopicstudiesofLocalGroupdwarfgalaxies.

With these stellar densities and spatial distributions, it would be possible, for example, tomeasureradialvelocitiesaccuratetobetterthan5kms−1foreveryAGBandredsupergiantstar,and nearly all RGB stars within 0.5 mag of the TRGB, in just one night on MSE, at mediumresolutionandwithSNR~10–20.Asinglepointing,withthreedifferentconfigurationsofthe3200 fibres, each observed for roughly a 1 hour, would provide a complete census of thesebright redstars.Withsamplesof>104 stars, sub-populations representativeof just1%of thetotalnumberofstarswouldbecomedetectablewithhighsignificance.ToobtainspectrafurtherdowntheRGBwouldtakemoretime,butnotunimaginablyso.Forinstance,atI=21,SNR~10perhourwouldbeachievedinthered/near-IRregion,meaningthathomogeneous,high-qualitykinematicandmetallicityresultsfor~30000starscouldbeachievedinonly10nights:i.e.,6–8hoursper setup,with10 configurationsneeded to reachevery target. If the shallower surveywere repeatedmultiple times over a period of several years, the sensitivity andmultiplexingwouldallowunprecedentedstudiesofvariabilityandevolutionarychangesforstarsinthelatephases of evolution. Similarly, intermediate-resolution radial velocities could be obtained foreverymain-sequencestarinthegalaxyyoungerthan∼400Myrinjust∼3nightsatSNR∼10−15forg<21.5.


90

2.4.2 ThechemodynamicaldeconstructionofthenearestL*galaxy

Figure39:Spatialdistributionof candidate redgiantbranchstars in theenvironsofM31andM33,as identifiedfrom colour-magnitude cuts from the Pan-AndromedaArchaeological Survey. Dashed circles highlightmaximumprojectedradiiof150kpcand50kpcfromM31andM33respectively.ThetilingstrategyforanMSEsurveyofthisregionisoverlaid.

WithMSE,wewillbeable to conducta systematic surveyof thekinematicsandchemistryofLocalGroupgalaxieswithin1Mpcwithalevelofdetailthathasneverbeenachievedbeforeandcannot be achieved anywhere else. On the scale of (large) spiral galaxies, Gaia is gatheringunprecedented astrometric measurements with the accuracies needed to produce astereoscopic and kinematic census of about one billion stars in our Galaxy. This dramaticadvance inour ability tomap the six-dimensional phase-space structureof a galaxywill likelyusher in a “GoldenAge” ofMilkyWay research. The consequences of this endeavor forMSEhavebeendiscussedinpreceedingsections.

Gaia will provide a very detailed view of the structure and dynamics of our Galaxy and thisinformationwillhavetobeputintocontextbycomparingtheresultswithobservationsofothergalaxies.Withoutthisessentialcomplementaryinformation,anyconclusionsthatwemaydrawabouttheprocessofgalaxyformationfromobservationsoftheMilkyWaywouldbepremature.It is in this context that the capability of MSE to provide the ultimate chemodynamicaldecompositionofthetwoclosestLocalGroupspirals(M31andM33)ismostimportant.Thesegalaxies represent theobvioussteppingstonesbetweenourhighlydetaileddescriptionof theMilkyWay,andthe low-resolutionstudiesofmoredistantgalaxies, intherealmofafewMpc


91

andbeyond,wherewecanobtainsignificantsamplesofgalaxies(asafunctionofgalaxytype,environment,mass, etc.). By~2020,M31 andM33will have been studied in detail bymanyexisting and imminent instruments/observatories. Nevertheless, a gaping chasm in ourunderstandingofthesestructureswillremain:i.e.,theirkinematicsandchemistryfromresolvedstellar populations. We note that Subaru/PFS intends to observe approximately 66 squaredegreesof thehaloofM31 (50 individualexposures; Takadaetal. 2014).However, it is clearfrom inspection ofFigure 39, discussed below, that a comprehensive understanding of thechemodynamicsofM31andM33requirefullspatialcoverageandhighcompleteness,especiallyinthetarget-richinnerregionsofthesegalaxies.

MSEwillmakeapivotalcontributiontothesubjectbyundertakingalargespectroscopicsurveyofthesetwoLocalGroupgalaxies.Thegoalwouldbetoobtainastatisticallysignificantsampleofstarsbelongingtotheirconstituentstructuralcomponents.Thescientificaimofthis“Galacticarchaeology” theme is tounderstandthegalactic formationprocess,which isclearlyanopen-ended endeavour. Our power to address the key issues (e.g., thick disk formation, haloformation, streams, chemical enrichment history), depend on the spectral resolution that thesurveyaffords.

Figure 39 shows the spatial distribution of candidate RGB stars in the environs ofM31 andM33, as identified by colour-magnitude selection from 400 square degrees of contiguous giimagingwithCFHT/MegaCamaspartofPAndAS.Colour-codingcorrespondstothecolouroftheRGBstars,suchthatredderRGBstars(likelyhighermetallicity)appearred,andbluerRGBstars(likelymetal-poor) appear blue.Dashed circles correspond tomaximumprojected radii of 50,100 and 150 kpc from M31, and 50 kpc from M33. PAndAS resolves point sources at thedistanceofM31(D~780kpc;McConnachieetal.2005)tog≃25.5andi≃24.5atSNR~10.Morethan107stellarsourcesareshowninFigure39.ThetypicalcolouroftipoftheRGBstarsin M31 is (g − i)~1.3 so gTRGB~22.5. PAndAS therefore reaches to (nearly) the horizontalbranchlevel,providingphotometryofsufficientdepththatpotentialspectroscopictargetsfora10mfacilitycouldbeselected.

InFigure39,theeffectivesurfacebrightnessofthefaintestvisiblefeaturesisoforder33magarcsec−2.Thiscorresponds literallytoafewRGBstarspersquaredegree.Notethatthediskofthe Milky Way is located to the North so there is increasing contamination in the colour-magnitude of the RGB locus by foreground dwarfs; the reddest RGB stars are particularlyaffected by this source of contamination. Young, blue stellar populations — and evenintermediate-agepopulationssuchasasymptoticgiantbranch(AGB)stars—arenotpresentintheouterregionsofM31inanysignificantnumbers.Thus,anyspectroscopicstudyoftheouterregionsoftheM31halowillnecessarilyconcentrateontheolder,evolved,RGBpopulation.Forthis reason,we consider separately surveys of theouter halo (characterized by a low surfacedensityofevolved, giant star candidates) and the innergalaxy (withahigh surfacedensityoftargetsfromamixtureofstellarpopulations):

• An outer halo survey of M31/M33: This program aims at obtaining a complete,magnitude-limited,spectroscopiccensusofeverystarintheouterregions(40–150kpc)of an L* galaxy halo to provide complete kinematics for every star, supplemented bymetallicity estimates for most stars and alpha-abundances for the brightest ones.


92

Ultimately, such a survey will allow us to deconstruct a nearby galactic halo into itsaccreted “building blocks”. Such a survey will provide the ultimate testbed of thehierarchicalformationofL*galaxiesandfurtheryieldtheoptimaldatasettoconstrainthedarkmattercontentofM31andM33;

• Faint stellar populations in the inner regions of M31/M33: This program has as itsfocusthestudyofthedisk/thick-disk/halotransitionregiontomeasuretheextentofthedisks, characterize the relationshipbetween these components, and todetermine theroleofmergersinthediskandinnerhaloevolution.

TherichnessofthisdatasetcanbeanticipatedbyconsideringwhathasalreadybeenachievedforM31 andM33using Keck/DEIMOS (hundreds of pointings in the vicinity ofM31 targetingRGB star candidates, for a total of > 10 000 stars, primarily by the SPLASH and PAndAScollaborations; e.g., Dorman et al. 2012, Collins et al. 2011, and references therein).Keck/DEIMOShasa5x16arcminfieldofviewandisabletoobserveoforder100targetsperpointing, making it the most well suited large aperture spectrograph currently available forstudiesofM31.

Figure40:A selectionof fields fromKeck/DEIMOSspectroscopic studiesof theM31disk regionalong theSouthWest major-axis of M31, shown in order of increasing (projected) radius. Gaussian fits indicating the thin and(whereapplicable)thickdiscsareshownasmagentaandbluecurves,respectively.Dashedandsolidverticallinesindicateselectioncriteriaforthekinematicidentificationofthickdiskstars.(FigurefromCollinsetal.2011).


93

Figure41:Leftpanel:Thepositionsoftwenty-fourKeck/DEIMOSmultiobjectslitmasksfromtheSPLASHsurveyofM31overlaidontheChoietal. (2002)KPNOBurrellSchmidtB-bandmosaic imageofM31.Notethat this imagecoversapproximately1.5degreesonaside,tobecomparedwiththe~25x25degreeimageshowninFigure39.Right panels:Maximum likelihood fits of a kinematically hot spheroid to eachof five sub-regionsof the survey,afterexcludingthevelocityrangeencompassingthegiantstellarstream(Ibataetal.2001)anditsassociatedtidaldebris (shaded pink). Violet lines show the cumulative region cold component; blue show the best-fit spheroidGaussian; green show the sum of these two components. The dashed lines show the systemic velocity ofM31relativetotheMW.IndividualsubregioncoldcomponentsareshowninorangeintheSSWregionpanel,but leftoutoftheotherpanelsforclarity.ThecomplexityofthediskregionofM31isclear. (FiguresfromDormanetal.2012).

Figure40andFigure41showvelocityhistogramsforRGBstarsinthevicinityofthediskofM31,fromspectroscopicstudiesofthisregionbyCollinsetal.(2011)andDormanetal.(2012).Figure 40 shows fits of twoGaussians to a large number of fields, that Collins et al. (2011)suggestisevidenceofathickdisk,withavelocitydispersionthatisnearly50%higherthanforthethindisk,andwithascaleheightthat isapproximately3timeslargerthanthethindisk.Adetaileddynamicalanalysisoftherelationbetweenthethinandthickdisksbasedonindividualstars has previously only been possible for the Milky Way, but here the perspective is verydifferentand,asaresult,verycomplementary.

Dormanetal.(2012)examinethediskofM31andconsiderablysmallerradiusthanCollinsetal.(2011),andshowthattherearesignificantcontributions fromthe innerspheroidofM31.ThelocationsoftheirfieldsandtheresultingvelocityhistogramsareshowninFigure41;itisclearthat this region is complex to analyseandmaximum likelihoodanalyses suggest thatmultiplecomponents are required per field. They find evidence that the spheroid rotates, and morecloselyresemblesthatofanellipticalgalaxythanatypicalspiralgalaxybulge.

Inmuchthesamewayasdiscussedpreviouslyforthedwarfgalaxies,alloftheseanalysesarebased on kinematic subsamples that are very sparse and highly incomplete. More than 10millionRGBcandidatesarepresent in thePAndASmapofM31shown inFigure 39, andyetmore than a decade of research using Keck/DEIMOS has barely sampled 1% of the possiblecandidates. It is interesting tonote that, evenwith theorderofmagnitudesdifference in the


94

stellardensityofM31intheinnerregionscomparedtothediffuseanddistantregionsshowninFigure 39,MSE isabsolutelynecessary:at small radius, themultiplexingofM31members isextreme,and in theoutskirts, thecontamination from foregroundstars isextreme,evenwithreasonablepreselection.

Dormanet al. (2012) note that “the literature is full of vocabulary such as ‘bulge,’ ‘spheroid,’‘innerspheroid,’‘outerspheroid,’‘disk,’‘thindisk,’‘thickdisk,’‘extendeddisk,’andsoon.Thereisnot yet a consensuson thebest combinationof thesenouns to representM31”.Given theextremedynamicalcomplexityofthisgalaxy,resultingatleastinpartfromavigoroushistoryofgalaxymergers, the nextmajor paradigm shiftwill result froma coherent, holistic viewof itskinematics and chemistry across its entire spatial extent, and whichMSE is uniquely able toprovide.


95

Chapter3 LinkinggalaxiestothelargescalestructureoftheUniverse3.1 Chaptersynopsis

• Connectingthepropertiesofgalaxiestothesurroundinglargescalestructure;• ConnectinggalaxydiversityinthelocalUniversetotheirevolutionacrosscosmictime.

ScienceReferenceObservationsLinkingGalaxiestotheLargeScaleStructureoftheUniverse• SRO-6:GalaxiesandtheirenvironmentsinthenearbyUniverse• SRO-7:Thebaryoniccontentanddarkmatterdistributionofthenearestmassiveclusters• SRO-8:Multi-scaleclusteringandthehalooccupationfunction• SRO-9:ThechemicalevolutionofgalaxiesandAGN

MSE will provide a breakthrough in extragalactic astronomy by linking the formation andevolutionofgalaxies to thesurrounding largescalestructure,across the full rangeof relevantspatial scales, from kiloparsecs through to megaparsecs. Within the ΛCDM paradigm, it isfundamentaltounderstandhowgalaxiesevolveandgrowrelativetothedarkmatterstructureinwhichtheyareembedded. ThisnecessitatesmappingthedistributionofstellarpopulationsandsupermassiveblackholestothedarkmatterhaloesandfilamentarystructurethatdominatethemassdensityoftheUniverse,andtodosooverallmassandspatialscales.

MSE is optimized for studying the evolution of galaxies, AGN and the environmental factorslikely to influence their evolution.MSEwill be the premier facility for unscrambling the non-linearregime(smallscaleclustering,mergers,groups,tendrilsandfilaments),thebaryonregime(i.e.,metallicity/chemicalevolution),andtheevolutionand interplayofgalaxies,AGN,andtheIGM, incorporating environmental factors and the key energy production pathways.MSEwillbuilduponthe impressive legacyofcurrentspectroscopicsurveys inthenearandfarfieldsbygoing several magnitudes deeper, covering an order of magnitude larger sky area, andconducting both stellar population studies of the nearest galaxies and precision dynamicalanalysesacrosstheluminosityfunctionandacrosscosmictime.

InthelocalUniverse,MSEwillmeasurethedynamicsandstellarpopulationsofeverybaryonicstructureintheVirgoandComaclustersdowntoextremelylowstellarmass.Thiswillallowthephysical association or otherwise of dwarf galaxies, nucleated dwarfs, globular clusters andultra-compact dwarfs to be probed through chemodynamical analysis of unprecedenteddatasets (both in size and completeness). Quenching processes operating in the clusterenvironmentcanbeexaminedforallsystemsatallradii,andinparticularanalysisofthestellarpopulationsofthousandsof lowmassgalaxiesall inthesameclusterwillgivedirectaccesstotherelativerolesofenvironmentallydrivenprocessesandinternallydrivenprocessesacrosstheluminosityfunctionandtoitsfaintestlimits.

MSEwillbeabletoconductthedefinitivespectroscopicsurveyofthelowredshiftUniversefordecades to come by obtaining spectra for every z < 0.2 galaxy in a cosmologically-relevantvolume spanning3200deg2, allowingeveryhalowithmassMhalo>1012M¤

tobedetectedwithfour or more galaxies. Follow up in select regions will result in a sample covering 100 deg2


96

spanning four decades in halo mass, accounting for the majority of stellar mass in the low-redshiftUniverse.ThiswillallowthemeasurementoftheshapeofthestellarmassfunctiontothescaleoftheLocalGroupdwarfgalaxies,overalargevolumeandarangeofenvironments.

Uniquely, thebroadwavelength rangeandextreme sensitivityofMSEallowsextrapolationofthis “low redshift” survey concept out to beyond cosmic noon. An ambitious set of photo-zselected survey cubeswill allowMSE tomeasure the build up of large scale structure, stellarmass, halo occupation and star formation out to redshiftsmuch greater than 2. By targeting(300 Mpc/h)3

boxes, each cube will measure “Universal” values for an array of potential

experiments. These survey volumeswill trace the transition frommerger-dominated spheroidformationtothegrowthofdisks,coveringthepeakinstar-formationandmergeractivity.Thiscombination of depth, area and photo-z selection is not possible without a combination offuture imagingsurveysandMSE.Assuch,MSEwillbeabletoproducethedefinitivesurveyofstructure,halosandgalaxyevolutionover12billionyears.

MSEwillbeunparalleledin itsabilitytoconnectthepropertiesofgalaxiestotheirsurroundinglarge scale structure. It will link the properties and diversity of galaxies observed in the localuniversetotheevolutionaryprocessesthathaveoccurredacrosscosmictime.Insodoing,MSEwill fulfillthreedistinctroles inthe2025+eraforextragalacticastrophysics:dedicatedscienceprograms,coordinatedsurveyprogramsandfeederprogramsfortheVLOTs.Allthreerolesarecritical, transformational, and will lead to major advancements in the area of structureformation, galaxy evolution, AGN physics, our understanding of the IGM, and the underlyingdarkmatterdistribution.

3.2 ExtragalacticsurveyswithMSE3.2.1 Galaxyevolutionandthelinktolargescalestructure

Thebasicmodel of galaxy formation has been in place formore than twenty years (White&Frenk 1991). Darkmatter dominates the evolution of structure in theUniverse, and it iswelldescribed by the gravitational growth of fluctuations observed in the Cosmic MicrowaveBackground. This gives rise to the cosmicweb of filaments and clusterswe see in the galaxydistribution.Theclustering,spatialstructure,andredshiftevolutionofthesegalaxiesissensitivetotheunderlyingcosmologicalmodel, theverynatureofdarkmatter,andthehighlycomplexmechanismofgalaxyformation.Whereasradiativecoolingofbaryonsleadstoefficientstarformation,variousmechanismsmustworktoreheat,orexpel,thisgasinordertomatchthelowefficiencyofstarformationobservedtoday. Large spectroscopic surveys, covering cosmologically relevant volumes withhomogeneous selection and good calibration, have demonstrated that the distribution andgrowthrateofgalaxiesislargelydecoupledfromthatofdarkmatterhaloes(Behroozi,Wechsler&Conroy2013;seeFigure42).Thisislikelyduetothehighlynonlinearnatureofcooling,starformation and heating processes such as photoionization, supernovae feedback, stellarwindsandenergyoutputfromsupermassiveblackholes.Thesephenomenaandothersoperatewithdifferent strengths, over different spatial scales, and their relative effectiveness evolves withtime.


97

Figure 42: The evolution of stellar mass fractions (stellar-to-halo mass ratio) as a function of halo mass. Thecomplex,non-linearnatureofthisplotshowsthegrowthofstellarmassislargelydecoupledfromthatofthedarkmatter(FigurefromBehroozietal.2013).

Whilethebroadpictureofgalaxyformationandevolution iscoming intofocus, it isclearthatwearestillfarfromunderstandinghowgalaxiesendupwiththemasses,structures,dynamicsandstellarpopulationsthattheyhavetoday.Linkingtheluminousmasstodarkmatterhalosiscritical,sincemanyoftheoutstandingproblemscanbegroupedbroadlyintworelatedthemes,namely,theabundanceandstructureofdarkmatterhalos,particularlyatlowmassscales,andstarformation,quenching,andtherolesofenvironmentandfeedback.

Ageneralclassofmodels,collectivelyknownas the“halomodel”,hasbeendevised toattackgalaxy formation fromastatisticalviewpoint.Amongthesemodels, theconditional luminosityfunction (CLF) and stellarmass function (SMF), both as a functionof halomass, and thehalooccupation distribution (HOD), are themost popular. Briefly, these approaches employ a fewsimplephysicalquantitiestomakepredictionsfortheclusteringproperties,luminosity(orstellarmass)functionsofgalaxies,aswellasgalaxy-galaxylensingresults.Modelsthatsimultaneouslyreproduce these observables are then representative of the underlying association betweengalaxiesanddarkmatterhalos.Thus,thehalomodelprovidesapowerful,statisticaldescriptionoftherelationshipbetweenthehalomassanditsbaryoniccontent.

AppliedtoSDSSdataatz~0,techniqueslikeHODmodeling(Zhangetal.2007)andabundancematching(Yangetal.2008),havedemonstratesthatgalaxyformationefficiencypeaksinhalosof

Mh~ 1012M

¤. At higher redshift, the constraints are poorer, as they either rely on either

photometric redshifts (e.g.,Wake et al. 2011; Leauthaud et al. 2012; Coupon et al. 2011), orspectroscopic surveys over small areas (e.g., Abbas et al. 2010; Knobel et al. 2009).Nevertheless,thesecanbeusedtotracetheevolutionofstarformationefficiencysinceaboutz~1overawide rangeofmasses.Thecosmic star formation rate isobserved tobe risingwithtimebetweenz~8andz~3,whereitremainshighandroughlyconstantuntilz~1(Hopkins&


98

Beacom2006; Seymour et al. 2008; Kistler et al. 2009;Magnelli et al. 2011; seeFigure 43).Since ΛCDMmodels predict a declining infall rate at all epochs, this observation has proveddifficult to explain theoretically (e.g.,Weinmann et al. 2011; Krumholz&Dekel 2011). Betterconstraints on halo occupation at these redshifts are needed, and these require the deep,spectroscopicsurveysthatMSEwillprovide.

Figure 43: The cosmic star formation history. Shown are the data compiled in Hopkins & Beacom (2006) (lightcircles)andcontributionsfromLyαEmitters(LAE)(Otaetal.2008).RecentluminousbluegalaxydataisshownfortwoUV luminosity function integrations:downto0.2L*z=3 (downtriangles;asgiven inBouwensetal.2008)andcomplete (up triangles).Swiftgamma-rayburst inferred rates arediamonds,with the shadedband showing therangeofvaluesresultingfromvaryingtheevolutionaryparameterbe-tweenα=0.6−1.8.AlsoshownisthecriticalstarformationdensityfromMadauetal.(1999;dashedlines,toptobottom)(FigurefromKistleretal.2009).

A key component of HODmodels, andmanyab initio galaxy formationmodels (Bower et al.2006), is that galaxies that are central in their darkmatterhalo evolvedifferently from thosethat are satellites. For central galaxies that are dominant in their dark matter halo, theirevolutionary history is strongly correlated with their total stellar mass at the epoch ofobservation (e.g., Brinchmann et al. 2004, Salim et al. 2007, Noeske et al. 2007, Daddi et al.2007,Rodighieroetal.2011,Whitakeretal.2012,Sobraletal.2014,Speagleetal.2014;seeFigure 44).However,satellitegalaxiesofagivenstellarmassappeartohaveceasedformingstars at an earlier epoch, and this leads to the well-established correlation between galaxyproperties (age,star formationrate,andperhapsmorphology)andenvironment (e.g.,Pengetal. 2010) that isobserved locally (e.g., Pengetal. 2010;Weinmannetal. 2010)andathigherredshifts(e.g.,Poggiantietal.2006;Cooperetal.2006;McGeeetal.2011;Cooperetal.2010;Baloghetal.2011).


99

Figure44:SFRmasssequenceforstar-forminggalaxiesfromz=0toz=2.5.Dottedlinesshowlinearfits,solidlinesshownon-linearslopes.Therunningmediansandscatterarecolor-codedbyredshift,withapower-lawfitabovethemassandSFRcompletenesslimits(solidlinesinbottom,rightpanel)(FigurefromWhitakeretal.2012).

Severaldifferentprocesses are thought to contribute to the central – satellitedistinction. Forexample, the amount of cold gas in galaxy disks is insufficient to sustain star formation atobservedratesforaHubbletime,anditisnecessarythatgalaxiesbe“fueled”bygasonlargerscales. For central galaxies that are dominant in their halo, this could come either throughcooling of a hydrostatic hot halo, or through “cold flows” i.e., cold, clumpy gas accreted forexamplealongfilaments.Bothmodesarelikelyimportant,atdifferentepochsandfordifferentmassgalaxies(Birnboim&Dekel2003;Keresetal.2005;Dekeletal.2009).However,satellitesarenotexpectedtohaveaccesstoasmuchofthismaterial,leadingtoarapiddepletionoffuelforstarformation(Larsonetal.1980;Cattaneoetal.2006).

Large-scale environment likely influences both the mass accretion rate (e.g., Wechsler et al.2002;Maulbetschetal.2006;Gao&White2007)andthewaythatgasisaccreted(Cattaneoetal. 2008; Dekel et al. 2009). The evolution ofmost galaxy properties (e.g., luminosity, colour,morphology) is sensitive to thegravitationalandhydrodynamic interactionsbetweengalaxies,andthedetailsoftheseinteractionsaredeterminedatleastinpartbytheenvironmentofthegalaxies. Theseeffectsaregenerally complex,nonlinearandoften subdominant tomore localeffectslikefeedback.


100

Figure45:Themeanstarformationrate(SFR)pergalaxyasafunctionofprojectedradiusforgalaxiesintheCNOC1cluster sample (solid squares) comparedwithmodel predictions. The horizontal dashed lines correspond to thefieldSFR,bracketedbyits1-σdispersion.(FigurefromBaloghetal.2000).

Distinguishingwhichofthemultitudeofprocessthataffectgalaxyevolutionaredominant,andwhere, remains a challenge, particularly since most of these effects are likely a function ofgalaxymass.Thus,acompleteunderstandingofgalaxyevolutionrequiresstudyingcentralandsatellite galaxies across a range of scale in environments, frommassive clusters (e.g., Park&Hwang2009)tosmall-mass,galactic-scalehalos(Annetal.2008).Acorrelationwithlarge-scaleenvironment might be expected because of the halo assembly bias seen in dark mattersimulations(e.g.,Maulbetschetal.2007;Gao&White2007),althoughthiseffecthasnotbeenobservedinthegalaxypopulationatlowredshift(Tinkeretal.2011).Acorrelationwithlocationwithinahaloisexpected,andobserved,asaresultofacorrelationbetweenaccretiontimeandpresent-day location (e.g., Ellingson et al. 2001; Balogh et al. 2000; seeFigure 45). Finally,independent correlationswith localdensity couldbe related to tidal interactionsandmergersbetweenclosepairsofgalaxies(e.g.,Linetal.2007;Ellisonetal.2008;Park&Choi2009;Pattonetal.2011).

Spectroscopic observations of millions of galaxies are needed to disentangle the effects ofenvironment from theotherparametersdrivinggalaxyevolution. Indeed, aftermanyyearsofspeculation based on small samples with poorly understood selection criteria, this field wastransformedbythelarge,homogeneoussurveydataprovidedbytheSDSS.Thesamplesizeandwavelengthcoverageallowedexplorationofcorrelationswithbothenvironmentandthestellarmassofthegalaxy.Thisimmediatelydemonstratedthatgalaxyproperties(suchasmorphology,age, metallicity, etc.) depend strongly on their stellar mass. Many of the trends with

0.5 1 1.5

0

0.5

1

1.5


101

environmentreportedpreviouslyturnedouttobeatleastpartlyduetothefactthatthestellarmassfunctiondependsonenvironment(Baloghetal.2001,Baldryetal.2006).Afteraccountingforthis,however,astrongenvironmentaldependenceremains:thefractionof“redsequence”galaxies (at fixed stellarmass) is larger in dense environments than in the general field (e.g.,Baldryetal. 2006). In fact, the fractionof redgalaxies iswelldescribedbya relatively simplefunctionofjuststellarmassandlocaldensity,asshowninFigure46.

Figure 46: The fraction of red (passive) galaxies is shown as a function of two different combinations of stellarmass,M,and localdensity, Σ. The symbols representdifferentmass ranges, from109<M/M*<10

10 inblue toM/M*>10

10.8inred.Theredfractiondependsonbothenvironmentandstellarmass,inasurprisinglysimpleway.(FigurefromBaldryetal.2006).

Since all galaxiesmust have been blue, star-forming systems at some point, the red fractionmustbelinkedsomehowtothedecline,orquenching,ofstarformationinbluecloudgalaxies.Tounderstandthenatureofthistransformation,onemusteitheridentifygalaxiesintheprocessofmovingbetweenthe twopopulations (e.g., in the“greenvalley”),and/or trace theredshiftevolutionofthesecorrelations.Locally,therelativelylowgasfractionsandstarformationrates,combinedwithlowinfallratesintogroupsandclusters,meanthatthefractionofgalaxiesintheprocessof transformation today isprobablyquite small. This couldbewhyobservationshaveshown the properties of star-forming galaxies at z < 0.2 to be atmostweakly dependent onenvironment (Balogh et al. 2004; Wolf et al. 2009; Vulcani et al. 2010; McGee et al. 2011;Wijesingheetal.2012).Thus,althoughthesetrendsarenowmeasuredquiteprecisely,westillhave very little insight into the physical mechanisms that might be driving any suchtransformation.Asimpleassumptionisthatsatellitegalaxiesaredeprivedofthegaseoushalosand cold flows that normally perpetuate star formation, an assumption that is at least partlysupportedbynumericalsimulations(e.g.,Kawata&Mulchaey2008;McCarthyetal.2008;Cen2011; Bahe et al. 2012). However, a simple implementation of this effect in ab initio galaxyformationmodels leads toapredictedenvironmental signature that ismuch toostrongwhencomparedwithobservations(e.g.,Weinmannetal.2006;2012).

Pengetal. (2010,2012)havetakentheobservationofBaldryetal. (2006)afewstepsfurther


102

andpresentedanespeciallyattractive,simplemodelinwhichgalaxyevolutionisdrivenbytwo“quenching rates” that are independent of one another; one related to stellarmass and theothertolocaldensity.Fromthistheyareabletopredicthowthestellarmassfunctionsofstar–formingandpassivegalaxieswillevolvetoz~ 3.ThisisshowninFigure47,wherethepointsshow the fraction of passive galaxies as a function of stellar mass for low and high densityregions,andthecurvesshowthepredictedbehaviorofthisquantitywithredshift.

Figure47:Thefractionofred(passive)galaxiesasafunctionofstellarmassforthelowandhighdensityquartilesofthedensityfield(leftandrightpanels,respectively).ThemodelparametersarefixedtoreproducetheSDSSdata(blacksquares)atz∼0(redcurve),whiletheothercurvesshowpredictionsforhigherredshifts.(FigurefromPengetal.2010).

It is important tomeasure these correlationswith comparable, or better, precision at higherredshifts in order to establish the timescales of transformation and solve this problem.Essentially,includingathirdaxisforcosmictimeinFigure46allowsadirectmeasurementofthe rate at which galaxies evolve in different environments (McGee et al. 2009). Evenmoredirectly,thehigherstarformationratesandinfallratesintoclustersatz>0.5shouldleadtoasubstantial,andmeasureable,populationofgalaxiescaughtinthemidstoftheirtransformation.In themodelofPengetal. (2010), forexample,normalizationof the“transitiongalaxy”massfunction, relative to that of the star–forming population, is equal to the transition timescaledividedbytheinverseoftheaveragespecificstarformationrateatM*.

Of course, such measurements have been attempted in the past, beginning with Butcher &Oemler (1978), and continuing with large redshift surveys today (e.g., Poggianti et al. 2006;Cooperetal. 2006;Park&Hwang2009;Cooperet al. 2010;McGeeetal. 2011;Baloghetal.2011;Raichoor&Andreon2012).Thebestconstraints todatehavecome fromthezCOSMOSand DEEP2 surveys due to their depth and relatively high sampling completeness. However,evencombinedthesesurveysconsistof70000redshiftscoveringonly~2deg2andspanning~9Gyrincosmictime.Thesamplesizealone(perGyr)is~50timessmallerthantheSDSS,andthesmallareasseverelylimittherangeofenvironmentsprobed.

MSEwillprovideabreakthroughinthefieldofgalaxyformationbylinkinggalaxyformationandevolutiontothesurroundinglargescalestructure,acrossthefullrangeofrelevantspatialscales(from kpc through to Mpc scales). The paradigm shift it will enable will likely exceed inimportancethatwhichresultedfromSDSS,dueinparttoitsabilitytoproducedatasetsofthe


103

statisticalpowerofSDSSperGyrofcosmictime,withsimilarcompletenessand limitingstellarmass. High completeness levels in MSE extragalactic surveys will be essential for studies ofenvironment to robustly identify structuresover a largedynamic range in spatial scale.Giventhe importance of correctly identifying neighboring galaxies when studying environmentaleffects(seeParketal.2008),samplingrateswillneedtobehigh:thestudyofstrong,small-scaleenvironmentsishighlycompromisedifthecompletenessis~50percentorless(Park&Hwang2009).

3.2.2 Spectroscopicredshiftsurveys

Understandingtheprocessesdiscussedintheprevioussectionnecessitatesobservingprogramsthat span redshift space, to reconstructdirectly the statisticalassemblyofgalaxypopulations.Theoverarchinggoalofgalaxyevolutionistolinkthehighredshiftregime–whereweobservegalaxiesatanearlierstageintheirevolution–tothediversityofgalaxypropertiesobservedinthelocalUniverse.Currentsurveysthatspanthelow(z<0.3,e.g.,GAMA/SDSS),moderate(z~1, e.g., zCOSMOS, Lilly et al 2007) and high redshift (z > 3, e.g., VVDS, Le Fevre et al. 2013)Universe come from hugely different telescopes. At low redshift, the extra-galactic field isdominatedbytheSDSS(2.5m,FoV~3degrees)andtheAnglo-AustralianTelescope(4m,FoV~2degrees).Athigherredshifts,amixtureof largefacilitiesdominate(≥8m,FoV<<1deg).Lowand high redshift surveys also tend to target photometric data from different facilities.Historically, co-movingvolumes forhigh redshift surveysare tiny (comfortablydominating theerror term for any “Universalmeasurement”) and tend tobehighly incomplete. Inparticular,surveys thatuseLyαtoobtain redshiftshaveverypoorvelocityaccuracy (dueto thecomplexnebulacomponent)andprobeaminorityofavailablegalaxieswithina surveyvolume (~20%,VVDS,LeFevreetal.2013).Asaresultofthiswork,however,wenowhaveagoodpictureofhowmassivegalaxiesevolverelativetothedarkmatterovertheredshiftrange0<z<1.

3.2.2.1 Capabilitiesatlowredshift

Inanyredshiftsurveythe localUniverseplaysaspecial role,as itwillalwaysbeherethat thelowest-mass galaxies can be observed at the highest physical spatial resolution. High-qualityobservationsofnearbygalaxies serveas theanchor forallhigher redshift surveys thataim tostudy galaxy evolution, and in many cases they provide the only way to connect statisticalpopulationparametersathigherredshiftwiththeunderlyingphysicalmechanisms.Inaddition,the nearest baryonic systems are generally spatially extended, and contain within them thefossil records of hierarchical structure formation in the local universe. Thus, they provide acomplementaryapproachtogalaxyevolutionfordecodingthehistoryofmergers,interactions,star formation and chemical enrichment that have operated in different environments, andacrossavastrangeingalaxymass.Here,theprimaryroleofMSEwillbetoobtainmoredetailedinformationnotaccessibleathigherredshift.

AnimportantaspectofalargefractionofextragalacticstudiesatverylowredshiftwithMSEisthe analysis of stellar populations. Good relative flux calibration between fibers is thereforecrucial.ThisisinrecognitionofthefactthatoneofthebiggestadvantagesthenearbyUniverseholdsforunderstandingthephysicsofgalaxyformationandevolutionisthelargeangularsizeofstellar systems. At first light, close-packing some fibers enables MSE to efficiently study the


104

spatiallyresolvedpropertiesoflowmassand/ordistant(z~0.1)galaxieswhosespatialextentislimited toa fewarcseconds.Full IFUcapabilities–anenvisionedsecondgenerationcapabilityforMSE – are required to fully exploit the large angular size of stellar systems in the nearbyuniverse. IFUcapabilitieswillallowMSEtotakefulladvantageof itsapertureandsignificantlyoutperformotherplannedIFUfacilitiesinprobingthegradientsandoutskirtsofnearbygalaxies(seediscussioninSection3.3.3).

Asdiscussedthroughoutthischapter,itisclearthatarangeofspectralresolutionsisparamountfor effective analysis of the properties of the closest galaxies: strong science cases exist forR~2000 (specifically, redshifts of faint galaxies and measurements of nebular emission linestrengths);R~5000 (dynamicsof galaxies); andR>>10000 (absorption lineabundanceanalysisanddynamicsofglobularclusters).Actualsciencesurveysconductedinthisareawillinterweavemultiple science programs, and so it will be desirable to have flexibility in the choice ofresolution,withthemaximumsurveyefficiencyachievedifmultipleresolutionscanbeobservedsimultaneously.Inaddition,broadwavelengthsensitivityisrequiredovertheopticalwindow:itiscrucial tomaintainsensitivityover3600–7500Å,toenablemeasurementofthe[OII]3727Åemissionlineforthemeasurementofmetalabundance,andtheHαemissionlineouttoz~0.1.Extending the wavelength coverage (for example to cover Paβ 1.282μm for amore accuratedeterminationofthedustattenuation,andthusstarformationrate,inthegaseouscomponentofgalaxies)isalsodesirable.


105

3.2.2.2 Capabilitiesathighredshift

Figure48:Thebottompanelshowstheshiftofkeyspectralfeaturesversuslookback-time(left-sideaxis)orredshift(right-side axis) with the baseline wavelength range for MSE indicated (grey rectangles indicate survey cubesrelatingtoDSC-SRO-08).Thetoppanelshowstherangecoveredbyothernotablefacilitiesandthemiddlepanelshowsthenightskyspectrum.

Whatare theneedsofMSEas they relate to studyinghigh redshift? i.e., at distances greaterthanafewhundredmegaparsecsandhenceepochspriortothepresent,z>0.1,look-backtime>1Gyr).Thisistheextra-galacticspatiallyunresolvedregimespanningabroadrangeofmasses(typically with stellar masses M* > 109M⊙), environments (pairs, groups, clusters, tendrils,filamentsandvoids),andepochs(0to~12.5Gyrs,i.e.,z<5.5).

TherelevantastrophysicalmeasurementsMSEprovidesinthisregimearelowandintermediateresolutionspectroscopyspanningtheoptical-near-IRrange.Processeswhichgenerateenergyatthese wavelengths are predominantly star-formation and accretion, potentially triggered viaenvironmental factors such as dynamical interactions, mergers and/or interactions with theinter-galactic (IGM) or intra-cluster medium (ICM). Complimentary external information willnaturally arise from panchromatic imaging facilities from the X-ray to the radio, providingpotentially high-spatial resolution imaging and potentially spatially resolved spectroscopy forwell-definedsub-samples.ThemostcompetitivefacilitiescomparedtoMSEforstudyinggalaxyevolutioninthisregimeareSubaru/PFSandVLT/MOONS.Atthe4maperturelevel,ESO/4MOSTand AAT/AAOmega will make important contributions. Within this context, MSE – as a highthroughput,largeaperturefacility–isbestplacedtoconductuniquesciencebeyondSubaruand

0 500 1000 1500

Wavelength / nm

01

23

4R

edsh

ift

03

57

89

1011

12Tr

avel

Tim

e / G

yrs

UV Absorption

G−band

MgNa

Lya OII

Hb

OIII

Ha

S1

S2

S3

S4

S5

S6

S7

S8

S9

00.

20.

40.

60.

81

Sky

Tran

sim

issi

onFa

cilit

ies

MSE

Subaru/FMOS

Subaru/PFS

VLT/MOONS

VISTA/4MOST

WHT/WEAVE

Mayall/DESI


106

VLT, best compliment LSST and Euclid andbest bridge from the 8m to 30+mgap as a feederfacility.Akeyaspectthatwecontinuallyemphasizeisthisdocumentforallscienceprogramsisto have close communications with the key imaging facilities, and in particular, LSST, Euclid,WFIRST,andtheSKA.

The z > 0.1 case does not drive the capabilities ofMSE towards high spectral resolution; R~3000 issufficientforalmostallastrophysicalobjectivesbecauseofthe inherentmotionwithinthe structures being studied (the exception to this is resolving the Lyα forest, although thisrequires extremely high resolution, R > 150000). Similarly, IFUs are not as critical as for thenearbyUniverse,sincetheresolutionofthemajorityofhigherredshifttargetsarecomparableto the native seeing of the site. However, extensive wavelength range, spanning the entireopticaltoNIRrangeisanimportantscienceenablingcapability,andthesciencecasesdescribedinthissectiondodrivetheredendofthewavelengthrangeintothenearinfrared.Figure48shows how key astrophysical tracers for extragalactic astronomy drift as a function ofwavelength.Fromascienceperspective,thereisnoobviousscience-drivenNIRcutoffinsofarasthe NIR cutoff effectively sets a redshift limit beyond which analyzing galaxies using ahomogeneoussetoftracers isnotpossible. Inadditiontowavelengthcoverage,thehighfiberdensityofMSEandtheabilitytoclosepackfibers is important,sincesurveyspeedscaleswithfiberdensity formanyof theenvisioned scienceprograms (suchasSRO-08, SRO-09), and it isimportanttoovercomefibre-collisionbias,aseriousconcernforthedrivingclusteringstudies.

3.2.3 Bringingittogether:theroleofMSE

Environmental considerationsona rangeof spatial scalesdemonstrate that,within theΛCDMparadigm, it is fundamental to understand howgalaxies evolve and grow relative to the darkmatter structure inwhich theyareembedded. At theheartofallofMSE’s investigations intogalaxyevolutionisadetailed,fullysampled,mappingofthedistributionofgalaxies,theirstellarpopulationsandsupermassiveblackholes,tothedarkmatterhaloesandfilamentarystructurethatdominatethemassdensityoftheUniverse,andtodosooverallmassandspatialscales.Thesemustbeanchoredtoprecisionmeasurementsoftheclosestgalaxies,sinceitisonlyinthenearby Universe that we can access the lowest mass galaxies, conduct high signal-to-noiseanalyses of galactic stellar populations, and obtain spatially-resolved insights into theirdynamicalandchemicalevolution.

The capabilities of MSE – its sensitivity, etendue, look-back time, wavelength range anddedicatedoperations–are such thata fundamental goal for the facilitymustbe toprovideaholistic view of galaxy evolution connecting the historically distinct regimes of the nearbyUniverse with the high redshift Universe. MSE is ideally suited to the design of observingcampaignswithequalco-movingvolumesfromlowredshift(z~0)tohighredshift(z~3).Atallredshiftslices,MSEcampaignswillreachtolowstellarmass,willcoverarepresentativevolumeoftheUniverse,willbespatiallycomplete,andwillhavestatisticaluncertaintiessub-dominantevenwhengalaxiesarebinnedbytheirproperties.Suchdatasetsdefinitivelyanswerahostofsciencequestionsover12Gyrsinlook-backtime:

• Detailedkinematicalandstellarpopulationanalysisofthecentralregionsofgalaxiestoexploresupermassiveblackholescalingrelationsacrossthefullmassandenvironment


107

regimesandexploringcorrelationsbetweenthepropertiesoflowmassstellarsystemssuchasglobularclustersandthelargescalepropertiesofthehostgalaxies;

• Spatially-resolved chemodynamic studies probing merger remnants at largegalactocentricradiuscombinedwithdetailedhalodynamicsforeachtargetgalaxy;

• SpatialclusteringanddynamicsoverscalesrangingfromkpctohundredsofMpc,tolinkgalaxies to the underlying dark matter distribution. Specifically, to identify galaxylocations and dynamics within bound haloes, and filamentary structures, and thedistributionofthosehaloesrelativetooneanother;

• Measurementofmergerratesasafunctionofredshift,galaxytypes,environmentandoverawiderangeingalaxymass.

• Properties of stellar populations, including ages, star formation rates, chemicalabundanceand initialmass functionparameters,particularlyat lowmasseswhere thegalaxymass–halomassrelationismostuncertain;

• TheoccurrenceandstrengthofAGNactivity,expectedtoplayasignificantroleingalaxyformationbyprovidinganimportantsourceofenergythatinhibitsstarformation;

In this chapter,wediscuss indetail someof the specific scientificopportunities thatMSEwillenable for understanding galaxy formation and evolution. We begin by discussing detailedstudies of the nearest galaxies across all mass scales (Section 3.3) before discussing theevolution of galaxies across cosmic time (Section 3.4). The science described provides acomprehensiveandtransformativecontributiontoourunderstandingoftherelationofgalaxiesto their surrounding large scale structure, and connects the galaxy diversity observed in thenearby Universe to their evolution observed across cosmic time. Some topics specific to theIntergalacticMedium and the detailed physics of black hole accretion are discussed in detailelsewhere,inChapter4.

3.3 Galaxiesacrossthehierarchy

ScienceReferenceObservation6(AppendixF)

GalaxiesandtheirenvironmentsinthenearbyUniverse

Tounderstandthephysicaldriversofgalaxyevolutionit isnecessarytomapthedistributionofstellar populations and supermassive black holes to the dark matter haloes and filamentarystructurethatdominatethemassdensityoftheUniverse.Thisrequiresdeep,spatiallycompletespectroscopicsurveysoverwideareas, ideallysuitedtothecapabilitiesofMSE. Weproposeawide-areasurveyS1-W,thatobtainsredshift-qualityspectraforeveryz<0.2galaxyovera3200deg2 area. This is efficiently achieved if target selection is basedondeep imaging cataloguesspanningtheNUVtoNIR,so thatprecisephotometric redshiftsareavailable. Thissurveywill,amongotherthings,alloweveryhalointheareawithmassMhalo>1012M¤

tobeidentifiedwithagroupoffourormoregalaxies. Weproposetofollowupselectregionswithinthissurveywithlonger exposures, resulting in a sample covering 100 deg2 to a depth of at least i<24.5, andspanningfourdecadesinhalomass.Thiswillallowustomeasuretheshapeofthestellarmassfunction to the scale of the Local Group dwarf galaxies, over a large volume and a range ofenvironments. A subset of fibers will be used to obtain long integrations on bright galaxies,yielding a final sample of >50000 high signal-to-noise ratio spectra, forwhich detailed stellar


108

populationsynthesisanalysis ispossible. Thiswillbe thedefinitivespectroscopic surveyof thelocalUniversefordecadestocome.

3.3.1 ThelowestmassstellarsystemsintheLocalUniverse

Figure49:Leftpanel:Distributionoflow-massgalaxiesassociatedwiththeMilkyWay.ThesampleofsuchsystemshasnearlydoubledinrecentyearsthankstotheDarkEnergySurvey(Imagecredit,A.Frebel,MIT).Rightpanel:Anultra-faintdwarfgalaxy–comparableinluminositytothelow-massGalacticsatellitesshowninthepreviouspanel–intheVirgoCluster.ManythousandsofsuchsystemsareexpectedtobediscoveredbyfutureimagingfacilitieslikeLSST,EuclidandWFIRST(FigurefromJang&Lee2014).

Themenagerieofknownstellarsystemshasgrownsignificantlyinrecentyears,withreportsofseveralnewclassesofcompact,low-massstellarsystems.Thesediscoveriesinclude“extended”or“diffusestarclusters”(e.g.,Brodie&Larsen2002;Huxoretal.2005;Pengetal.2006),“ultra-faintdSphgalaxies”(e.g.,Belokurovetal.2006;Figure49)and“ultra-compactdwarfs”(UCDs;Drinkwater et al. 2003). Other types of low-mass, compact stellar systems include the morefamiliarglobularclustersandnuclearstarclusters(see,e.g.,therecentcompilationofMihosetal.2015shownintheleftpanelofFigure50).Atthistime,areliablecensusofsuchstellarsystemsinthelocaluniverse—aprerequisiteforunderstandingtheirpropertiesandorigin—islacking,butitisabundantlyclearthatthesefaintand compact objectswill play a key role in the coming decade in answering some importantopenquestions inastrophysics.Unfortunately, their faintnessandcompactnaturemeans thatthey are difficult to identify and study. Although clever and efficient techniques have beendevelopedtoefficiently identifycandidatessystemsfromhigh-resolutionUV-visible-IR imaging(e.g.,Munozetal.2014;Liuetal.2015),confirmationandcharacterizationofthesesystemswillalwaysrequireopticalspectroscopy.


109

Figure 50: Left panel: Magnitude–size–surface brightness relations for a collection of dynamically hot stellarsystems,mostly intheVirgoandFornaxclusters (FigurefromMihosetal.2015).Rightpanel:Comparisonoftheobservedandsimulated two-dimensionalvelocitydispersionmap forM60-UCD1,oneof themostmassiveUCDscurrently known. The central velocity dispersion “spike” is the signature of a 21 million solar mass black holeembeddedinthislow-mass(1.2×108M

�)andcompact(Re=24pc)stellarsystem(FigurefromSethetal.2015).

The next generation of imaging facilities (LSST, Euclid, WFIRST) — which collectively offercomplete SED coverage over the optical and IR regions, good sensitivity and high angularresolution—canbeusedtoselect largeandwell-characterizedsamplesofnearbycandidates,but,asyet,noplannedfacilitywillbecapableofdetailedandefficientspectroscopicfollowup.Withitslargefieldofview,highsensitivity,highmultiplexing,andrangeofspectralresolutions,MSEwouldspectacularly fill this role,andwouldallowastronomers in the2020s toaddressanumberofimportantopenissues.Theseinclude:

i. TheAbundanceofCompactStellarSystems.Asubsetofcompactsystems(theultra-faintdSphs)arethoughttocorrespondtothelow-mass,darkmatterhalospredictedbycosmologicalsimulations.Atpresent,theobservednumbersofsuch systems fall well short of the predictions of “standard” cosmologicalmodels, which haspotentiallyprofound implications for thenatureofdarkmatter (i.e.,whether it is cold,warm,mixed,decaying,etc.)andforgascooling,star formationandfeedback in the low-masshalos.Thespacedensitiesofothertypesofsystems,suchasUCDs,bearsdirectlyontheefficiencyoftidal stripping in low-mass galaxies. Unfortunately, spectroscopic confirmation of carefullyselectedtargetsisessentialifwearetobuildacompletecensusofsuchsystemsoverarangeofenvironment.Evenintherichestnearbyclusters,suchasVirgo,thisisamonumentaltaskthatrequires low-resolution spectra (with SNR~10 at R ~ 2,000 and g~24 – 25) forhundreds ofthousandsofcandidatesdistributedover~100deg2fields.

ii. GalaxyDisruptionandTransformation


110

Thereismountingevidencethatmany,andperhapsmost,UCDshaveformedviatidalstrippingof low-massgalaxiesthatinitiallycontainedacentralnuclearstarcluster. Independentlinesofevidenceadditionallysuggestthatat leastsomenuclearstarclusterswere, inturn,assembledthroughmultiplemergersofglobularclustersinlow-masshalos.Finally, it ispossiblethattidaleffects acting on initially compact star clusters might inflate their velocity dispersions andeffectiveradiisothattheycouldmasqueradeasultra-faintdSphs.Acommonelementinthesedifferentscenariosisthetransformationandpossibledisruptionoflow-massstellarsystemsbytidaleffects.Testingthesesuggestionsacrossarangeofmassandenvironmentwillhingeonourabilitytoacquirespectroscopyforlargeandrepresentativesamplesofcompactstellarsystems.In this case, desirable data products would include radial velocities (for membership anddynamical studiesof sub-populations),agesandmetallicities (for testsof formationscenarios)andinternalvelocitydispersions(fordynamicalmassestimatesandtestsofdarkmattervs.tidalheating).iii. ChemicalEnrichmentandStarFormationinExtremeEnvironments

Some types of stellar systems are noteworthy in being rare and extreme environments. Forinstance, UCDs and nuclear star clusters are among the densest environments known in thelocal universe and can provide insights into how star formation and chemical enrichmentproceedsundersuchcircumstances.Theyarealsoextremeenvironmentsforexploringtheformof the initial mass function, which mounting evidence suggests may vary with mass and/orenvironment.iv. DarkMatteronLargeandSmallScales

Orbitingwithin the gravitational potentials of their host galaxies, compact stellar systems aredynamical testparticles thatcanbeused toprobe thedistributionofdarkmatterwithin low-and high-mass galaxies. The technical requirements for spectroscopy (most notably, spectralresolution)willdependstronglyonthechoiceofhostgalaxy,withhigherresolutionneededforthe lowest-mass systems. In a related issue, spectra for any compact sources associatedwithfine structures seen in somegalaxyhalos (such as tidal streams, plumes and shells), canhelpshedlightontheaccretionhistoriesofthesegalaxies.v. SupermassiveBlackHolesandTheirEnvironments

Theexcitingdiscoveryofa supermassiveblackhole (BH) inM60-UCD1,one themostmassiveUCDs in the Virgo Cluster (Seth et al. 2015), has provided compelling evidence that tidalstrippingoflow-massgalaxiesisaviableformationchannelforUCDs.Equallyimportant,BHslikethe one in M60-UCD1 provide a rare opportunity to explore the BH mass function and theoccupationfunctioninlow-andintermediate-massgalaxies.WhilethesecuredetectionofaBHinaUCD,globularclusterornuclearstarclusterwillalwaysrequireAO-assistedorspace-basedIFU spectroscopy (from, e.g., TMT or JWST), the pre-selection of candidate systemsmust bemadeon thebasisofmoderate-SNR, intermediate-resolution, integrated-light spectroscopy; apowerfulnicheofMSE.


111

3.3.2 Theevolutionoflowmassgalaxies

Figure51:ThelocalstellarmassfunctionasmeasuredbySDSS,2dFGRSandGAMA.GAMAisthedeepestofthesesurveys,andprobesrobustlydowntothescaleofM32;lowermassmeasurementsareactuallylower-limitsduetosurfacebrightnesslimitations.TheshadedregionshowstheprecisionandstellarmasscoverageanticipatedwithMSEthroughSRO-06.

In the past 15 years, imaging and spectroscopyof large, unbiased samples of nearby galaxieshave provided a good, empirical description of massive galaxy formation and evolution.Specifically,theseobservationshaveallowedtheidentificationanddetailedcharacterizationoffundamental scaling relations, such as those between stellar mass, SFR andmetallicity. Thelargesamplesizesallowedthemeasurementofthescatter intheserelationships,andoftheirdependence on other parameters, such as morphology, AGN activity and environment(Kauffmann et al. 2003; Tremonti et al. 2004; Balogh et al. 2004). Measurements of galaxyclustering,weaklensing,andvelocitieshaveenabledtherelationshipbetweengalaxiesandtheirdarkmatterhaloes tobemapped insomedetail (e.g., Leauthaudetal.2012;vanUitertetal.2016). Through observations like these, which inform and constrain theoretical models, ourunderstandingofthephysicsdrivinggalaxyevolutionhasundergonetremendousgrowth.

However,theextensionofthesestatisticalstudiesintothedwarfregimeisrelativelyunchartedterritory,duetotheflux-limitednatureofmostsurveys.Ourknowledgeoftheverylowestmassgalaxies isrestrictedtothose intheLocalGroup(seediscussion inBoylan-Kolchinetal.2012),andthelocalobservationalcensusislikelyincomplete(Tollerudetal.2008,Koposovetal.2008,McConnachie2012).Shallow,wide-areasurveyssuchastheSDSS(Yorketal.2000)andthe2dFgalaxy redshift survey (2dFGRS; Colless et al. 2001) areonly sensitive tomassive galaxies andintermediate-massdarkmatterhalos(Ekeetal.2006;Yangetal.2007).Andwhileahandfulofvery low-mass systems have been found in both these surveys, neither the 2dFGRS nor SDSSwereabletoidentifyhosthalosbelowMh~1012M¤

,therebybarelysamplingthedominanthalo


112

massattheirsensitivitylimits.

TheGAMAsurvey,withspectroscopycompletedin2014,improvedupontheSDSSintwomainrespects. First, the increaseddepthof thesurveyextended the limitsof themeasuredstellarmassfunctionbymorethananorderofmagnitude,fromM*∼5x108M¤

tolessthanM*∼107M¤

(seeFigure51).Secondly,GAMAobtainedspectrafornearly100%ofthemagnitude-selectedtarget sample, independent of small-scale clustering. This sampling is critical to robustly linkgalaxieswith theirdarkmatterhaloes, and todistinguish satellitegalaxies from thedominantgalaxy ineachhalo. Thoughcoveringonly300squaredegrees, thissurveydemonstratedthatlow-mass galaxies evolve in a fundamentally different way from their more massivecounterparts(Lara-Lopezetal.2013a,2013b,Baueretal.2013).Thehighspatialcompleteness,moreover, enabledanunprecedented studyof satellite galaxyevolution (Prescott et al. 2011,Figure52),andmergerrates(Robothametal.2014).

Figure52:Theredfractionofsatellitegalaxiesasafunctionofprojectedradius,forfourdifferenthoststellarmassrangesintheGAMAsurvey(Prescottetal.2011).Theseresultsshowedthatquenchingofstarformationismostefficientinmassivesystems,andactswithin∼500kpcofthehost.Thispointstorelativelyslowmechanismsofgasremoval,likestrangulation(Larsonetal.1980;Baloghetal.2000)asthemaindriverofthisdifferentialevolution.(FigurefromPrescottetal.2011).

Building on these new insights into the dwarf galaxy regime, the next generation of galaxysurveys in the nearby universemust target large, homogeneous samples of the lowestmassgalaxies.Thecombinationofhighsensitivity,widefieldofview,andsurveycapabilityprovidedbyMSE isperfectlysuitedtoplaythisrole. Toachieveanunderstandingof low-massgalaxiesthat rivals what we know for their more massive counterparts, MSE should conduct acomprehensivesurveyofseveralthousandgalaxiesinthemassrange107<M*/M¤

<109,spanninga range of environment (from rich cluster cores to filaments and voids) and including allmorphologicalclasses.

SRO-06describesapossibleobservingprogramwithMSE thatmeets theabove requirementsandthatallowsfortransformativeadvancesinmanyareas.SRO-06extendswellintotheregime


113

ofdwarfgalaxies(i.e.,tostellarmassesofM*∼107M¤atz<0.1anddowntoanultimate,low-zmasslimitofM*~105.5M¤

),enablingstudiesofhalooccupancy,halodependentmassfunctions,halosubstructure,andstarformationefficiency,describedinmoredetailbelow.Galaxieswillbeselectedusingwell-defined criteria tominimize selection bias,with complete sampling of themassfunctiondowntoverylowmassesandverysmallseparations.Throughsuchasurvey,MSEwill impact a vast range of science areas pertaining to galaxy evolution, particularly at lowmasses.SRO-07,discussedlater,describesasurveyofgalaxyclustersthataugmentsmanyofthescience goals of SRO-06 through extension to more extreme environments, including ournearestneighbours,ComaandVirgo.

3.3.2.1 Halooccupationandtheformationofthestellarmassfunction

i. StellarmassfunctionThemost fundamentalmeasurement thatMSEwill make as part of SRO-06 is to extend theobserved stellarmass function tomasses below108M

¤,with a cosmologically representative,

unbiased,spatiallycompletespectroscopicsample. Figure 51showsthecurrentconstraintson the z < 0.1 galaxy SMF fromGAMA (Baldry et al. 2012), the SDSS (Baldry et al. 2008), the2dFGRS (Cole et al. 2001), and the predicted performance forMSE (SRO-6). The stellarmassfunctionisofparticularinterestbecausenumericalsimulationshavelongarguedthathalossuchastheLocalGroupshouldcontainthousandsofgalaxies,ratherthanthe~100thatareknown(Moore et al. 1999; Klypin et al. 1999). Indeed, observations find that the Local Groupmassfunctionisrelativelyflat(Pritchet&vandenBergh1999,Driver&dePropris2003,Ferrareseetal. 2016). There are two principal mechanisms that have been proposed to reconcile thisapparent discrepancy (Benson et al. 2003), namely supernova feedback (whereby SNIIexplosionsheatandexpelasignificantfractionofthebaryonsfromlowermasshalos),andearlyphoto-ionization (in which the primordial baryons in low-mass halos are heated by externalsources such as AGN and Pop III stars, and are thereafter unable to cool due to low gasmetallicities).Thesetwomechanismsgivequitedistinct,qualitativepredictionsforthefaint-endslopeofthegalaxySMF,astheformermechanismleavesarelicstellarpopulationineveryhalowhilethelatterdoesnot.Thus,byextendingexistingworktoobservationsofgalaxiesmorethananorderofmagnitude lower inmass,MSEwill provide leverageondecoupling the effects oftheseheatingmechanismsfromoneanother.

ii. DarkmatterhalooccupationIntheΛCDMparadigm,galaxyformationtakesplacewithindarkmatterhalos.Measurementofthe overall abundance of such halos as a function of mass, dN/d(logM) is among the mostimportantoutstandingissuesinmoderncosmology.Thisquantityissensitivetotheformoftheinflationaryperturbationsandtothenatureofthedarkmatterparticleitself:i.e.,whetheritiscollisionalor collisionless,decaying, self-interacting,warm (light), or cold (massive).Moreoverestablishing its functional form at low redshiftwith high precision is critical formeasuring itsevolution,whichinturnissensitivetothecosmologicalparameters.Linking the luminous mass to dark matter halos is critical, both to understand the haloabundance and to understand galaxy formation. A key focus for SRO-06 is therefore a


114

comprehensive sampling of both the halomass functionand the stellarmass function to thefaintestpossiblelimits.Forafullysampled,r≤24magnominalsurvey,itshouldbepossibletodetect~200halosper100deg2withlog(Mh)/M¤=10.75±0.5dex.Thisis50timesgreaterthanthedensityfoundintheGAMAsurvey(Robothametal.2011)—thedeepestprecursorforhalomeasurements (although the small area limits its utility for the purpose of cosmologicalconstraints)— andmore than an order ofmagnitude beyond themass range probed by theSDSSand2dFGRSdatasets(Rinesetal.2007;Tempeletal.2014).

3.3.2.2 Galaxygrowthandtransformation

Thehierarchicalassemblyofgalaxiesthroughmergingisafundamentalpredictionofcolddarkmatter models, and plays a particularly important role in the evolution of galaxy sizes anddensities(e.g.,Trujilloetal.2006).AsdiscussedinSection3.3.4,mergerratescanbeestimatedfromthefrequencyofclosepairs,potentiallyincludingmorphologicalinformation(Robothametal.2014;Casteelsetal.2014).ThehighlycompletespectroscopicsurveydescribedinSRO-06isideal for these analyses, in order to identify close pairs in velocity as well as space. Suchmeasurementsprovidenotonlyatestoftheunderlyingdarkmattermodelprediction,butalso,togetherwiththein-situstarformationrate,acompleteempiricaldescriptionofgalaxygrowthratesatagivenepoch.

Inadditiontomerging,galaxiesowemuchoftheirpresentmasstoin-situstarformation.Howstar formation is regulated, or even stopped, in a galaxy is an unanswered question offundamental importance. Formassivegalaxieswehavedevelopedagoodpictureofhowthestar formationratedependsonstellarmass,halomassandepoch (e.g.,Behroozietal.2013).However,thereisevidencethattheserelationshipschangeatlowmasses, inawaythatisnotpredictedbycurrentstate-of-the-artsimulations(Figure53).MSEwillextendtheanalysisshownin Figure 53 by at least an order of magnitude in stellar mass, over a much larger area.Furthermoreitwillbepossibletomeasureanydependenceonhalomass,whichisapredictionofmostsatellitequenchingmodelsbutremainscontroversialobservationally(e.g.,McGeeetal.2009;Wetzeletal.2013;Vulcanietal.2010;Rasmussenetal.2012;Luetal.2012).

Twomajorphenomena–AGNandstellarfeedback–arethoughttobearprimaryresponsibilityforthelate-timequenchingofstarformation,withadditionalphysicalprocessesplayingaroleinhigh-densityenvironments.ThefeedbackofAGNhasbeenproposedasapossibleoriginofthequenching of the star formation activity inmassive galaxies (Bluck et al. 2014), while that ofsupernovae and massive stars dominates in dwarf systems (but see discussion below). Thekineticenergy injected in the ISMby thenuclearblackhole inmassivegalaxiesandby stellarwindsindwarfsisabletoremovethegaseouscomponentfeedingthestarformationprocess.


115

Figure 53: Predictions of the sSFR-mass relationship from the EAGLE hydrodynamic simulations, comparedwithobservations at low redshift. The purple and green solid lines are low and high-resolution simulation results,respectively; the dashed lines indicate where resolution effects may be important. There is indication of adiscrepancywiththeSDSSStripe82analysisofGilbanketal.(2010)attheloweststellarmassesprobed. (FigurefromFurlongetal.2015).

i. DodwarfgalaxieshostAGN?

The roleofblackholes in galaxy formation isprobablyoneof themostdebated topics in thefield(seetheAnnualReviewarticlebyKormendy&Ho2013).Currentobservationsarelargelylimitedtohighmassgalaxies,wherethemajorityofgalaxiesareconsideredtohostablackhole.However,severalhundredlowmassand/orbulgelessgalaxiesarenowalsosuspectedofhavingcentral black holes (e.g., Greene & Ho 2004, Reines et al. 2013, Satyapal et al. 2014). Asdemonstrated by the work at higher masses, large samples are required to study thedemographicsof thispopulation.Currentwork is focusingon fine-tuning selection techniquesand obtaining follow-up observations to confirm successful identification of an AGN (e.g.,Secrestetal. 2015). Thecomingyearspromise topindown theoptimal selection techniques,likelythroughacombinationofopticalspectroscopyandmid-IRimaging(e.g.,fromWISE).New,largespectroscopicsamplesoflowmassgalaxies,withgoodqualityimaging,willbeneededtoexploitthisgroundwork.

From analysis of emission line ratios with MSE, it is possible to determine the occupationfractionofAGN in lowmassgalaxies,anddeterminewhether supermassiveblackholes relatemorecloselytothespheroidal(bulge)componentortootherpropertiesofthehost(e.g.,totalgravitationalmass).ThelargesampleofobjectsidentifiedbyMSEwilloffertheopportunitytostudytheoccupationfractionofAGNinlowmassgalaxies,determinewhetherSBHsrelatemore


116

closely to the spheroidal (bulge) component or other properties of the host (e.g., totalgravitationalmass),anddiscoverhowtheAGNaffectsthestructure(onbothglobalandnuclearsclaes)ofthehostgalaxy.

ii. Theroleofenvironment

In high-density environments, such as clusters and groups, gravitational interactions orinteractionswith the hot and dense ICM trappedwithin the potentialwell of the over-denseregion (such as ram pressure stripping and thermal evaporation) can efficiently remove thegaseouscomponentofgalaxies, transformingstar formingsystems intoquiescentobjects.Thestudy of these physical processes is crucial for understanding galaxy evolution. However,analysisofsatellitesintheLocalGrouphasdemonstratedthatquenchingofstarformationhasacomplex dependence on stellar mass that is not understood. Figure 54 (from Fillingham et al.2015;seealsoGehaetal.2012)showsthatwhiletheeffectivenesswithwhichstarformationisquenchedinsatellitegalaxiessteadilydeclineswithdecreasingstellarmassinthemassregimeprobedbyalllargespectroscopicsurveys,atlowermassesthetrendappearstoreverse.Whilethe effect is subtle, this differential measurement (comparing satellite galaxies with centralgalaxies)isapotentiallypowerfulwaytoconstrainfeedbackparameters(e.g.,McGee,Bower&Balogh2014;Baloghetal.2016). AstatisticallysignificantsampleofgroupswithLocalGrouphalo masses is therefore a critical complement to these local studies, to determine theuniversality,orenvironmentaldependence,of the remarkably lowefficiencyof star formationobservedintheverylocalUniverse.

OnlyMSEisabletomakethismeasurementoveracosmologicallyrelevantvolume,andwithahomogeneousselectionofgalaxiesoverthefullmassrangeasdescribedinSRO-06.

Figure54:Dependenceoftheinferredstarformationquenchingtimescaleofsatellitegalaxiesasafunctionoftheirstellarmass. Shortquenchingtimescalesleadtohighfractionsofgalaxieswithlittleornostarformationattheepochofobservation.LargespectroscopicsurveyslikeSDSSandGAMAhavebeeninstrumentalinuncoveringthetrend in decreasing timescale with increasing stellar mass shown as the colored lines at the high-mass end.However, this and other analyses of the Local Group show that the lowest-mass satellites may respond verydifferentlytoenvironment(FigurefromFillinghametal.2015).


117

3.3.2.3 Stellarpopulations

WhilerelativelylowSNRspectraareusefulforprobingdynamicsandlargescaleenvironment,asignificantadvanceinourcurrentunderstandingrelatingtotheevolutionofbaryonsingalaxiesrequireshighqualitydatatomeasurethestarformationandenrichmenthistoriesofthestellarpopulations. Colors are useful for distinguishing currently star-forming galaxies from passiveones, but to break the age-metallicity degeneracy and so obtain more precise ages of thedominantredpopulation,oneneedsabsorptionlineindicesfromhighSNR(>25)spectraforasignificant sample of the galaxies. Thus, SRO-06 will transform our understanding of severalrelatedissuesconcerningthemassdependenceoffundamentalscalingrelationsofgalaxies,andtheroleoffeedbackinregulatingtheevolutionofbaryonsingalaxies.Highlightsinclude:

i. ChemicalenrichmentThemass-metallicityrelationderivedfromSDSSgalaxiesextendstologM*/M¤

= 8.5(Tremontietal.2004).Pushingtolowermasseshaspreviouslyonlybeenpossiblewitheitherahandfulofnearbygalaxies (e.g., Leeetal.2006)or stackinganalyses (Andrews&Martini2012), andhasextended this relationdownbya further factorof ten inmass, to suggest that themetallicitycontinues to fall with mass. However, at low masses, the chemical enrichment is likely tobecomeincreasinglystochastic,necessitatinglargesamplesofindividualmeasurementsthatarecarefully modeled (and not just stacked) to account for, e.g., varying levels of nitrogenabundance,which can be elevated in dwarfs (Berg et al. 2011). Intriguingly, the lowestmassLymanBreakGalaxiescurrentlydetectableatz= 2alsohaveanomalouslyhighN/Oabundances,sothatunderstandingtheoriginforthesechemicaldeviationshasapotentiallybroadimpact.

Atlowmasses,weexpectchemicalenrichmenttobecomeincreasinglystochasticandsensitivetofactorssuchaswinds,infallandenvironment(Kirbyetal.2013),potentiallywithametallicity“floor”whereself-enrichmentisdrivenbyafewgenerationsofstars(Sweetetal.2014).Here,again,theenvironmentaldependenceofthesescalingrelationsismass-dependent(Ellisonetal.2009; Cole et al. 2014). Quantifying the role of these factors requires individual metallicitymeasurements for a cosmologically-relevant sample of dwarfs. Moreover, outliers in mass –metallicityspacecanhelpusquantifythefractionofdwarfsthathaveatidalorigin(e.g.,Croxalletal.2009),whichmayrepresentasignificant(butcurrentlypoorlyconstrained)fractionofthedwarfpopulation(e.g.,Ploeckingeretal.2014andreferencestherein).

ii. Trackingthevariationintheinitialmassfunction

The role of the initial mass function (IMF) for studies of galaxy formation and evolution canhardlybeoverstated,as theassumptions regarding its slopeandmass rangeare thebasis fortheinterpretationofmostgalaxyintegratedobservablepropertiesatanyredshift.Recentwork(van Dokkum & Conroy 2010; Capellari et al. 2012; Smith et al. 2012a) has suggested theintriguingpossibilitythattheslopeofthestellarIMFvariesasafunctionofgalaxymassand/ormorphological type (seeFigure 55).But thisevidence remains controversial, andbecauseofthe small samples for which such measurements are available today, it is unknown to whatextent the IMF depends on spatial scale, galaxy type, mass, redshift or environment. It ispossible to put useful constraints on the high-mass end of the initial mass function using


118

combinationsofHαemissionlinesandgalaxycolour(Kennicutt1983;Baldry&Glazebrook2003;Gunawardhanaetal.2011). AndifspectroscopywithSNR>100isavailable,ameasurementofthe IMF shapeat the low-mass endalsobecomespossible via thedwarf-sensitiveNaI 8183Å,8195ÅandFeH9916Å (Wing-Ford) lines (vanDokkum&Conroy2010;Cappellari etal. 2012;Smithetal.2014).

Figure 55: Compilation of observed relations (black solid lines) for how the slope of the IMF varieswith galaxycircularvelocity,andtheirextrapolations(blackdashedlines).AlsoshownareindividualmeasurementsforsomeLocal Group galaxies, based on resolved star counts. This remains a tremendously important and controversialtopic, with data limited to small samples of inhomogeneously selected galaxies. (Figure from McGee, Goto &Balogh2014).

3.3.2.4 Theevolutionoflowmassgalaxies:summary

MSE will be the preeminent facility for studies relating to missing satellites, group finding,mergerrates,thestudyofthegalaxypopulationswithrespecttohalomass,andthehalomassfunctionandSMF. Itallowsthesamplingof thegalaxystellarmass functionrightdowntotheJeans limitofM*~105.5M¤withinenvironments ranging fromrichclusters togroup, filament,andvoids.Criticalfactorsforsuccesswillbecompletesamplingoverrepresentativevolumesataspectral resolution sufficient to obtain ± 20 kms-1 accuracy, in established regions with high-quality,multi-wavelengthdata.SuitableregionsincludetheequatorialGAMA9h,GAMA12handGAMA15hregions (Driveretal.2011)thatcontaindeepGALEX,VST,VISTA,WISE,HERSCHEL,GMRT,andASKAPdata.Highcompletion iscriticalbecausemany low-masshaloswill typicallycontainonly a fewgalaxies and all of these galaxieswill need tobe sampled to attain robustvelocity dispersions (i.e., halo masses). Inevitably, this will require a multiple-pass survey tounpicktheclosepairs, tripletsandcompactgroups.Areasonableresolution(i.e.,R∼2000) isneeded to measure the velocity dispersions of the lowest-mass groups (note that the Local


119

Groupvelocitydispersion is60kms-1).For fullcharacterizationof thegalaxypopulationwithineach halo, ancillarymulti-wavelength datawould be required such as that assembled by theGAMA team (for anoverview, seeDriver et al. 2016), and towhich a largenumberof futurefacilitieswillcontribute,asdiscussedindetailinChapter1.

3.3.3 Spatiallyresolvedstudiesofgalaxies

Integral field spectroscopy is arguably themost powerful way to capture the complexities ofgalaxy formation, to examine how the darkmatter, stars, neutral gas, ionized gas and super-massiveblackholesall interactwitheachother.Therearea largenumberofobservablesthataredifficultor impossibletoobtainfromsinglefibresurveysandthatallowdirecttestsofthelatest galaxy formation theory. Observables from integral field spectroscopy include gas andstellarinternalandbulkkinematics,thespatialdistributionofstarformation,stellarmetallicityand abundance gradients, stellar age gradients, gas phase metallicity distributions, resolvedionizationdiagnosticsandmanyothers.

Table5:AsummaryofcurrentandfutureMOSIFUinstruments.Foreach instrumentandtelescopecombinationwelisttheactualorexpecteddateofcommissioning,wavelengthrange,telescopeaperture,field-of-view,numberof fibres, fibresize (diameter), thenumberof IFUsandthespectral resolution (R=λ/dλ inFWHM).KMOS is theonlysystemusingimageslicersratherthanfibres,sowegivethenumberofspatialelementsintheslicerandtheslicer spatial sampling at *.

The power of integral field spectroscopy has beenwidely recognisedwithin the astronomicalcommunity with most large telescopes now having integral field spectroscopy capabilities ofsometype.Wearealreadyatthestageofhavingcompletedatleasttwogenerationsofmajorintegral field surveys (e.g., SAURON/ATLAS3D, Cappellari et al. 2011a;CALIFA, Sanchez et al.2012) and the first generation ofmajormulti-object integral field surveys are now underway(SAMI,Croometal.2012;MANGA,Bundyetal.2015;KMOS,Stottetal.2016).Futureprojectsarealreadyatanadvancedstage(WEAVE,Balcellsetal.2010;HECTOR,Bland-Hawthorn2015),butthecombinationoflargeaperture,widefield-of-view,highspectralresolutionandexcellentimagequalitymakeMSEauniqueprospectintheintegralfieldspectroscopydomain.

Asignificantnumberoflarge-scalemonolithicIFUsarenowavailableonlarge-telescopes,andsoa singlemonolithic IFUusing all the fibres ofMSE is unlikely to be competitive. For example,compare the~3400 fibers of MSE to the~90000 spectra per shot of the MUSE IFU on VLT(Bacon et al. 2015; however, we note that the limited field of view of MUSE and otherinstruments could mean that very large monoliths could still prove of significant interest).However, with a large number of deployable IFUs MSE would be an immensely powerful

Instrument Telescope DateWavelengthrange(nm)

Telescopediameter(m)

Fieldofview(sq.deg)

Numberoffibres

Fibresize(")

NumberofIFUs

Spectralresolution

FLAMES VLT 2003 370-950 8 0.42 300 0.52 15 11000-39000

SAMI AAT 2012 370-740 3.9 1 819 1.6 13 1700-4500KMOS* VLT 2013 778-2456 8 0.12 4704* 0.2* 24 2800-4800MaNGA SDSS 2014 360-1000 2.5 3 1423 2 16 2000

HET VIRUS 2017 350-550 10 0.37 33600 1.3 150 800WEAVE WHT 2018 370-1000 4.2 2 1000 1.3 20 5000HECTOR AAT 2020 370-900 3.9 3 8500 1.6 100 4000

~2025 360-1300 11.25 1.5 3400 ≤1.2 50 6500MSE


120

instrument. InTable 5 we list the main characteristics of current and planned multi-objectintegral field spectrographs. VIRUS on the HET, although not a deployable IFU system, has ahugemultiplexthatgivesitsomeofthesameadvantagesasadeployablesystem.FLAMESistheonlycomparablesystemwithtrulydeployableIFUsonan8–10mclasstelescope,andthishasanorderofmagnitudelessfibresandanorderofmagnitudesmallerfield-of-view.AsaresultaMOSIFUsystemonMSEsitsinatrulyuniquepartofparameterspace.Suchasystemisplannedas a second generation capability and here we discuss some aspects of the motivation andimplementation.

3.3.3.1 TheIFUperspectiveofgalaxies

GivencurrentinstrumentslikeMUSEofVLTandKeckCWIwithmonolithicIFUcapabilities,itistheMOS IFUcapabilityona largeaperture thatwill setMSEapart.Assuch, thesciencecasesdescribedbelowfocusonthesciencecaseforMOSIFU,ratherthanalargemonolithicIFUs(thatcouldinprinciplealsobefittedtoMSE).

Thecurrentandnear-futuregenerationsofmajor IFUsurveysonsmaller telescopeswill likelyhave targeted of order ~100 000 local, z < 0.1, galaxies by 2025. In this context the majoradvantagesforMSEwillbehighsurfacebrightnesssensitivitycoupledwiththeimagequalityofMaunakeathatallowsMOSIFUspectroscopytopushbeyondthelocalUniverseformorethanjust the highestmass galaxies. A second substantial focus will be on local lowmass galaxies(log(M*/M¤

) < 108), that are the building blocks of higher mass galaxies and are the mostinfluencedbytheirenvironment.

i. Theevolutionofgalaxymorphology

Figure56:λRvs.concentration(C=R90/R50)fortheSAMIgalaxies(Fogartyetal.2015)thatareslowrotators(red),fast rotators (blue) and spiral galaxies (orange). The black tracks are from Bekki & Couch 2011 and show the


121

evolutionarypathofspirals(topleft)subjecttomultipleinteractionsandtherebytransformedintoanS0(bottomright),5Gyrlater.(FigurefromFogartyetal.2015).

Themorphologicalstructureofgalaxiesisoneoftheirmostfundamentalproperties.Weknowthat environmentmodifies galaxymorphology, thatmass andmorphology are connected andthatgalaxymorphologiesevolveovertime.However,tacklingtheunderlyingphysicalprocesseswhich drive these trends is hard. Integral field spectroscopy, particular through stellarkinematics,has thepower touncover thephysics.Afterall, thephotometricmorphologyofagalaxyisdefinedbythedistributionoforbitsthatitsstarsliveon.

Figure 57: λR vs. ellipticity (left) and concentration (C = R90/R50) (right) for CALIFA galaxies, colour coded bymorphology (circles). These are compared to remnant galaxies in merger simulations (diamonds). (Figure fromQuerejetaetal.2015).

We know that high density environments contain more S0 galaxies and fewer spirals (e.g.Dressler,1980)andthetransformationofspirals intoS0smustplayan importantrole.TheS0fraction inclustersdoubles from~20%atz~0.8 to~40%bythepresentday (Dressleretal.,1997;Couchetal.,1998),withacorrespondingdeclineinthespiralfraction.Thetrendissimilar,but twice as strong, in groups (Just et al., 2010), suggesting that groups efficiently drivetransformations.Revealingthephysicsbehindsuchtransformationsrequiresstellarkinematicsasonlythiscandiscriminatebetweenhydrodynamicalmechanisms(thatshouldnotproduceanincrease in random support) and gravitational interactions (that will build-up pressuresupported systems). Simulations of group interactions or minor mergers already providepossible pathways from spirals to S0s (Bekki & Couch, 2011; Bournaud et al., 2005), and arequalitatively similar to observations (Figure 56; Fogarty et al. 2015). However, alternativeinterpretationsarepossible,suchasmajormergers(Querejetaetal.2015;seeFigure57).

A fundamental challenge currently is thatweare comparing current spiral galaxies to currentS0s. To properly understand this transformationwe need to look at earlier spiral galaxies, astheirpropertiesmaychangewithcosmictime.Directscalingfromcurrentsurveyssuggeststhatitwillberelativelystraight-forwardtomeasurestellarkinematicsinL*orgreatergalaxiesouttoz ~ 0.3 which already covers 3.5 Gyr of dynamical evolution. Higher mass galaxies can betargetedatevenhigherredshift.MSEwillhavetheuniqueabilitytomeasurestellarkinematics


122

for large samples of galaxies beyond the local Universe and as a result directly examine theevolutionofdiskstructure.AsaresultwewillbeabletodirectlymapearlierspiralstocurrentspiralsandS0s.

WorkfromATLAS3Dhasshowsthatmorphologymaybebetterdefined intermsofdynamicalproperties (Emsellem et al., 2011; Cappellari et al., 2011b). Early type galaxies can now beclassified as fast rotators (rotation dominated) and slow rotators (dispersion dominated). Theglobal fraction of slow rotators is relatively constant at ~15% of early types, although theyincreaseinnumbertowardsthecentreofclusters(Cappellarietal.,2011b;Fogartyetal.,2014).Slow rotatorsmay undergomoremajormergers and build up a higher fraction of theirmassfrom the accretion of satellites (Khochfar et al., 2011), but the cluster environment is notoptimalforthistohappen.Again,inordertounderstandthegrowthofslowrotatorsdirectlyweneedtoexaminehowthepopulationevolveswithredshift.Thiswillonlybepossiblewithlarge-scaleintegralfieldsurveysbeyondthelocalUniverse.Thefrequencyofslowrotatorsindifferentregimes can be used as a strong test of current cosmological hydrodynamic simulations thatstruggle to converge on the amount of angular momentum in simulated galaxies (e.g.,Scannapiecoetal.,2012).

Size evolution is another key featureof the galaxy population, particular in themostmassivegalaxies.Onceagain,kinematicsallowsustounderstandthestructureofsuchgalaxies.Herethechallenge is to obtain sufficient spatial resolution to be able to resolve rotation curves andminimize thetransferof rotationalvelocity tomeasureddispersionviabeamsmearing.This ispossibleforthemostmassivegalaxiesintheoptical,asdemonstratedbyvanderWel&vanderMarel(2008)usingFORSontheVLT,howevertomakethisarealityonMSEsmallerfibresthatbettersampledtheseeingwouldberequired.

ii. Modulatingstarformation

Numerousphysicalprocessesareabletomodulatestarformationacrosscosmictime.Thiscanbroadlybeseparatedintofactorsthatareassociatedwithgalaxyenvironmentandgalaxymass(Pengetal.,2012).However,weneedtomovebeyondthesebroadcategorizationstowardsamorephysicallymotivatedpicture,whereweunderstandthedifferingcontributionsfromeachphysicalprocess.

Environmentalquenchingisthesuppressionofstarformationinhighdensityregions(e.g.Lewisetal.,2002).Whiletheexistenceofsuchquenchinghasbeenknownforsometime,thedetailedphysical mechanisms remain hard to discern. The different processes that can contribute toquenchingincluderam-pressurestripping(e.g.,Gunn&Gott,1972),strangulation(Larsonetal.,1980, Bekki, 2009), dynamical interactions and mergers. Crucially, these processes havedifferentobservationalsignatures.Ram-pressurestrippingleadstoconcentratedstarformation,as gas is removed from the outer disk first, and may cause shocks and extra-planar gas. Incontrast,strangulationleadstosuppressionofstarformationacrossthewholedisk.Interactinggalaxiescanhavetheirgasfunneledintothecentralregions,triggeringnuclearstarformation,andtheymaybedynamicallydisturbed.Theprocessesarelikelytodependonwhetheragalaxyisthecentralgalaxyorasatelliteinitsparenthalo,aswellasthemassoftheparenthaloandlocalgalaxy–galaxyinteractions(e.g.Daviesetal.2015).


123

IFU spectroscopy from MSE can provides (1) spatially resolved Hα emission to map starformationontime–scalesof107yr;(2)Balmerabsorptiongradientsfromstellarpopulationsthattraces star formation on 109yr time–scales; (3) stellar and gas-phase kinematics to identifygalaxies that are dynamically disturbed or misaligned; (4) shock detection via emission linediagnosticratiossuchas[NII]/Hα,[OIII]/Hβand[SII]/Hα(e.g.,Kewleyetal.2006).Thisneedtobedoneintandemwithdetailed3Denvironmentalmeasurementssothatlarge-scaleandsmall-scaleenvironmentcanbedefined,aswellasclosepairsofgalaxiesthatareinteracting.

Mass-relatedquenchingofstarformationmaybecausedbyAGN(Crotonetal.,2006)orrelatedtomorphologywithbulgesprovidingsupportagainsttheformationofcentralgasdisks(Martiget al., 2009). IFU spectroscopy allows us to pick apart the separate components of a galaxy(Johnstonetal.,2014)andmeasuretheseparatestellarpopulationcharacteristicsofbulgeanddisk independently. We can also identify the sphere of influence of an AGN and determinewhetherthestellarpopulationsandcurrentstarformationareinfluencedbythepresenceofanAGN.

MSEwill have the surfacebrightness sensitivity todetectHα to trace star formationuntil theemission line hits the spectral range of the instrument (1.3 microns) at z ~ 1. As the starformation rate increasesback tohigher redshift (Hopkins&Beacom,2006)wewillbeable todirectlytestwherethisstarformationishappening.Istheincreasedinthenormalisationofthestarformationmainsequencedrivenprimarilybystarformation indisks?When,and inwhichcircumstances, does star formation become more clumpy (as seen in current infrared IFUsurveys;Genzeletal.2011)?

iii. Theroleofoutflowsandinflows

Figure58:SAMImapsofwindgalaxyGAMAJ115927.23-010918.4.ThedashedredcircleintheSDSS3-colourimage(rightpanel) shows the SAMI field-of-view (15arcsecdiameter). The increased velocitydispersion,σgas, and lineratios,[SII]/Hα,offthediskplane,aswellasthedisturbedvelocityfield,areallindicativeofgalacticwinds.(FigurefromHoetal.2016).

15 10 5 0 -5 -10 -15� R.A. [arcsec]

�15

�10

�5

0

5

10

15

�Dec.[arcsec]

SDSS H↵ log([SII]/H↵)

5 0 -5

�5

0

5

Vgas �gas

1

2

3

[10�16/ergs/s/cm2]

�60

�30

0

30

60

[km/s]

0

20

40

60

[km/s]

-0.4

-0.3

-0.2

-0.1


124

Manystarforminggalaxiesshowingevidenceforoutflows(e.g.Heckmanetal.,1990;Veilleuxetal.2005),buttheirfullimpactontheprocessofgalaxyformationinstilluncertain.NewlocalIFUsurveyshavealreadystartedtoaddressthisissue(e.g.,Hoetal.2016,seeFigure 58).Ahighfractionofoutflowsarebeing foundusingacombinationofgaskinematicsandgas ionizationdiagnostics. Someof thesehavecurrent star formation ratedensitiesbelow thecanonical0.1M

¤ kpc-2yr-1 expected to drivewinds, and stellar ages and star formation time–scales appear

consistent with the winds being driven by bursty star formation episodes. Central to suchanalyses is high spectral resolution to separate out star forming and wind components fromcomplex emission line profiles. Shockmodels (e.g., Rich et al. 2011) can be used to estimategalacticwindpropertiessuchasshockluminosity,theshockvelocity,thedynamicaltimescaleofthewind,andthecontributionofshockstotheglobalenergybudget.

Theroleofgasaccretionisfundamentaltothebuild-upofgalaxies.Itisgasaccretionthatfuelsthe ongoing process of star formation. Themisalignment between gas and stellar kinematicsprovidesapotentmeanstoidentifyandcharacterizegasaccretionbecauseinternallyproducedgas(e.g.,viastellarmass loss)will followthekinematicsofthestars,whileexternallyaccretedgasneednot follow themotionof the stars.Misalignmentshavebeenanalyzed inearly–typegalaxies(Davisetal.2011),andthisisnowbeingextendedtoallgalaxytypesbyongoinglocalIFU surveys. The gas misalignment distribution and can be used to constrain the dynamicalrelaxation, depletion and accretion timescales of gas in galaxies (Davis& Bureau 2016).WithMSE pushing such measurements beyond the local Universe we will be able to directly seewhether the increased star formation to high redshift is driven by a increase in externallyaccretedgas,andinwhichsituationsthisismostsignificant.

A complementary approach to studying gas accretion is via gas-phase chemical abundancegradients.Normalspiralgalaxiesgenerallyshowclearmetallicitygradients,becomingenrichedbya factorof10 from~ 8–10kpctotheircentral regions.Gas inflows,either fromexternalaccretionortriggeredbyinteractions,candrivelowmetallicitygastowardsthecentralregionsofagalaxy.Thisresultsinflattermetallicitygradients(Kewleyetal.,2010;Richetal.,2012).

iv. Theroleofdynamicalinteractionsandmergers

In the standardhierarchical pictureof galaxy formationandevolution, themassof galaxies isbuilt up hierarchically through successivemergers and through the accretion of gas from theintergalacticmedium.Athigh redshift galaxymergersmaybe theprimary assembly route formassivegalaxiesanddriverapidstarformation.Theoreticalpredictionsofthemergerratearehighlyuncertain (e.g., Jogeeetal.2009,Hopkinsetal.2010)dueto the inabilityofmodels toprecisely map galaxies onto dark matter halos and sub-haloes (see Lotz et al. 2011 for adiscussion).Measurementofdynamicaldisturbancecanbearoutetoidentifyingmergers(e.g.,Shapiro et al. 2008). This is currently being done locally with nearby samples, and at highredshiftwithstar-forminggalaxiesintheinfrared(e.g.,withKMOS),butMSEwillhavetheabilityto trackdynamical disturbanceuniformly from z = 0– 1 inHα, uncovering the importanceofmergersanddiskinstabilityintheevolutionoftheirstarformationdensity(Hopkins&Beacom2006). Kinematic data can also be used to calibrate and compared other approaches such asclose pairs of galaxies (e.g., Robotham et al. 2014) and morphological disturbance (e.g.,Conseliceetal.2003).


125

Cosmological hydrodynamical simulations suggest that it is possible to connect higher–orderstellarkinematicmomentsh3 (skewness)andh4 (kurtosis) ingalaxiestotheirassemblyhistory(Naabet al., 2014). These simulationspredict that fast rotating galaxieswith gas–richmergerhistoriesshowastrongh3–(V/σ)anti-correlation,whilefastrotatorsthatexperiencedgas–poormergers do not, due to the absence of a dissipative gas component. The higher-order stellarkinematics of from MSE can be used to diagnose the formation pathways for the galaxypopulation.Aswellasextendingthisworktohigherredshift,MSEcanpushdowninstellarmassforlocalgalaxies.Asaresultwecanstarttounderstandwhetherdwarfgalaxieshavethesameformation routes asmoremassivegalaxies in termsof theirmerging,ordiscernwhere in thehierarchyphysicalprocesseschange.

v. Cosmologyviascalingrelations,peculiarvelocitiesandlensing

Scalingrelationshavelongbeenusedasacosmologicaltool.Thereisagrowingrealisationthatusing scaling relations (such as the fundamental planeor Tully-Fisher relation) tomap cosmicflows has great potential to probe the gravitational growth of structure and test generalrelativity. Such tests are largely orthogonal to othermethods targeted at dark energy usuallybasedonhighredshiftspectroscopicsurveys(e.g.,Subaru/PFS,Mayall/DESI,etc.).Integralfieldspectroscopy allows us to reduce the scatter in scaling relations and generatemore physicalquantities, such as moving from the fundamental-plane to the mass-plane (Cappellari et al.,2013, Scott et al., 2015, Cappellari, 2016). The ability to capture both bulge and diskcomponentsmeansthatcombinedscalingrelationssuchass0.5(Kassinetal.,2007)canbeusedforallgalaxies,asdemonstratedbyCorteseetal.(2014).

Integral field spectroscopy of galaxy velocity fields has great potential as a lensing tool (deBurgh-Dayetal.,2015a,b).Measurementofshearonvelocityfieldscanbemadetomuchhigherprecisionthatnormalweaklensingbasedonshapemeasurements.Asaresultindividualgalaxy-galaxylensedsystemscaninprincipleprovidedirectmeasurementsofdarkmatterhaloprofiles.AsecondcombinationofIFUspectroscopyandlensingistousetherotationcurvesderivedfromIFU spectroscopy to reduce the shape noise in lensing measurements by comparison to theTully-Fisherrelation(Huffetal.2013,Blain2002).IntegralfieldspectroscopyofSNhostgalaxiescanhelptoshedlightonthedifferencesbetweenSNpopulations(e.g.Sullivanetal.,2006),andwhether this is dependent on the global properties of a galaxy or the local properties at thelocationoftheSN.Understandingsuchdifferences iscritical toreducingthesystematics inSNcosmologymeasurements.Chapter 4providesamoredetaileddiscussiononnextgenerationcosmologysurveyswithIFUs.

3.3.3.2 Instrumentandsurveydesign:practicalconsiderations

Ifweassumeabaselineof3400 fibres, then thiswouldmap to∼56 IFUs ifeachonehad61fibres (~11 arcsec diameter assuming 1.2” arcsec fibres and a 75% fill factor). In practice avarietyof IFU sizeswouldbemost appropriate (e.g.Bundyet al., 2015) tomatch the sizesofgalaxiesinanydesignreferencesurvey.AresolutionofR=6500overmuchoftheopticalwouldin general be much preferred, compared to the lower R = 2000. This is because the higherresolution allows a variety of science including the kinematics in low mass galaxies,measurement of dispersion in gas disks and separating different kinematic components in


126

outflows.

Supportingdata:TheconnectionofIFUdatafromMSEtogetherwithavarietyofancillarydatawillgreatlyenhancethescienceoutcomes.Ideallytargetsshouldbeselectedfromregionsthathavewelldefinedenvironmentalmeasurements,coverabroadrangeinstellarmass,andhavesubstantial multi-wavelength data available. The GAMA survey (Driver et al., 2011) currentlyprovidesthebestregionsinthisregard,withdeep(r<19.8)andhighlycompletespectroscopyand carefully matched 21–band photometry (Driver et al., 2016) from UV (GALEX) to far-IR(Herschel).However,alarge-scaleMSEIFUsurveywillneedtoreachbeyondthecurrentGAMAsurvey. Spectroscopic programs like WAVES (wavesurvey.org on the VISTA telescope using4MOSTwillprovidesuitabletargetcatalogues,butofcourseMSEgalaxyredshiftsurveyscouldalsodothis.TyinganMSEIFUsurveytonewHIsurveysusinginstrumentssuchasASKAPwillbeimmenselyvaluable,allowingustoaccessgas fractionandthe large-scaleangular-momentumof galaxies beyond the optical radius. At the same time newdeeper, higher resolution imagedatawillbecomeavailablefromLSSTandEuclid.

Sensitivitylimitations:ThetwolargestconstraintsontheinformationcontentinIFUdataisthesurfacebrightnessofthesourceandthespatialsampling.Higherspatialsampling(tothelimitoftheseeing)ispossiblebyhavingsmallerresolutionelements(i.e.smallerfibres).However,withsmallerfibrestheSNRperelementisreduced,andinsomecasescanreachthepointatwhichdataisread-noiseratherthanbackgroundlimited.AsoneoftheuniqueaspectsofMSEwillbeitsabilitytopushtohigherredshiftoveralargefield-of-view,thisnaturallypushesthedesigntothesmallestfibresthatarestillbackgroundlimited.Thenthesurfaceareathatcanbetargeteddependson thenumberof fibres thancanbe takenby the spectrographs.Smaller fibres takelessroomonthespectrographslit,butnotsufficientlylesstofullycompensateforthereducedfield- of-view with smaller fibres. As a result the sky area that can be sampled by the IFUsdirectly depends on the number of spectrographs in the system. The area where spatialresolutionismostdemandedisatthehighestredshift.AttheextremeendoftheMSEredshiftrange for Hα the sensitivity to stellar continuum will be poor (due to surface brightnessdimming). Ifonlytargetingemissionlinesatthisredshift,thenfibrescouldbesmallenoughtooptimallysampletheseeing,asthesurfacebrightnesssensitivitylimitswillbelesssevere.Thispoints to two sets of fibres, one thatwill be focussed on surface brightness sensitivity in thecontinuum(say~1”diameter)andasecondthatoptimallysamplestheseeing(say~0.2−0.3”).Detailed modeling of the target galaxy populations should be done to optimise this fully.Reachingtheseeinglimitofthetelescopealsoplacesconstraintsonthedesignofthetelescopeandcorrectoroptics.

Positioning IFUs in the focal plane: With less than 100 IFUs and a field-of-view of 1.5 degdiameter,someconsiderationneedstobegiventohowbesttopositionIFUsinthefocalplane.Withatypicaltargetredshiftofz~0.2−0.3therewillbesubstantialclusteringacrossthefieldofview,whichpointstoaflexiblepositioningsystemwhereeachIFUhasarelativelylargepatrolregion.DetailedsimulationswillberequiredtoexaminehowdifferentpositioningtechnologiesimpacttheefficiencyofIFUsurveys.


127

3.3.4 Thegrowthofgalaxiesthroughmergers

In the hierarchical structure formation paradigm, galaxymerging is expected to play a crucialrole in the build-up of galaxymass. However, the role of galaxymergers extends far beyondtheir nominal delivery of mass, thanks to the dynamical processes triggered during theinteractionandcoalescencephasesofthemerger.Inbrief,tidaltorquesleadtotheformationofabar,whichpresentsamechanismthroughwhichangularmomentumcanbedrainedfromthegas,which is then funneledtowards thegalacticcentre. In turn, this increasedcentralgasdensity has long been predicted by simulations to trigger starbursts, alter the distribution ofmetalsand induceAGNactivity (e.g.,Barnes&Hernquist1991;Springeletal.2005;Coxetal.2008;Hopkinsetal.2006,2010;Torreyetal.2012).Inthelowredshiftuniverse(z<0.2),largespectroscopic surveys of galaxies have led to a detailed assessment of the role of mergersamongst relative massive galaxies. For example, there is now unequivocal evidence thatmergerscanleadtoincreasedstarformationrates(e.g.,Lambasetal.2003,Nikolicetal.2004,Scudderetal.2012,Pattonetal.2013,Scott&Kaviraj2014)andcantriggerAGNactivity(Ellisonetal.2011,Khabiboullineetal.2014,Satyapaletal.2014).

Nonetheless,manyquestionsremainunansweredbythecurrentgenerationofgalaxysurveys,including the role of lowmass galaxies (eitherwith one another, or as accreted satellites) totheseprocesses,themechanismthatleadstogalaxyquenching(andthetimescaleforthisshut-downofstarformation)andhowtheseeffectsevolveouttohigherredshifts.Addressingthesequestions requires extending large, homogeneous, highly complete spectroscopic surveys tohigherredshifts,andacarefullyplannedsynergywithfuturemulti-wavelengthfacilities.MSEisperfectly positioned to contribute to this potentially high impact field of galaxy evolution forseveralreasons.First,itshighspatialcompletenesswillyieldahighlyeffectiveidentificationofspectroscopicgalaxypairs,andavoidmanyofthebiasespresentin,forexample,theSDSS(e.g.,Patton & Atfield 2008). Second, the sensitivity to lower stellar masses will permit a morecompleteviewofthemergerprocessatlowredshift,byallowingustoquantifytherelativeroleof “minor” mergers, and dwarf-dwarf mergers in a statistically significant way. Finally, acombination of deeper detection thresholds and coverage of the near-IR will permit ahomogeneousassessmentofproperties suchas star formation ratesandmetallicities,usingaconsistentsetofdiagnosticsacrossabroadwavelengthrange.


128

Figure 59: Examples of early-type galaxies from the ATLAS3D survey that are currently involved in a tidalinteractionwithanearbymassivecompanionandexhibitingprominenttidaltails.Clockwise,compositetruecolourimagesofNGC770,NGC680,NGC2698/99,NGC5507andNGC5574/76(FigurefromDucetal.2015).

3.3.4.1 Theidentificationofmergersamples

Inthemostnearbygalaxies,on-going,orrecentlycompletedinteractionsbetweengalaxiesarereadilyrecognizedthroughmorphologicaldisturbances,e.g.,Figure 59. Thankstoverydeepimaging, which is necessary in order to detect faint surface brightness features and stellarsubstructures,contemporarysurveysarerevealingthatsuchstructuresarecommoninnearbygalaxies(e.g.,McConnachieetal.2009,Taletal.2009,Martinez-Delgadoetal.2010,Ducetal.2015). Detailed modeling of the morphologies of shells and tidal features permits areconstructionoftheprogenitorsandorbitaldetailsof themerger (Barnes2016,Mortazavietal. 2016, Privon et al. 2013, Barnes & Hibbard 2009). However, these detailed studies ofindividual, or small samples, of very nearby galaxies do not provide a global picture of themergerphenomenon,nordoesdeepimaging(currently)representafeasiblemechanismforthewholesaleidentificationoflargesamplesofmergers.

In order to more generally characterize both the frequency of galaxy interactions, and theirimpact on galactic properties, automated selection techniques that can be applied to thecurrent generation of large imaging and spectroscopic surveys have been developed. Theseselection techniques generally fall into two categories. The first approach is to look formorphologicallydisturbedgalaxies,throughnon-parametricassessmentofgalaxymorphologies,


129

usingmetrics suchassuchasGini-M20andasymmetry (Conseliceetal.2003,Lotzetal.2004,Pawliketal.2016,Pethetal.2016). Althoughtraditionallyappliedto imagesobtainedintheoptical, these techniques can equally well be applied to galaxy images obtained at otherwavelengths, such as the IR or even radio,where they are potentially evenmore effective atdiscerning persistent morphological disturbances (Lelli et al. 2014, Psychogyios et al. 2016).However,despitetheirwidespreaduse,theautomatedmorphologicalidentificationofmergershasseveraldrawbacks.Forexample,the identificationofamerger issensitivetotheboththedepthof the imageandthenature(e.g.,massratioandorbitalgeometry)of themerger itself(Lotz et al. 2010a,b, Ji, Peirani&Yi 2014). Indeed, somemergers, suchas those that are gaspoor may have very weak asymmetries and minor mergers may barely perturb the moremassivecomponent.Moreover,thereisnodefinitiveseparationinasymmetryparameterspacebetweenmergersand`normal’galaxies,suchthatcontaminationbyclumpyorirregulargalaxiesisalwaysproblematic.

The alternative approach is to identify kinematic (or spectroscopic) pairs of galaxies fromspectroscopic surveys. This techniqueselects likelymergersbasedoncloseprojectedpairsofgalaxieswithsimilarline-of-sightvelocities(e.g.,Pattonetal.2002,Carlbergetal.2000,Woodset al. 2010, Patton et al. 2011,Wong et al. 2011, Lambas et al. 2012, Kampczyk et al. 2013,Davies et al. 2015). The availability of spectroscopic data is a further boon to thecharacterization of galaxies involved in the merger, permitting measurements of the starformationrate(SFR),gasandstellarmetallicitiesandclassificationofAGN.

3.3.4.2 Lessonsfromthelowredshiftuniverseandfuturedirections

Thepropertiesoflowredshiftgalaxypairsarenowwellcharacterized,thankstoacombinationoflargesamplesandsupportingmulti-wavelengthdata.Figure60summarizesseveralofthekeyresultsthathaveemergedfromstudiesofSFR(Nikolicetal.2004;Alonsoetal.2007;Ellisonetal.2008;Scudderetal.2012;Pattonetal.2011,2013),gasphasemetallicity(O/H,e.g.,Ellisonetal.2008;Michel-Dansacetal.2008;Scudderetal.2012;Gronnowetal.2015)andAGNfractions(e.g.,Alonsoetal.2007;Woods&Geller2007;Kossetal.2010;Ellisonetal.2011;Sabateretal.2013; Satyapal et al. 2014). In brief, these results have lent observational support to thepredictionsof simulations inwhichmetal-poorgas from thediskoutskirts flows to thecentresimultaneously diluting the gas phasemetallicity, triggering new star formation andprovidingmaterialforAGNaccretion.

Figure60:Resultssummarizingthechangesexperiencedbygalaxiesinmergersatz<0.2,basedonspectroscopicpairs selected fromtheSDSS,asa functionofprojectedseparation. Comparisonsaremadedifferentiallywitha


130

control sample of non-interacting galaxies matched in stellar mass, redshift and local environment. The greyshaded panel shows the results for fully coalesced “post-mergers”. Left panel: The SFR enhancement in pairsrelativetocontrolsonalogscale.ThetypicalenhancementofSFRinclosepairsisafactorofafew,peakinginthepost-mergers. Middle panel: The gas phasemetallicity, relative to a control sample on a log scale. There is agradualdecreaseinthemetallicitytowardssmallerseparations,withthemostextremedilutionsoccurringinthelatestagepost-mergers.Rightpanel:ThefrequencyofopticallyselectedAGNinpairsrelativetocontrols.Again,theexcessincreasessteadilywithdecreasingpairseparation,peakinginthepostmergers.(FigurefromEllisonetal.2013).

Despitetheprogressinourunderstandingoftheimpactofmergersongalaxyevolution,therearesomeaspectsofthisfieldthatarepoorlyaddressedwithexistingdatasets.Forexample:

i. Whatistheroleofmergersamongstlowmassgalaxies?

The limiting spectroscopic magnitude for the SDSS is approximately 18 in the r-band.Consequently,themajorityofgalaxiesforwhichdetailedpropertiessuchasSFRandO/HhavebeenmeasuredarelimitedtotypicallylogM*<9.Moreover,thelimiteddynamicrangeoftheSDSS means that minor mergers are relatively under-represented, with the pairs sampledominatedbythosewhosemassesarewithinafactorofafewoftheircompanion(Ellisonetal.2008). These limitationsmean that SDSSprovidesagoodviewofneither themajorityof theminormerger(orsatelliteaccretion)regime,northecharacterizationofmergersbetween lowmassgalaxies.Andyet,thedominanceoflowstellarmassesbynumberindicatesthatmajorityof mergers should be dominated by such interactions at all redshifts (De Lucia et al. 2006;Fakhourietal.2010).Basedonasmallsampleof~100lowmass(logM*<9.7)galaxiesinpairs,Stierwaltetal.(2015)haverecentlyshownthat“dwarf”interactionscanalsoleadtoenhancedSFRs, at a level similar to their more massive counterparts. Although the GAMA survey canimprove upon these statistics, being sensitive to deeper magnitudes, a full census of thecontribution to cosmic star formation ideally requires extending observations into the true“dwarf” regime. With its far deeper spectroscopic limits,MSEwill provide the same level ofstatisticalrigourasSDSS,butforthecharacterizationofmergereffectsdowntomasseslogM*~6.5.Asurveyofdwarf-dwarfmergerswouldalsoaddresswhetherinteractionsplayasignificantroleinthecreationofcentralsupermassiveblackholes(SMBH),orcantriggeraccretionontoanexistingSMBH.

ii. Whatistheimpactonthegalacticgasreservoirduringandafteramerger?ThecombinationofopticalsurveyswithallskymissionsatdifferentwavelengthsrangingfromtheUV (e.g., GALEX) to themid-IR (e.g.,WISE) has enabled a scientific yieldwell beyond thecapabilities of a single wavelength regime. However, a significant piece of the puzzle hasremained elusive, themeasurement of atomic gas on industrial scales. Although theAreciboALFALFASurvey(Giovanellietal.2005)hasprovidedamajorbooninthisdirection,detectionsat21cmcontinuetobecountedinthethousands,ratherthanthehundredsofthousands,andthe vastmajority of low redshift galaxies in surveys such as the SDSS, 2dF andGAMAdonothave HI masses available. This is a major short-coming, since the delivery, availability,consumptionandpossibleexpulsionofgasarelikelytobekeyfeaturesofacompletepictureofgalaxy evolution (e.g., Bothwell et al. 2013). A further impediment for the study of HI gas ingalaxymergersisthattheworkhorsetelescopefor21cmsurveys,Arecibo,hasabeamwidthofalmost4arcmin,fartoolargetoresolveclosepairs.TargetedVLAobservationsofgalaxypairsis


131

currentlyfeasible,butextremelyexpensiveandalsodoesnotprovidealargecontrolsampleforcomparison(e.g.,Scudderetal.2015).The adventof the SKA (and its pathfinders)will revolutionize the fieldof galaxy formationbyprovidingunprecedentedquantitiesofHImassesforthegalaxiessurveyedbyMSE.Theimpactforgalaxymergerstudieswillbetheabilitytoinvestigatetheroleofinteractionsonthegalacticgasreservoir.Somesimulationshavepredictedmergersmayberesponsibleforreplenishing,orevenboosting(throughhalogascondensation)thecoolgascontentofgalaxies(e.g.,Mosteretal.2011;Tonnesen&Cen2012;Rafieferantsoaetal.2015). Ontheotherhand, theburstsofstarformationknowntotakeplace inmergerswouldbeexpectedtodepletethegas,andthetriggeredAGNcouldprovidesignificantfeedback.Itisthereforefarfromclearwhethermergersre-invigorate star formation, or may ultimately quench it. So far, no change in the HI gasfraction of mergers has been detected in the SDSS (Ellison et al. 2015). However, 21cmmeasurementsforSDSSmergersnumberinthetens,andareaffectedbybeamconfusion.WithMSE and SKA, it could be possible to determine resolved HI mass for thousands of galaxymergers(andpost-mergers)andallowameasurementofgasconsumptiontobemade.

iii. MergersathighredshiftAfterthesuccessofsurveyssuchasSDSSforcharacterizingmergersatz<0.2,thenextobviousstepistheextensiontohigherredshifts.ThebeststatisticalassessmentofmergerinducedSFRsbeyond z > 0.2 comes from the GAMA survey (Davies et al. 2015), but the redshift range ofGAMAextendsonlytoz~ 0.5.Athigherredshifts,samplesofspectroscopicpairsarefrequentlysmaller (e.g.,Xuetal.2012;Schmidtetal.2013;),or relyoneithervisualclassificationor lesscertain redshifts (e.g., Chou et al. 2011; Wong et al. 2011). Beyond z~ 0.5, there is theadditionalcomplicationthatmanyofthekeyemissionlinesusedforinferringSFRs,O/HandthepresenceofanAGNareshiftedintotheIR.Althoughalternativediagnosticsexist,homogeneityiscrucialforuncoveringtheoftensubtlechangesobservedinmergers;ideallywewouldliketobeabletostudytheevolutionofmerger-inducedstarformationwithauniformsetofdatafromz~ 0 to z~ 1andbeyond. MSE ispotentially superblypositioned to completeourpictureofgalaxy mergers to such redshifts, where different gas fractions and morphologies couldpotentiallydramaticallychangetheefficiencyofgasflows.

3.4 Steppingthroughthehierarchy:thenearestgalaxygroupsandclusters3.4.1 Clustersaslaboratoriesofgalaxyevolution

ScienceReferenceObservation7(AppendixG)

BaryonicstructuresandthedarkmatterdistributioninVirgoandComa

Galaxy clusters act as (nearly) closed boxes, retaining the entire thermal and gravitationalhistory of their assembly and growth. Here we propose an ambitious survey targeting allbaryonic structures (mostly galaxies and globular clusters) brighter than g ~ 24.5 within thevirialradiiofthetwonearestlargeclusters,VirgoandComa.Thebrightestsubsetofsources(g<22; some 50 000 sources in Virgo) will also be observed at high resolution. With thisunprecedented complete dataset, we will map the dark matter distribution of halos on the


132

largest scales by examining the line of sight velocities of all baryonic tracer particles in theseclusters. Thiswill provide aprecisemeasurement of the darkmattermass distribution in theclusters, and reveal coherent structure inphase space that canbeused to infer theaccretionhistorywithintheclusterpotential.TheMSEVirgo+Comasurveywillbeunrivalledforexploringall aspects of galaxy evolution, especially with respect to stellar populations and low massgalaxies.ThisisbecausetheVirgoandComaclustersaffordusthebestopportunitytostudythegalaxypopulation in its totality, fromquiescenttostar forminggalaxies, fromthehighdensitycorestothelow-surfacebrightnesshaloes,acrossawidemassrange,onallspatialscales, inacontrolledenvironmentinwhichallotherbaryonicsubstructuresarealsowellcharacterized.

Galaxygroupsandclustersareextraordinarilyimportantastrophysicallaboratories.AstheonlyplacesintheUniversewhereallrelevantcomponents—darkmatter,hotgas,coldgasandstars—canbestudiedinaself-consistentmanner,groupsandclustersaffordustheopportunitytoobserve the interactions between galaxies and the assembly of large scale structure, thusestablishing an important link between predictive ΛCDM theory and the complex (butobservable)hierarchicalprocessofgalaxyformation(Figure 61).

Galaxy groups – both within the infall regions of massive galaxy clusters, and more isolatedsystems– represent theearliest stage in thishierarchy.Within infallinggroups,galaxiesmaketheimportanttransitionfromcentralgalaxiesintheirownhalo,togas-starvedsatellitegalaxiessubjecttoawiderangeofnew,externalforces.Attheextremeofacontinuumdistributionofhalo masses, galaxy clusters accrete individual galaxies and larger subclumps from theiroutskirts. With the most recent arrivals found predominantly near a cluster’s virial radius,investigating the galaxy population as a function of clustercentric distance provides away toreconstructthehistoryofstructureassemblyanditseffectongalaxies(e.g.Baloghetal.2000;McGeeetal.2009).

Clustersandgroupshavebeenextensivelystudiedforover30years,andthefieldisnowreadyfor the next technological/observational breakthrough. The situation is especially critical forgalaxygroups:whilethevastmajorityofstellarmassatthepresentdayiscontainedingroups(withmassiveclustersrepresentingjust2%ofthestellarmassinthepresent-dayuniverse,e.g.Eke et al. 2005), we know very little about the galaxy dynamics, stellar populations or thetransformational physics that drive their evolution. In clusters, testing cosmologicalhydrodynamical simulations hinges on collecting detailed information — ages, metallicities,systemicvelocitiesaswellasinternalkinematics—forbaryonicstructuresfromtheclustercoreout to several virial radii. As anexample, the simulationspredict adepletionof bothhot andcoldgasandadeclineinthestar-formingfractioningalaxiesasfaroutas5virialradii(e.g.Bahéet al. 2013), but observations that extend to such large distances are limited to a handful ofclusters or superclusters (e.g.,Merluzzi et al. 2010,Mahajan et al. 2011, Smith et al. 2012b,Hainesetal.2011).


133

Figure61:Cosmologicalhydrodynamicalsimulationsarenowresolvinggalaxieswellintothe“dwarf”regime(M~108M

¤),matching thedepthof the latest state-of-the art deep imaging surveys. The left twopanels show two

massive(virialmassinexcessof1014M¤)clustersintheIllustrissimulations(Mistanietal.2016),withvirialradius

indicated by the dashed circle. The panels to the right show, at the top, a mosaic image of the Virgo cluster,extendingouttoonevirialradius,obtainedwithCFHT/MegaCam;and,atthebottom,theHSTimageofthecoreofoneofHubbleFrontierFieldclusters,Abell2744(z~0.308;forscalethisregioncorrespondstotheareawithinthegreenboxinthesimulatedclustertotheleft).Byestablishingclustermembership,studyingthestellarpopulations,chemicalenrichmenthistoriesandinternaldynamicsofthousandsofclustermembers,MSEwillallowforseamlesscomparisonbetweenthesimulationsandthedata.

The much needed observational breakthrough is the availability of high quality, deep,homogeneous spectroscopy for massive datasets. The ability of MSE to reach faint objects,coupledwithavastmultiplexingfactor,makesMSEanidealtoolforunderstandingthegrowthof the hierarchy of galaxy environments. Cluster environments are a key component of themajor science theme ofMSE of linking galaxies to the large scale structure of the Universe;mongstthescienceinvestigationsMSEwillenableare:

• Amembershipcensusingroupsandclusters

Spectroscopic redshifts are essential in building clean samples of baryonic structures


134

(galaxiesandglobularclusters)belongingtogroupsandclusters.Initslowresolutionmode(R ∼ 2000), MSE can efficiently accomplish the task thanks to its massive multiplexingcapability. Detailed cluster dynamical analysis is not only fundamental for understandingthe interactionsbetweengalaxies, the intraclustergasand thedarkmatterhalo.Galaxiescan be used as tracers to measure the cluster mass and velocity anisotropy profiles,providinginsightintothenatureofdarkmatter,constrainingalternativetheoriesofgravity(Bivianoetal.2013),andexploringthehaloformationprocessandhistory(e.g.Klypinetal.2014).Finally,dynamicalmassestimatesofclustersplayanimportantroleinunderstandingthe relationship between various observational mass tracers (e.g., X-ray, Sunyaev–Zel'dovich,optical richness)and themassof theunderlyingdarkhalo, thus improving theaccuracywithwhichthegalaxyclusterabundanceevolutioncanbemeasuredandrelatedto the predicted evolution of the dark matter halo mass function. These themes areexploredinmoredetailinChapter4.

• A study of integrated stellar populations, star formation rates (SFR) and enrichmenthistories,extendingfromthemostmassivesystemstodwarfgalaxies.

In high mass galaxies, known scaling relations between stellar mass, SFR and metallicityhave long been used as important constraints to theoretical models. High SNR, mediumresolution (R∼ 6500)MSE spectroscopy will extend these scaling relations to the dwarfgalaxies regime (M< 109M

¤), the building blocks fromwhichmoremassive systems are

formed. By targeting a statistically significant sample of thousands of dwarf galaxies,spanning a range of environments, it will be possible to explore the processes (e.g.,dynamical interactions, black hole accretion, supernova feedback) driving chemicalenrichment and regulating star formation. Internal galaxydynamics—using IFUsor fiberbundles available onMSE as second generation capabilities—will elucidate the roles oftides, past merger events, and stripping processes in the evolution of cluster galaxies.Section3.3.3hasafulldiscussionofthegalaxyevolutionsciencecaseforIFUcapabilitiesonMSE.

• AprecisionmeasurementoftheInitialMassFunction

Knowledgeof the IMF isasinequanon condition for the interpretationof the integratedspectra of galaxies. A survey of rich clusters in combinationwith the survey described inSRO-06 will efficiently provide a definitivemeasurement of IMF-sensitive line indices forgalaxiesasafunctionoftheirmass,morphology,andacrossallpossibleenvironments.

• High resolution (R ∼ 20000) abundance analysis of compact stellar systems — compactgalaxies, stellar nuclei, Ultra Compact Dwarfs (UCDs) and globular clusters (GCs). Severallinesofevidencepointto(someof)thesesystemsastheend-resultofextremeevolution—tidalstrippingandthreshing—withinaclusterenvironment.MSEwillallowustotestthishypothesisbyprovidingpreciseestimatesofages,metalabundance,recentstarformationand AGN activity for thousands of compact systems, investigating possible links amongstbaryonicsubstructuresthathavetraditionallybeenthoughofasdistinctfromeachother.

Inthenextdecade,MSEwillbeastransformationalforclusterstudiesasSDSSwasinthedecadepast.Mostofourmodernunderstandingaboutnearbygalaxyclusterscomesdirectlyfromthe


135

SDSSand2dFGRS:thesurveyshomogeneityandcontrolofsystematicsarewhatallowedustobegin to understand the interplay between galaxies and their environments. However, thelimiting magnitude, moderate signal-to-noise ratio, moderate resolution and spatialincompletenessatsmallseparationsofSDSSareseverelimitationsintheutilizationofSDSSdatafor studies of galaxy clusters. Although other surveys (e.g. GAMA) have probed deeper thanSDSS, theyonly targeted relatively small areas of the sky or a selected few clusters.MSEwillovercome these limitations.Most critically, its fieldof viewandmultiplex capabilitywill allowMSE to directly couple detailed information on kpc scales or smaller (associated with starformationandgalacticwindsforexample)tostructureonscalesofseveralMpcandlarger—atask that has historically posed severe observational challenges. Finally,MSEwill provide thespectroscopicdatasettomatchthenextgenerationofdeepimagingsurveys—Euclid,LSSTandeROSITA— and complement current and next generation of ground based observatories, inparticularALMA,SKA,andJWST.

3.4.2 TheVirgoandComaClusters

Figure62:TheleftpanelshowsthespatialdistributionofgalaxiesthatareclustermemberlocatedwithinonevirialradiusoftheVirgocluster,colorcodedaccordingtotheirtotalmagnitudeasindicatedbythebarattheright.TheCFHT image of the core of the cluster (a 2 square degree area surroundingM87) is shown to the right, wheregalaxiesare identifiedbyyellowor redellipsesaccording towhether theywereknownornotprior to theCFHTNGVSsurvey.

InaddressingtheproblematicsoutlinedinSection3.4,localgalaxyclusters—inparticulartheVirgoandComaclusters—havespecialsignificance.Accordingtosimulations,themajorityofgalaxies thatarealreadymassiveandevolvedatz≥2willendup inhaloeswithmassesMh>1014

M

¤today(Poggiantietal.2013);richclustersatlowredshiftsarethereforetherepositoryof themajorityof thedescendantsof themassivegalaxypopulationstudiedathigh redshifts,and it is where we should look to trace their evolution. Additionally, while studies at highredshifts are largely confined tomassive galaxies (M* > 1010.5

M¤), it is only through detailed

studiesofthe localvolumethatwecanhopetounderstandtheroleofgasdynamics,cooling,starformationandfeedbackinthehierarchicalassemblyofbaryonicsubstructures.


136

TheVirgoandComaclustersaffordusthebestopportunitytostudythegalaxypopulationinitstotality,fromquiescenttostarforminggalaxies,fromthehighdensitycorestothelow-surfacebrightnesshaloes,acrossawidemassrange,onallspatialscales,inacontrolledenvironmentinwhichallotherbaryonicsubstructuresarealsowellcharacterized.Atadistanceof16.5Mpcandwith a gravitating mass of M200 = 4.2 ×1014 M¤

, the Virgo Cluster is the dominant massconcentration in the local universe, the centre of the Local Supercluster, and the largestconcentrationofgalaxieswithin~35Mpc.Itsproximityallowsgalaxiestobestudiedonscalesofonly100pc (1arcsecat thedistanceofVirgo).ComahasagravitatingmassofM200=2.6×1016M

¤(Kuboetal.2007)withinitsvirialradius(r200=2.8Mpc=1.6deg,comparedto1.6Mpc=5.5

deg for Virgo), but given its larger distance, a factor~5 farther away than Virgo, subtends asmallerareaonthesky.InComa,MSEcouldobservemostofthevirializedregionwithinasinglepointing,andthisgreaterefficiencynearlycompensatesforthelongerintegrationtimesneededto reach the samephysical depth as inVirgo. InVirgo,~70 separateMSEpointingswouldbeneededtocovertheclusterouttoonevirialradius.

BothVirgoandComahavebeenextensivelystudiedatvirtuallyallwavelengths.ForVirgo,anewcatalogueoflikelyclustermembers(basedonphotometricandmorphologicalcriteria)reachingg= −7mag (stellarmasses~5 ×105

M

¤) is now available based on CFHT/MegaCamdata (the

Next Generation Virgo Cluster Survey, NGVS, Ferrarese et al. 2016, Figure 62). Detailedstructural parameters, reaching surface brightnesses of 30mag arcsec-2

(g-band), exist for all

clustermembers.A90%completecensusofGCs,aswellasmanyUCDs,isavailablewithinonevirialradius(Figure 63).Intheoptical,Comahasbeenimagedbybothground(CFHT,Subaru)andspace-based(HST)telescopes(Adamietal.2006;Yamanoietal.2012;Carteretal.2008),althoughnot to thedepthorspatialcoverage thatexists forVirgo.Extensivemultiwavelengthdataisalsoavailableforbothclusters.

Given the spatial extent of galaxies in Virgo and Coma, and the different density of variousbaryonicsubstructureswithinthecluster,arangeofMSEobservingmodesandconfigurationswillbeneededtocoverallsciencegoals.ThesesciencegoalsaredescribedindetailinSRO-07,andinclude:


137

Figure63:Leftpanel: thesurfacedensityofGCswithinonevirial radiusoftheVirgocluster (Durrelletal.2014).Orbiting in theclustergraviationalpotential,GCscanbeusedasdynamical tracers tostudytheunderlyingmassdistribution.(FigurefromDurrelletal.2014).Rightpanel:Gemini/GMOSIFUdataforasampleofcompactgalaxiesintheVirgocluster(Guerouetal.2015).Basedontheages,metallicitiesandinternaldynamicsinferredfromtheIFUdata, the authors suggest a tidally strippedorigin for these systems. These data are at present limited to ahandfulof(mostly)massivegalaxies;extendingthemtoasampleofseveralhundredgalaxies,allthewaytothedwarfgalaxyregime,wouldallowustoidentifyevolutionarytrendslinkinggalaxiesacrossawidemassrangeandasafunctionofenvironment.(FigurefromGuerouetal.2015).

• The star formation and enrichment history of dwarf galaxies (M*~ 107– 109M

¤). As the

buildingblocksfromwhichmoremassivesystemsareassembled,dwarfgalaxiesarekeytoourunderstandingofhierarchicalstructure formation.However, theprocessesdrivingthechemicalenrichment,thelevelandfrequencyofAGNactivity,andtheroleoffeedbackandenvironmental effects on their evolution are still largely unknown. The effective radii ofgalaxies in themass range of interests are 10 to 20 arcsec in Virgo, and 2 to 4 arcsec inComa. Single fiber, moderate resolution (R ∼ 6500) observations would yield gas phasemetallicities(providedthenecessarywavelengthcoveragecanbeachieved),starformationand enrichment histories, as well as provide diagnostic tests to assess the presence ofAGNs.Thiswouldallowustoextendtheknownscalingrelationsbetweenstellarmass,SFRand metallicity to the dwarf galaxy regime, and provide renewed constraints oncosmologicalmodels.

• Thenatureofcompactstellarsystems. InVirgo,asingleMSEconfigurationperfieldwouldyield systemic velocities and high resolution spectra for all stellar nuclei, GCs and UCDsbrighterthang=22mag.Thisincludes400stellarnucleiingalaxieswithmassesaslowas106M

¤andspanningarangeofenvironments,allknownUCDs,andover50,000GCswithinonevirialradiusofthecluster.AtthehighestMSEresolution(R≥20000)thespectrawouldyieldadetailedkinematicalandstellarpopulationanalysisandallowustoseekevolutionaryconnectionsbetweenthenuclei,globularclusters,UCDs,andthe largescalepropertiesofthehostgalaxies.

• Mapping the dark matter distribution. Faint objects orbiting in a galaxy gravitational


138

potential, includingdwarfgalaxies,GCs,planetarynebulaeandUCDs,arepowerfultracersoftheunderlying(total)massdistribution.Atlargedistancesfromthegalaxycores,wherethelowsurfacebrightnessofstellarhaloesmakesspectroscopicobservationsprohibitivelyexpensive, these tracers can be used to map the outer kinematics. A map showing thespatial distribution of over 50,000 globular clusters within the virial radius of the Virgocluster is shown in the leftpanelofFigure 63 (Durrelletal.2014).Sytemicvelocities forthese clusters (which would be a natural by-product of the program described in thepreviousbullet,butcouldalsobeachievedwithalowresolution,singlefibercondifuration)wouldleadtoaprecisemeasurementofthedarkmattermassdistribution,andalsorevealcoherentstructureinphasespacethatcanbeusedtoinfertheaccretionhistorywithintheclusterpotential.

• Astudyof the internalkinematicsofclustergalaxies.Hereand in the followingexamples,wewill refer specifically to Virgo; the generization to Coma is trivial and indeed a Comasurveywill generally benefit from the smaller projected extent of the cluster and higherdensityoftargets.TheVirgoclustercontainsabout700galaxies(DMsubhalos)withstellarmasseslargerthanM*~10

9M☉(seeFigure 62).Theinternalvelocitydispersionsofthesegalaxiesvaryfromσv~15–20kms-1atM*=10

9M¤ toσv~300kms-1

forthemostmassive

galaxies.Spatiallyresolved,mediumresolutionspectroscopy(R∼2000forgalaxiesthataremoremassive thanM*~ 6 × 10

9M¤ and have σv~ 75 km s-1, and R∼ 6500 for the less

massive systems), possibly using an IFU or fiber bundle, covering the galaxies from theircoretothe low-surfacebrightnessregionsouttoseveraleffectiveradii,wouldprovide2Dmaps of the age, metallicity and internal kinematics (see right panel ofFigure 63). Theinformation, available for several hundred galaxies spanning nearly three orders ofmagnitude in mass and a range of cluster environments, would be transformational. Itwould elucidate evolutionary trends along theHubble sequence and allowus to emplorethenatureoftheperturbingmechanisms(tidalstripping,threshing,ram-pressurestripping,starvation,etc.)thatshapegalaxieswithinaclusterenvironment.

To appreciate the power of MSE for a spectroscopic survey of this sort, consider the timerequiredtoconductanequivalentsurveywithexistingorsoontobecommissionedmulti-objectspectrographs. As detailed in SRO-07, a high resolution abundance analysis of all baryonicstructures brighter than g ~ 22.0 within a virial radius of the Virgo cluster (100 deg

2) wouldrequireabout400hrsofdarktimeandyieldspectrafor~50,000sources.Itwouldthereforebepossible to complete this survey in, atmost, twoobserving seasonswithMSE, assuming thatVirgo is observable during dark time for an average of ~ 200 hrs per year. A comparableprogrammewiththetwoVLThighresolutionspectrographs,VLT-FLAMES/GIRAFFE(132fibres,0.126deg2FOV)andVLT-MOONS(1000fibres,0.14deg2FOV)wouldrequire24and22years,respectively.ThesameprogramusingMMT-Hectochelle(240slits,0.013deg2FOV)orMagellan-M2DF(128slits,0.02deg2FOV)wouldbeunfeasibleduetothesmallFOVoftheinstruments.

3.5 Galaxiesacrossredshiftspace3.5.1 Redshiftevolutionofclusteringandhalomodels

ScienceReferenceObservation8(AppendixH)

Multi-scaleclusteringandthehalooccupationfunction


139

Weproposeanambitiousdesignof ninephoto-z selected survey cubes thatwill allowMSE tomeasurethebuildupof largescalestructure,stellarmass,halooccupationandstarformationouttoz=5.5.Bytargeting(300Mpc/h)3

boxes,eachvolumewillmeasure“Universal”valuesfor

anarrayofpotentialexperiments.At lower redshiftswewilldirectlyobservehaloabundancesbelow1012M

¤,whichmeanswecanmeasuretheoccupationofhalosandtheirabundanceover

afourdecaderangeinhalomass,accountingforthemajorityofstellarmassinthelow-redshiftUniverse.Athigherredshiftsoursurveyvolumeswilltracethetransitionfrommerger-dominatedspheroid formation to the growth of disks, covering the peak in star-formation and mergeractivity. This combination of depth, area and photo-z selection is not possible without acombination of LSST andMSE. As such, MSE will be able to produce the definitive survey ofstructure,halosandgalaxyevolutionover12billionyears.

Thedarkmatterhalosthatarebelievedtohostmassivegalaxiesmakeupmorethan85%ofthegalaxy’stotalmass.Anessentialaspectofgalaxyevolution is theunderlyingcomplexbaryonicprocesses that take place within these dark matter halos. Processes such as gas infall andoutflow, star formation, metal enrichment, negative feedback to star formation, black holegrowth,andgalaxymerging,shouldallworkdifferentlyfordifferentdarkmatterhalomassesatdifferent cosmic times. However, directly measuring dark halo masses either requiresobservations that are very challenging (especially at high redshift, e.g., Conselice et al. 2003;Swinbanketal.2006;Mandelbaumetal.2006),orrelyonunusualandrarecircumstancessuchasstronggravitationallensingevents(e.g.,Treu2010).

Thankfully, halo clustering is a strong function of halo mass, with more massive halos beingmore strongly clustered. Galaxies of different stellar masses and different star formationhistories will correlate with dark matter halos of different masses and this relationship willevolvewithtime(e.g.,Foucaudetal.2010;Wakeetal.2011;Linetal.2012).Studiesofgalaxyclusteringinlarge,localsurveyshaveshownhowclusteringatz~0doesdependsignificantlyonseveral specific properties. These include luminosity (e.g., Norberg et al. 2002a; Zehavi et al.2005),colourorspectraltype(e.g.,Norbergetal.2002b;Zehavietal.2002),morphology(e.g.,Guzzoetal.1997),stellarmass(e.g.,Lietal.2006)andenvironment(e.g.,Abbas&Sheth2006).Inrecentyears,deepspectroscopicsurveyshavemadeitpossibletoextendtheseinvestigationstoz≤2,yieldingresultsonhowthesedependenciesevolvewithtime(e.g.,LeFevreetal.2005;Coil et al. 2006;Meneux et al. 2008, 2009).More luminous andmassive galaxies tend to bemoreclustered,andthissegregationseemsstrongeratz~0thanatz~1.


140

Figure64:Leftpanel:Halomassestimates,MminandM1, forallgalaxysamples fromtheCFHTLS-widesurvey,asfunction of luminosity threshold, corrected for passive redshift evolution to approximate stellar mass selectedsamples. The dashed lines correspond to the relation between central galaxy luminosity,𝑳𝒄! , and the halomassstatedinZehavietal.(2011).RightPanel:Light-to-halomassratios,𝑳𝒄! ,/Mmin,withidenticalparametersasthosefittedwiththesamerelationbutasfunctionofhalomass.(FigurefromCouponetal.2012).

Theseresultssupportascenarioinwhichthestellarmassofagalaxyisessentiallyproportionaltothemassofthemostrecentdarkmatterhaloinwhichitwasthecentralobject(Conroyetal.2006; Wang et al. 2006). Investigating such results at higher redshift is essential, as therelationship between the galaxies and their halos is expected to be more straightforward.However,samplesatz>2arelimitedandbasedmainlyonmulti-colourphotometricselectionsor photometric redshifts and suffer from limitations in sample size, area covered andincompleteness.Inordertocomparethecorrelationfunctionξ(rp,π)anditsprojectionwp(rp)athigh redshifts (z > 2) with similar measurements at z < 2, a deep, accurate, complete andunbiasedsampleofspectroscopicredshiftsoveralargeareaisrequired.

It isnatural touseahalo-basedprescriptionwheregalaxies formby thecoolingofgaswithinboundandvirializedclumpsofdarkmatter.GalaxiesoccupydarkmatterhalosfollowingaHODmodel.Inturn,theHODfullydescribesthebiasinthedistributionofgalaxieswithrespecttotheunderlyingdarkmatterdistribution, intermsoftheprobabilitydistributionthatahaloofvirialmassMh containsN galaxies of a given type (Berlind&Weinberg 2002). HODmodels can beused to directly investigate the changing relationship of darkmatter and luminousmatter, toseparatecontributionsfromsatelliteandcentralgalaxies,andrelatethemtothemassesofthedark matter halos. Until now, the majority of analyses conducted using HOD modeling tointerpret galaxy clustering have been based either on large, low-redshift surveys such as theSDSS (e.g., Zehavi et al. 2011), or at higher redshifts, smaller (~1deg2) deep fields such asCOMBO-17, VVDS and COSMOS (e.g., Phelps et al. 2006; Abbas et al. 2010; Leauthaud et al.2012). Photometric redshifts enable clustering measurements in the framework of the halomodeloveralargerredshiftbaselineandlargerarea(e.g.,Couponetal.2012),butsuchstudiessuffer froma lackofaccuracyatall scales required toconstrain themodels. Inparticular, thedarkmatterhalosinthelower-massregimesufferfromthelackofprecisionontheline-of-sight


141

dimension due to the reliance on photometric redshifts. Such imprecision tends to bias theresults for faint galaxies toward being satellite galaxies in large halos, rather than centralgalaxiesinlow-masshalos.

Theclusteringapproachtoextract informationonthedarkmatterhalomasseshasprovenanefficientwaytoidentifythephysicalprocessesdrivinggalaxyevolutionatthedark-matterscale.For instance, it is now established that the fraction of luminous-to-dark matter is a strongfunctionofdark-matterhalomass(e.g.,Foucaudetal.2010;Wakeetal.2011;Leauthaudetal.2012). This fraction decreases toward lowermasses, because the dynamical time is long andsupernovaefeedbackisstrong,preventingefficientstarformation.Thefractionalsodecreasestowardhighermasses,wherestar formation ispreventedbythecoolingtimeandstrongAGNfeedback.Thus,theefficiencyofstarformationpeaksatanintermediatemassrange(~1012M

¤

atz~0),andthelocationofthispeakevolvestohighermassesathigherredshift.Figure64,extractedfromCouponetal. (2012),demonstratesthatHODmodelingcanprobethe light-to-halo-massratio.Thisfigurealsoshowsthatthetechnique is limited inthe lowermassregime,especiallyathighredshifts,duetothelackofreliableredshiftsatthefaintend.

Figure65:Therelationshipbetweenthecorrelationlength,r0,andspecificstarformationrateofsBzKgalaxiesinthreedifferentstellarmassbins:9.0< log(M*/M¤

)<9.5(bluecircles),9.5< log(M*/M¤)<10.0(greentriangles),

10.0<log(M*/M¤)<10.5(orangesquares).(FigurefromLinetal.2012).

Interestingly,when looking at the clusteringproperties of galaxies selected according to theirstar-formation rates, evenmore complex phenomena can be explored.Figure 65 illustratestheevolutionof thecorrelation lengthwithdifferent specific star formation rates (sSFR).Twodistinct regimes are apparent, with galaxies having sSFR ≤ 2 × 10-9 yr-1 showing increasingcorrelation power with decreasing specific star-formation rates, and galaxies above thisthresholdshowtheoppositebehaviour.ThepositivecorrelationobservedathighsSFRreflectsthe“mainsequence”evolutionofgalaxies,withgalaxies indeeperdarkmatterhalos showingmoreefficientstarformation.Ontheotherhand,theanti-correlationobservedatlowsSFRscanbeunderstoodasanenvironmentaleffectsimilartothatseenatlowerredshifts,i.e.,theseare


142

likely to be galaxies that are falling into the denser environmentswhere their star formationactivity ismore effectively suppressed. Large samples of galaxieswith accurate spectroscopicredshiftsarerequiredtomakefurtherprogressinunderstandingthesetrends.Forinstance, inthecontextofHODmodels,thepreviouscorrelationbetweensSFRandcorrelationlengthcouldbe understood as a physical processes in action for satellite and central galaxies. It is alsoessentialtoprobehigherredshiftstobeabletounderstandwhen,andhow,suchrelationshipswereestablished.

Figure66:Leftpanelpresentsthedistributionofdarkmatterandstellarmassasafunctionofhalomass.Itisclearthatweexpect it tobehighlydominatedby1012M*halos.Rightpanelpresents theexpectedgalaxyoccupationfrequencyasafunctionofhalomass.Ourproposedsurveywillsamplethe1012M

¤halomassregimewithbetter

than5galaxiesperhalo,allowingforper-halodynamicalmassmeasurements.

Thecriticaldarkmatterhalomassrangeintermsofstellarmasscontentisthedecadearound1012M

¤(see leftpanelofFigure 66).ThishalomassrangecontainsourownMilky-Wayhalo

and that of our nearest large spiral galaxyM31. To robustly detect groups with a low false-positiveratewerequireatleast5galaxiestobeobservablewithinagroup/halo,andtomeasuredynamicalhalomasswithina factor~3accuracy requires10ormoregalaxies tobeobservedwithinagroup/halo(Robothametal2011).Also,itisimportanttoreduce“cosmicvariance”toanegligible level (Somerville et al. 2004) and so include a fair sample of rare, high- and low-densityregionsinthesurvey.

Combiningthesciencecasesmentionedabovetocoexistwithinthesamesurveyvolume(bothsky coordinates and redshift window aligned) requires a large sky area and a robust photo-zselectioninordertoimprovetheefficiencyofthevolumeoverlap.Withoutaphoto-zselection(i.e. only an apparentmagnitude selection)wewould naturally findmoremassive clusters athigher redshiftsonly, andwewouldhavea comparatively small volume that contains the fullrangeof1012M

¤–1015M

¤halos(seeFigure16inRobothametal2011).Also,withoutaphoto-z

selection itbecomesextremely inefficient tosample faintgalaxiesat lowredshift,with fainterapparentmagnitudesnaturallypushingthepeakinthen(z)distributiontohigherredshifts.

SRO-08describesapossibleobservingprogramwithMSEthatenablesavast rangeofsciencegoals simultaneously in addition to tackling the key issues of HOD studies and galaxyenvironments,asafunctionofredshiftto largelook-backtimes,thatemploysahomogeneousand consistent survey strategy across all of redshift space. The basic goal is to observe 3200

1010 1011 1012 1013 1014 1015

Halo Mass / Msun

Dark MatterStellar Mass

00.

20.

40.

60.

81

Com

ovin

g m

ass

dens

ity (a

rbitr

ary

units

)

1010 1011 1012 1013 1014 1015

Halo Mass / Msun/h

101

102

103

104

105

106

107

108

Num

ber o

f Hal

os d

0.5d

ex

Intinrisc HMF1+ galaxies per halo2+ galaxies per halo5+ galaxies per halo10+ galaxies per halo

MSE survey: 3,438 sq deg / 0<z<0.22 / i<25.3

Gal N = 16M


143

square degrees of the Northern extra-galactic sky (overlapping fully with the proposed LSSTsurveyregionandsubstantiallywiththecurrentSDSSfootprint)betweenredshift0and0.2anddown to a limitingmagnitude of iAB=23 selected from LSST standard depthmulti-year survey(see SRO 7 survey S1-W). To efficiently probe down to lowmass halos it is advantageous toobserve as large an area as possible, and 3200 sq deg over the suggested redshift extentminimally containsa300x300x300Mpc/h co-moving cube.Thismeansour volume is largeenough thatwe reachUniversal homogeneity in all three dimensions (Scrimgeour et al 2013,Driver&Robotham2010).Byobservingto i=23withinz=0.2weexpect tobedeepenoughthatessentiallyall1012M

¤masshalosarebothdetectedandhaveareasonablemassestimate,

with10ormoregalaxiesidentified.

Byrepeatingthebasicsurveydesign(inparticularthetargetvolume)forthislowredshiftHODandhaloabundancefocusedsciencecase,wecanopenupanewsuiteofscience.Figure 67shows a basic possible observing strategy,where a co-moving 300Mpc/h cube is targeted ineightseparateredshiftwindowsusingamixtureofLSSTphoto-zselection(S1toS7)andEuclidphoto-zselection(S8).Eachsurveyregionproposedwouldbetargeteddowntoadifferenti/Y-band limit butwith~100% completenesswithin each volume (certainly better than 95%, i.e.~SDSS).Suchasuiteofsurveysallowsdetailedhalooccupationmodellingouttoz=5,spanningtherapidincreaseandslowdeclineinuniversalstar-formation(toppanelofFigure67),theeraofmergerdominatedmassbuild-up(z>2),thetransitionintogalaxydiskformationandin-situstar-formation dominating build-up (z < 2) and the epoch of rapid large-scale structureformation (z>2).Forobtaining robust“Universalvalues” formerger ratesandstar formationhistorytheco-movingvolumeanalysedandthestellarmassdepthobservediskey.Byselectingcommon300Mpc/hcubeswewillhavesubper-centsamplevarianceindependentoflook-backtime.Byaimingtobecomplete tostellarmassesat least1dexbelowM* forS1 toS4wecandefinitelyexploretheinterplaybetweenmassbuildupthroughmergerandstarformation(seeRobothametal2014foraz=0.2versionofthisexperiment).Beyondthisrangewearelimitedbythehighqualityphoto-z i<25.3sampleprovidedbyLSST.Despite this,wecanstillprobe thedominantcomponentofstellarmass(M*)anditshalooccupationdistributionouttoz=2withS5 and S6. This takes our galaxy evolution analysis out to 10 Gyrs in consistent co-movingvolumes that have high statistical quality. Further, OII becomes visible out to z = 3.8 i.e. itbecomesacommonemission feature thatS1 toS8canallobserve,offeringaconsistent star-formation tracerovernearly12Gyrs froma single facility.Access tobothLy-αandOIImakespossibledirectmeasurementsofthefeedbackofgasintotheinter-stellarmedium.WewillalsobeabletoobserveHα,andconstructafullBPTdiagramouttoz=2(i.e.S1toS6).


144

Figure67:Aproposedrangeofequi-comovingMSEsurveycubes(S1toS8).ThetoppanelshowsthecosmicSFHfromHopkins&Beacom2006.The lowerpanel shows theproposedsurveys, showing thegrowth in large-scalestructureinTAOsimulated(50Mpc/h)3surveycubes.

Itisclearthatsuchasurveysuiteopensupavastamountofsciencethatisnotaccessibletoanyotherfacility.Anon-exhaustivesummaryisgivenbelow:

i. Measurehalooccupation(throughgroupfinding)below1012M¤.Thiscombinedwithan

HODanalysiswilldefinitivelyuncovertheinterplaybetweenstellarmassandhalos,andwillputanydetailedstudyofindividualhalos(thefutureMilkyWayandM31archeologystudiesdescribedinChapter2intoapropercosmologicalcontext.

ii. Directlyobservetheevolutionofthemassiveendofthehalomassfunctionouttoz=1(halftheageoftheUniverse).Thisisstepbeyondclustercountcosmology,andwillbeconductedwithahomogenousselectionwithLSST(theselectionfunctionisoneofthemajorlimitationsofcurrentclustercosmologywork).

iii. Study the evolution of the large-scale structure and cosmic web out to z = 5 with aconsistentgalaxytracerwithhomogenousbias.

iv. Study the merger rate and star-formation history for all galaxies down to M*/10(coveringtheconvergedmajorityofstellarmass)outtoz=1.

v. Studythecloseclusteringandmergerrateofmassivegalaxies(1011M¤)outtoz=5.

vi. WithphotometryfromLSSTwewillbeablethemajorityofthestellarcomponentofthecosmicspectralenergydistribution(CSED)outtoz=2.

vii. Togetherwithothernextgenerationtelescopes(mostsignificantlyLSST,Euclid,WFIRST,SKA,theVLOTs),thesesurveyswillhaveavitalrolein:

a. Measuringmorphologicalevolutiontoz=1b. Studyingtheinterplaybetweengas,dust,starsandenvironmentc. Providing IFU targets for 30+m telescopes for well understood samples at

multipleepochsd. OfferingthesamplesupersetforhighS/Nobservationsofindividualgalaxiesin

ordertoascertaintheevolutionofmetalsthroughcosmictime.

ThemaincompetitiontoMSEindoingsuchacombinedsurveyisSubaru/PFS(Takadaetal2013)


145

andVLT-MOONS(Cirasuloetal2012).PFSwillhaveaspectralrange380–1300nm,anditwillnominally be conducting z < 1, i < 21.5 and 1 < z < 2, i < 23.9 photo-z pre-selection surveys(basedonshallowerHSCdata).Inbothregimes,thescienceprogramsMSEwillundertakeareatleast a magnitude deeper (a factor~3 in stellar mass) and over much larger areas (fixed 16square degrees for PFS, 400 to 30 square degrees over the same range for MSE. The hugeadvantageofMSEcomparedtoalloftheproposedPFSsurveysisthatwewillbeextendingtothescaleofhomogeneity inall threedimensions.Formanyof thesciencecases laidouthere(andforeshadowedonasmallerscalebyPFS)thedefinitivemeasurementwillbemadebyMSE,with no further appreciable improvements to be made by moving to larger survey volumes(since many extragalactic measurements are dominated by sample variance not Poissonstatisticsofthesampleitself).

VLT/MOONSwillhaveaspectralrange680–1800nmandisfocusingonsurveysatz>1;tyingtogether low and high redshift surveys using equivalent tracers at all redshifts is therefore auniquestrengththatispossibleonlywithMSE.Further,MSEhasapotentiallyhugeadvantageinitsability touse theLSSTandEuclid/WFIRSTphotometricsourcecatalogues.These telescopeswill only become available post 2020, and offer a paradigm shift in available optical imagequalityover largeareas. Inparticular,usingLSSTallowsustoconductsurveysS1(3200squaredegrees) to S7 (22 square degrees) with identical photometry and photo-z technique. Suchhomogeneitywillimprovetherobustnessofalmosteverytypeofanalysismade.GiventhefastsurveyspeedofMSE,onceLSSTandEucliddatabecomeavailablenootherfacilitywillbeabletokeep up with the observing speed of MSE, i.e., it will dominate the next generation ofspectroscopicsurveysfrom0<z<4.

3.5.2 ThechemicalevolutionofgalaxiesacrossCosmictime

ScienceReferenceObservation9(AppendixI)

ThechemicalevolutionofgalaxiesandAGNoverthepast10billionyears

We propose a high-SNR (SNR>30) spectroscopic study of carefully selected volume-limitedsamples from the present epoch to the peak of star-formation 10 billion years earlier. Eachvolume-limited sample should contain 10k galaxies uniformly spanning the broadest possiblestellar-massrange,andinregular1Gyrintervals(i.e.,10volume-limitedsamplesof10kgalaxiesresultinginatotalsurveyof100kgalaxies).Thevolume-limitedsamplescouldbeselectedeitherfromphoto-zcataloguesidentifiedbyLSST,EUCLIDorWFIRST,orfromaredshiftpre-surveyofthe selected regions (i.e, such as that proposed by Science Reference Observation 8). Keyrequirementsforthissurveyarehighthroughout,near-IRspectralcoverage(extendingasred-wardasphysicalpossible),butrelatively lowtomodestresolvingpower(R~1000). Integrationtimesarelikelytoextendupto~50hrsforthefaintesthighestredshiftsystems.Hence,thefullstudy will require 50 – 100 nights of observations (assuming a multiplex factor i.e.,~1000targets per field-of-view), resulting in a legacy sample formultiple and diverse studies of thechemicalevolutionofgalaxiesandAGN.

The chemical evolution of galaxies and AGN over cosmic time, and in particular the gas- andstellar-phasemetallicities, isatopicofsignificantcurrentattention(e.g.,Zahidetal.2014,andreferences therein). Relatively strong evolution has been reported in the stellar mass –


146

metallicity relation (Tremonti et al. 2004) fromnearby samples to those close to the peak ofcosmicstar formationhistory.See, forexample,samplesassembledbyZahidetal. (2013)andFigure 68.However, thesamplesizesarerelativelymodestandarguablybiasedtowardsthehigheststar-formingsystemsateachepoch.Thisisbecausetheselectionistypicallymadeshort-wardsofthe400nmbreakfortheveryhigh-redshiftintervals,andassuchprobesonlythemostUV-bright galaxies. Locally we know that the specific star-formation rate, stellar mass, andmetallicity form a fairly complex surface (e.g., Lara-Lopez et al. 2013a).Moreover, gas-phasemetallicity (usually the only measurement accessible in high-z surveys) is harder to interpretthanthestellar-phasemetallicity,astheformerareverysensitivetorecentinfallandoutflows.IdeallyonewouldwishtostudythesSFR–stellarmass–metallicitytrifectausingbothgasandstellar-phasemetallicities,forrepresentativepopulationsatregulartimesteps.Byassumingtheearlierpopulationevolvesintothelaterpopulationonecanbuildupacompletepictureoftheglobalevolutionofboththegasandstellar-phasemetallicitesovertime,andtheimpactofbothsSFRandstellarmassonmetallicities(andvice-versa)atarangeofepochs.Therequiredinfallandoutflowmodels can thenbedetermined through comparisons to simple analyticmodels,withpotentiallysomeadditionalconstraintscomingfromthecomparisonoftheemissionline-widths(sensitivetodynamicsandoutflows)totheabsorptionline-widths(sensitivetojustthedynamics)usingmulti-Guassianline-fitting(e.g.,McElroyetal.2015).

Figure68: Themass-metallicity relation seenbyGAMAandwith the redshift rangesdiscussed in SRO-09 shown(0.1Zsolliesatthelowerlimitoftheplot).MSEwillprobemasslimitssimilartoGAMAatz<0.1andz<0.2butouttoz<1andz<2respectively.Thez<0.1mass–metallicityrelationfromSDSS+GAMA(Lara-Lopez,etal.2013a) isdisplayedasthecyanline,themagentalineshowthesamerelationnormalizedtothe0.25<z<0.39sampleandthegreenpolygonshowsthestackedgalaxyz~1.6relationfromZahidetal.(2013).Theseobservationsclearlypredict

108 109 1010 1011

M*/Msun

88.

59

9.5

12+l

og(O

/H)

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

●●

●●

●

●

●

●●

●

●●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●●

●

●

●

●

●

●

● ●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●●

●

●

●

●●

●

●

●

●

●

● ●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

● ●

●

●

●●

●

●

● ●

●

●

●

●●

●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●●

●

●●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●●

●

●

●●

●●

●

●

●

●

●

●●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●●

●●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●●

●

●

●

●

●

● ●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

0.06<z<0.10 (GAMA − 3081)

●●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●●

●

●

●●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

● ●

●

●

●

●

●●

●

●●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

● ●

●●

●

●

●

●

●

●●

●

●

●

●

●

●

●

● ●

●

●●

●

●

●●

●

●

●●

●

●

●

●

●

●

● ●

●

●

●

●●

●

●●

●

●

●

●

●●

●●

● ●

●

●

●

●●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

● ●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●●●

●

●

●

●

●

●

● ●●

●

●●

●

●

●

●

●●

●

●

●

●

●

●

●

● ●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●●

●

●

●

●

●

●

●

●

●●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

● ●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●●

●

●

●

●●

●

●

●●

●

●●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

● ●

●

●

●

●

●

●

●●●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●●

●

●

●

●

●

●

●●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

0.15<z<0.17 (GAMA − 1821)

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●●

●

●

● ●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●●

●

●

●

●●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●● ●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

● ●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

●●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●●

●

●

●

●

●

●

●

● ●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●●

●

●

●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●●

●●●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

● ●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●● ●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●●

●

●

●

●

●●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●●

●

●

●

● ●

●

●

●●

●●

●

●

●

●

●

●

●

●

●

●●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●●

●

●

●

●

●●

●●

●

●

●

●

●●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

0.25<z<0.28 (GAMA − 1646)

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●●

●

●

0.37<z<0.39 (GAMA − 382)

●●●

●

●

●

●●

●

●●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

0.50<z<0.53 (GAMA − 86)z>0.53 (GAMA − 0, MSE − 50k)

MSE z<1 ~m

ass limit

MSE 1<z<2 ~m

ass limit

z~1.6

z<0.1

z~0.3?


147

a strongevolution in theM–Z relationasa functionof redshift.MSEwill probe this evolutionusing individualgalaxies,inacomparativelyunbiasedsampleouttoz~2.

The measurement of stellar- phase metallicities, essential for the above analysis, requiresignificantlyhigher-S/N spectra thanare typically attained in surveys suchas zCOSMOS,VVDSand Deep2, since they are reliant on absorption rather than emission line measurements.Modelingof line-broadeningthroughdynamicsandoutflowsalsorequiressignificantsignal-to-noise levels inthespectral featuresandforthefeaturestobeclean(i.e.,awayfromnightskylines).Asonepushes towardshigher-redshift this lateraspectbecomesharderbecauseof thepreponderance of telluric features, etc. This can potentially be overcome by using UVmetal-lines (e.g., Sommariva et al., 2012;Figure 48 showed thedrift of key spectral featureswithlook-backtime).MSEisabletoprovideessentialdatatofurtherthisfieldbyconstructingahigh-SNRsampleofaseriesofwellselectedandsufficientlyextensivesamplestoaddressanumberofcompellingquestionsby:

i. Providing a fully empirical description of the sSFR-stellarmassmetallicity relation forsystemswithmassesgreaterthanM*~ 109M

¤tothepeakofthecosmicstar-formationhistoryatz~ 2(seeFigure 43).Thisempiricalrelationwouldincludemeasurementsofboth the gas and stellar-phase metallicities, removing any biases due to recent gasinfall/outfall.Suchanempiricalresultwouldprovideanidealbenchmarkforthefurthercalibrationanddevelopmentofnumerical(hydro-dynamical)andsemi-analyticmodels,whichasyethaveveryfewhigh-zmetallicityconstraints.

ii. Quantifying the decline in the abundance and frequency of AGN activity within thenormalgalaxypopulationatallredshiftstoz~2(theepochofpeakAGNactivity)usinghighSNRemissionlinediagnostics,suchasBaldwin,Phillips&Terlevich(BPT)diagramsand WHα versus [N II]/Hα (WHAN) line strength diagnostics. This would enable theexploration of the co-evolution of galaxies and AGNs by measuring both the starformationhistoryofthegalaxypopulation,thedeclineintheluminositydensityofAGNandanycouplingbetweenthesetwo.

iii. Measuringandcomparingthegas-phaseandstellar-phasemetallicitiestoconstrainthedegree of pristine gas infall/outflow as a function of look-back time. Some modelssuggest galaxies initially evolve according to a closed-box model after which freshpristinegasarrives significantly reducing thegas-phasemetallicity.Theclear signatureofaclosed-boxmodel isa stellar-phasemetallicityapproximatelyhalf thatof thegas-phase.Detailedcomparisonsofthestellarandgasphaseswillprovideconstraintsonthedegreeofinfallandoutflowthatcanthenbemonitoredovertherangeoftimesampled.

iv. Performingdetailed stellarpopulationmeasurements todetermine theprogressionofstellar-evolution.Earlierepochsofstar-formationaretypicallyhiddenwithinlaterwavesof star-formation, however detailed analysis of the complete spectrum on a pixel-by-pixelratherthanlinebasiscanrevealthefullstar-formationhistories.Thesehistoriesoflater volume-limited samples need to mesh with what is seen in the earlier volume-limited samples enabling detailed test of our star-formation models. It is also worthnothing that such an analysis as outlined in iii and iv, combined with ASKAP/SKA HImeasurements (at the lower-z end), will provide a comprehensive test of the star-formation history and baryonic composition of individual galaxies, placing strongconstraintsonlikelygalaxyevolutionscenarios.

v. Usingmultiplestar-formationtracers(e.g.,HαvUVvmidorfar-IR)tostudytheonsetof


148

star-formationwithdynamicalseparationand/orenvironment.

To identifyAGN, star-forming, and inert galaxies,measure their chemical composition, and toachievetheobjectives listedabove,requiresnotonlyarobustredshiftmeasurement,butalsosufficient signal-to-noise spanning a broad range of emission and absorption features. Torobustly distinguish AGN from star-forming galaxies requires SNR > 5 in the rest-frame 370 –680nm (OII through to Na), to measure gas-phase metallicity (0.1 Z

¤), and to measure star

formationratestomodestlevels(1M¤/yr),requiresSNR~ 10–15(Choietal.2014).Torecover

stellar-phasemetallicitesusing,forexample,Lickindices(400–650nm)orUVmetalslines(130–200nm)to0.1Z

¤,werequireS/N~ 20–30.Here,MSEcanbeagamechanger.

Massivesamplesarenotcritical,althoughinordertofullymapouta3DstructureinsSFR,stellarmass andmetallicity does require volume-limited samples of~10k systems. This assumes anaverage of 20 galaxies per bin and~8 bins each in SF, stellar-mass and metallicity. SRO-09describes a possible MSE observing program where S/Ncontinuum~30 spectra from 0.4 – 1.8micronareobtainedfor10sampleseachcontaining10kgalaxies.Thesesamplesarespacedatregular Gyr time steps from the present epoch to a look-back time of 10 Gyrs. This samplesprovies fromthepeakof thestar-formationactivityat z~2, to thepresentepoch. In total, theproposedprojectwould target100000galaxies,whichwouldbeeither spectroscopicallypre-surveyedwithMSE itself,or selected fromtheupcomingphoto-zsurveysofLSST,EUCLIDandWFIRST.

ThepotentialofMSEhereistoprovidesignificantstatisticstohaveatransformationimpactandfullymaptheevolutionofthemass-metallicityrelationtothepeakofthestar-formationactivityat z~2. Existing facilities are capable of constructingmodest samples but only with the highthroughput, high fiber density and extensive spectral coverage thatMSE can provide, do wehavethepotentialtomovebeyonddetectionstorobuststatistics.Theseobservationswillspan,not just themost luminous systems, but those systemswhich contain the bulk of the stellarmassatthecurrentepoch(andhencethebulkofmetalsintheUniverse).


149

Chapter4 IlluminatingtheDarkUniverse4.1 ChapterSynopsis

• TheUniverseinabsorption:linkingstarsandgalaxiestotheirsurroundingmedia• Theevolutionofsupermassiveblackholesthroughmulti-objectreverberationmapping• ExploringtheUniversethroughmassivetransientdatasets• Aprecisionmachinetomeasurethedynamicsofdarkmatterhalosonallspatialscales.

ScienceReferenceObservationsforIlluminatingtheDarkUniverse• (SRO-03:TheformationandchemicalevolutionoftheGalaxy)• (SRO-04:Unveilingcolddarkmattersubstructurewithprecisionstellarkinematics)• SRO-10:Mappingtheinnerparsecofquasarsthroughreverberationmapping• SRO-11:LinkinggalaxyevolutionwiththeIGMthroughtomographicmapping• SRO-12:Apeculiarvelocitysurveyoutto1GpcandthenatureoftheCMBdipole

MSEwill cast lighton someof thedarkest componentsof theUniverse. Thebackboneof thelarge scale structures in which galaxies are embedded is dark matter, that is invisible inelectromagnetic radiation. Embedded alongside galaxies in these large scale structures is theintergalacticmedium,amajorreservoirofbaryonsintheUniversebutonewhichisnotoriouslydifficulttoobservedirectlyinthediagnostic-richopticalwavelengthrange,otherthanthroughits absorption effects on background galaxies. Within galaxies, the interstellar medium issimilarlydifficulttodetect,andthecentralsupermassiveblackholesaregenerallyonlyprobedthrough indirect techniques. MSE will provide unprecedented perspectives on each of thesecomponentsofthedarkUniverse,andprovideasuiteofcapabilitiesideallysuitedtoprovidingcriticaldatafornextgenerationcosmologicalsurveys.

Specifically,MSEwillprovidetomographicreconstructionofthethreedimensionalstructuresoftheGalacticinterstellarmedium(ISM)andtheintergalacticmedium(IGM)throughhighsignal-to-noise absorption studies alongmillions of sight lines, that are orders ofmagnitudes largerthan that which is currently possible. Whether focused on stellar or galactic science, therelationsbetweenstarsandgalaxiesandtheirsurroundingmediaisakeyissueinunderstandingthecyclingofbaryons. In theextragalactic regime,MSEcan target the IGMat z~ 2–2.5, toreconstruct the dark matter distribution and associated galaxies at high-z, allow detailedmodeling of galactic scale outflows via emission lines, investigate the complex interplaybetween metals in galaxies and the IGM, and provide the first comprehensive, moderateresolutionanalysisofalargesampleofgalaxiesatthisepoch.

MSEwill allow for anunprecedented, extragalactic timedomainprogram tomeasuredirectlythe accretion rates and masses of a large sample of supermassive black holes throughreverberationmapping. This information is essential for understanding accretion physics andtracing black hole growth over cosmic time. Reverberation mapping is the only distance-independentmethodofmeasuringblackholemassesapplicableatcosmologicaldistances,andcurrentlyonly~60local,relatively low-luminosityAGNhavemeasurementsoftheirblackholemasses based on this technique. MSE will conduct a ground breaking campaign of ~100


150

observationsof~5000quasarsoveraperiodof several years (totaling~600hourson-sky) tomap the inner parsec of these quasars from the innermost broad-line region to the dust-sublimationradius.Theseobservationswillmapthestructureandkinematicsoftheinnerparsecaroundalargesampleofsupermassiveblackholesactivelyaccretingduringthepeakquasareraand provide a quantum leap in comparison to current capabilities. Other extragalactic timedomain programs can be expected to produce some of the largest spectroscopic datasets ontransientphenomenathatexist,thatwillbeessentialresourcesastheinternationalcommunitydevotesincreasingobservingtimetothisrelativelyunexploreddomain.

MSE is theultimate facility forprobing thedynamicsofdarkmatteroverevery-and-all spatialscalesfromthesmallestdwarfgalaxiestothelargestclusters.FordwarfgalaxiesinthevicinityoftheGalaxy,MSEwillobtaincompletesamplesoftensofthousandsofmemberstarstoverylarge radius and with multiple epochs to remove binary stars. Such analyses will allow theinternaldarkmatterprofiletobederivedwithhighaccuracyandwillprobetheoutskirtsofthedark matter halos accounting for external tidal perturbations as the dwarf halos orbit theGalaxy.Anewregimeofprecisionwillbereachedthroughthecombinationoflargescaleradialvelocity surveys using MSE with precision astrometry for tangential velocities in the centralregionsusingtheVLOTs.IntheGalactichalo,largescaleradialvelocitymappingofeveryknownstellar streamaccessible fromMaunakeawithhigh velocityprecisionwill reveal theextent towhichthesecoldstructureshavebeenheatedthroughinteractionswithdarksub-halosandsofor the first time place strong limits of themass function of all subhalos (luminous and non-luminous) in an L* galaxy. On cluster scales, MSE will use galaxies and globular clusters asdynamicaltracersinexactlythesamewayasstarsindwarfgalaxies,toprovideafullyconsistentportraitofdarkmatterhalosacrossthehalomassfunction.

Finally,MSEprovidesaplatformforenablingnextgenerationcosmologicalsurveys,particularlythroughlargescalesurveysoftheinternaldynamicsofgalaxiesaspartoffuturemulti-objectIFUdevelopments.Suchsurveyscanprovidedynamicaltestsprobingthescaleofhomogeneity,cantestthebackreactionconjectureinGeneralRelativity,andprovidesanovelmeansofmeasuringgravitational lensingviathevelocity field, thatcomplementsstandardmeasurementsbybeingsensitivetoverydifferentsystematics.

MSEwilltransformthestudyofsomeofthedarkestcomponentsoftheUniverse.Absorptionlinestudies, reverberation mapping programs, massive transient datasets and precision velocitystudies of stars and galaxies as tracer particles utilize aspects of the performance of MSE –sensitivity, stabilityandspecialization– thatareunique to the facilityandwhichconsequentlyallowuniqueinsightsintokeyaspectsofthebaryoncycle,blackholephysics,thetimedomain,and cosmology. MSE represents a suite of capabilities that is an essential platform for thedevelopment of cosmological programs that go beyond the current generation andwhich arerelevantonthe2025+timescale.


151

4.2 TheUniverseinabsorption4.2.1 TheInterstellarMedium

Figure 69: Themidplane of theMilkyWay near the constellation of Perseus. 21cmobservations ofHI from theCanadianGalacticPlanesurvey(Tayloretal.2003).OverlaidonthesehydrogendataareIRASimagesofthedustthat intersects our line of sight through the disk. The point-like sources (WRST) aremainly distant galaxies andAGN.(FigurefromJ.English,http://www.ras.ucalgary.ca/CGPS/press/aas00/pr/pr_14012000.html).

MSEisauniqueandpowerfultoolforstudyingtheinterstellarmedium(ISM)oftheMilkyWay.It will play an essential, central role in revolutionary three dimensional ISM mappingachievements that will be boosted by the ESA Gaia parallax distances. In parallel, and incombinationwiththe3Dmapping,itwillallowunprecedentedspecificstudiesoflong-standingandnewquestionsaboutthenatureoftheISM.Inallcases,thepowerofMSEforthisscienceliesinthecombinationoflargeaperture,highthroughput,highspectralresolution,widefieldofview,andhighmultiplexing:

• Thelargeapertureandhighthroughputincreasesthenumberofweaktargetsthatcanbe individually studied and the distance coverage. In the case of large-scale ISMmapping,MSE’shighobservingefficiency reduces surveydurations to tractable levels,puttingitfarbeyondthereachofanyotherexistingorplannedfacility/instrument.

• The high spectral resolution improves the determination of ISM absorbing columnsand/or stellar types used to derive the reddening, and it allows coupling absorptionswithemissionspectrabymeansofvelocityidentification;

• The wide field of view facilitates the required coverage over large fractions of theGalaxy;

• Thehighmultiplexing increases thenumberof targets along a given sightline and thespatial resolutionof themaps. Fordetailed studyof individual structures,MSE’swidefieldanddensefibrecoverageiswellmatchedtomanyGalacticfeatures,suchasstellarstreams, astrospheres, and individual interstellar clouds, targets thatwouldotherwisebeaccessibleintheradioregionalone.


152

4.2.1.1 TowardsdetailedthreedimensionalmapsoftheGalacticISM

Multi-wavelengthobservationsoftheGalacticISMhaveneverbeensodetailednorsoabundantasnow.High-qualityemissionmapsofthegasanddusthavebeenproducedthankstoawiderangeofground-orspace-basedfacilitiesthatoperateatalmostallwavelengths,includingtheγ–ray (FERMI), X–ray (ROSAT, Chandra, XMM, Suzaku), UV (GALEX), IR (ISO, Spitzer), sub-mm(Herschel, Planck), mm (JCMT, APEX) and radio (DRAO, LOFAR) regions.Figure 69 shows acompositeimageofthemidplaneoftheGalaxyneartheconstellationofPerseus,createdfrom21cmdata fromtheCanadianGalacticPlaneSurvey (Tayloretal.2003),with IRASandWRSTdataoverlaid.Typically,datasetsusuallycomeintheformoftwo-dimensionalimagesinspecificspectral bands, or in data cubes (i.e., two spatial dimensions plus a spectral, or polarimetric,dimension). Yet, paradoxically, our understanding of the ISM structure remain surprisinglylimited,evenat the largest spatial scales,primarilybecausewe lack theaccuratedistancesofthe features that are responsible for the observed emission and their three-dimensional (3D)distribution.

Building accurate 3D maps of the ISM requires (i) massive amounts of distance-limitedabsorption data, i.e., stellar surveys at high spectral and spatial resolution covering largefractionsof the sky,and (ii) accompanying informationon the targetdistances.MSEandGaiawillprovidesuchapowerfulsynergy.Moreprecisely,therewillbetwolevelsofmapping:

• an initial distance assignment for intervening clouds based on all types of absorptiondata,dustreddening,gaseouslinesordiffuseinterstellarbands.Theresulting3Dmapswillbeageneraltoolofwideuse;

• differentialmapping,i.e.,comparisonofthedistributionsofthevariousabsorbers,e.g.,DIBs and dust. This will opens a variety of studies, in particular on the ISM physicalpropertiesandthemulti-phasestructure.

Today our understanding of the origin, interplay and evolution of the various phases of theGalacticISMisseriouslyhamperedbythelackofrealistic3Ddensityandvelocitydistributions,since it prevents the construction of quantitative, physically-motivated models. It is hard tobelieve that the distance to the most prominent features of the X-ray, sub-mm and radiocontinuumfull-skymaps–theso-calledNorthPolarSpurandLoop1features–variesfromthesolarneighborhoodtothedistanceofthegalacticcenter,dependingonauthors.Consequently,thislackof3DperspectiveimpactsnegativelyonalargenumberofresearchfieldsdirectlylinkedtotheMilkyWayISM:starformationanditslinkwithparentalISMproperties,includingtheroleof turbulence,magnetic field,metallicity or dust content; cavities carvedby stellarwinds andsupernovaeand theirevolutions;mixingof stellarejecta; large scaleevolutionarymodelingofthe Milky Way structures, bulge, bar, spiral arms, halo properties; tests of accretion versusinternalevolution;etc.Forallthesefields,the3Dstructurewouldbeaninvaluableingredienttomakefullprofitofall2Dphotometric,spectralandpolarimetricdatasets.

This situation is also regrettable since a number of major projects making use of prominentastrophysical facilities and/or surveys from ground or space require an increasingly detailedunderstanding of the Galactic ISM. For instance, interpreting observations of the cosmicmicrowavebackground(CMB)anditspolarizationisstronglydependentonoptimalmodelingof


153

Galactic dust emission and its spectral characteristics. These, in turn, depend on the spatialdistributionandtemperatureofthedust,andontheintensityoftheinterstellarradiationfieldthat heats it. To produce realistic models of this radiation field, it is therefore necessary tocompute the propagation of photons through the 3D dust and gas distributions. Similarly,understanding theGalactic cosmic rays (GCRs) spectra and their spatial variability requires adetailed knowledge of the 3D distribution of the multiphase ISM. In turn, the analysis andmodelingofdiffuseγ–rayemission (which is generatedby the interactionof cosmic rayswithISMnuclei and by up-scattering of the interstellar radiation field by cosmic electrons) doublyrequires a 3D knowledge of the ISM distribution; on the one hand, because it is needed tocompute thecosmic ray trajectoriesaccurately,andon theotherhand,because the radiationfield itself is governed by this distribution. Finally, obscuration by dust poses a significantobstacle to our understanding of stellar populations within the Milky Way. Extinction andreddening are not always able to be estimated to the level of precision that is needed tocharacterizethepropertiesofdistantstars.

Studies of specific astrophysical targets would also benefit from detailed 3D ISM maps. Atpresent the absence of knowledge on their distances, foregrounds, environments andbackgrounds limits the identification of their own spectral features, or the studies of theinteractions with the ambient medium. In particular, all newly discovered structures aroundstars (bow-shocks, trails, ionization cavities) would strongly benefit from knowledge of thephysical and dynamical properties of the ambient medium. Indeed, searches for particularcategoriesofobjectscouldbemoreefficientlyconductedifonecoulduse3Dmapstoidentifythemostfavourableskyregions.

4.2.1.2 TheMSE–Gaiaperspective

Figure70:3Dmappingbasedonreddeningmeasurements.Planarcutinthe3Dopacitydistributioninvertedfrom23 000 (wide + narrow)-band reddening measurements. The Sun is at (0,0). Units are parsecs, red is very lowdensity and violet indicatesdense areas. External areaswith ahomogenous color correspond to the absenceofconstraining(distance-andlongitude-distributed)targetstars.Thelocalcavityatthecenterissurroundedbywell-known cloud complexes and other cavities, including a huge region devoid of dust in the third quadrant. The


154

accuracyonclouddistancesisontheorderof5pcto1kpc,buttheangularresolutionisofthesameorder,i.e.verypoor.(FigurefromLallementetal2014).

3Dmapspublishedtodatearebasedonvariousinversiontechniquesthathavebeenappliedtoabsorptionorextinctiondata.A full three-dimensionalBayesianmethodadaptedto individualmeasurementsofgascolumns(orextinctions)hasbeendevelopedbyVergelyetal.(2001)andrecentlyappliedtoreddeningmeasurementsindividuallyrecordedtowards23000nearbystarstoproducelocalmaps(Lallementetal.,2014,Figure70).Marshalletal.(2006)hasproducedthefirstlargescalereddeningmapbyadjustingamodeledGalacticdistributionofstarsanda3Ddustdistributionto2MASSphotometricdata.Duringthelasttwoyearsthesituationhasstartedtoevolveveryrapidlythankstothemassivespectroscopicandphotometricstellarsurveysthatarecomingonline.Unprecedentednumbersofdistance-limiteddatahavebeenmadeavailablethatcanbeusedtoproduce3Dmapsofdustorgas(forexample,seeSaleetal.2014forIPHAS-basedmaps; Zasowski et al. 2015 for SDSS/APOGEE DIB-basedmaps; Green et al. (2015) forPanstarrs-1 reddening-based maps, Figure 71; the special issue led by Monreal-Ibero &Lallementon3Dmaps,Mem.Dell.Soc.Ital.,2015).

Itisclearthatthissituationisabouttochangedramaticallywhenallcurrentandfuturemappingefforts start tobenefit from theESAGaiamission.Asdiscussed inChapter 1andChapter 2,Gaiawillmeasure thepositionsof all stars down toG~20mag, and create anextraordinarilyprecise three-dimensional map of more than a billion stars throughout the Milky Way. Thisremarkable dataset — once combined with new information on absorption lines, DIBs,reddening and extinction from optical/infrared spectroscopy — has the potential torevolutionize studiesof the ISMand its three-dimensional structure. In all present and futuremaps, accurate distances will replace photometric estimates, and reddening will be moreaccurately determined simply from removing one unknown parameter in Bayesian reddeningcomputationsappliedtothephotometricdata.Mostnotably,theESAGaiamissionwill inturnbenefit immeasurably from detailed Galactic extinction maps that will allow to breakdegeneraciesbetweenreddeningandtemperatureforallfainttargets.Clearly,thekeysynergyforMSEinthestudyoftheGalacticISMiswithGaia.


155

Figure 71: 3D mapping based on reddening measurements. Dust distribution based on 800 millionPanstarrs1/2MASS wide-band photometric data. Shown is reddening towards stars located in three differentdistanceintervalsintheCepheus-Polarregion.Theangularresolutionisexcellent(5arcmin)butradialresolution(distancetoclouds)ispoor.BothtypesofmapswillstronglybenefitfromGaiaparallaxesandadditionalmassivespectraldatawithMSE.(FigurefromGreenetal.2015).

However,buildingveryaccuratemapsoflargesvolumesintheMilkyWayandinparticulartheirassociated dynamics will be possible only once high resolution massive surveys with highmultiplex spectrographs likeMSEwill enter in action. As amatter of fact, statisticalmethodsbased on photometry have limitations in spatial/angular resolution because they require astatistically significant number of targets per unit volume, they suffer from biases associatedwith the absence or insufficient knowledge of the targetmetallicity, or of the reddening lawapplicable to the intervening dust. Most important, they do not provide any kinematicinformation. In contrast,MSE spectraof individual targetseach carryprecious informationontheline-of-sightvelocitystructure,andofferthepossibilityofcouplingwiththeemissionspectra


156

(throughthevelocityidentifications).Spectraonlyprovidediagnosticsofthephysicalpropertiesofthecloudsbymeansoflineratios,and,mostimportantly,correspondtoangularandspatialresolutionthatareonlylimitedbythenumberofavailableindividualtargets.Fromthetechnicalpointofview,newmethodsofinversionwillneedtobedevelopedtoadapttoMSEspectra,thatwillbeordersofmagnitudemorenumerousthanexistinghighresolutionspectraldata.

4.2.1.3 MSEandotherfacilities

Figure72:Mappingpotentialities fromspectroscopyandphotometry:Left: InterstellarNaIdoublet (top)andthe6284Ådiffuseband(bottom)extractedfromaV=14.7targetstaratD=2.8kpc.TheR=48000VLT/UVESspectrum(red) ismodeledwiththeproductofasyntheticstellarspectrum(yellow),an interstellarprofile/DIBmodelwithtwovelocitycomponents (lightblue)andasynthetic telluricabsorption(green).Right:applicationofsuchfittingmethods to radial cloud mapping based on the strengths of two DIBs (6284Å, 8620Å; top and middle panels,respectively) andextinction (A0; bottompanel).Distances arederived from the spectroscopic stellarparametersandphotometricdata.DIBsandextinctionvary ina similarwayand trace the localArmandPerseus. Individualspectral data show a stronger variability that may correspond to different physical properties of the clouds orsmall-scalestructure, illustratingthepotentialofhighresolutionspectroscopycomparedtophotometry.(FiguresfromPuspitarinietal.2015)

Building3Dmapsof the ISM in thepost-Gaia era requires spectroscopy for a vast numberoftargets, ideallydistributedovera rangeofdistances, inasmanydirectionsaspossible,andatspectral resolutions and SNR levels that are high enough to enable the detection andmeasurement of key ISM absorption features. Clearly, this is not feasible using single-objectspectrographsastheobservationswouldbefartootimeconsuming.Itwouldalsobeimpossibleusing small-fieldmulti-object spectrographs, sinceobservationsneed tobe carriedout over asignificant fraction of the sky. In short, a high-resolution mapping of the three-dimensional


157

structureofgasanddust in theMilkyWaywill requirea telescope/spectrographcombinationthathashighthroughput,wide-field,ahighlevelofmultiplexing,andhighresolutions.MSEwillliterallyproduceaparadigmshiftforthisfield.

3Dmappingof the ISMbasedonspectroscopicGaia target stars ismostefficaciouswhen thenumberofsightlines ismaximized.This inturndependsonthenumberoffibres,thefieldandsurvey size, and the telescope/instrument efficiency (i.e., system throughput). High spectralresolution(R≥20000)isapre-requisiteforISMstudies:highresolutiondisentanglesISMlinesorbandsfromstellarfeatures,allowkinematicanalyses,andstronglydecreasesthelowerlimitonthequantityofabsorbingmatterthatcanbedetectedsincemostlinesarenarrow.ExamplesofspectralanalysessimilartowhatwillberoutinelyandmassivelypossiblewithMSEareshowninFigure 72. Lines and diffuse bands extracted from hot or cool star spectra recorded withFLAMES/GIRAFFEorUVESat theESO-VLT trace the line-of-sightcloudstructureat largescale,and profiles agree with those derived from photometric data (in much larger number, asdiscussed above). Figure 72 also shows how at R ≥ 20 000 it is possible to follow howabsorptionvelocitiesvarywithdistance.

The competition to MSE in this field in the immediate future will primarily come fromAAT/HERMES,althoughMSEismuchmoreefficientthankstoitsgreaterwavelengthcoverage(afactorof~1.5),multiplexing (a factorof~8)and throughput (a factorof~6.5).TheGaia/RVS,Subaru/PFS andMayall/DESI have spectral resolutions that are too low by factors of~2 – 5while VLT/MOONS focuses on the 0.8 – 1.8μm spectral region, thereby missing most of thetraditional ISM diagnostics. There is, however, a powerful synergy between MSE andVISTA/4MOST:whileVISTA/4MOST is~5timesslower thanMSE in reaching the requisiteSNRbasedon thedifference in telescopecollectingarea, this is roughly compensatedby its largerfield of view (5 square degrees). The fact that these two facilities are located in differenthemispheresisanimportantconsiderationwhenmappingthelarge-scalestructureoftheISM.

AcompellinganduniqueaspectofthissciencetopicisthatanyspectralobservationofGalacticand extragalactic targets obtained by MSE contains information on the ISM through theabsorption features that are imprinted by interveningmatter and through the reddening (ordifferentialabsorption).TheserendipityofthisprogramthereforeenablesMSEtoobtainaverylargeandusefuldataset for ISMstudies. Inaddition, specificobservationsandsurveyscanbeenvisionedwithMSE tomake unprecedented progresses in themapping of the ISM and ourunderstanding of the relevant physics. SRO-03 describes a Galactic Archaeology program forMSEthathas,asabasiccomponent,theserendipitousstudyofMSEusingmorethanamillionsightlinesspreadacrossthesky.ThisISMsurveywouldbeextremelyefficientsincehigh-qualityspectra(i.e.,SNR~100)couldbeobtainedinjust~15minexposuresforg=16starsofspectraltypesO,BandA.Forreference,asurveyof4950squaredegreescoveringaswathof20degreescenteredontheGalacticplanewouldrequire~1000hrsofMSE,orabouttwoweeksperyearoverthecourseofadecade-longsurvey.Suchobservationscouldbeundertakeninconditionsofveryhighskybrightness.

Inordertounderstandthemulti-phasestructureoftheGalacticISM,MSEwillaccessabsorptionby different tracers to provide the tools to model the physical, dynamical and chemicalpropertiesofthemolecular,diffuseandionizedGalacticISMphases.IneachMSEindividualstar


158

spectrum, interstellar gaseous atomic (NaI, CaII, KI, …) and molecular (CH, CH+, CN, C2, ...)absorption linesanddiffusebands(over400bands intheoptical)canpotentiallybeextractedaccording to thewavelength range and the target type, and in parallel the reddening can bedeterminedfromthespectral/photometricclassificationandthedistance.Aninterstellartracerdoesnotneedtobestrictlyproportional to thecolumndensityof interstellarhydrogen,or tothetotalcolumnofdust,tobeusefulinlocatinginterstellarclouds,sincedistanceassignmentsarebasedonthegradientsoftheabsorbingcolumns.It isforthisreasonthatabsorptionlinesarisingfromanyinterstellarspecies,anddiffusebandsofallkinds,canpotentiallybeusedinacombined analysis (in addition to the reddening). On the other hand, their independentdeterminationswillprovideuniqueinformationoftheinterplaybetweenthephases,theroleofradiation,energeticparticles,magneticfield,etc.Emissiondata(mainlyHI,COandHα)canalsobe used to help phase disentangling. Since absorbing lines contain information on the gasdynamics,bothradialvelocitiesandcolumns(includingvolumedensitiesfrominversion)canbeusedinmodels.Angularvariationsovertheskyofthelinecenterscanalsobeusedtoconstrainthetransversemotions.

Dependingontherangeofcolumndensities,sometracersmaynotbefullyappropriate.Stronginterstellar lines suffer from saturation close to the disk or towards very opaque clouds, andshould thus be replaced by other, weaker, lines for large distances or columns. For nearbytargets,informationcanbeextractedfromthe"classical"andstronglines:NaI(5890–5896Å),CaII(3934–3968Å),CaI(4227Å)forthegasandfromthereddeningforthedust.Thestrongestdiffusebands5780,5797,6284,6196,6614Åarealso themostappropriate.While thestronglinescanstillbeusedabovetheGalacticplaneor indirectionswherecloudsaredistributedinvelocities,weaker linesaremoreappropriatefor largeopticaldepths(toavoidsaturation), forexampleKI(7699Å),andtheweakerdiffusebands(numerousanddistributedabove4300Å).DIBs—whosecarriersarestillunknown,butarestronglysuspectedtobemacro-molecules inthegasphase—alsotracetheamountof interstellarmatteralongagivenlineofsight.Whiletheyhaveneverbeenconsideredseriouslyformappingpurposes,recentworkshowsthattheyareubiquitous,beingpresenteven in thespectraofdistantcoolstars (e.g.,Chenetal.2012),andextractionsoflinesandDIBsarenowabletobeperformedforanystellartype.Thisallowsus to benefit from all targets and build maps at the highest possible spatial resolution. Forreference,we list inTable 6 themajorgaseous linesandthe interstellargasphases theyaretracing. While the ionized and diffuse atomic ISM fraction is best traced in the optical, themolecularanddenseatomicphaseisbesttracedwhenaccesstotheblueisprovided.MolecularcloudsaredetectedwithCH(4300Å),CH+(4230Å),CN(3870Å)andalsometalliclinessuchasTiI(3630Å), FeI (3860Å),MnI, NiI, AlI (3940Å). Alternatively themolecular phase can be probedwiththeinfraredbandsofC2around8780Å.


159

Table6:MajorISMdiagnostics

An illustrative example of the MSE potentialities already discussed above is the multi-componentabsorptionstudyappliedtooneofthefieldsoftheGaia-ESOSpectroscopicSurvey(GES) and shown inFigure 72. VLT/FLAMES (R = 18 000, 48 000) spectrawere adjusted byproductsofastellarsyntheticspectrum,aDIBorlinemodelandatelluricabsorptionmodel,asshownintheleftpanels.Thevelocitystructureandtheabsorptionstrengthsareevolvingwithdistance,tracingthespiralarmsandtheirkinematics,asshownintherightpanels.MSEsurveydata,byfarsuperiorinangularandmagnitude/distancecoverageandusingGaiadistances,willprovidethekinematicsoftheISMwithunprecedenteddetailsalongwithits3Ddistributions.

4.2.1.4 MSEandtheISM:specificstudies

In addition to the large scalemapping of the ISM described in the previous section, focusedstudies in specific regions of the Galaxy with MSE can also provide unique ISM science. Wedescribebelowseveralexamples,cognizantthatotherfieldsofparticularinterestdoexistorwillariseatthetimeofMSEfirstlight:

i. Smallscalestructuresseeninpolarization,theinterfacebetweenstarsandtheISMDetailed information on small-scale (~1”) structures in the ISM is now available from radiopolarization data (e.g., Landecker et al. 1999; Ransom et al. 2010; Wolleben et al. 2010;Wolleben 2007). Someof these structures correspond to shells generated by stellarwinds orsupernovae.Otherstructures trace the interfacesbetweenevolvedstarsorplanetarynebulae(PNe) and the ambient ISM, including “tails” behind fast-moving PNe. Such interfaces, albeitmuch larger, are also seen at other wavelengths, most notably in the infrared (e.g., Spitzer“arcs” inOrion)or in theUV (i.e., the spectacular caseofMira’s tail;Wareinget al. 2007). Inmany(andperhapsmost)cases,theoriginsofthedetectedpolarizationsignaturesarefarfromclear.Thelackofinformationonthedistanceandtheextentoftheseobjectsmakestheirtruenatureobscureandundermineseffortstoconstructquantitativemodels.

Observations withMSE are ideal for advancing this area of study. Dedicated observations oftarget stars located in the field and associated absorption measurements would yield more

Feature Wavelength Probe Feature Wavelength Probe5890 CH 43005896 CH+ 42303934 CN 38703968 C2 8780

CaI 4227 32305780 32405797 33806284 TiI 36306196 FeI 38606614 MnI 3940?

NaI 3300 NiI 3940?KI 7699 AlI 3940?

Molecular

TiII

Metallic

NaI

Strong

CaII

DIB

Faint


160

accurate distances to the radio polarization sources, and thus provide accurate physicaldimensions.Moreover, theabilityof thenewobservations todetect variations in thevelocitystructure at very small angular scales would allow for direct comparison to hydrodynamicsimulationsofthestar–ISMinteraction,andprovidecriticalinformationfortheconstructionofrotationmeasuremodels.Thesemodels,inparticular,wouldgiveimportantinformationonthemagnetic-fieldstructureintheinteractionandtailregions,andcouldleadtotheconstructionofbettermagneto-hydrodynamicmodelsof the star – ISM interaction.Moreover, emission linesmight also be detected and their variability studied as a function of the location in the bow-shockandtailstructures,whenpresent.Multi-fibrespectrawouldbeparticularlyadaptableforsuchobservations.

ii. SupernovaRemnantsIncreasinglydetailedradio,opticalandX-rayobservationsofsupernovaremnantshaveraisedanumberof importantquestions.Whyaretheresuchlargedifferences intheratiosbetweenX-rayandradioemissionsforsupernovaeofthesametypeandwithcomparableradioemissions?Why are there radiative recombination continua (interpreted as freely expanding andrecombininggas)inX-rayspectraofmixed-morphologysupernovaremnantsthatareknowntobe interacting with molecular gas (e.g., Miceli 2010)? The nature of the remnant, and thedetailed distribution of the ISM surrounding it, are key parameters in understanding theseinteractions which play amajor role in galactic ISM recycling. Observations of the field starsaround supernova remnantswithMSEwill allow themeasurement of absorption or emissionfeaturesthatholdcluestotheISMstructure,enablecorrelationstudies,andultimatelyleadtomodelsthatdescribetheejectaexpansionandshockproperties.

iii. SmallscalestructureofdiffusemolecularcloundsandtheCH+problemWhether or not diffuse interstellar clouds are “clumpy” remains a hotly debated issue. Inparticular,anystructureinthespatialdistributionofthemajormolecule,H2,isverydifficulttoidentify since the column density of this species can bemeasured only in the far UV, whichrequiresspectralobservationsofbright,backgroundstarswithOorBspectraltypes.ThedegreeofH2 clumpiness is critically important formodeling theabundanceswithin these cloudsas itaffects the penetration of UV photons and thus photo-destruction processes. Because H2 iscloselycorrelatedwithCH(Federman1982),observationsofthebluelinesofthisradicalcanbeusedasasurrogateforH2.Byselectingappropriatecloudsatintermediatelatitudeswheretheconfusionisminimalandthesurfacedensityofbackgroundstars isstillhigh,onecanmapthedistributionofCHandinferthatofH2.Theuseofbackgroundstarsmakeitpossibletoprobethespatialstructureoverasurprisinglybroadrangeofscales,withadynamicrangeofabout3000.While the largest separations (~1.5degrees or~3 pc for a cloud at 100 pc) can be used todelineate the overall geometry of the cloud and its boundaries, the smallest ones (a fewarcseconds,orabout0.001pc)wouldprobethestructureatverysmallscales.MSEspectra takenata resolutionofR≥20000wouldbesufficient to reach theseobjectivesbecause, apart from the main transition at 4300Å, several additional weaker features areavailable around 3890Å, allowing the measurement of line opacities and, hence, CH columndensitiesevenif lineprofilesareunresolved.Suchobservationswillprovidesimultaneous,and


161

invaluable,informationonthe“CH+problem”.ThehighabundanceofCH+intheISMispresentlynotunderstood:thereactionpresentsanenergybarrierof4600KandisinsufficienttoproducetheobservedamountsofCH+intheconditionsprevailingindiffusemoleculargas.Similarly,thelargeabundanceofH2inJ>2rotationalstatesrequiresenergeticprocessesthathaveyettobeidentified and which can play an important role in the physics and chemistry of diffusemoleculargas.CH+isthusanimportantspeciesinthatitcanbeusedtoinvestigatethenatureofnon-thermal processes that may play a key role in the physics of interstellar clouds. Severalscenariosinvolvingeithershocks(PineaudesForêtsetal.1986),vortices(Godardetal.2009)orcloud/intercloud interfaces have been suggested as the additional energy source required toovercometheCH+formationenergybarrier.Allthesescenariosimplythepresenceoflocalizedregions heated to higher temperatures but with very different geometries (i.e., shocks andinterfacesareessentially2-Dstructures)whileinthescenarioinvolvingturbulence,vorticesaredistributedoverthewholecloudvolume.Thus,adetailedstudyofthespatialdistributionofCH+with a facility like MSE will allow astronomers to identify conclusively the process at work.Furthermore,inthissamespectralrange,linesfromCNarealsopresent,andthesecanbeusedtoprovideadditionalconstraintsoncloudmodels.iv. Thecarriersofthediffuseinterstellarbands(DIBs)

Through detailed mapping and ISM-phase assignment of the hundreds of diffuse bandsdetectableintheoptical,MSEwillbringnewconstraintsonthespeciesthatareresponsiblefortheDIBsandtheirconditionsofformation.Specifically,a largenumberoftheabsorption linesbetween4200Åand9000ÅareDIBs.SomeDIBsareverytightlycorrelatedwiththecolorexcess,somewiththeatomicgas,andsomewiththemoleculargas.Forthemostpart,DIBsarebroaderthan gaseous lines, but they contain important kinematic information as they are Doppler-shiftedaccordingtotheradialvelocityoftheircarriers.Theycanalsobequitenumerous:opticalspectrahavebeenshowntocontainasmanyas400DIBs(see,e.g.,Hobbsetal.2009;Friedmanet al. 2011). Depending on their relationship to the extinction or the gas columns, theirdetectioncanrequirehighSNRand/orhighlyextinctedstars.

At present, no definitive identification exists for any of the carriers, despite decades of study(althoughlargegaseouscarbonmoleculesarestronglyfavored).Mostobservationshavetendedtofocusonmeasuringcorrelationsbetweenspecificbandsandgasordustcolumns.Whilesomefamiliesthatsharesomepropertieshavebeenfound,thereisastrongvariabilityinthedegreesof correlation. Recent studies have shown that the strength of someDIBs is governedby thestellar/interstellar radiation field, reflecting their degree of ionization. The cloud history andphysicalstate(i.e.,shocks,ionization,cooling,condensation,shieldingagainstUVphotons,etc.)must play a role not only in the cloud chemistry and molecular content, but also in theformationofDIBs.

Buildingonthesestudies,MSEwouldofferanewwaytoapproachthelongstandingproblemofthe identificationofDIBs, aswell as characterizing thepropertiesof the clouds towhich theybelong.DeterminingthedetailedspatialdistributionoftheDIBcarrierswithinagivencloudathigh spatial resolution, through spectroscopic measurements of multiple targets, will giveinformation on how those carriers evolve within the cloud. Although challenging, suchobservationsmayalso revealDIBs thatoriginate fromthesametypesofmoleculessince they


162

should display the same spatial structure. No such studies have yet been attempted becausethey require not just DIB measurements but also absorption data (needed to constrain thephysicalandchemical structureof the ISM).MSEwill thereforebe the ideal facility for suchastudy. Targets should in this case include cool, evolved stars in order to maximize spatialresolution.

4.2.2 TheIntergalacticMedium

ScienceReferenceObservation10(AppendixJ)

LinkinggalaxyevolutionwiththeIGMthroughtomographicmapping

Weproposetosimultaneouslyusehigh-z(z>2.5)Quasi-StellarObjects(QSOs)andbrightgalaxysight-linestoprobetheLyman-αforestandmetalcontentoftheIGMatz~ 2-2.5,andtargetphotometrically selected faint galaxies within~1Mpc of each sight-line to directly identifysourcesassociatedwiththeIGMstructures.Moderateresolution(R~5000),deepspectrawouldbeobtained forall sources to targetwavelengths fromLy-α to [OIII], including interveningUVabsorptionlines,atz~ 2–2.5.ThiswillpushthecapabilitiesofMSE,requiringgoodsensitivityand moderate spectral resolution from 3600Å to 1.8μm – with the primary instrumentationrequirementofexcellentsensitivityat3600–5400Å.Suchastudywouldallowareconstructionofthegasspatialdistributionandoftheunderlyingdarkmatterdistribution,detailedmodelingof galactic scale outflows via emission lines, investigations of the complex interplay betweenmetals in galaxies and the IGM, and the first comprehensive,moderate resolution analysis oflargesamplesofgalaxiesatthisepoch.TheIGMisthereservoirofbaryonsforgalaxyformation,andathighredshiftcontainsmostofthe baryons in the Universe. Galaxies emit ionizing photons and expel metals and energythroughpowerfulwindswhichdeterminethephysicalstateofthegasintheIGM.Therelevantphysical scale is of order 1 Mpc or less (~2 armins at z~2.5 using standard cosmologicalparameters);onlargerscales,thegasisinthelinearregimeandprobeslargescalestructures.

Deepobservationsofhigh-zquasarandbrightgalaxyspectradisplaynumerousabsorptionlinesblue-wards of the Lyman-α emission from the quasar (e.g., Rauch 1998). Such features areproducedviaabsorptionfromline-of-sightinterveninggasinawarmphotoionisedIGMandtheresulting“Lyman-αforest”isaprimarydiagnosticformappingthedistributionandcompositionoftheIGM(Cauccietal.2008,Leeetal.2014a,b,Cisewskietal.2014).ObservationsoftheLy-αforest have driven numerous investigations of galaxy evolution and cosmology. For example,comparisons of line statistics are a critical test of cosmological simulations: i.e., the IGM’sthermalhistoryandopacity is adirect testof reionizationofbothheliumandhydrogen (e.g.,Becker & Bolton 2013) while the detection of metal species allows us to probe chemicalenrichmentand theextentof thecircumgalacticmedium(CGM;Araciletal.2004,Chenetal.2014).

At high redshift (z > 2), the IGMcontains thebulkof thebaryons in theUniverse.As such, itprovidesthereservoirofgasavailableforanygalaxytoevolve,withIGMaccretionfuellingmassgrowthviastar-formation.Bysimultaneouslyprobing the IGMstructureandcomposition,and


163

galaxydistributions,nebulargasdynamicsandmetallicitywecanbuildacompletepictureoftheinterplay between galaxies and their larger scale surroundings. The spatial distribution of theIGMathigh-zisalsodirectlyrelatedtotheunderlyingdarkmatterdistributionandassuch,byfullymappingtheIGM,wewillallowacompletereconstructionofthematterdensityfield(thecosmicweb)atagivenepoch.

State-of-the-artworkconnectinggalaxiesandabsorbersathighredshifthascomefromtheKeckBaryonicStructureSurvey(e.g.,Steideletal.2010)andsimilarworkdoneattheVLT.Todate,thedistributionof galaxies around a combined sampleof twodozenhigh-z quasars has beenmappedwiththesetwotelescopes(seethesummaryinRudieetal.2012;Crightonetal.2011);Thesequasarspectrahavehigh-resolution(R≥45000)andthegalaxyspectrahaveresolutionR~300 to 1,000. Perhaps surprisingly, it has been found that the flux transmission in the Ly-αforest correlates with Lyman break galaxies up to a scale of 5 Mpc (Adelberger et al. 2003;Crighton et al. 2003). These programs have also revealed the connection between metalabsorptionlinesandtheCGM,againtothesurprisingscaleofmanyhundredsofkpc(Rudieetal.2012;seeFigure 73 forLyαabsorptionwithin±1000kms−1ofthesystemicredshiftof10galaxieswithin aprojecteddistanceof 100kpcof the lineof sight to their backgroundQSOs).Such studies are clearly in their infancy, but are indicators of the powerful combination ofwholesalepairingsofhigh-qualityQSOspectrawithgalaxysurveys.

MSEisapowerfulfacilityforthestudyoftheIGMandthegalaxiesembeddedwithinit.Buildingon the surveys at Keck and the VLT, the main requirement for increasing the number ofabsorbers (for association with the foreground galaxies) is the large number of quasars withhigh-resolution (R ≥ 40000) and high-SNR spectra. In particular, it is important to have widewavelength coverage so that large numbers ofmetal species can be associated with a givenabsorber. In theVLT/UVESandKeck/HIRESarchives,high-resolutionspectra for~400quasarsareavailable.Therefore,acampaignwithMSEtoamassspectraforlargenumbersofgalaxiesinthesefieldswouldyielda10to20foldincreaseinthenumberofsightlinesstudied.SinceBOSSwill observe a large number of quasars brighter than 20 magnitude, it would be possible toidentify fieldswithanexcessofquasars forMSE to target.Atpresent, there isnoplannedorproposedfacilitythatequalsthecapabilitiesofMSEinthisarea.

For larger-scale IGMmapping, a large field of view is required to enable a large number oftargets to be simultaneously observed over a large area. This is difficult for non-surveyinstrumentsorfacilities,andhereMSEhasaclearadvantage.Otherlargespectroscopicsurveyinstruments either lack the sensitivity (e.g., 4m class, VISTA/4MOST, WHT/WEAVE andMayall/DESI) or resolution (e.g., R = 2000 at~4000Å for Subaru/PFS). Subaru/PFS, at its veryupper limits, could attempt an IGM tomography experiment, however, this does not directlyformpartofthePFSextragalacticsciencecase(Takadaetal.,2014),althoughIGMtomographycouldbeundertakenasabi-productoftheproposedPFSgalaxyevolutionsurvey.SuchasurveywillnothavesufficientresolutiontodeterminethemetalcontentoftheIGM,willnot identifythefaintgalaxiesassociatedwiththeIGMstructureat2<z<2.5(withaproposedJ<23.4limit)andwillnothavethesensitivityandwavelengthcoveragetodeterminethestellar-orgas-phasemetallicity of individual galaxies at this epoch. It is thereforeunlikely that this sciencewill besuccessfullyundertakenuntiltheconstructionofMSEandnoplannedinstrumentwillbeabletocompletesuchaprojecttothehighfidelitylevelofMSE.


164

Figure 73: Lyα absorptions within ±1000 km s−1 of the systemic redshift of 10 galaxies located at a projecteddistance less than 100kpc of the line of sight to background QSOs, from the sample by Rudie et al. (2012).Keck/HIRES data are in black, while the red shows Voigt profile decomposition of the HI absorption near theredshift of the galaxy. The continuum and zero level of the spectrum are shown in dashed and dotted lines,respectively.Thesystemic redshiftofeachgalaxy ismarkedby theverticaldashed lineat0kms−1. (Figure fromRudieetal.2012).

The SDSS-III Baryon Oscillation Spectroscopic Survey (BOSS; and its successor the SDSS-IV“ExtendedBOSS”)andMayall/DESIaimtomeasurethebaryonacousticoscillation(BAO)scaleusingtheLy-α forest (amongstothertracers)atz~ 2–3.5 (Delubacetal.2015).The(e)BOSSsurveyinparticularwillcollectspectraof250000quasarsover10000deg2,whileDESIproposestosurvey14000deg2andtargetaboutfivetimesasmanyquasarsathigherresolution(R~3000–4800)withthe4mMayalltelescope(Schlegeletal.2009).Moderatetohighresolutionspectra


165

are not required to accurately measure the ~150 comoving Mpc BAO scale (the trade-offbetween quasar number density and SNR is insteadmore important, e.g.,McQuinn&White2011), and soBOSS andDESIwouldnot represent a significant competitor toMSEbecause itchosetoundertakeaverylargeLy-αforestsurveyatlowresolution.Asmallersampleofhigh-resolutionspectrawillenablecorrelationsinthesmall-scaleIGMstructuretobeprobedbyLy-αabsorptions (as induced by, e.g., ionization or temperature fluctuations during He IIreionization).ThiswouldbeanexcitingareainwhichMSEcouldcapitalize.

SRO-11describesascienceprogramthathasmultipleaimswithinthefieldofIGMstudies,withparticular focus on connecting the distribution and properties of the IGM to those of thegalaxiesembeddedwithinit.BysimultaneouslyprobingtheIGMandgalaxiesat2<z<2.5,SRO-11buildsacompletepictureofboth the IGMandgalaxy,distributionandcompositionat thisepoch.Wewill probe the build-up ofmetals in both environments, andwitness the complexinterplaybetweengalaxiesandtheirsurroundings.WewillperformthefirstdetailedhighSNRandmoderateresolutionanalysisofaextensivesampleofgalaxiesatz>2–increasingsamplesizesbyordersofmagnitude,andfullymapthe IGMdistributiononscales inaccessible toanycurrentorproposedinstrumentotherthanMSE.

WenowdiscussinmoredetailthesekeyareasofMSEIGMscience.

4.2.2.1 TheinterplaybetweengalaxiesandtheIGM

Figure 74:Mean Lyα forest transmissivity in quasar spectra plotted as a function of distance from the nearestLyman Break Galaxy from the sample of Crighton et al. (2011). Power-lawmodel correlation functions are also


166

overlaid.AdecreaseinHItransmissionatseparationsof2–7h−1Mpcisseenthatisdescribedbyapowerlawwithindexγ=−1.(FigurefromCrightonetal.2011).

TheIGMisknowntobethemajorreservoirofbaryonsathigh-redshifts.Whilemassgrowthathigh-z is likely to be increasingly dominated bymergers (e.g., Conselice 2014), recent resultshaveindicatedthatinthemajorityofcasesthisdoeslittleaffectstar-formation(Robothametal.2013,Daviesetal.2015).Therefore,theprimarymechanismthroughwhichtypicalgalaxiesformnew stars is through IGM gas accreting onto their halos. Conversely, star-formation in thesegalaxiesemitsionisingphotons,whichheattheIGManddrivesuperwinds,whichexpelsmetalsout of the galaxy (e.g., Heckman et al. 1993, Ryan-Weber et al. 2009). In general, however,details of the cyclingof gas into, through, andout of galaxies remainspoorly understood.Animportant missing element is an understanding of the gas belonging to the CGM, which isbelieved tobeboth the repositoryof the inflowinggasand the receptacleof the feedbackofenergyandmetalsgeneratedwithinthegalaxy.

Adelberger et al. (2005) used Lyman-Break techniques to select z~3 galaxies in the fields ofquasars. In thisway, theywereable to study thecross-correlationsof LyαandC IVabsorbersseen inquasar spectrawith thegalaxypopulation in the field, finding thatC IV isdetected inmostofthegalaxieshalosuptoafewhundredofkiloparsecs.Theyfurtherfoundthatthecross-correlationfunctionofgalaxiesandCIVsystemsappearstobethesameastheautocorrelationlengthof thegalaxies, consistentwith the idea thatCIV systemsandgalaxies reside in similarparts of the universe. Similarly, they established that Lyα is associated with galaxies up todistanceswithin5–6h-1Mpc.Moresurprisingly,theirobservationshintedatadecreaseofthiscross-correlationatsmalldistanceswithin1h-1Mpc,possiblyduetoenhancedstarformationinthesegalaxiesheatingtheirsurroundingIGM.

Crightonetal. (2011)measured the cross-correlationbetweenH I gas causing the Ly-α forestand1020Lymanbreakgalaxiesatz~3.InagreementwiththeAdelbergeretal.(2003)power-law relation, theymeasured the cross- correlation between C IV absorbers and Lyman breakgalaxies for scales 5 – 15h-1 Mpc. This is shown in Figure 74, where an increase in H Iabsorption compared to the mean absorption level within~5h-1Mpc of a galaxy can beobserved,inagreementwiththeresultsofAdelbergeretal.(2005).Crightonetal.(2011)arguedthattheLyα–Lymanbreakgalaxycross-correlationcanbedescribedbyapowerlawonscaleslarger than 3h-1Mpc. After taking galaxy velocity dispersions into account, the results at <2h-1

Mpc are also in good agreement with the results of Adelberger et al. (2005). There is littleindicationofaregionwithatransmissionspikeabovethemeanIGMvaluethatmightindicatethepresenceofstarformationfeedback.

Morerecently,Rudieetal.(2012)studiedasampleof886star-forminggalaxies,locatedclosetobackgroundquasarsightlines, intherange2.0≤z≤2.8.UsinghighresolutiondatafromKeck,theirprogramaimedtostudythegasproperties inthesesystems,suchascolumndensityandtemperature.FromtheirNHImeasurements,theyfoundthatabsorberswithlogNHI>14.5cm-2are tightly correlated with the locations of galaxies, while the absorbers with lower columndensitiesarecorrelatedwithgalaxypositionsonlyon≥Mpcscales.

Given thediffering results in studiesof the transmissivityof theLy-α forest,anewand largersample of quasar absorbers and galaxy spectra fromMSEwould represent an important new


167

addition to the field. SRO-11 describes a program to obtain moderate-resolution spectra forgalaxiesat2≤z≤2.5, lyingclosetothesightlinesofbrightquasarstofirmlyestablishthelinkbetween foreground galaxies and gas detected in absorption in the quasar spectra. Goodcoverage isneeded forall galaxieswithin theprojectedphysical separationof1Mpc.For highspatial completeness, this will require visiting each field several times in order to cover thegalaxiesaroundthequasars.

SRO-10will tackle the keyquestionof understandinghow the galaxies embedded in the IGMstructureinteractwiththelargescalebaryonicdistribution.ByprobingthedynamicsofnebularmaterialingalaxiesinthevicinityoftheIGMsight-lines,SRO-10willbuildapictureofhowtheyexchangematerialwiththesurroundingenvironment.Detailedanalysisofnebularemissionlineprofiles such as Lyman-α and [OII] (e.g., Verhamme et al. 2008, Weiner et al. 2009) incombination with stellar absorption lines, will allow the identification of galactic scaleinflows/outflowsandanestimateofthemassexchangebetweengalaxiesandtheirimmediatesurroundings(e.g.,Pettinietal2001;Boucheetal.2013).Previously,thelargesamplesofhighsignal-to-noiseandmoderateresolutionspectrarequiredforthisanalysishavebeenconstrainedtoa small numberof sources.As such,ourunderstandingofbothoutflows fromsuper-windsand gas accretion is limited. However, the observing strategy required for IGM tomography(Section4.2.2.2) lends itselfdirectly toacombinedstudyof thegalaxieswithin the IGM.Highsignal-to-noise, moderate resolution and good sensitivity over a wide wavelength range areessentialforbothprojectsandassuchtheycanbecompletedsimultaneously.Withadditionofthe measurements of galaxy metallicities (Section 4.2.2.3), it will also be possible to directlycomparethechemicalcontentofgalaxiesandthesurroundingIGM,probingthepollutionofthecircumgalactic medium and the buildup in metals outside of galaxies. With comparableresolution spectra between the IGMmapping and galaxy studies, we will be able to directlyassociate these galaxieswith absorption features in the line-of-sight galaxy/QSO spectrum toinvestigatethepropertiesofindividualneutralhydrogenabsorbersandtheregionstheyinhabit.

4.2.2.2 Tomographicmappingofthelargescaledistributionoftheintergalacticmedium

By targetingasufficiently largenumberofbright,backgroundsources it ispossible touse theintervening Lyman-α absorption features to fully reconstruct the IGM distribution at a givenepoch using Bayesian inversion techniques – so called, IGM tomography (Caucci et al. 2008,Cisewski,et al. 2014, Lee et al., 2014ab, Ozbek et al. 2016). The IGM distribution is directlylinked to thedarkmatterdistributionandas such, it ispossible to reconstruct the fullmatterdensityfieldonscalesoforderofthemeanseparationoflines-of-sight(e.g.,Pichonetal2001,Cauccietal.,2008).Figure4showssuchatomographicreconstructionofasmallvolumeoftheIGM demonstrating the feasibility and power of this approach, and through SRO-10 thistechniquewillcomeofage.


168

Figure75:Tomographicreconstructionof3DLyαforestabsorptioninCOSMOSfieldusingLRISdatafromLeeetal.(2014a), shown in three redshift segments in 3D (top) and projected over three slices along the R.A. direction(bottompanels).ThecolorscalerepresentsreconstructedLyαforesttransmissionsuchthatnegativevalues(red)correspond to overdensities. Square symbols denote positions of coeval galaxies within the map; error barsindicate the σv~300 km s−1 uncertainty on their redshifts. Pink solid lines indicatewhere three of the skewersprobe thevolume,with insetpanels indicating the corresponding1Dabsorption spectra that contributed to thetomographicreconstruction(FigurefromLeeetal.2014a).

ThepowerofSRO-10for largescalestudiesof the IGMwillbesignificant.TheLy-α forest isapowerfulprobeofstructureformation,buttherearesomeimportant inconsistenciesbetweencurrent, complementary datasets. For example, the amplitude of the power spectrum, σ8,inferredfromtheLy-α forestfluxpowerspectrumissomewhatlargerthanthevalueobtainedfromtheCMB.Thismaybeanindicationthatthereareasyetunidentifiedphysicaleffectsthatinfluencetheindividualmeasurements,suchastheuncertainthermalstateoftheIGM(Vieletal. 2009). A large, independent sample ofmoderate-resolution quasar spectra in the redshiftrange2.1<z<4.5(correspondingtoabout0.38<λobs<0.67μmforLy-αintheobservedframe)would probe the flux power spectrumdown to the Jeans scale and allow significant progressrelativetoexistingstudies.ThesampleofspectraobtainedinSRO-10willallowanexplorationofpossiblesystematiceffectsandoffervaluableinsightswhencomparedtolow-resolutionSDSSdataandthemuchsparser,higher-resolutionspectra(seeRollindeetal.2013).

AnadditionalapplicationoftheLy-αforestobservationsobtainedaspartofSRO-10willbethe


169

measurementof itsopacity,which is sensitive to the ionizationandthermalstateof the IGM,andprovidesanindirectprobeofthesourcesresponsibleformaintainingtheIGMinitshighlyionized,post-hydrogen-reionizationstate.PrecisemeasurementsoftheredshiftevolutionofthemeantransmissionthroughtheLy-αforest,𝐹 ≥ 𝑒!! >≡ 𝑒!!!"" ,where𝜏!"", istheeffectiveoptical depth, provide ameans tomeasure the intensity of theUV background (Bolton et al.2005, Faucher-Giguere et al. 2008). This is an important probe of the ionizing emissivity ofsourcesintheearlyUniverse,andisthereforeakeyconstraintonmodelsoftheHIreionizationera.Inaddition,theepochofsinglyionizedheliumreionizationisthoughttocompletebyz~3,roughlycoincidingwiththepeakinquasarnumbers(Furlanetto&Oh2008).AnincreaseintheIGMtemperatureduetotheassociatedHe IIphoto-heating isobservedatz<4 (Beckeretal.2011,Parisetal.2011),andwillalsoimpactontheobservedLy-αforestopacity.

Figure 76: The Lyα forest effective optical depth plotted against redshift. The filled circles with error bars (1σstatisticalonly,offsetforclarity)displaythemeasurementsofFaucher-Giguereetal. (2008)fromasampleof86ESI/HIRESKeckspectraat2≤z≤4.2.However,otheranalysesfindnoevidenceforthis feature(e.g.,Parisetal.2011). For comparison, simulatedMSE data (at R~5000, SNR = 20) drawn from two different high-resolutionhydrodynamicalsimulationsareshownastheredandblue1σerrorbars.Thespectraaredrawnfrommodelswithdifferent IGMthermalhistories(see inset)whichresult indifferenttemperaturesatmeandensity,T0,andhencedifferent τeff . Each bin contains a total Lyα forest path length of 4000 proper Mpc (corresponding to~200independentquasarspectraforasingle∆z=0.1binatz=3.4),andthestatisticaluncertaintiesshowthestandarderroronthemeanobtainedbyaveragingoverthreeproperMpcchunks.Alarge,independently-selectedsampleofmedium-tohigh-resolutionMSEspectrawouldbewellsuitedforexaminingthepresenceofthisfeature,aswellasprovidingprecisemeasurementsoftheeffectiveopticaldepth.

Measurements of the𝜏!"" redshift evolution are therefore valuable for improving existingconstraints,exploringsystematicuncertainties,andinvestigatingthethermalhistoryoftheIGM.Thelatterhasaneffecton𝜏!"" throughthetemperaturedependenceoftheHIIrecombinationrate,suchthatthenumberdensityofneutralhydrogen𝑛!" ∝ 𝑇!!.!.Atz~ 3.2,a“dip”inthe


170

otherwise monotonic redshift evolution of𝜏!"" has previously been measured from spectraobtained with the SDSS survey (Bernardi et al. 2003) and more recently in ESI/HIRES Keckspectra (Faucher-Giguere et al. 2008). It has been suggested that this featuremay be due toheatingof the IGMduringHeII reionization, although it is verydifficult toexplain this featuretheoretically (Bolton et al. 2009). The large, independently-selected sample of moderate- tohigh-resolutionspectra thatwillbeobtained inSRO-10willenable furtherexaminationof thiskeyissue,particularlythroughtargetedfollowupatevenhigherresolution(seeFigure76).

Asecond,relatedareaismeasuringthetemperatureoftheIGM,whichissensitivetothephoto-heatingof the IGMduringreionization.Directlymeasuringthe IGMthermalstaterequirestherelativelynarrowthermalwidthsofLy-αabsorptionlinestoberesolved:

𝑏!" = 12.9𝑘𝑚𝑠!!𝑇

10!𝐾

!.!

Thisrequireshigh-resolution(R~40000)spectroscopy;lowerresolutionstudiesataroundR~20000stillresolvesthelargerscaleonwhichtheIGMispressure(Jeans)smoothed.This“Jeanssmoothing” scale is setpredominantlybyphotoheatingduring reionization. Inprinciple,whencombinedwithmeasurementsoftheinstantaneoustemperature,constraintsontheIGMJeansscale from quasars can thus yield insights into the timing and duration of the reionisationhistory.Theuncertainthermalhistoryisfurthermoreanimportantsystematicwhenattemptingto measure the instantaneous temperature of the IGM (Becker et al. 2011). The thermalstructureof the IGM isalsoexpected tobepatchyon largescalesdue to the inhomogeneousnatureofHeIIreionisation(McQuinnetal.2009).Studyingtheline-of-sightvariationinthedatawouldthusprovideanovelwaytotestbothHIandHeIIreionizationscenarios.

ThelownumberdensityofsufficientlybrightquasarsthatcanbeobservedatmoderateSNRinthe high resolution mode of MSE would clearly preclude a very large program such as thatdescribedinSRO-10.However,a largenumberofquasarsobtainedinaP.I.-ledprogramcouldstillbeusedveryeffectivelytoprobethepressuresmoothingscaleintheIGMalong—and,forsufficientlyclosequasarpairs,transverseto—thelineofsight.

4.2.2.3 Thebuild-upofmetalsatz>2

ThecampaigntomapthedistributionoftheIGMthroughLyαdescribedinSRO-10andSection4.2.2.2alsoallowsustoprobeitsmetallicity.MetalabsorptionlinesfromthelineofsightIGMare seen in the spectraof high redshift quasars and canbeused tomap thedistribution andevolution of metals in the Universe (Petitjean & Aracil, 2004, Scannapieco et al. 2006). Byobserving systems with sufficient resolution and sensitivity, we can distinguish these metalabsorptionlinesfromthoseoftheLyman-αforestatthesameepochandmapthedistributionofmetalsintheIGM.

WiththehighsignaltonoiseandmoderateresolutionobservationsdiscussedinSection4.2.2.2,wewillobtainhighly robust spectraofa largenumberofhigh redshift sources.Thesespectrawill allow the first detailed analysis of both the stellar- and gas- phase metal content ofindividual galaxies at high-z. Stellar-phase metals of the brighter sources can be obtained


171

through rest-frame UV-continuum stellar absorption lines (using process similar to thatdiscussed inSommarivaetal.2012),while spectralobservationsout to1.8μmwill enable theidentificationoftheemissionlinefeaturesrequiredtodeterminegas-phasemetallicitiesviatheR23diagnostic(e.g.,Perouxetal.2013).

Currentstateoftheartobservationsofgalaxiesatthisepochconsistofeithertensofsourcesobserved with low resolution spectrographs (R~200, e.g., Popesso et al. 2009, Henry et al.2013), which rely on stacking analysis to identify stellar absorption lines and are limited tostatisticalanalysesofthefullpopulation,orsmallnumbersofwellstudiedsourcesatmoderateresolution (R~ 1000 – 3000, Belli et al. 2013, Maier et al. 2014). With the proposed IGMmappingobservationsof SRO-10,wewill target~140000 sources at z >2with the signal-to-noiserequiredtofullyinvestigatestellarabsorption/nebularemissionlinesinindividualgalaxies.Targetspectracouldbefurtherbinnedinresolutionelementstoincreasesignaltonoise,whileretainingsufficientresolutiontoidentifykeyfeaturesrequiredtodeterminestellarmetallicities.In combination with deep photometric data to derive stellar masses (for example fromSubaru/HSC and LSST), we will probe the mass – metallicity relation (e.g., Lara-Lopez et al.2013a) for a large sample of individual galaxies. Though such a studywewould, for the firsttime,produceadetailedlargestatisticalstudyofthebuildupofmetalsatz>2andwitnesstheformationofthemass–metallicityrelationinarobustsampleofgalaxies.

4.2.2.4 Galaxy-scaleabsorbers:DampedLyαsystems(DLAs)

Figure77:CumulativecosmologicalmassdensityofneutralgasinDLAsasafunctionofmaximumcolumndensityfrom the study by Noterdaeme et al. 2009 . The apparent flattening of the curve at logN(H i) � 21.7 impliesconvergence.(FigurefromNoterdaemeetal.2009).

DampedLyαsystems (DLAs)areobservableat thehighesthydrogencolumndensitiesandare


172

expectedtotracegalaxies.ThelargesthaulofDLAstodatecomesfromtheSDSS(Noterdaemeet al. 2012), permitting an estimate of the cosmic density of neutral gaswhich is thought totrace theevolutionofneutral gas reservoirs for star formation (seeFigure 77).However, tomovebeyondthepurestatisticsofDLAnumbersandHIgascontent,highresolutionfollow-upwithspectrographson8–10m-classVLT/Kecktelescopesisrequired.Thisisbecausethemetallinesused todetermine themetallicities andelement ratios areweak, narrowand cannotbeeasily detectedwith resolutions of a few thousand. Therefore, although some 1000DLAs areknown, detailed properties such as chemical abundances, kinematics and element ratios areknown for just~10%of these (Zafar et al. 2015).MSEwould provide a sample ofmoderate-resolution spectra that would allow astronomers to increase dramatically the number ofmetallicityestimatesforlargesamplesofDLAs.

MetallicitystudiesofDLAshavemanydifferentapplicationsinastrophysics.Theratioofalpha-to-ironpeakelementsisknowntodependonstarformationhistory,andverymetal-poorDLAsoffer an alternative to extreme metal-poor stars as diagnostics of early nucleosynthesis.Similarly, theprimordialabundanceofdeuteriumconstrainsBigBangnucleosynthesismodels,while studiesof specificelementsprovide independentconstraintsonnucleosyntheticorigins.Finally,finestructurelinesprovideindependentinformationonstarformationrates,andallowtestsforpossibletimevariationsinphysicalconstants.Forthesereasons,theefficientstudyofeverlargerDLAsamplesislikelytobeapriorityinthecomingdecade,and,aswenowexplain,MSEcanplayaleadingroleinsuchefforts.

Although the last two decades has seen a steady increase in the use of DLAs as probes ofchemical abundances at high redshift, the very largenumberofDLAs thatwouldbe availablefromMSE surveyswould allow rare examplesof extreme-metallicityDLAs tobe identified.Atlow-metallicities, extremely metal-poor DLAs would provide an excellent complement to theabundancestudiesbasedonmetal-poorGalactichalostars.ThemostmetalpoorDLAsalsoofferexcellentindependenttestsoflow-metallicityyields(e.g.,Akermanetal.2005).Thediscoveryofacarbon-enhancedmetal-poorDLA(Cookeetal.2011)iscurrentlytheonlyoneofitskind,butpotentiallyrepresentsthemissinglinkbetweenPopulationIIIenrichmentandstandardchemicalevolutionmodels.Attheotherendofthescale,extremelymetal-enrichedDLAsrepresenttherare cases of galaxies that have self-enriched to close to solar metallicity already by z~2(Kaplanetal.2010).

For high-resolution observations at the bluest wavelengths, there would be further excitingopportunities.Mostnotably,atR≥20000,it ispossibletosearchformoleculargas,whichhasbeen detected in only a handful of DLAs to date (Noterdaeme et al. 2008). MSE can detecthundredsofmolecule-bearing (mostlyH2) galaxies, and there aremultiple scientific dividendsfrom such detections. Obviously, this would be the first truly statistical census of H2 at highredshift.MeasurementofthevariousJlevelpopulationsalsoprovidesoneofthefewwaysthatthe internal temperatures and densities can bemeasured (Noterdaeme et al. 2010) and thedetection of CO gives a measurement of the CMB temperature (Srianand et al. 2008,Noterdaemeetal.2011).Perhapsofgreatestphysicalimportance,follow-upobservationsofthemolecule-bearingDLAscouldbeusedtosearchfortimevariationinthefundamentalconstants(e.g.,Murphyetal.2008).Todate,suchresearchhasreliedonjustafewindividualabsorbers.


173

ThebestscienceresultsforMSEwillrequireacarefulmatchingofthewavelengthssothatthemetalsandHIarecovered for thesameredshift range.Forexample,anexcellent line for thiswork is SiII 1808Å: it is relatively strong, but usually not saturated, and Si is relativelyunrefractory (comparedwith iron, for example). For absorbers in the range1.8 ≤ z ≤ 3.5, thebluelow-resolutioncoverage(forLyα)wouldneedtobeapproximately3400–5500Å.Thehighresolutioncoveragewouldneedtobe~5000–8000Å.Thistypeofstudyisfurthermotivationforgoodbluesensitivity(inadditiontostudiesofH2).

Inordertocalculatetheapproximateabsorptionlinenumbersthatwillbecoveredinasurveycovering~10000deg2,weadoptatypicalquasarcolourofg–i=0.2.Thebrightsurveylimitisthereforeapproximately i=19.5.Combinedwiththequasar luminosityfunction,thisaredshiftpath ishenceΔz~1persquaredegreeforasurveywith i<19.5.Atredshiftsofz~2–3,thenumberdensity(i.e,thenumberperunitredshiftpath)ofDLAsisn(z)~0.2.So,foranareaof10000deg2,wecanexpect10000x1x0.2=2000DLAs.

WiththenewlyavailablelargenumberofspectraandimagesfromSDSSinthelastfewyears,ithas been possible to undertake a “statistical” approach to some aspects of quasar absorbersstudies, including stacking of spectra with low SNR. Examples of this include looking at thelensingeffect (Menard&Peroux2003),dust content (Murphy&Liske2004;Yorket al. 2006;Frank&Peroux2010;Khareetal.2012;Noterdaemeetal.2015)incompositespectra,stackingimageswithabsorbers(Zibettietal.2005)orstatisticallycorrelatinggalaxies-absorbers(Boucheetal.2004;Gauthieretal.2009;Lundgrenetal.2009).Withmuchenlargedsamplesofthiskind,onewouldbeabletodividesurveysintodatasubsetsthatcouldbeusedtostudysomespecificaspectsinmoredetails(e.g.,influenceofmetallicityonDLAdustcontent).Studiessuchasthesehave laid the foundation for our understanding of the properties of gas in distant galaxies.However,our currentunderstandingof the ISMathigh redshifthas clearly reached thepointwhere it is limitedby statistics, and this iswhereMSE canmakeanenormous impacton thisfield.

TheuniquecontributionsofMSEtothisfieldaretwo-fold.First, itwouldbethefirstfacilitytoprovidesurveylevelstatisticsathighspectralresolution.Forinstance,abright-timesurveyof10000deg2atthehighestresolutionswouldyieldSNR~ 20spectra(in1hr)atalimitofg~20.4.However, in order to convert the measured metal line strengths into useful chemicalabundances, the HI content of the absorber must also be measured. The Ly-α line isconsiderablybluer (at1216Å) thanmostof themetal lines, andwithequivalentwidths (EWs)typicallyinexcessof10Åcanbemeasuredinmuchlowerresolutionspectra.ThesecondbenefitoftheMSEismadethroughprogramslikeSRO-10,atlow/moderateresolution,tomeasureHIinDLAs over a redshift range of z ≥ 2. The combination of coverage in both the blue (at lowerresolution)andatredwavelengthsathigherresolutionwouldyieldanextraordinarygaininthestatisticsofhigh-redshiftDLAs.

4.3 Timevariableextragalacticastrophysics4.3.1 Reverberationmappingofblackholesandtheinnerparsecofquasars

ScienceReferenceObservation11(AppendixK)


174

MappingtheInnerParsecofQuasarswithMSE

The centreof everymassivegalaxy in the localUniversehosts a supermassiveblackhole thatlikelygrewbetweenaredshiftof1to3throughactiveaccretionasaluminousquasar.Despitedecades of study, the details of the structure and kinematics of the inner parsec of quasarsremain elusive. Because of its small angular size, this region is only accessible through time-domainastrophysics.Thepowerfultechniqueofreverberationmappingtakesadvantageofthechanging emission-line properties of gas near the black hole in response to variations in theluminosity of the black hole's accretion disk to measure the sizes and velocities of the line-emitting regions; with this information, we can map the quasar inner parsec and accuratelymeasureblackholemasses.Thisinformationisessentialforunderstandingaccretionphysicsandtracing black hole growth over cosmic time; reverberation mapping is the only distance-independentmethodofmeasuringblackholemassesapplicableatcosmologicaldistances.Weproposeaground-breakingMSEcampaignof~100observationsof~5000quasarsoveraperiodofseveralyears(totaling~600hourson-sky)tomaptheinnerparsecofthesequasarsfromtheinnermostbroad-lineregiontothedust-sublimationradius.Withhighqualityspectrophotometryandspectralcoveragefrom360nmto1.8μm,thisunprecedentedreverberation-mappingsurveywillmapthestructureandkinematicsoftheinnerparsecaroundalargesampleofsupermassiveblack holes actively accreting during the peak quasar era. In addition, a well-calibratedreverberation relation for quasars offers promise for constructing a high-z Hubble diagram toconstraintheexpansionhistoryoftheUniverse.

Active Galactic Nuclei (AGNs) and quasars (by which wemean themost luminous broad-lineAGNs) are now thought to lurk at the hearts of all sufficiently massive galaxies. When weobserve AGN, we know that the emission comes from metal-enriched gas surrounding asupermassive black hole. MSE will be the world’s premier wide-field spectroscopic surveytelescope for AGN; a legacy that is ensured by its large collecting area, high throughput andmultiplexing, wide spectral coverage, and overall efficiency for survey operations. We havepreviouslydiscussedAGNandblackholescienceasitrelatestogalaxyevolutioninChapter3,andSRO-08andSRO-09inparticulardetailsambitioussurveystargetingbothgalaxiesandAGNat high SNR.While their importance to galaxy evolution is significant, AGN, quasars and thesupermassive black holes that provide the energy source are fundamentally interestingastrophysicalphenomenaintheirownright,andherewefocusontheuseofMSEtoprobethesupermassive black holes and accretion processes that provide the engine to these energeticphenomena.

At the present epoch, supermassive black holes are ubiquitous in the centres of massivegalaxies.Theblackholesgrewpredominantlyaroundredshiftsfrom1to3,whentheuniversewasapproximatelyafifthtoahalfofitscurrentagethroughactiveaccretionasluminousactivegalacticnuclei(AGN),alsoknownasquasars.Remarkably,subparsecquasaraccretiondiskscanoutshinetheirhostgalaxies–thousandsoftimeslarger–bytwotothreeordersofmagnitude.Despite their small size, black holes are fundamentally linked to their host galaxies as shownthrough the strong scaling relations between theblack holemass andhost galaxy properties.Energy injection during the quasar phase in the form of feedbackmay regulate these scalingrelations; inanycase,supermassiveblackholegrowthclearlyoccursalongsidethebuild-upofstellar mass in galaxies. Measuring accurate black hole masses and understanding the inner


175

structureofdistantquasarsisessentialtoadvancingourknowledgeofsupermassiveblackholegrowth,AGNphysicsandphenomenology,themechanismsfor launchingquasaroutflows,andtheco-evolutionofsupermassiveblackholesandtheirhostgalaxies.

Figure78: Left:An imageof aquasarwhere the surroundingbroad-line region structure is thatof adiskwithaspiral densitywave.Right: Timedelay vs. velocitymap for the Lyα (red), C IV (green), andHe II (blue) emissionlines.Whenthecontinuumemissionfromtheaccretiondiskvaries,thegasgivingrisetothebroademissionlinesresponds similarly after a characteristic time delay. The spread of velocity components in each emission line isgeneratedbydifferent locationswithin thebroad-lineregion,andtherefore the lineasawhole respondswitharangeoftimedelays.WithsufficientSNR,timesampling,velocityresolution,andfluxcalibration,thevelocity-delaymap can be inverted to reconstruct the disk image. (Image credit: K. Horne [star-www.st-andrews.ac.uk/astronomy/research/agn.php])

Though quasars have been studied in radio through X-ray wavelengths for decades, thereremain fundamental, open questions about accretion physics. For example, the well-knownShakura & Sunyaev (1973) prescription for describing the light distribution of the UV-opticalemittingregionofaccretiondisksunderestimatestheirsizesbyfactorsofseveral(Blackburneetal.2011;Jimenez-Vicenteetal.2012;Edelsonetal.2015).Thegeometryandkinematicsoftheregiongeneratingthebroademissionlines–themostprominentfeaturesofquasaroptical-UVspectra – are still poorly constrained. These distant, cosmic powerhouses have such smallangular sizes that they cannot be resolved with existing or near-term technologies; our onlyaccess to constraining their structure empirically is through time-domain astrophysics. Inparticular,wecanusetimeresolutiontosubstituteforangularresolutionbymeasuringtherest-frame time lag of the response of a broad emission line’s flux to changes in the continuumilluminatingthebroad-lineregion(Blandford&McKee1982).

Strong,broad,resonancelinesareseenfromionswithalargerangeofionizationstates,fromOVI toMg II; the gas is photoionizedwith lower-ionization gas at larger radii out to the dust-sublimationradius.Therecombinationtimesinthebroad-lineregionareshort,andsothetimelagbetween thecontinuumandbroad-line fluxvariability corresponds to the light travel timebetween the central, ionizingUVcontinuumsourceand thebroad-line region.Measuring thistime lag thus provides a characteristic distance between the broad-line region gas and thesupermassiveblackhole,Rline.Thebroademissionlines(withwidthsofthousandsofkms−1)alsoreveal theDopplermotionsofdensegasclosetothecentralblackhole. Withacharacteristicdistance(fromthetimelag)andvelocity(fromthelinewidth),theblackholemassisgivenbyMBH = f(∆v)2Rline/GwhereG is the gravitational constant,∆v is ameasure of thewidth of the


176

emissionline,andfisafactoroforderunitywhichaccountsforthegeometryandkinematicsofthebroad-lineregion(see,e.g.,Peterson2011,andreferencestherein).

Furthermore, for an individual quasar, the time lags and the average continuum luminositiesgenerateanRlinevalueforeach line. Comparisonofthecharacteristictimelagsfromdifferentlinesputspowerful constraintson the structure, kinematics, andphysical conditions (e.g., gasdensityandionizationparameter)ofthebroad-lineregiongas(Korista&Goad2004).Inthebestcases,withappropriatevelocityresolutionandtimesampling,high-fidelityvelocity-delaymaps(lineresponsivityasa functionof line-of-sightvelocityandtimedelay;seeFigure 78)canbeobtainedwithareverberation-mapping(RM)campaignwithadurationTdur>6τline,whereτlineisthe time delay for a given line. Such data from multiple broad emission lines with varyingionizationpotentialshave thepower to imageevencomplexstructures in thecentral regions,forexamplespiralarmsinadiskofaccretingmaterial.

Currently,only~60local,relativelylow-luminosityAGNhaveRM-basedmeasurementsoftheirblackholemasses(Bentz&Katz2015),andonly~20%ofthesetargetshavehigh-qualitydatasufficientforproducingvelocity-delaymaps.Thelargestopticalmappingcampaigntodateistheongoing(2yrstodate)SDSS-BOSSprogramtomonitor849i<21.7AGNswiththe2.5-mApachePoint Telescope in a single 7 deg2 field (Shenet al. 2015).Dedicated time-intensiveprogramsthatmonitortherest-frameUVthroughopticalofahandfulofAGNarecurrentlyunderwaywithe.g., HST COS (NGC 5548; P.I. Peterson) and the VLT X-Shooter (P.I. Denney). Though theseprogramswillbefoundationalfortakingthenextstepforwardinRMscience,MSEpromisestogo significantly further. Cosmological redshifting means that optical spectroscopic RMcampaignsofquasarsarefundamentallylimitedbyamismatchbetweenthehighqualitydataatlow-zandwhatisaccessibleforhighz.Thelinessampledinthesetworegimes–e.g.,HβandCIV for low and high z, respectively – arise from significantly different parts of the broad-lineregion,andthereislargescatterintheblackholemassesderivedfromthesetwolinesinsingleepochspectraasaresultoftheirdistinctcharacteristics(Denney2012).Furthermore,localRMAGNoccupyfundamentallydifferentregionsofparameterspacethantypicalhigh-zquasars intermsofluminosity,theshapeoftheirionizingcontinua,andlikelytheirhostgalaxyproperties.Accurate reverberation-mapping black hole mass measurements for a large sample of z~2quasarswouldenablehigh-accuracycalibrationofblackholemassesfromsingle-epochspectraforthefirsttime.

SRO-11 describes a dedicated reverberationmapping campaign usingMSE thatwill provide atransformationaladvanceinthisfieldbyobtaining~100observationsof~5000quasarsoveraperiod of several years (totaling~600 hours on-sky). These observations will map the innerparsec of these quasars from the innermost broad-line region to the dust-sublimation radius.With high quality spectrophotometry and spectral coverage from 360 nm to 1.8 μm, thisunprecedented reverberation-mapping survey will probe the structure and kinematics of theinner parsec around a large sample of supermassive black holes actively accreting during thepeak quasar era. Themajor advantages ofMSEover other plannedRMprograms (e.g., SDSS,OzDES,or4MOST)aresensitivity(~ 2–3magsdeeperin1hr)andnear-IRcapabilitybecauseofthe planned mirror size and instrumentation, and the exquisite Mauna Kea site. The onlyplanned instrument that would be competitive with MSE on these terms is Subaru/PFS;however,thePFSwavelengthrangeonlyextendsto1.25mmandtherearenocurrentplansto


177

invest the requisite time toamapping campaignasoutlined in thisproposal.The inclusionofnear-IRcapability(to1.8mm)toreverberation-maptherest-frameUV-opticalfor~5000high-zquasarswillmakeMSEagame-changerinthisfield(seeFigure79).

Figure79:Theobservedwavelengthofbroademissionlinesof interestoverlaidonarepresentativeatmospherictransmissionspectrumforMaunaKea.Acrossthepeakofthequasarepoch(z=1.3–2.7;boundedbyhorizontallines)emissionlinesfromOVItoHαareaccessiblewithwavelengthcoveragefrom360nmto1.8μm;atleastonehydrogenlinewillbeataregionofhighatmospherictransparency.Thisrangeoflinesprobessizescalesfromtheinnermostbroad-lineregion(oforder light-daysto light-weeks)tothedustsublimationradius(~10times larger;Mor&Netzer2012).

Inadditiontothiscorescience,SRO-11willprovideawell-calibratedreverberationrelationforquasars. This offers the promise for constructing a high-z Hubble diagram to constrain theexpansion history of the Universe. Though the demographics of the quasar population havechanged remarkably since z~ 3, the structure of luminous quasars shows surprisingly littleevolution in fundamentals such asmetallicity and spectral energy distribution. They are thuspromisingobjects for constructingahigh-zHubblediagramgivenanappropriate independentestimateofluminositysuchasawell-calibratedRline–Lcontrelation(Bentzetal.2013).Thesizeofaline-emittingregioncanbemeasuredfromreverberation;thisthenyieldstheaveragequasarluminosity. From themeasured fluxand the redshift, aHubblediagram to z~3 canbemadefromtheproposedMSEhigh-zquasarreverberation-mappingcampaign,constrainpossibletimevariationinthedarkenergyequationofstate,whichwouldchangethedistance-redshiftrelation(Watson et al. 2011, King et al. 2014). Based on SN Ia, the distance-redshift relation isconstrained only out to z~1.7. Of course, considerable work remains to be done over thecomingdecadetoensurethatthislimitcanbereachedinalargefractionofrandomlyselectedquasars.

4.3.2 SupernovaeIaandmassivetransientdatasets

When fully operational in the 2020s, LSST and SKA1will give us unprecedented views of the


178

transient universe. LSST will produce orders of magnitude more transients than all of ourground-basedspectroscopicassetscanhopetofollow-up,withestimatesof105–106transientalertspernight.

Transients are a diverse and largely unexplored area of astronomy. Nevertheless, we alreadyhaveaforetasteoftherichesonthehorizonfromrecentwide-fieldtransientstudiesthatworktoshallowerdepthsthanLSST(e.g.,<16–17mags)suchasPan-STARRs,theAll-SkyAutomatedSurvey for Supernovae (ASAS-SN), Intermediate Palomar Transient Factory (iPTF), and theCatalinaReal-TimeTransientSurvey(CRTS).EssentialtounderstandingtheeventsdiscoveredbyLSSTandon-goingall-skytransientsurveysisspectroscopicfollow-up,especiallyinthecasesofrare or previously unknown transients. The importance of this area to the future of theastronomical communitywashighlightedby the recentNOAOworkshop “Spectroscopy in theEraofLSST”,heldinTucsoninApril2013.

Fortransientobservations,mostofthecurrentfollow-upprogramsareperformedonobject-by-object basis in view of their low number densities and the limited flexibility of multi-objectspectrographs.Amassivetransientfollow-upsurvey,potentiallylinkedwithdeepspectroscopicsurvey projects, is only practical on highly multiplexed, rapidly reconfigurable spectrographswithaccesstolargeinputcatalogues.ItisherethatMSEcandominatetheidentificationoffainttransient or variable astrophysical sources of scientific interest, and these observationsrepresentanextremeobservingmodeforthefacility.Ratherthandirectlytakingadvantageofitsmassivemultiplexingcapabilities,transientstudieswithMSEtakeadvantageofthededicatedspectroscopicnatureofthefacilitytoaccumulatelargedatasetsovertimewithonlyaverysmallfractionofsciencefiberseverdedicatedtotargets.IfNtargetsineveryMSEobservation(thatusesatotalofNfibers)wereobservingrecently identifiedtransients(say,fromLSSTortheSKA),the surveyefficiencydecreaseof (N/Nfibers)%would yield theequivalentofN10-m telescopesdoing nothing but transient spectroscopy with their single object spectrographs. This effortwould be distinct from target-of-opportunity observing that interrupts ongoing observations;suchprogramsarenotwellsuitedtoasurveyfacilitylikeMSE.

A key area of transient science for MSE is in the follow up and characterization of Type Iasupernovae(SNeIa).ThesearewellestablishedcosmologicalprobeswhoseluminositydistancesprovidekeyconstraintsontheDarkEnergyequation-of-state,w.SNeIaaregeometricalprobesthatareindependentoflarge-scalestructuregrowth.This,eveniffutureimprovementsontheDarkEnergyFigure-of-Merit (FoM)fromSNe Iaarenotbeas largeas fromotherprobes (e.g.,weak lensing, BAO, etc.), more data at higher redshifts will be useful for comparing basiccosmological assumptions (particularly if the Universe is not simply described by ΛCDM). Ofcourse, there ismore to cosmology thanmerely improvements to theDark Energy FoM, andhavinglargerandmoredistantSNeIasampleswill,forinstance,allowincreasinglyprecisetestsofisotropy.


179

Figure80:PredictedratesofTypeIaandnon-IasupernovaeasafunctionoftheredshiftdetectedbyLSSTintheDEEPsurveyfield(leftpanel)andthemainsurvey(rightpanel).(FigurefromTheLSSTScienceBookv2).

Improvements on w to the~1% level, using SNe Ia of redshifts z ≤ 0.8, must involve bettercontrol of systematic effects: e.g., improved calibrations andbetter knowledgeof the physicsdriving SN explosions. The analysis of nearby SNe spectra suggests that standardization withspectra alone may be possible. This promises to reduce the intrinsic SnIa dispersions usingaccuratespectralsub-classifications.

While the number of high-redshift SNe Ia (z > 0.8) is still quite small, samples will improvedramaticallywith thenextgenerationof imaging surveys (e.g., Euclid, LSST). LSST is aiming toperform high-cadence observations of several “deep drilling fields” in order to obtain well-sampledlightcurvesfor~50000SNeIaperyear.Follow-upspectroscopyfortheseobjectswillbeessentialforobtainingtheirphysicalproperties.Figure 80showsthepredictednumberofsupernovaeofvarioustypesintheLSSTasafunctionofredshiftinboththedeepdrillingfields(left panel) and themain survey area (right panel). Given their faintness, a wide-field, large-aperturetelescopewithahighly-multiplexedspectrographwillbeessential.Mostofthesedeepdrillingfieldswilllikelybeatequatoriallatitudestofacilitatethisessentialspectroscopicfollowup.Onaverage,eachdeepdrillingfieldwillcontainroughlyathousandSNeIaperyear,sothatasinglepointingwithMSEwillcontain150SNeIaoverthecourseofayear.Furthermore,EuclidwillidentifyhundredsofSNe(Astieretal.2014)forwhichancillaryground-basedfollow-upwillprovide checkson systematics in typingand redshiftdetermination.MSE ispoised tobecomethedefactoSNe-followuptelescopeforLSSTandothersurveys

Additional,excitingtransientscience(i.e.,beyondSNeIacosmology)willcertainlybenefitfromenormous numbers of spectra taken with a wide-field, 10m telescope. Future surveys willdeliverthousandsofcandidates,andefficientspectroscopicfollowupwillbeessentialifwearetobetterunderstandthephysicsdrivingtheseevents.Manyofthosesourceswillberun-of-the-mill AGN, typical SNe, or unremarkable examples of known types of stellar variables. Butvariable-source surveys are already producing exciting science on unusual supernovae (e.g.,Chornocketal.2013),tidaldisruptionsofstarsbysupermassiveblackholes(e.g.,Chornocketal.


180

2014,Holoienetal.2014),extremestellar flares (e.g., Schmidtetal.2014),etc.The transientclassestargetedwillevolvetokeepthesciencecutting-edge.Fromthestart,typicalstellarandquasarvariabilitywillbeignoredsothatthestarsandquasarsbeingtargetedinvarioussurveysdonotdominatethetransientsbeingfollowedup. It is interestingtonotethatLSSTalonewillimage~1500deg2pernightaccessibletoMSE,inwhichareatherewillbe~300newSNeand>75000variablestars(Ridgwayetal.2014).Observingprogramscouldbeestablishedtomonitorsingle-pointing deep spectroscopic fields during their MSE observing windows. MSE willdominateanewregimeoffainttransients(probingveryearlyorverylatestages,aswellaslargedistancesandlowluminosities), includingfasttransients,as itwillbeabletodedicatemoreofitstimetotransientspectroscopythananyother10-mclasstelescope.

4.4 ThedarksectoroftheUniverse

Over the past century, cosmology has evolved from a theoreticallymotivated science to onedrivenmainlybyobservations.Withthediscoveryofthelate-timeaccelerationoftheUniverseexpansion (Riess et al. 1998; Perlmutter et al. 1999)possiblydrivenby anew formof energywithsufficientlynegativepressure,astronomersandphysicistshaveconcludedthat96%oftheenergy density of the universe is in a non baryonic— or “dark”— form. There is a growingrealizationthatunderstandingthesenewcomponentsoftheuniverse(i.e.,thedarkmatteranddarkenergy)mayrequirefundamentallynewphysics.Numerous ideashavebeenadvancedtoexplaintheaccelerationandpredictitsredshiftevolution(e.g.,seethereviewofFriemanetal.2008).

Determining the cosmologicalworldmodel has thus become one of the hottest questions ofcontemporary cosmology. Current and expected technologies (in terms of telescopes andinstrumentation)offer thepossibility toprobeavery large fractionof theobservableuniverse(both in imaging and spectroscopy) to a steadily increasing depth. Indeed, the exponentiallygrowingvolumeofdataisnowthekeyfactorinsupportingtheopportunitytotestnewtheoriesincosmology.Inthe2020sandbeyond,thespecificconstraintofcosmologicalparameterswillbeconductedondedicatedfacilities/experimentsthatareoptimisedtoprovidemeasurementstobetter than1%viawide-areadeepopticalandnear-IR imagingfacilities (e.g.,4MOST,DESI,Euclid, WFIRST, LSST), via combinations of weak lensing (high-resolution imaging), baryonicacoustic oscillations (BAO,mainly using spectroscopic redshifts), and redshift space distortion(RSD)analysis.Intheseareas,thenumberofsourcesratherthanredshiftaccuracyorspectrumcontinuumsignal-to-noiseisgenerallyparamount.

The resulting three-dimensional mapping of galaxies provides the information needed toaccurately determine the galaxy power spectrumand its time evolution, thereby constrainingthe various cosmological parameters. Specifically, the isotropic power-spectrum of any large-scale structure tracer — such as galaxies, quasars or the Ly-α forest absorbers — probesuniversegeometry,theaccelerateduniversalexpansion,thematterandenergycontentoftheuniverse, the law of gravitation on large scales, neutrinomasses, and the level of primordialnon-Gaussianity.Theanisotropicgalaxypower-spectrum iscausedbyaneffectcalledRedshiftSpaceDistortions (RSDs),whichalsoprovideagood test forour theoryofgravity. Inorder toreach sub-percent precision, the accuratemodeling of astrophysical processes that affect thecosmologicalmeasurementsiscapital.


181

As described in the rest of this document, the forte of MSE is in the detailed study ofastrophysics, and it is this that drivesmuch of the design of the facility. As such,MSE has asignificantandimportantroletoplayforotherdedicatedcosmologicalsurveys,intermsofthequantification of the bias of tracers, themeasurement of cosmic variance, the calibration ofphotometric redshifts, and other factors relating to essential calibration. For example, thecalibrationofphotometric redshifts for theEuclid spacemission requiresdeep,ground-based,spectroscopicobservationsof105galaxiestoadepthofRIZAB=24.5(Euclidphotometricband),andthesegalaxiesmustberepresentativeofthesampleofgalaxiesusedforweaklensing.Thisisinadditiontofollowupspectroscopyrequiredtoensurethatthe“purity”ofthesample(i.e.,the fraction of correct redshifts that are measured using on-board slitless spectroscopy) isknown to a value better than 0.1%. Strategies to implement this test using the Euclid deepsurvey require ground-based confirmation; large and complete spectroscopic surveys will beparticularlyusefulforminimizingthisimportantsystematicuncertainty(seeFigure81)

Figure81:Thepredictedsystematicerrordividedby themarginalisedstatisticalerror forparametersw0andwa,fromEuclidBAOmeasurements(includingthePlanckdata),requiredtonormalisetheBAOscale(solidlines).ThesearecalculatedfordifferentvaluesofthesystematicerrorplacedonthemeanredshiftinabinofwidthΔz=0.1.Therequirement for Euclid is 0.1% (shows by the vertical dotted line) and can be verified with Euclid slitlessspectroscopy using a sample of 140 000 galaxies with purity >99%. Ground based observations of a significantfractionof these samegalaxies fromahighly complete spectroscopic surveywill playan important role. (FigurefromLaureijsetal.2009).

Fromastand-aloneperspectiveaswellasacalibrationone,MSEalsohasthepotentialtobeacriticaltoolinthedevelopmentofourcosmologicalmodel.Itswidefieldofview(toprobelargecosmologicalscales), itshigh levelofmultiplexing(numeroustracerstargetedsimultaneously),itsshortreconfigurationtime(observingefficiency),largeapertureandhighsystemthroughput(to detect faint sources and increase sample size) are all vital elements to address forefrontcosmological questions. But competition is intense, especiallywith the advent ofMayall/DESIandSubaru/PFS.


182

Intherestofthissection,weexplorethepotentialimpactofMSEinthefieldofcosmology.Inkeepingwith the focus of this document on science that is uniquely enabled and compellingbecauseof the capabilitiesofMSE,wedonot include furtherdiscussionofBAOexperiments.BAOisakey(ifnotdriving)componentofmanycurrentprograms incosmology.SuchsurveysareclearlypossiblewithMSE;indeed,thededicatedoperationsofMSEmakesittheideal10mclassfacilityforsuchasurvey.Buttheexpectedsciencegainoverwhatwillbeavailableinthemid2020sishardtojudgeinsuchacompetitivefield.ThecapabilitiesofMSEwillneverthelessbeavailableifresultsfromcurrentlyplannedsurveysrequirefutureBAOsurveystotakeplace.

Asrelatestothetwoelusivecomponents,DarkMatterandDarkEnergy,fortheformerMSEisthe Astronomical Dark Matter Machine. No facility comes close to obtaining a morecomprehensive and complete understandingof theproperties of darkmatter from sub-dwarfgalaxy scales up to the largest cluster scales. For the latter, the capabilities ofMSE offer thepotentialofgoingbeyondthecurrentandplanneddarkenergyexperimentsthroughdedicatedprograms.MSE–whenequippedwith IFUs–uniquelyoffers thepotential for investigationofthe velocity field of stars in distant galaxies, that is essential for advancing RSD analyses andwhich, when combined with weak-lensing shape measurements, will increase significantlycosmologicalconstraintsfromweaklensing(Reyesetal.2010).

4.4.1 Thedynamicsofdarkmatterhalosfromsub-luminoustoclustermassscales4.4.1.1 Identifyingthedarksub-halopopulationthroughprecisionvelocities

Figure82:Theeffectofdarkmatterhalosub-structureson low-massstarstreams. Inasmoothhalo (leftpanel),low-mass stellar streams follow narrow paths on the sky, confined to a narrow range in distance and velocity(shownincolour).Inthepresenceofdarkmattersubstructure(rightpanel),thepotentialbecomesunevenandthestream path and velocity structure become complex.With Gaia, such studies will now finally be within reach.(FigurefromIbataetal.2002).

Cold dark matter cosmology predicts that galaxies contain hundreds of dark matter sub-structures (Klypin et al., 1999) with masses similar to dwarf satellite galaxies. Only a smallfraction of these dark satellites can be identified with the observed population of satellites,however, which raises the question of where the missing satellites are. A large body oftheoreticalworkhasdemonstrated that it ispossible tohide thevastmajorityofdarkmattersatellites by having their baryons expelled during the era of reionization (see, e.g., Kravtsov2010,and references therein).However, it remainsa fundamentalpredictionofΛCDMtheorythat the dark matter clumps exist in large numbers. Possibly the best means to test this


183

prediction is toexamine thedynamical influenceof suchstructuresdirectly innearbygalaxieswherewe possess the richest datasets. Indeed, Ibata et al. (2002) and Johnston et al. (2002)demonstratedthatmassivedarksatellitescanstronglyperturbfragilestructuressuchasstellarstreams. The halo substructures change the host galaxy from a smooth force-field in theabsenceofCMD lumps (lefthandpanelofFigure 82) intoa“choppysea”wherethestreamanditsprogenitoraretossedhitherandthither(righthandpanelofFigure 82).Theeffectofthisisthatthestreampathbecomestwisted,whichleadstodensityvariationsalongthestream(Siegal-Gaskins&Valluri2008,Yoonetal.2011,Carlberg2012),kinksinthestreamtrack(Erkal& Belokurov 2015), and accompanying dynamical heating (Ibata et al. 2002, Carlberg 2009).MeasurementsofthiseffectallowthemassspectrumofthepopulationofM≲108M

¤subhalos

tobedetermined.

The globular cluster stream of highest contrast that is currently known is that of Palomar 5,which is a structure that can be seen directly in SDSS star-countmaps of blue point sources.AnothergoodtargetistheGD-1stream(Grillmair&Dionatos2006),along,thinstreaminthenorthern hemisphere. However, even for thesemost favourable of objects, only a handful ofstarsinthestreamsarebrightenoughtobedetectablebyGaia(theirmainsequenceturnoffslieatg=20.2andg=18.5,respectively).ThusitisveryunlikelythatwithGaiaalonewewillbeabletodetectmanydistanthalostreamsandexaminetheirkinematicsinenoughdetailtodetectthepresence of dark-matter subhalos. Clearly this is also beyond the capabilities of surveysundertakenon4m-classtelescopes.

However,MSEwillallowustosolvethisproblem.Newstellarstreamsmaybefounddirectlyinlarge MSE spectroscopic surveys, or indeed by following up candidate structures found forinstancewithEuclidorLSSTphotometry.AswithPalomar5andGD-1,itislikelythat(post-facto)afewgiant-branchmemberscanbeassociatedtoGaiastarswithpropermotionmeasurements,whichwouldprovideanaccurate(approximate)orbitforthedetectedstream.

SimulationsinthepresenceofalargepopulationofexpectedCDMmini-halosshowthattherewillbelocalisedheatingoftheglobularclusterstreamstarsbytypicallyafactoroftwooverandabovetheintrinsicvelocitydispersionofthestream(whichinthecaseofPalomar5isabout2km/s,Kuzmaetal.2015)andchanges inthemeanvelocityalongthestreamofupto10km/s(Carlbergetal.2015)duetoencounterswiththe largerdarkstructures.Thespatialscaleoverwhich the increased dispersion will be present depends on the physical size of the biggeststructurethathasheatedthestream,butaretypicallylargerthanabout1kpc.

In addition, if ΛCDM is correct, there will also be effects from the hundreds of smaller darkmatterstructuresthatwillmaketheGalacticpotentiallocallyquiteirregular.Withthemini-halomass function adopted in Ibata et al. (2002), these cause a heating of the stream of~5km/s/Gyr, so that the velocity dispersionwill increase approximately uniformlywith distancealongthetidaltail.

Findinglongstreamswithoutsmall-scalevelocitystructurewouldprovideincontrovertibleproofthatdarkmatterissmoothonsmall(sub-galacticscales),whilediscoveringevidenceforheatingbydarkclumpswouldbeavindicationoftheexistenceofthedarkstructuresandatriumphfordarkmattertheory.EitherresultwouldbealegacyforMSE.SRO-04describesanobservational


184

program,perhapsundertakenaspartoftheGalacticArchaeologysurveyofMSE,thatwillrealizethistransformationalgoal.

4.4.1.2 Dwarfgalaxiesasdarkmatterlaboratories

Giventheirproximity,largedynamicalmass-to-lightratiosandearlyformationtimes,theLocalGroup's dwarf galaxies have become attractive targets for investigating the nature of darkmatter.Cosmologymakespredictionsregardingthenumberandinternalstructureofthedarkmatter halos that host dwarf galaxies, while particle physicsmakes predictions regarding theemissionofhigh-energyphotonsfromdarkmatter'snon-gravitationalinteractions.Combinedwith data from X-ray and gamma-ray observatories, MSE's ability to gather large stellar-kinematicsamplesforfaintdwarfgalaxieswillbecrucialfortestingallofthesepredictionsandrevealingthenatureofdarkmatterinbothcosmologicalandparticlephysicscontexts.

4.4.1.2.1 CosmologicalStructureFormation

Figure 83:Darkmatter density profiles andeffects of baryon-physical processes. Left: Logarithmic slopes (α) ofdarkmatterdensityprofiles,measuredatradiusr=500pcandredshiftz=0inobserved(squares;Ohetal.2011)and cosmologically-simulated (crosses) field dwarf galaxies, as a function of stellar mass. The simulations usehydrodynamicsandstarformationprescriptionstostudytheimpactofbaryonphysicsonthestructureofcolddarkmatter (CDM) halos; the solid line indicates the theoretical predictionwithout baryon physics (i.e., CDM only).(FigurefromGovernatoetal.2012).Middleandright:Darkmatterdensityprofilesatz=0forsimulatedGalacticsatellitesinCDM+baryons(solidblacklines)andCDM-only(bluedashedlines)runs.ThemostluminoussatellitesLV < 10

7LV,¤; similar to Fornax) develop shallower central density profiles,while lower-luminosity satellites LV <106LV,¤;similartoDraco)retaincentrallycuspedprofiles.(FigurefromZolotovetal.2012).

Thecurrentparadigmforcosmologicalstructureformationisbuiltonthehypothesisthatdarkmatteris“cold”andcollisionless,enablinghierarchicalassemblyofdarkmatterhalosbeginningon subgalactic scales (White&Rees1978,Blumenthal et al. 1984,Davis et al. 1985,Navarro,Frenk&White1997,Springeletal.2005). Indeed,calculationsofthematterpowerspectrumassociated with popular “weakly interacting massive particle” (WIMP; e.g., neutralinos)candidatesforthedarkmatterparticleimplythatgalaxiesliketheMilkyWayshouldhost~1015satellites in the form of individual, self-bound dark matter subhalos, sub-subhalos, ..., and“microhalos”withmasses~1Earthmass(Hofmannetal.2001,Greenetal.2004,Diemandetal.2005). Yet thesmallestdarkmatterhalos thathavebeen inferred fromobservationshavemasses~ 105M

¤ within their luminous regions (often extrapolated to virial masses Mvir

~ 108M¤), leaving open the possibility of more-complicated dark matter models that would


185

inhibitstructureformationonsmallerscales(Spergeletal.2000,Bodeetal.2001,Dalcantonetal.2001,Gilmoreetal.2007).

Thusforastrophysicalpurposes,darkmatterparticlecandidatesandassociatedcosmologiescangenerallybeclassifiedaccordingtowhetherthedarkmatterparticle'sproperties(e.g.,massandcorresponding free-streaming scale, non-gravitational interactions, etc.) play a role in galaxyformation.Forthestandardcosmology,basedoncollisionless,colddarkmatter(CDM),theydonot:non-gravitationalinteractionsarenegligibleandfree-streamingscalesaretinycomparedtoeventhesmallestgalaxies.Asaresult,theinternalstructureofCDMhalos(i.e.,densityprofile,power spectrum) is similar on all scales (Navarro, Frenk & White 1996, 1997). In contrast,cosmologiesbasedon“warm”darkmatter(WDM)invokelargerfree-streamingscalesthatcaneffectively truncate the matter power spectrum on scales sufficiently large to inhibit theformationofdwarfgalaxies.Moreover,non-gravitationalself-interactions(e.g.,viaYukawa-likepotentials,(Loebetal.2011)canaffectthespatialdistributionofdarkmatterwithinindividualhalos(Vogelsbergeretal.2012,Rochaetal.2013).

Figure 84: Left, center: Spectroscopically-derived estimates of half-light radii and dynamical masses enclosedtherein,fortwochemo-dynamicallyindependentstellarsub-populationsintheFornaxLV~10

7LV,¤)andSculptorLV~106LV,¤)dwarf spheroidals. Overplotted lines indicatecentraland thereforemaximum slopesofmassprofilesthathaveacoreofconstantcentraldensity(γ=3,dotted)andthecusp(γ=2,long-dashed)characteristicofhalosformedinCDM-onlysimulations.Right:posteriorprobabilitydistributionfortheslope.Resultsareconsistentwithcores,whichhaveγ≤3atallradii,butdisfavorcuspsγ≤2atallradii(FigurefromWalker&Penarrubia2011).

Helping to movitave such alternative dark matter models are several apparent discrepanciesbetween CDM-simulated and observed universes, particularly on the smallest galactic scales.Around galaxies with Mvir~ 1012M

¤, cosmological N-body simulations that consider only

gravitational interactions among CDMparticles generally form~10 timesmore subhaloswithMvir~ 108M

¤ thanhavebeendetectedas luminousdwarf-galacticsatellitesofeithertheMilky

WayorM31(Klypinetal.1999,Mooreetal.1999). Moreover,themostobvioussolutionstothis “missing satellites” problem – invoking galaxy formation criteria that suppress starformationinallbutthemostmassivesubhalos–arelimitedbytheinabilityofthemostmassive


186

simulated subhalos to serve as plausible hosts of themost luminous Galactic satellites. Thelatter have dynamical masses (estimated within their half-light radii) that are systematicallysmallerthanthemasses(evaluatedatthesameradii)oftheformer(Boylan-Kolchinetal.2011,2012).This“toobigtofail”problemislikelythesymptomofafurtherdiscrepancybetweentheshapesofsimulatedandobservationally-inferredmass-densityprofiles,ρ(r).CDMhalosformedin N-body simulations have central “cusps” characterized by 𝑙𝑖𝑚!→!𝜌 𝑟 ∝ 𝑟!! , whereasrotationcurvesandstellarkinematicsofdwarfgalaxiesoften favor“cores”ofuniformdensity(left-handpanelofFigure83andFigure84;Mooreetal.1994,Floresetal.1994,Kuzioetal.2008,deBlok2010,Walker&Penarrubia2011).

Do thesediscrepancies falsify theCDMhypothesis, requiring amore-complicateddarkmattermodel? Not necessarily. The relatively recent incorporation of hydrodynamics – along withprescriptions for star formation and energetic feedback therefrom – has demonstrated that“baryonphysics” canaffect the internal structure andevennumberof subhalos in auniverseotherwise dominated by CDM. Given sufficiently vigorous and bursty star formation, strongwinds frommassive stars and supernovae drive outflows that remove ISM rapidly comparedwithlocaldynamicaltimescales(Navarroetal.1996b,Readetal.2005,Governatoetal.2010,Pontzen et al. 2011, 2014). As a result, the orbits of dark matter particles expand non-adiabatically,potentially transformingcentralcusps intocoresand loweringmasseswithinthehalf-light radii of dwarf galaxies (Figure 83). This transformation also lowers the bindingenergiesofsubhalos, leavingthemmorevulnerabletotidaldisruption. SeveralN-body+hydrosimulationsindicatethatthisscenariomightreconcilestandardCDMwithobservations,therebyshiftingfocusfromthenatureofdarkmattertogalaxyformationBrooksetal.2013,Wetzeletal.2016).

However,N-body+hydrosimulationsagreewithback-of-the-envelopecalculationsinpredictingthat baryon-physical processes become inefficient in the least luminous, most dark-matter-dominatedgalaxies,suchthatdwarfgalaxieswithL<106L

¤shouldretaintheirprimordialCDM

cusps(Penarrubiaetal.2012,Garrison-Kimmeletal.2013;Figure83).Oneimplicationisthatquestionsaboutthenatureofdarkmattercanfinallybeseparatedfromconsiderationsofgalaxyformation,providedobservationscanrevealthedensityprofilesofgalaxieswithL<106L

¤.The

current census of Local Group dwarf galaxies includes~40 such systems,~10 of which hostmorethan103starsbrighterthanV ~23(seeChapter2ofthisdocument,alsoFigure 2. Forthesegalaxies,anMSEsurveywilldeliverstellar-kinematicsamplesas largeasthosecurrentlyused to distinguish darkmatter cores from cusps in dwarf galaxieswith L < 106L

¤ (Walker&

Penarrubia 2011). ThusMSEwill have unprecedented power to constrain the nature of darkmatterviaobservationsofdwarf-galacticstructure.


187

4.4.1.2.2 ParticlePhysics

Figure85:Upperlimitsonthecrosssectionfordarkmatterannihilationintofinalstatesofquarks(left)andleptons(right), derived fromanalysisof gamma-raydataand stellar spectroscopyof15Galactic satellites. Dashed linesshowthemedianexpectedsensitivity,whilebandsindicate68%and95%quantiles.Dashedgraylinesindicatethecrosssectionexpectedforathermally-producedweaklyinteractingmassiveparticle(Steigmanetal.2012).(FigurefromAckermanetal.2015;seealsoGeringer-Samethetal.2015).

Thus far, all empirical evidence for dark matter's existence is derived via interpretation ofastronomicalobservationswithintheframeworkofgeneralrelativityand/oritsNewtonianlimit.Confirmation of dark matter's particle nature will require detections of non-gravitationalinteractionswithknown(standardmodel)particles.PossibilitiesincludeproductionattheLargeHadron Collider, “direct” detection of nuclear recoils during rare scattering events inunderground mines, and/or “indirect” detection of high-energy photons released as pairs ofdarkmatterparticlesself-annihilateorindividualparticlesdecay.Duetotheirlargedarkmatterdensities and their lack of the astrophysical backgrounds (e.g., supermassive black holes,scattering of cosmic rays off ISM, etc.) that plague searches near the Galactic center, dwarfgalaxies represent thecleanestavailable targets in searches forannihilationanddecay signals(Gunnetal.1978,Lakeetal.1990).

Thefluxofphotonsreceivedfromannihilationanddecayeventsisproportionaltotheproductof a particle physics factor (annihilation cross-section and decay rate, respectively) and theintegral of the darkmatter density profile raised to the power n, where n is the number ofparticlesparticipatingintheinteraction(n=2forannihilationandn=1fordecay).Thus,givenameasurementofphotonflux,orevenanon-detection,onecanusestellar-kinematicestimatesofthedarkmatterdensityprofiletoinfertherelevantparticlephysicsproperties.Forexample,Figure 85 showsthemostrecentconstraintsondarkmatter'sannihilationcrosssectionasafunctionofparticlemass (Ackermanetal.2015;seealsoGeringer-Samethetal.2015). Theseupper limitsarederivedbycombiningnon-detectionsofgamma-raysfromtheFermi-LATwithdensity profiles estimated from stellar-kinematics of fifteen of theMilkyWay's dwarf-galacticsatellites. For particlemassesMχ < 100GeV, these limits are beginning to rule out the crosssectionthatwouldnaturallygivethecosmologically-requiredΩDM~0.2inthecaseofthermally-producedWIMPs(Steigmanetal.2012).Stellar-kinematicdatahavealsobeenusedtoevaluatethesignificanceofreporteddecaysignals inX-rayobservationsof individualdwarfspheroidals(Loewensteinetal.2010,Boyarskyetal.2010,Jeltemaetal.2015,Ruchayskiyetal.2015).

101 102 103 104

DM Mass (GeV/c2)

10�27

10�26

10�25

10�24

10�23

10�22

10�21

h�vi

(cm

3s�

1)

bb

4-year Pass 7 Limit

6-year Pass 8 Limit

Median Expected

68% Containment

95% Containment

Thermal Relic Cross Section(Steigman et al. 2012)

101 102 103 104

DM Mass (GeV/c2)

10�27

10�26

10�25

10�24

10�23

10�22

10�21

h�vi

(cm

3s�

1)

⌧+⌧�

4-year Pass 7 Limit

6-year Pass 8 Limit

Median Expected

68% Containment

95% Containment

Thermal Relic Cross Section(Steigman et al. 2012)


188

Regardless of whether an unambiguous photon signal is ultimately detected, clearly theresulting inferences about particle properties are only as good as estimates of dark matterdensitiesderivedfromstellarkinematics.ThisdependencehighlightstheimpactMSEwillhaveon efforts to determine darkmatter's particle nature. Among the knowndwarf galaxies, themostattractivetargetsforannihilation/decaysearchesarealsotheleastluminous(LV<103LV,¤),primarilybecausethesehappenalsotobethenearest.However,whilethedynamicalmassesofthese“ultrafaint”systemsarealsoconsistentwithextremelylargedarkmatterdensities(Martinet al. 2007, Simonet al. 2007, 2011), such results areparticularly uncertaindue to thepaltrysamples (N << 100) and poorly-constrained systematics like inflation of stellar velocitydispersions by binary stars (McConnachie&Cote 2010). Improving upon existing samples bymorethananorderofmagnitude(seeFigure35fromSection2.4.1ofthisdocument),MSEwillhaveamajorimpactinthisarea,whichrepresentsoneofthebestopportunitiestoresolvedarkmatter'sparticlenature.

4.4.1.3 Unveilinggalaxyclusterstoz=1

ThemassfunctionofgalaxyclustersanditsredshiftevolutionoffersacriticaltestofDarkEnergyand/or alternative gravity models because it combines a measurement of the cosmologicalvolumeelementwiththeobservedgrowthrateofstructure.Standardlineargravitationaltheorywithinanexpandinguniversepredictsthatoverdensitiesshouldgrowaccordingto

𝛿 + 2𝐻𝛿 − !!Ω!𝐻!𝛿 = 0, 𝛿 = 0.

Importantlythough,atz<1,theuniversemakesacriticaltransitionfromamatter-dominatedtoa Lambda-dominated state. Although gravity drives the development of overdensities, thegrowthofmassivehalosisinhibitedbythefactthatHevolvestoaconstantvalueinaLambda-dominated universe (an effect known as “Hubble drag”). The opposing effects ofmatter anddarkenergyformsthebasisofcosmologicaltestsbasedupongalaxyclusters.

Agalaxyclusterconsistsofadarkmatterhalo(5/6ofthetotalmass),anintra-clustermedium(ICM)consistingofanatmosphereofhot,diffusebaryons(approximately1/6ofthetotalmass)andapopulationofmembergalaxies(approximately1/10thoftheICMmass).Figure86showsavisualization of 6 massive clusters where the galaxies, ICM and “dark mass” are all shown.Galaxyclustersmaybedetectedusingobservational techniquessensitivetoeachmainclustercomponent: weak-lensing is sensitive to the total cluster mass; the Sunyaev-Zeld’ovich (SZ)decrement and X-ray imaging are sensitive to the cluster ICM; galaxy overdensity studies aresensitivetothemembergalaxypopulation.


189

Figure 86: Composite images of 6 galaxy clusterswith X ray images fromChandra (pink) andmass distributionsfrom lensing studies (blue) overlaid on optical HST images. Figure fromhttp://chandra.harvard.edu/photo/2015/dark/. Image credit: X-ray:NASA/CXC/Ecole Polytechnique Federale deLausanne, Switzerland/D.Harvey & NASA/CXC/Durham Univ/R.Massey; Optical & Lensing Map: NASA, ESA, D.Harvey(EcolePolytechniqueFederaledeLausanne,Switzerland)andR.Massey(DurhamUniversity,UK).

MSEispoisedtobeanessentialtoolforcosmologicalstudiesbasedongalaxyclustersthroughhighcompletenessredshiftsurveysthatwillprovidethepurestsampleofclusterswithdetailedinformation on their dynamical state for comparison to SZ, X-ray, and galaxy overdensitysurveys. Indeed, observations of the cluster abundance to z = 1 and beyond are the currentfocusofseveralhigh-profileprojectsutilizingthesetechniques:

• Projectswhichdetectgalaxyclusterstoz=1andbeyondviatheirSZdecrementtowardstheCMBhaveachievedmaturity,detectingsamplesofseveralhundredclustersdrawnfrom observations covering up to 2500 square degrees (Bleem et al. 2014). CurrentupgradestoexistingSZfacilities,e.g.,SPT-3GandACT-3G,willpermitclustersurveystobeundertakenatgreatersensitivityandwithfastermappingspeedsinthetimeframe2016-2020;

• ModeratelydeepX-rayimagingsurveysforgalaxyclusterstoz=1currentlycoverupto50squaredegreesof theskyandgeneratesamplesofseveralhundredgalaxyclusters(e.g.,theXXLclustersurvey;Pierreetal.2016,Pacaudetal.2016).Suchstudieswillbetransformed in September 2017 with the launch of the Russo-German spacemissionSpektr-RGcarrying theeROSITAX-ray telescope.This facilitywill image27000 squaredegreesoftheskyandwilldetectoforder100000galaxyclusterstoz=1.

• Blind weak-lensing surveys to detect galaxy clusters are currently in their infancy.


190

However,weak-lensingstudiesof individualgalaxyclusterscurrentlyprovide themostaccuratemethodofdeterminingmeanmassesofsamplesofgalaxyclusters(Applegateet al. 2014). The application of weak lensing to detect cosmic structures will betransformed by the launch of the Euclid space mission. Euclid aims to measure theevolutionofthemasscorrelationfunctionviaso-called“cosmicshear”(i.e.weaklensingshear induced by mass distributions typically on scales larger than that of individualclusters). As previously discussed, thismissionwill image 20 000 square degrees in asingle, broad optical-NIR filter with a PSF of~0.16 arcseconds. In addition, the NISPinstrument will generate images in three IR bands at lower PSF to determine galaxyphotometricredshiftsandperformlowresolutionslitlessspectroscopyat1–2microns.The Euclid photometric catalogue will also be used to identify galaxy clusters basedupon galaxy overdensities in the pseudo-3D space of sky position and galaxyphotometricredshift(e.g.,Sartorisetal.2016).

Thelandscapeofgalaxyclusterstudiesoverthenexttwodecadeswillthereforebedominatedby large samples of clusters identified using multiple, complementary techniques fromobservations executedovermany thousandsof squaredegrees. The aimof such studies is toobtainpreciseandaccurateknowledgeofourcosmologicalmodel,particularlythedarkenergyequation of state, in addition to compiling a detailed picture of how the mass and physicalhistoryofgalaxyclustersinturnaffectstheevolutionaryhistoryoftheirmembergalaxies.

Therearetwocriticalissuesthataffecttheextenttowhichsuchgalaxyclustersamplescanbeusedtotestcosmologicalmodels:

1) Therelationshipbetweentheobservableusedtoidentifyeachclusteranditstruemass(inthiscasemeasuredwithrespecttoacommonoverdensityscale,e.g.,M500,c).

2) Therelationshipbetweenasampleofclustersidentifiedusingagivenobservationalmethod(e.g., SZ, X-ray, galaxy overdensity) and the true population of clusters existing in theuniverse.

Cluster mass — the key parameter whose evolution is predicted by theory and N-bodysimulations—isaproblematicparameterbecauseitcannotbeobserveddirectly.Traditionally,cluster masses for large samples of clusters could only be inferred statistically, via various“observable”–massscalingrelations.PopularobservablesincludeX-raytemperature/luminosity,SZ decrement, and cluster richness, Ngal. However, there is currently no mass proxy that issimultaneously accurate (unbiased) and precise. Cluster cosmology benefits hugely fromknowingtheaverageclustermassaccurately (for theamplitudeof themass function)andtherelativemassesof clustersprecisely (toget the shapeof themass function). For theabsolutecalibration,weak lensingcurrentlyachievesthehighestaccuracy.Togetrelativemasses,X-raygasmassand/ortemperaturearemoreusefulbecausetheyhaveasmallerintrinsicscatterthanWL.

SowhatcanMSEbringtothis fieldofstudy?Putsimply,redshifts.Spectroscopicredshiftsareaccurateto0.001orbetter.Theycanbeusedtofixindividualgalaxiesinspaceandthusidentifyclusters,theycanbeusedtodetermineclustervelocitydistributions,andtheycanbeemployed


191

to determine the recent star formation history ofmember galaxies viable the observation ofnarrowemissionandabsorptionfeaturesgeneratedwithinthephotospheresofstarsor intheinterstellargas.

Identifying galaxy clusters from large spectroscopic plus photometric data sets generatescatalogues which display the arguably the highest purity and lowest contamination of anyclusteridentificationmethod.Thereasonforthisisthatthebackgroundofnon-clustergalaxiesalong the lineof sight,while large inprojectedor2Dstudies,becomesvery lowwhenspreadalongtheredshiftaxis.

Identifyinggalaxyclusterswithawide-fieldextragalacticMSEsurvey(suchasthatdescribedinSRO-08)overlappingexistingSZ,X-rayorEuclidskyareasoffers thepossibilityofgeneratingaclustercataloguethatcanactasaRosettaStoneforthisfield,inthatitallowstherelativebiasofSZ,X-rayand/oroverdensitycluster identificationmethodstobeunderstood.Forexample,doX-rayselectedclustersrepresentadynamically-relaxedsubsetoftheclusterpopulationaboveagiven mass threshold? Are SZ detected clusters biased toward merging systems? Do weak-lensingidentifiedclustersdisplayanorientationbiaswhichbooststheprojectedmassoftriaxialsystems? Each of these physical questions determines how different cluster identificationmethods view the true cluster population (virialized halos exceeding a givenmass threshold;Anguloetal.2012).

In addition to identifying galaxy groups and clusters to z = 1 and heyond, MSE will alsodetermine the velocity structure of their member galaxies. The observation of thousands ofgalaxy redshifts in themostmassive,nearbyclustersoffersanopportunity tomeasureclustervelocity dispersions to a few percent, translating into a ~10% accuracy inmass. This wouldprovide a highly competitive mass calibration among the various cluster mass proxies. Thestatistical accuracy could be further improved by “stacking” clusters of similar values ofobservables(e.g.,richness,X-raytemperature).

Eveninthenon-idealcaseswhereMSEmeasuresonlyafew10sto100sofgalaxyvelocitiespercluster,valuableinformationonthevelocitydistributionwillstillbeextractedandwillprovideavaluable diagnostic of the relaxation state in each cluster. A simple physical definition of thecluster relaxation state is that it describes the time elapsed since the last cluster-scalemajormerger. Mergersdepositkineticenergy fromeachprogenitor into thevelocitydistributionofthe merged cluster dark matter halo, the ICM and the member galaxies. Each componentdissipatesthisenergyviadifferentmechanisms(dynamicalfriction,gascooling)andondifferenttimescales. Therefore, using MSE spectroscopic redshifts to obtain quantitative measures ofclusterrelaxation(forexample,usingdeparturesfromaGaussianvelocitydistributionaswellasmoresophisticatedanalyses)offersanimportantdiagnosticofthephysicalsourceofscatterinclusterscalingrelations.

MSEwilladdtwoimportantnewdimensionstoourunderstandingofhowgalaxyclusterstracethe underlying mass field: 1) it will provide probably the best (highest purity and lowestcontamination) cluster catalogue from which “wavelength bias” engendered by SZ, X-ray orcolour-overdensity searches can be understood, 2) MSE will add a quantitative measure ofclusterrelaxationstatetoaidourunderstandingofthesourceofphysicalscatterinclustermass


192

observables. Clearly, this field is an excellent example of the synergies of MSE with nextgenerationmulti-wavelengthfacilities.

4.4.2 Peculiarvelocitiesandthegrowthrateofstructures

ScienceReferenceObservation12(AppendixL)

Dynamicsofthedarkandluminouscosmicwebduringthelastthreebillionyears

Weproposeagalaxyredshiftandpeculiarvelocitysurveycoveringupto24,000squaredegrees,i.e.¾ofthefullskyapartfromMilkyWay.Thegoalistomeasureboththecosmologicalredshiftofabout2milliongalaxiesandforasubsettoobtaintheirradialpeculiarvelocities.Thesurveywillmeasureboth early- and late-typegalaxies up to redshift z~ 0.25using the FundamentalPlane and optical Tully-Fisher techniques. The unprecedented depth of 1 Gpc is significantlydeeperthansurveysthatwillbecompletedin2025,suchasthetwoAustraliansouthernsurveys:the optical multi-fiber TAIPAN and the radio SKA Pathfinder WALLABY survey. This field ofresearchiscurrentlyundergoingarevival,andthissurveywillachievethesamegalaxynumberdensity as in the contemporary pioneering proof-of-concept surveys as Cosmicflows-2 and6dFGSv catalogs, while multiplying the covered universe volume by a factor 150. Only thesescalesenablestrongtestofgravitationmodels.

Threemajorsciencegoalsdefinethisprogram:(i)linear-growthrateofcosmicstructuresatlowredshift,obtained throughvelocity-velocity comparisonand the luminosity fluctuationmethod,whicharefreeofcosmicvariance.Thisallowsonetodiscriminatebetweenmodifiedtheoriesofgravityat1%levelandyieldasignificantcomplementtothehigh-redshiftconstraintsprovidedby the coeval spectroscopic surveys achieved e.g., by Subaru/PFS, Mayall/DESI, Euclid, andWFIRST; (ii) galaxy formationwill be investigated in regard to unprecedented velocity-cosmic-web’s6Dphase-space; (iii) dynamical testsprobing the scaleofhomogeneity,and test for theaveraging problem in General relativity and the consequent backreaction conjecture, whichpotentiallyprovidestermsdynamicallyequivalenttobothdarkmatteranddarkenergy.

Thepeculiarvelocitiesofgalaxiesare inducedby thegravitationalpullof themasscontentofthenearbyuniverse,andareadirectindicatorofthedistributionofdarkandluminousmatter.Onlargescales,thesepeculiarmotionsbecomecoherentbulkflowstowardover-denseregionsand away fromunder-dense regions. The velocity bulk flowsdependon the amplitudeof thematter fluctuations and imprint distinct anisotropic features in the distribution of galaxies inredshift space, generally referred to as redshift space distortions (RSD; Peebles 1980; Kaiser1987).Asdependingontheamplitudeofthematterfluctuations,peculiarvelocitiesallowustomeasure the growth-rate of structures, and provide one of the most powerful probes todistinguish between dark energy andmodified gravitymodels (ESA-ESOWG on FundamentalCosmology, 2003; Astro2010 Panel Reports “New Worlds, New Horizons in Astronomy andAstrophysics”,2011).Whethertheobservedlarge-scalebulkflowsareconsistentornotwiththeΛCDM model is still under debate. Hereafter, we use the phrase “RSD” to refer to peculiarvelocityanalyseswherenoancillarydistance information isavailable,and“peculiarvelocities”torefertoanalyseswherebothredshiftanddistancesareknownindependently.


193

RSDs are receiving significant attention as a potential cosmological probe. As a tracer of thegrowth rate of structure, RSDs are in principle capable of distinguishing dark energy frommodified gravity models (e.g., Guzzo et al., 2008, Percival & White 2009). Observationally,several projects carried out during the last decade have used RSDs to constrain cosmologicalparameters: the 2dFGRS (Hawkins et al. 2003), the SDSS LRG (Cabré & Gaztanaga 2009,Samushiaetal.2012),2SLAQ(Rossetal.2007)andtheVVDS(Guzzoetal.2008).Mostrecentlymeasurements have been obtained byWiggleZ (Blake et al. 2012, Contreras et al. 2013) andVIPERS (de la Torre et al. 2013). All major future cosmological spectroscopic surveys nowhighlightRSDsasoneoftheirprobes.

The standardmeasurement of RSD (by the quadrupole anisotropy in the redshift-space two-pointcorrelationfunctionsofgalaxies)isbasedongalaxyredshiftsurveysthatprobethematterdensityfieldtracedbya(single)biasedpopulationofsourcesandthenitmeasurestheratioβ=f/b between the linear growth rate f and the luminosity bias of the tracer, b. Neglectingsystematicerrors,thesample-variancelimitforstandardRSDanalysescanbeapproximatedbyσ ln β ≲ Ab-!.!V-!.!exp(B/b!n), in which n is the expectedmean number density of galaxieswithin the survey volumeV, andAandBnumerical constants (Bianchi et al. 2012).All futurespectroscopic surveys will provide tighter cosmological constraints than previously achievedusingRSDbymeasuringahighdensityofsourcesathighredshift(0.8<z<2)andoveramuchwiderfield-of-viewthancurrentRSDmotivatedsurvey.AMSEredshiftsurveycouldsupersedetheSubaru/PFSresults(1500deg2,4Mgalaxies)andattainsimilarprecisiontoMayall/DESI(14000deg2,25Mgalaxies)andEuclid(15000deg2,30Mgalaxies)ifcollectingmorethan10Mhigh-redshiftgalaxieswith luminositybiasb≥1.3over10000deg2 (seeFigure 87, inwhich twoconfigurations are shown: in the redshift range 0.5 ≤ z ≤ 2, and 1 ≤ z ≤ 2, both with n =0.0003h3Mpc−3, leadingrespectivelyto15Mand12Mspectra).ThisanalysissuggeststhatRSDswithMSEareapromisingareainwhichtodevelopkeyprograms,althoughit isfundamentallynecessaryforMSEtoprovidenewconstraintsthatothersurveyscannotdo;clearlyMayall/DESIandEuclidarepowerfulcompetitorswithRSD.

Figure87:Relativeerroronβasafunctionofthegalaxybiasforcurrentandplanned(orfinalstage)surveys(solidanddashedlines,respectively)andfortwoMSEconfigurationsdescribedinthetext.

However,forallRSDanalysesbasedsolelyonredshifts,theaccurateknowledgeofthebiasisa


194

central requirement. Moreover, the standard RSD technique is further limited by cosmicvariance, especially relevantwhen dealingwith small (local) volumes;multi-tracer techniques(McDonald& Seljak 2009) or higher-order statistics (Marín et al. 2013) are generally used toovercometheseproblems.Instead,owingtothetightrelationbetweenthepeculiarmotionsofgalaxies and the underlying mass density fluctuations that depends upon the backgroundcosmology and the growth rate of fluctuations, if distance information is also availableindependently from redshift, then one can use the full peculiar velocity information to probestructure formationusing the so-called “velocity-velocity comparison”.Although thesekindofmeasurements are limited to low redshift, this opens up tremendous scientific opportunities,definitelycompetitivewithhigh-redshiftmeasurementsaimedatprobingthedepartures fromGeneral Relativity. As described in SRO-12, this programwill require a redshift survey of~2milliongalaxiesforwhichthedistanceswillbe independentlyestimated,usingtheTully-FisherandFundamentalPlanetechniques.Itisimportanttostressthatwhilesomeprogresswiththistypeofobservationcanbemadeusingsinglefibers,SRO-12isbestsuitedtotheMSEIFUmode,anenvisionedearly-lightcapability.

Figure 88: Linear growth rate at low redshift by velocity-velocity comparison. The predictedMSE FishermatrixconstraintsforthesurveydescribedinSRO-12areshownaroundthefiducialΛCDMPlanckcosmology(solidline)atredshiftz<0.05andz<0.1for10000deg2andgalaxynumberdensityn=0.03h3Mpc−3(left)andn=0.003h3Mpc−3(right). Thickand thin (red)errorbars account fordensity-only (RSD)and jointdensity-velocityMSE constraints,obtainedusing the samenumberof galaxies.Dashed (solid) lines correspond to f~Ωγ

mwith γ =0.5, (0.55), 0.6,0.65,0.7downward.

Ahugeadvantageofthe“velocity-velocitycomparison”mentionedabove isthat itcompletelyeliminatescosmicvariance.Indeed,inthecaseofthehypotheticalsituationofperfectdata,justa comparison of a few points is sufficient to determine the relevant parameterswithout anyassumption about the galaxy bias. This comparison is at theheart of peculiar velocity studies(e.g., Davis et al. 2011), andprovides an assessment of gravity anddark energy theories; seeFigure88.AsshownbyCarricketal.(2015),themeasurementofthenormalizedlineargrowth-ratefσ8,basedon~5000peculiarvelocitieswithatypicaldepthofz~0.02,hasaprecisionof5%;with amuch deeper andmore comprehensive peculiar velocity and redshift survey fromMSE,onemightexpecttoreducethisuncertaintybyafactorof~5,yieldingaprecisionof1%.Inadditiontonotbeingaffectedbycosmicvariance,peculiarvelocitysurveysarealsotheideal


195

framework to investigate primordial non-Gaussianities and general-relativistic (gauge) effectsspecificoflargescales(Jeongetal.2012;Villaetal.2014),andtotestthescale-dependenceofthegrowthrateofstructure,afeatureprevalent inmodifiedtheoriesofgravityandclusteringdark-energymodels(Parfreyetal.2011;seealsothereviewbyCliftonetal.2012).

Following Koda et al. (2014) for a Fisher matrix forecast (see also Johnson et al. 2014), themeasurementoffσ8orβfromangle-averagedauto-andcross-powerspectraofgalaxies’densityandpeculiarvelocitycanattainarbitrarilyhighprecision.UnlikeconstraintsfromRSDanalysisbyredshiftsurveys,therelativeerroroftheseparametersmonotonicallydecreaseswiththemeannumberdensityofsources,n.Figure88showshowaminimalMSEvelocitysurveycovering10000 deg2 provides terrific constraints on the growth-rate𝑓 ≈ 𝛺!

! , yielding one of the moststringenttestsofGeneralRelativity(γ≈0.55)atthelevelofδγ~0.05.Whilesimilarprecisionwillbeattainedbyfuturespectroscopicredshiftsurveysathigh-redshiftsuchasSubaru/PFS(notshown)andEuclid(whichwillbenefitfromamuchlargervolumebutprobethegrowth-rateinamorechallengingregime),onlyMSEcanprovidesuchtightconstraintsatlowredshift.Table7reports theFisher-analysis constraints that canbeachievedby SRO-12withn=0.003h3Mpc−3and n = 0.03h3Mpc−3; relative errors can attain the 1% levelwith a deep and large survey. Itillustrates the benefit of the velocity-velocity comparisonmadepossible by the knowledge ofpeculiar velocities (the small, thick error bars in Figure 88) with respect to standard RSDpredictionsobtainedbasedontheknowledgeofthedensityonly(larger,thinerrorbars).

Table7:Fishermatrixconstraintsontherescaledlineargrowth-rate,β=f/b(wherebisthelinearbiasparameter),and on the amplitude ofmatter fluctuations, σ8, fromdensity-only (RSD) and joint density-velocity (VEL) powerspectra.Resultsforskycoverageof10000deg2(marginalizationovertheotherparametersandusingtheWMAP-9referencemodel).ValuesfortheminimalandoptimalsetupforSRO-12areboldfaced.

n=0.003h3Mpc-3 n=0.03h3Mpc-3dβ/β dσ8/σ8 dβ/β dσ8/σ8

z<0.1RSD 0.151 0.137 0.145 0.132VEL 0.065 0.06 0.024 0.024

z<0.20RSD 0.054 0.049 0.051 0.047VEL 0.036 0.032 0.015 0.014

z<0.25RSD 0.038 0.035 0.037 0.033VEL 0.028 0.026 0.013 0.012

z<0.30RSD 0.029 0.026 0.028 0.025VEL 0.023 0.021 0.011 0.01

SRO-12alsoallowstheapplicationoftheluminosityfluctuationmethod.Spatialvariationsintheluminosityfunctionofgalaxies,withluminositiesestimatedfromredshiftsinsteadofdistances,are correlated with the peculiar velocity field. Comparing these variations with the peculiarvelocities inferred fromthe redshift surveyscouldalsobeapowerful testofour fundamentalcosmologicalparadigmoncosmologicalscales.ThemethodhasbeensuccessfullyappliedtotheSDSSDataRelease7(Feixetal.,2015).AnapplicationtoMSEwith its largerskycoverageandlargernumberofgalaxiesshouldyieldresultscompetitivetoRSDanalysis.


196

4.4.3 Probinghomogeneity,isotropyandtheback-reactionconjecture

The cosmological standardmodel relies on the hypothesis that theUniverse is homogeneousand isotropic on large scale, without giving any information about this scale. The theory ofprimordialinflation(whichisincorporatedintotheΛCDMmodel)actuallypredictsacertainlevelofdensity fluctuationsonall scales,withapowerspectrumofscalar fluctuationswhich isnotscale invariant (ns ≠ 1) but colored as confirmed by the recent Planck measurements. As aconsequence, the transition to the large-scale homogeneity and isotropy is expected to begradual, attaining the level δΦ/c2~ 10−5 at the epochof last scattering, as inferred from thecosmic microwave background (CMB) radiation. At redshift z ~ 1100, the large-scalehomogeneity iswellsupportedbythehighdegreeof isotropyoftheCMB(Fixsenetal.1996),underthehypothesisofthecosmologicalCopernicanprinciple.

Atlowredshiftsuchsymmetriesarelessevident.Asfortheisotropy,usinghardX-raydatafromtheHEAO-1A-2experiment,Scharfetal.(2000)demonstratedhowthedipoleanisotropyofthedistantX-rayframe(z~1)canconstraintheamplitudeofbulkmotionsoftheuniverse,probingevidence foramildanisotropy that indeedcouldbeofGalacticorigin. In the localuniverse,arecent analysis of peculiar velocity data (Hoffman et al. 2015) estimated the bulk flowof thelocalflowfieldandshowedthat it isdominatedbythedata(asopposedtotheassumedpriormodel)outto200h−1Mpc,TheyalsoshowedthatitisconsistentintheX–Ysupergalacticplanewith the CMB dipole velocity down to a very few kms-1, as recovered by the Wiener filtervelocityfieldreconstructedfromtheradialvelocitydata.Withavolume150times largerthanthatcurrentlyprobed,thuslargelyencompassingthehomogeneityscale,MSEwillbeuniqueinbeingansweringthequestionofwhetherthecosmologicalbulkflowexistson800Mpcscales,asfoundbysomeclusterstudies.ThesurveydescribedbySRO-12canobtaintherequireddata.

Isotropyitselfisnotsufficienttodeducehomogeneity,unlessprobingitasafunctionoffluxorredshift.Basedonsimplecountingofsourcesinthree-dimensionalregionsinordertoestimatethefractal(Minkowski-Bouligand)dimensionanditsdeparturefromtheEuclideanvalueD2=3,Hogg. et al (2004) assessed the homogeneity scale for the LRG sample from the SDSS at~70h−1Mpc intheredshift range0.2<z<0.4.ThisvaluehasbeenconfirmedbyWiggleZdata(Scrimgeour et al. 2012), proving the transition to the large-scale homogeneity around~70−80h−1Mpcatredshiftz=0.2−0.8,withanupperboundaround300h−1Mpc.Thisvalue isconsistent with fractal analysis, anomalous diffusion and Shannon entropy methods basednumerical-simulation-based studies (e.g., Yadav et al. 2010, Kraljic 2015, Pandey 2013). Veryrecent measurements based on the BOSS DR12 quasar sample (Laurent et al. 2016) set thehomogeneityscalefrom250to1200h−1Mpcatredshiftz=2.2–2.8.

Themost important implicationof inhomogeneity is the long-standing“averagingproblem” inGeneral Relativity (Ellis 1984), accounting for the role of (small-scale) perturbations on theexpansion rate of the spacetime once they enter the nonlinear regime.Whilst in Newtoniancosmology,at the linearorder, theireffectvanishesonaveragebyconstruction, in fullGRthedensity fluctuationspotentiallyshouldhavesomeback-reactioneffectonaveragedquantities.According to the Buchert formalism (Buchert 2000, Ellis & Buchert 2005), they drive neweffective source terms in the Friedmann-like equations accounting for the expansion rate ofeveryspatialdomain,D,overwhichaveragedquantitiesarecalculated.Auniquecosmological


197

applicationofpeculiarvelocity surveys is thedirectmeasurementof the so-calledkinematicalback-reaction term,

€

QD =Var[div v ] − 23 σ

2 QD = Var divv - !!< σ! > , the mean and

variance being calculated over the domain D (here σ2 is the amplitude of the shear velocitytensor). Depending on its sign, this term is dynamically equivalent to dark matter (onsmall/galacticscales)ortodarkenergy(onlarge/cosmologicalscales).Thisquestionisdelicateand far from be settled. In the era of precision cosmology, where the immediate goal isvalidatingordisprovingthecosmologicalstandardmodelandthetheoryofGeneralRelativity,aquantitative measurement of kinematical back-reaction term is mandatory; a non-vanishingvalue would definitely indicate the necessity to revisit the standard FLRWmodel. A peculiarvelocity survey such as that described by SRO-12 provides the key ingredient to probe thisconcept;aslongasthevelocityfieldinredshiftspaceisirrotational,andsomaybederivedfroma velocity potential (e.g., Nusser & Davis 1994), one can directly infer the kinematical back-reaction term by measuring the Minkowski functionals of the velocity fronts defined by theexcursion-setofthevelocitypotential (Buchert&Carfora2008).Becauseofthe limitedsizeofpast and current velocity surveys, this program has not been addressed on real data yet.Instead,thevelocitysurveydescribedinSRO-12,owingtoitslargevolume,isexpectedtoyieldrobustenoughresultstoassesstheback-reactionconjecture.

4.4.4 Cosmicwebdynamicsandgalaxyformation

Figure89:ThematteroverdensityfieldδcomputedfromthedivergencefieldinCosmic-flows-2dataset(Tullyetal.2013) isplotted incolor.Galaxies identifiedby redshift surveysareplottedaswhitedots.Thenewtechniqueofdefiningfilamentsusingcosmicvelocityfieldsgivesanewinsighttothemorphologyofthelargescalestructure.Afilamentofcoherentmotionofmatterishalfthe“Arch”seenrunningfromPerseusPiscestoPavo-IndusstructuresinredshiftsurveysandinthecosmicV-websheartensor.Onlytheanalysisofwatershedsallowsonetoidentifythe“splitting”zoneofthisstructure.(FigurefromTullyetal.2014).

ItisworthreiteratingthatthedrivingsciencethemeforMSEinrelationtogalaxyformationandevolutionisinconnectinggalaxiestothecosmicwebinwhichtheyareembedded(Chapter3).


198

Peculiarvelocitysurveys–withnon-redshift-baseddistanceinformation–provideameanstoprobethelargescalestructure’sdynamicswithMSEthroughthekinematicsofthecosmic-web.SRO-12detailsananalysisofextragalacticflowsbyMSEthatwillgiveunprecedentedinsightsintothedarkmatterroleonalargevarietyofscales,fromplanesofdwarfgalaxies,togroupsofgalaxies,galaxyclusters,filaments,voidsandsuperclusters(seeFigure89foranexamplebasedontheCosmicflows-2dataset;Tullyetal.2014).WithanaccuratemeasurementofpeculiarvelocitieswithMSEsuchasdescribedinSRO-12,onecouldapplythemethoddevisedbyHoffmanetal.(2012)toreconstructthevelocity-basedcosmicweb,orV-web,whosecomponents(sheets,filaments,knots,andvoids)arekinematicallyidentifiedbythelocalvalueoftheeigenvaluesofthevelocitysheartensor.BasedonN-bodysimulations(Figure90),thistechniquehasbeenproventoresolvecosmicstructuresdownto<0.1h−1Mpc,providingmuchmoredetailsthanasimilarclassificationalgorithmbasedonthegravitationaltidaltensorobtainedfromthedensityfield(dubbedthe“T-web”;Hahnetal.2007,Forero-Romeroetal.2009),whichallowsonetoachieveonly~1h−1Mpcresolutionscale.Thehigherresolutioncanbeexplainedbytheslowerevolutionofthevelocityfieldawayfromthelinear-regimethanthedensityfield,whichretainslessmemoryoftheinitialconditions.

Figure90:Cosmic-webcomponents(voids, filaments,sheets,andknots inwhite, lightgray,darkgray,andblack,respectively)obtainedfromthevelocitysheartensor(left,theV-web)andgravitationaltidaltensor(center,theT-web). The latter, reconstructed from the density field (right), allows~10 times lower resolution scale. (FigureadaptedfromHoffmanetal.2012).

Moreover,once the large scale structure (V-web,T-web,density field,…)hasbeenaccuratelydrawn,exploitingthehigh-resolutionspectroscopiccapabilitiesofMSEonecaninvestigatethealignmentofthegalaxyspintothecosmic-webfilaments,eventuallyrelatedtothetidaltorquetheory and its extensions (e.g., Codis et al. 2012; Codis, Pichon & Pogosyan 2015), thesegregationofsinglegalaxiesandgroupswithrespecttotheirenvironmentanddynamics,andthe implications for galaxy formation (e.g., Hahn et al. 2010). It is a natural extension of thefundamentalthemeofrelatinggalaxiestothelargescalestructuresdescribedinChapter3.

4.4.5 Kinematiclensing

Gravitational lensing shears the velocity field of disk galaxies in a non-trivialway, yielding ananglebetweenthemaximum-andnull-velocityaxesthatdiffersfromtheexpected90°without


199

lensing.Likewise, inpresenceoflensing,thevelocityandmorphologicalpositionanglesdonotcoincideanymore.Using7-fibresbundles(IFUs)withtotaleffectivediameter1.8–2.7”(or19-fibresbundlesforhigh-valueobjects)tomapthevelocityfieldofspiralgalaxiesatredshiftz~0.3–1.0,onecandirectlypoint-wiseestimatethegravitationalshear,attainingvaluesaslowastypicalofcosmologicalweak-lensing,orlargervaluesasduetogalaxyclusterstypicallylocatedatz≲0.5.Thistechnique,sofarproposedintheradiodomainandbasedonhigh-resolutionHIvelocitymaps,canprovideanindependentmeasurementofgravitationallensingofgalaxies.Itsaim is an improvement in themassmodelingof the large-scale structureand clustersusingasmallernumberofgalaxiesthanrequiredbytraditional(statistical)methods.

Gravitationallensing,oneofthebestprobestomapthespatialdistributionofdarkmatterandexplore cosmology, is traditionallymeasured looking at the deformation andmagnification ofbackground images induced by the intervening matter. In weak and intermediate lensingregimes(κ,γ~0.01–0.1),therequiredprecisionisachievedbystatisticalprocedure,averagingthe signal over a large number of background sources and looking at the correlation of theirshapes. Thismethod is the coreof themajorweak-lensingprojects, e.g., CFHTLenS, RCSLenS,CS82, KiDS, DES, HSC, andwill be extraordinarily improved by LSST and ultimately Euclid andWFIRSTinthevisibleandNIRwavelengthdomain.

Figure91:Kinematiclensing.Left:Theeffectofgravitationallensingonthevelocityfieldofthebackgroundgalaxiesis to modify from perpendicular the angle between the null velocity axis and maximum velocity axis. Right:Relationshipbetweenthelensingshearγandconvergenceκoftheluminosityprofile,andtheangleαbetweenthekinematic axes; ameasurement of the latter with 5 – 7° accuracy (dark-light contours) allows for a point-wiseestimateofγof~2–3%,almostregardlessthevalueoftheκ.

Alternatively, onemay consider the effect of gravitational lensing on the velocity fieldof thebackground sources, in particular late-type galaxies and fast rotators; gravitational lensinganisotropicallystretchesthemajorandminoraxesoftheluminosityprofileofextendedsources,and their velocity field. This is illustrated in Figure 91. Here, we consider the backgroundsourcetobeanideal,late-typegalaxy(i.e.athincirculardisk)ofangulardiameter~10’withanon-trivial inclinationwithrespecttothe line-of-sight. It thereforeappearsasanellipse inthesourceplane (black line inFigure 91, leftpanel) andhasa fairly simple rotationcurve (hereapproximatingthatofNGC3145).Lensingcausesthemaximum-velocityaxis(a.k.a.the“velocityposition angle”, definedby the locations of themaximumandminimumof the velocitymap)


200

and the null-velocity axis (colored lines in Figure 91, left panel) to not be perpendicularanymore. Blain (2002) shown that the angle α between the two axes depends on theconvergenceκandabsolutevalueoftheshearγas

α = 2arctan1−κ +γ1−κ −γ

≈π2

+ 2γ − 2κγ.

Thisapproximation isvaliduptothe intermediate-lensingregime(κ,γ≈0.1), typicalofweak-lensingbygalaxyclusters.Toovercomethedifficulty indetermining thenull-velocityaxis,onecaninsteadmeasuretheanglebetweenthevelocityandphotometricpositionangles,whichintheimageplanediffersfromzeroinpresenceoflensing(blackarrowinFigure91,leftpanel).According to this equation, for a typical accuracy of the velocity positionangle,~5°, andprovided thenull-velocity axis is determinedwith the sameaccuracy,onemight expect tobeabletomeasureabsoluteshearassmallasγ∼0.03–0.035(Figure91,rightpanel),whichistheorderofmagnitudeexpectedforcosmologicalweak-lensing.

With a perfect velocity measurement based on the full map of the velocity field, such asprovidedbyradiosurveys(e.g.,SKA),onlytwolensedgalaxiesaresufficienttoexactlyretrievetheintrinsicgravitationalshearinweak-lensinglimit(Morales2006).Alowerresolutionvelocitymapasprobedbyhigh-resolutionfibre-bundles(IFUs)inopticalbands,eventuallyimprovedbya stacking technique of close galaxies, can still bring off the same goal in the weak- andintermediate-lensingregime.Gravitationallensingestimatedbyvelocityfieldwillbesensitivetototallydifferentsystematicswith respect tostandardmeasurementsbasedonlyon luminosityprofileofbackgroundsources.

It isworthnoting that it isnot just in the regimeofweak lensing thatMSEwillhavea strongscienceimpact.Inthestronglensingregime,MSEwillexploitstrongsciencesynergies.InSDSS-I,the Sloan Lens ACS Survey (SLACS; Bolton et al. 2006, 2008a) discovered over 100 stronggravitational lenssystemsthroughthespectroscopicsignatureoftworedshiftsalongthesameline of sight. Follow-up imaging of spectroscopic lens candidates with the Hubble SpaceTelescope and high-resolution ground-based facilities has enabled a unique experimentalapproach to the study of the structure and dynamics of massive galaxies at relatively lowredshift (e.g.,Koopmansetal.2006,Boltonetal.2008b).TheBOSSEmission-LineLensSurvey(BELLS: Brownstein et al. 2012, Bolton et al. 2012) is successfully extending this technique tosignificant cosmological look-back time at redshift z~ 0.4 – 0.6, thereby probing directly thestructuralanddynamicalevolutionofLRGs.

AfuturespectroscopicsurveywithMSEcanallowthistechniquetobeextendedtoredshiftsofz~1andbeyond(forELGlenses),yieldingadirectmeasurementofmassivegalaxystructureoveranunprecedentedcosmictimebaseline.ELGlenseswouldalsoprovideaccuratemeasurementsofthetotalaperturemassesofgalaxiesbeyondz=1,forwhichmassesarecurrentlyavailableonly throughrelativelyuncertainstellar-populationsynthesis techniques.Thiswouldprovideapowerfulnewdiscriminantbetweencompetingmodelsfortheassemblyandevolutionofgalaxymass.


201

References

Abbas,U.,&Sheth,R.K.2006,MNRAS,372,1749Abbas,U.,etal.2010,MNRAS,406,1306Ackermann, M. & Fermi-LAT Collaboration. 2015,PhysicalReviewLetters,115,231301Adami,C.,etal.2006,A&A,459,679Adelberger,K.etal.,2003,ApJ,584,45Adelberger,K.L.,etal.2005,ApJ,629,636Akerman,C.,etal.,2005,A&A,440,499Alonso,M.,etal.2007,MNRAS,375,1017Andrews,B.,Martini,P.,2013,ApJ,765140Angulo,R.,etal.2012,MNRAS,426,2046Ann,H.B.,Park,C.,&Choi,Y.-Y.2008,MNRAS,389,86AokiW.etal.2013AJ145,13AokiW.,etal.2014,Science,345,912Aoki,W.,etal.2009,A&A,502,569Applegate,D.,etal.2014,MNRAS,439,48Aracil,B.,etal.,2004,A&A,419,811AsplundM.,etal.2006,ApJ,644,229Astier,P.,etal.,2014,A&A,572,80AustralianAcademyofScience2015,“Australiaintheeraofglobalastronomy”Bacon, R., Brinchmann, J., Richard, J., et al. 2015,A&A,575,A75Bahe,Y.M.,McCarthy,I.G.,Crain,R.A.,&Theuns,T.2012,MNRAS,424,1179Bahe,Y.M.,etal.,2013,MNRAS,430,3017Balcells, M., Benn, C. R., Carter, D., et al. 2010, inProc. SPIE, Vol. 7735, Ground-based and AirborneInstrumentationforAstronomyIII,77357GBaldry,I.&Glazebrook2003ApJ593,258Baldry, I. K., Driver, S. P., Loveday, J., et al. 2012,MNRAS,421,621Baldry,I.K.,etal.2006,MNRAS,373,469Baldry, I. K., Glazebrook, K., & Driver, S. P. 2008,MNRAS,388,945Balogh, M. L., Christlein, D., Zabludoff, A. I., &Zaritsky,D.2001,ApJ,557,117Balogh,M.L.,etal.2004,ApJL,615,L101Balogh,M.L.,etal.2011,MNRAS,412,2303Balogh,M.L.,Navarro,J.F.,&Morris,S.L.2000,ApJ,540,113Balogh,M.L.,etal.2016,MNRAS,456,4364Barklem,P.,etal.,2005,A&A,439,129Barnes,J.&Hibbard,J.,2009,AJ,137,3071Barnes,J.,2016,MNRAS,455,1957Barnes,J.,Hernquist,L.,1991,ApJ,370,65Bate,N. F.,McMonigal, B., Lewis, G. F., Irwin,M. J.,Gonzalez-Solares, E., Shanks, T., & Met- calfe, N.2015,MNRAS,453,690Battaglia,G.,etal.2006,A&A,459,423Battaglia,G.,etal.2008,MNRAS,383,183Battaglia,G.,etal.2011,MNRAS,411,1013

Bauer,A.E.etal.2013,MNRAS434,209Bechtol,K.,&TheDESCollaboration.2015,ApJ,807,50BeckerG.D,BoltonJ.S.,2013,436,1023Becker G.D, Bolton J.S., Haehnelt M.G., & SargentW.L.W.2011,410Behroozi,PeterS.,R.H.Wechsler,&C.Conroy,2013,ApJ770,57Bekki,K.2009,MNRAS,399,2221Bekki,K.,&Couch,W.J.2011,MNRAS,415,1783Belli,S.,etal.,2013,ApJ,772,141Belokurov,V.,etal.,2007,ApJ,654,897Belokurov, V.,Walker,M.G., Evans,N.W.,Gilmore,G., Irwin,M. J.,Mateo,M.,Mayer, L., Olszewski, E.,Bechtold,J.,&Pickering,T.2009,MNRAS,397,1748Belokurov,V.,Zucker,D.B.,Evans,N.W.,Gilmore,G.,Vidrih,S.,etal.2006,ApJ,642,L137BensbyT.,etal.2010A&A,521,57BensbyT.,etal.2011A&A,533,134Benson, A. J., Bower, R. G., Frenk, C. S., et al. 2003,ApJ,599,38Bentz,M.C.,&Katz,S.2015,PASP,127,67Bentz,M.C.,Denney,etal.2013,ApJ,767,149Berg,D.,Skillman,E.&Marble,A.2011ApJ738,2BergeronP.,etal.2005ApJ625,838Berlind,A.A.,&Weinberg,D.H.2002,ApJ,575,587BernardiM.etal.2003,AJ,125,32Bianchi,D.,etal.,2012,MNRAS,427,2420BigielF.,etal.2010,AJ,140,1194Birnboim,Y.,&Dekel,A.2003,MNRAS,345,349Biviano,A.,etal.,2013,A&A,558,1Blackburne, J. A., Pooley, D., Rappaport,S., &Schechter,P.L.2011,ApJ,729,34Blain,A.,2002,ApJ,570,51Blake,C.,etal.,2012,MNRAS,425,405Bland-Hawthorn,J.2015,inIAUSymposium,Vol.309,Galaxiesin3DacrosstheUniverse,ed.B.L.Ziegler,F.Combes,H.Dannerbauer,&M.Verdugo,21–28Blandford,R.D.&McKee,C.F.1982,ApJ,255,419Blandford,R.,etal.2010,NewWorlds,NewHorizonsinAstronomyandAstrophysicsBleem,L.,etal.2015,ApJS,216,27Bluck,A.,etal.,2014,MNRAS,441,599Blumenthal,G.R.,Faber,S.M.,Primack,J.R.,&Rees,M.J.1984,Nature,311,517Bode,P.,Ostriker,J.P.,&Turok,N.2001,ApJ,556,93Bolton J.S., Haehnelt M.G., Viel M., & Springel V.2005,MNRAS,357,1178Bolton J.S.,Oh S.P.,& Furlanetto S.R. 2009,MNRAS,396,2405Bolton,A.S., Burles, S., Koopmans, L.E.V., Treu, T.,&Moustakas,L.A.2006,ApJ,.638,703


202

Bolton, A.S., Burles, S., Koopmans, L.E.V., Treu, T.,Gavazzi,R.etal.2008a,ApJ,682,964Bolton,A.S.,etal.2012,AJ,144,144Bolton, A.S., Treu, T., Koopmans, LV.E., Gavazzi, R.,Moustakas,L.A.2008b,ApJ,684,248Bonaca,A.,Geha,M.,Kallivayalil,N.,2012,ApJ,760,12Bonifacio,P.,etal.,2009,A&A,502,1Bonnivard, V., Combet, C., Daniel, M., Funk, S.,Geringer-Sameth,A.,Hinton,J.A.,Maurin,D.,Read,J.I., Sarkar,S.,Walker,M.G.,&Wilkinson,M. I.2015,MNRAS,453,849Borucki,W.J.,etal.,2010,Science,327,977Bothwell,M.,etal.2013,MNRAS,433,1425Bouche,N.,etal.2004,MNRAS,354,25Bouche,N.,etal.2013,Science,341,50Bournaud,F.,Jog,C.J.,&Combes,F.2005,A&A,437,69Bouwens,R.,J.,etal.,2008,ApJ,686,230Bovy,J.,2016,ApJ,817,49Bower,R.G.,etal.2006,MNRAS,370,645Boyarsky,A.,Ruchayskiy,O., Iakubovskyi,D.,Walker,M. G., Riemer-Sørensen, S., & Hansen, S. H. 2010,MNRAS,407,1188Boylan-Kolchin, M., Bullock, J. S., & Kaplinghat, M.2011,MNRAS,415,L40—.2012,MNRAS,422,1203Boylan-Kolchin, M., Bullock, J. S., & Kaplinghat, M.2012,MNRAS,422,1203Boylan-Kolchin,M.,etal.,2015,MNRAS,453,1503Brinchmann,J.,etal.,2004,MNRAS,351,1151Brodie&Larsen2002,AJ,124,1410Bromm,V.,&Loeb,A.2003,Nature,425,812BrooksA.M.,etal.2009,ApJ,694,396Brooks,A.M., Kuhlen,M., Zolotov,A.,&Hooper,D.2013,ApJ,765,22BrottI.,etal.2011,A&A,530,116Brown,W.R.,Geller,M.J.,Kenyon,S.J.,&Diaferio,A.2010,AJ,139,59Brownstein,J.,etal.2012,ApJ,744,41Buchert,T.,Carfora,M.,2008,CQGra,25,5001Buchert,T.,etal.2000,PhRvD,62,3525BuchhaveL.A.,&Latham,D.2015,ApJ,808,187BuchhaveL.A.,etal.2012,Nature,486,375BuchhaveL.A.,etal.2014,Nature,509,593Bundy, K., Bershady, M. A., Law, D. R., et al. 2015,ApJ,798,7Butcher,H.,&Oemler,A.1978,ApJ,219,18Cabre,A.,&Gaztanaga,E.2009,MNRAS396,1119Cappellarietal.,2011b,MNRAS,416,1680Cappellari,M.2016,ArXive-prints,arXiv:1602.04267Cappellari,M.,&Copin,Y.2003,MNRAS,342,345Cappellari, M., Emsellem, E., Krajnovic, D., et al.2011a,MNRAS,413,813

Cappellari, M., McDermid, R. M., Alatalo, K., et al.2012,Nature,484,485Cappellari, M., Scott, N., Alatalo, K., et al. 2013,MNRAS,432,1709Carlberg,R.G.,etal.2000,ApJL,532,L1Carlberg,R.,2009,ApJL705,L223Carlberg,R.,2012,ApJ748,20Carlberg,R.,etal.2015,ApJ808,15Carlin, J. L.,Grillmair,C. J.,Muñoz,R.R.,Nidever,D.L.,&Majewski,S.R.2009,ApJ,702,L9CarolloD.,etal.2014ApJ788,180CarreroG.,etal.2015AJ149,12Carrick,J.,etal.2015,MNRAS,450,317Carter,D.,etal.2008,ApJS,176,424Casteels,K.etal.2014,MNRAS445,1157Cattaneo,A.,Dekel,A.,Faber,S.M.,&Guiderdoni,B.2008,MNRAS,389,567Caucci,S.,etal.,2008,MNRAS,386,211Cayreletal.2004,A&A,416,1117Cen,R.2011,ApJ,741,99Chapman,S.,etal.2006,ApJ,653,255Chen, C., Zhang, J., & Vogeley, M.S. 2009, MappingtheGlobalImpactofSloanDigitalSkySurveyChen,H.-C.,etal.2013,A&A,550,62Chen,H.-C.,etal.2014,MNRAS,438,1435Chengetal.2012,ApJ,746,149Chornock,R.,etal.2013,ApJ,767,162Chornock,R.,etal.2014,ApJ,780,44Chou,R.,etal.,2011,AJ,141,87Cirasuolo, M., Afonso, J., Bender, R., Bonifacio, P.,Evans,C.,etal.2012,arXiv:1208.5780Cisewski,J.,etal.,2014,MNRAS,440,2599Clark, P.C., Glover, S.C.O., Smith, R. J.,et al. 2011,Science,331,1040Clifton,T.,etal.,2012,PhR,513,1CodisS.,etal.2012,MNRAS,427,3320CodisS.,etal.2015,MNRAS,452,3369CohenJ.G.etal.2013ApJ778,56Coil,A.L.,etal.2006,ApJ,644,671Cole,A.etal.2014ApJ,795,54Cole, S., Norberg, P., Baugh, C. M., et al. 2001,MNRAS,326,255Colless, M., Dalton, G., Maddox, S., et al. 2001,MNRAS,328,1039Colless,M.,etal.2001,MNRAS,328,1039Collins,M.L.M.,Chapman,S.C.,Rich,R.M.,Ibata,R.A.,Martin,N.F.,Irwin,M.J.,Bate,N.F.,Lewis,G.F.,Peñarrubia,J.,Arimoto,N.,Casey,C.M.,Ferguson,A.M. N., Koch, A., McConnachie, A. W., & Tanvir, N.2013,ApJ,768,172Collins,M.,etal.,2011,MNRAS,413,1548Conroy, C.,Wechsler, R. H., & Kravtsov, A. V. 2006,ApJ,647,201Conselice, C. J., Bershady, M. A., Dickinson, M., &Papovich,C.2003,AJ,126,1183


203

Conselice, C. J., Bershady, M. A., Dickinson, M., &Papovich,C.2003,AJ,126,1183Conselice,C.,2014,ARAA,52,291Contreras,C.,etal.2013,MNRAS,430,924Cooke,R.etal.,2011,MNRAS,417,1534Cooper,M.C.,etal.2006,MNRAS,370,198Cooper,M.C.,etal.2010,MNRAS,409,337Cortese, L., Fogarty, L. M. R., Ho, I.-T., et al. 2014,ApJL,795,L37Couch, W. J., Barger, A. J., Smail, I., Ellis, R. S., &Sharples,R.M.1998,ApJ,497,188Coupon,J.,etal.2012,A&A,542,5Cox,T.J.,etal.,2008,MNRAS,384,386Crighton,N.2011,MNRAS,414,28Crighton,N.,etal.,2003,MNRAS,345,243Croom,S.M., Lawrence, J. S.,Bland-Hawthorn, J., etal.2012,MNRAS,421,872Croton,D.J.,Springel,V.,White,S.D.M.,etal.2006,MNRAS,365,11Croxall,K.,etal.,2009,ApJ,705,723Daddi,E.,etal.2007,ApJ,670,156Dalcanton,J.J.&Hogan,C.J.2001,ApJ,561,35DatsonJ.etal.2014MNRAS,439,1028Davies,L.J.M.,Robotham,A.S.G.,Driver,S.P.,etal.2015,MNRAS,452,616Davies,L.,etal.,2015,MNRAS,452,616Davis,M., Efstathiou,G., Frenk, C. S.,&White, S.D.M. 1985, ApJ, 292, 371 de Blok, W. J. G. 2010,AdvancesinAstronomy,2010Davis,T.A.,&Bureau,M.2016,MNRAS,457,272Davis,T.A.,Alatalo,K.,Sarzi,M.,etal.2011,MNRAS,417,882Davis,T.,etal.,2011,ApJ,741,67deBlok,W.J.G.,&Walter,F.2006,AJ,131,343de Burgh-Day, C. O., Taylor, E. N.,Webster, R. L., &Hopkins,A.M.2015a,MNRAS,451,2161—.2015b,PASA,32,40delaTorre,S.,etal.2013,A&A,557,54DeLucia,G.,etal.2006,MNRAS,366,499DeRosa,G.etal.2015,ApJ,806,128DeSilva,G.,etal.2007,AJ,133,1161DeSilva,G.,etal.2009,A&A,500,25Deason,A.J.,Belokurov,V.,Evans,N.W.,Koposov,S.E., Cooke, R. J., Peñarrubia, J., Laporte, C. F. P.,Fellhauer,M.,Walker,M.G.,&Olszewski,E.W.2012,MNRAS,425,2840DekelA.,etal.2009,ApJ,703,785Dekel,A.,etal.2009,Nature,457,451Delubac,T.,etal.2015,A&A,574,59Denney,K.D.2012,ApJ,759,44Denney,K.D.etal.2010,ApJ,721,715Diemand,J.,Moore,B.,&Stadel,J.2005,Nature,433,389Flores,R.A.&Primack,J.R.1994,ApJ,427,L1Dorman,C.,etal.,2012,ApJ,752,147

DrakeA.J.,etal.2013ApJ763,32Dressler,A.1980,ApJ,236,351Dressler,A.,Oemler,Jr.,A.,Couch,W.J.,etal.1997,ApJ,490,577Drinkwateretal.2003,Nature,423,519Driver, S. P., Hill, D. T., Kelvin, L. S., et al. 2011,MNRAS,413,971Driver, S. P., Hill, D. T., Kelvin, L. S., et al. 2011,MNRAS,413,971Driver,S.P.,Wright,A.H.,Andrews,S.K.,etal.2016,MNRAS,455,3911Driver,S.,&DePropris,R.2003,A&AS,285,175Driver,S.,&Robotham,A.,2010,MNRAS,407,2131Driver,S.,etal.,2013,MNRAS,430,2622Driver,S.,etal.,2016,MNRAS,455,3911Drlica-Wagner, A., & DES Collaboration. 2015, ApJ,813,109Duc,P.A.,etal.,2015,MNRAS,446,120Durrell,P.,etal.2014,ApJ,794,103Edelson,R.,etal.2015,arXiv:1501.05951EisensteinD.J.,etal.2006ApJS,167,40Eke, V. R., Baugh, C. M., Cole, S., Frenk, C. S., &Navarro,J.F.2006,MNRAS,370,1147EkstromS.,etal.2012,A&A,537,146El-Badryetal.,2016,ApJ,820,131Ellingson, E., Lin, H., Yee, H. K. C., & Carlberg, R. G.2001,ApJ,547,609Ellis,G.F.R,Buchert,T.,2005,PhLA,347,38Ellis, G.F.R., 1984, in General relativity andgravitation,p.215-288Ellison,S.L.,Patton,D.R.,Simard,L.,&McConnachie,A.W.2008,AJ,135,1877Ellison,S.,etal.,2011,MNRAS,418,2043Ellison,S.,etal.,2013,MNRAS,435,3627Ellison,S.,etal.,2015,MNRAS,448,221Ellison,S.,Simard,L.,Cowan,N.,Baldry,I.,Patton,D.,McConnachie,A.2009MNRAS,396,1257Elmegreen,D.etal.2015,OptimisingtheUSGroundBasedOpticalandInfraredAstronomySystemsEmsellem, E., Cappellari, M., Krajnovic, D., et al.2011,MNRAS,414,888Erkal&Belokurov2015,MNRAS450,1136Fakhouri,O.,etal.2010,MNRAS,406,2267Faucher-Giguere C.-A., Prochaska J.X., Lidz A.,HernquistL.,&ZaldarriagaM.2008,ApJ,681,831Federman,S.R.1982,ApJ,257,125Feix,M.,etal.,2015,PhRvL,115,1301Ferramachoetal.2014,MNRAS,442,2511Ferrarese,L.,etal.2016,ApJ,arXiv:1604.06462Fillingham,S.,etal.2015,MNRAS,454,2039FirnsteinM.,PrzybillaN.,2012A&A,543,80Fischer,D.&Valenti,J.,2005,ApJ,622,1102Fixsen,D.,etal.,1996,ApJ,473,576Flemingetal.2015,AJ,149,143Fliri,J.,&Vals-Gabaud,D.2012,Ap&SS,341,57


204

Fogarty,L.M.R.,Scott,N.,Owers,M.S.,etal.2014,MNRAS,443,485—.2015,MNRAS,454,2050Forero-Romero,J.,etal.,2009,MNRAS,396,1815Foucaud,S.,etal.2010,MNRAS,406,147Francois,P.,etal.2007,A&A,476,935Frank,S.&Peroux,C.2010,MNRAS,420,1731Fraternali, F., Tolstoy, E., Irwin, M. J., & Cole, A. A.2009,A&A,499,121FrebelA.etal.2014ApJ786,74FrebelA.,NorrisJ.E.,2015,ARA&A,53,631Frebel, A., Kirby, E. N., & Simon, J. D. 2010, Nature,464,72Freeman,K.,&Bland-Hawthorn, J.2002,ARA&A,40,487Friedman S,D., York, D.G.,McCall, B.J., Dahlstrom, J.Sonnentrucker,P.,etal.2011,ApJ,727,33FrielE.D.,etal.2010AJ139,1942Frieletal.,2014,A&A,563,117Frieman, J.A., Turner, M.S., & Huterer, D. 2008,ARA&A46,385FrinchaboyP.etal.2013ApJ777,1Frinchaboy,P.M.,Majewski,S.R.,Muñoz,R.R.,Law,D. R., Łokas, E. L., Kunkel,W. E., Patterson, R. J., &Johnston,K.V.2012,ApJ,756,74Fulbrightetal.2007,ApJ,661,1152FulbrightJ.P.,etal.2006,ApJ,636,821FurlanettoS.R.,&OhS.P.2008,ApJ,681,1Furlong,M.etal.2015MNRAS450,4486Gao,L.,&White,S.D.M.2007,MNRAS,377,L5GarciaPerezA.E.,etal.2013ApJ767,9Garrison-Kimmel,etal.,2014,MNRAS,438,2578Garrison-Kimmel, S., Rocha, M., Boylan-Kolchin, M.,Bullock,J.S.,&Lally,J.2013,MNRAS,433,3539Gauthier,J.-R.,etal.2009,ApJ,702,1Geha,M., Blanton,M., Yan, R.& Tinker, J. 2012ApJ757,85Geha,M.,Guhathakurta,P.,Rich,R.M.,&Cooper,M.C.2006,AJ,131,332Geha, M., van der Marel, R. P., Guhathakurta, P.,Gilbert,K.M.,Kalirai,J.,&Kirby,E.N.2010,ApJ,711,361Genzel, R., Newman, S., Jones, T., et al. 2011, ApJ,733,101Geringer-Sameth, A., Koushiappas, S. M., & Walker,M.2015,ApJ,801,74Geringer-Sameth, A., Koushiappas, S. M., & Walker,M.G.2015,Phys.Rev.D,91,083535GhezziL.,etal.2010ApJ,720,1290Gilbank,D.G.,etal.2010,MNRAS,405,2419GilmoreG.,etal.2012,ESOMessenger,147,25Gilmore,G.,Wilkinson,M.I.,Wyse,R.F.G.,Kleyna,J.T., Koch,A., Evans,N.W.,&Grebel, E.K.2007,ApJ,663,948Giovanelli,R.,etal.2005,AJ,130,2598

Godard, B., Falgarone, E., & Pineau Des Forets, G.2009,A&A,495,847Gonzalezetal.2011A&A530,54Gonzalez,G.,1997,MNRAS,284,403Governato,F.,Brook,C.,Mayer,L.,Brooks,A.,Rhee,G.,Wadsley, J., Jonsson, P.,Willman, B., Stinson,G.,Quinn,T.,&Madau,P.2010,Nature,463,203Governato, F., Zolotov, A., Pontzen, A., Christensen,C., Oh, S. H., Brooks, A. M., Quinn, T., Shen, S., &Wadsley,J.2012,MNRAS,2697Green, A. M., Hofmann, S., & Schwarz, D. J. 2004,MNRAS,353,L23Green,G.,etal.2015,ApJ,810,25Greene,J.E.,&Ho,L.C.2004,ApJ,610,722Greif, T.H., Springel, V., White, S.D.M., et al. 2011,ApJ,737,75Grillmair&Dionatos2006,ApJ643,L17Gronnow,A.,etal.2015,MNRAS,451,4005Guerou,A.,etal.2015,ApJ,804,70Gunawardhanaetal.2011,MNRAS415,1647Gunn,J.E.,&Gott,III,J.R.1972,ApJ,176,1Gunn, J. E., Lee,B.W., Lerche, I., Schramm,D.N.,&Steigman,G.1978,ApJ,223,1015Guzzo,L.,Pierleoni,M.,Meneux,B.,Branchini,E.,LeF\`evre,O.,etal.,2008,Nature,451,541Guzzo,L.,Strauss,M.A.,Fisher,K.B.,Giovanelli,R.,&Haynes,M.P.1997,ApJ,489,37Hahn,O.,etal.2007,MNRAS,375,489Hahn,O.,etal.2010,MNRAS,405,274Haines,C.,etal.2011,MNRAS,412,127HansenB.etal.2004ApJS155551HansenB.etal.2007ApJ671380Harbeck,D.,Grebel,E.K.,Holtzman,J.,Guhathakurta,P., Brandner,W., Geisler, D., Sarajedini, A., Dolphin,A.,Hurley-Keller,D.,&Mateo,M.2001,AJ,122,3092Hawkins,E.,etal.2003,MNRAS,346,78Heckmanetal.,1993,ASSL,188,455Heckman,T.M.,Armus,L.,&Miley,G.K.1990,ApJS,74,833HegerA.,WoosleyS.E.,2010ApJ724,341Hempel,M.etal.,2014,ESOMessenger,155,29Henry,A.,etal.,2013,ApJ,776,27HerwigF.,2005,ARA&A,43,435Higgs,C.,R.,etal.,2016,MNRAS,458,1678Hill, V., & the DART collaboration, 2012, in GalacticArchaeology:NearFieldCosmologyandtheformationoftheMilkyWay,p297Ho, I.-T., Medling, A. M., Bland-Hawthorn, J., et al.2016,MNRAS,457,1257Ho, N., Geha, M., Munoz, R. R., Guhathakurta, P.,Kalirai, J., Gilbert, K. M., Tollerud, E., Bullock, J.,Beaton,R.L.,&Majewski,S.R.2012,ApJ,758,124Hobbs, L.M., York, D.G., Thornburn, J.A., Snow, T.P.,Bishof,M.,etal2009,ApJ,705,32Hoffman,Y.,etal.,2012,MNRAS,425,2049


205

Hoffman,Y.,etal.,2015,MNRAS,449,4494Hofmann,S.,Schwarz,D.J.,&Stöcker,H.2001,Phys.Rev.D,64,083507Hogg,D.,etal.2004,ApJ,601,29Hogg,D.,etal.2016,arXiv:160105413Holoien,T.,etal.2014,MNRAS,445,3263Hopkins,A.M.,&Beacom,J.F.2006,ApJ,651,142Hopkins,A.M.,&Beacom,J.F.2006,ApJ,651,142Hopkins, P. F., Bundy, K., Hernquist, L.,Wuyts, S., &Cox,T.J.2010,MNRAS,401,1099Hopkins,P.F.,Croton,D.,Bundy,K.,etal.2010,ApJ,724,915Hopkins, P. F., Richards, G. T., & Hernquist, L. 2007,ApJ,654,731Horne,K.,etal.2004,PASP,116,465Howes,L.,etal.2015,Nature,527,484Huff, E. M., Krause, E., Eifler, T., George, M. R., &Schlegel,D.2013,ArXive-prints,arXiv:1311.1489HunterI.,etal.2009,A&A,496,841Huxoretal.2005,MNRAS,360,1007Ibata,R.,etal.2002,MNRAS332,915Ibata,R.,etal.,2001,Nature,412,49IvansI.I.,etal.2003ApJ592,906Jang&Lee2014,ApJ,795,6Jeltema,T.&Profumo,S.2016,MNRASJeong,D., Schmidt,F., Hirata,C., 2012, PhRvD, 85,3504Jietal.,2016,Nature,531,610Ji,I.,Peirani,S.,&Yi,S.2014,A&A,566,97Jimenez-Vicente,J.,etal.2012,ApJ,751,106Jogee, S., Miller, S. H., Penner, K., et al. 2009, ApJ,697,1971JohnsonC.I.,etal.2014AJ148,67Johnson,A.,etal.,2014,MNRAS,444,3926Johnston, E. J., Aragon-Salamanca, A., & Merrifield,M.R.2014,MNRAS,441,333Johnston,K.,etal.1999,ApJ,512,109Johnston,K.,etal.2002,ApJ570,656Just, D. W., Zaritsky, D., Sand, D. J., Desai, V., &Rudnick,G.2010,ApJ,711,192Kaiser,N.,1987,MNRAS,227,1KaliraiJ.S.,2012,Nature,486,90Kalirai,J.,etal.2006,ApJ,648,389Kampczyk,P.,etal.,2013,ApJ,762,43Kaplan,K.etal.,2010,PASP,122,619Kaplinghat,M.,&Strigari,L.E.2008,ApJ,682,L93Kassin, S. A.,Weiner, B. J., Faber, S.M., et al. 2007,ApJL,660,L35Kauffmann,G.etal.2003MNRAS346,1055Kawata,D.,&Mulchaey,J.S.2008,ApJL,672,L103Keller,S.,etal.,2014,Nature,506,463Kennicutt,R.1983ApJ272,54Keres,D.,Katz,N.,Weinberg,D.H.,&Dave,R.2005,MNRAS,363,2

Kewley, L. J.,Groves,B.,Kauffmann,G.,&Heckman,T.2006,MNRAS,372,961Kewley, L. J., Rupke, D., Zahid, H. J., Geller,M. J., &Barton,E.J.2010,ApJL,721,L48Khabiboulline,E.,etal.2014,ApJ,795,62Khare,P.etal.,2012,MNRAS,419,1028Khochfar, S., Emsellem, E., Serra, P., et al. 2011,MNRAS,417,845KilicM.,etal.2012,MNRAS,423,132KilicM.,etal.2005ApJ633,1126King,A.L.,etal.2014,MNRAS,441,3454King,A.L.,etal.2015,arXiv:1504.03031KirbyE.etal.2008,ApJ,685,43KirbyE.etal.2010,ApJS,191,352Kirby, E. N., Boylan-Kolchin, M., Cohen, J. G., Geha,M.,Bullock,J.S.,&Kaplinghat,M.2013,ApJ,770,16Kirby, E. N., Cohen, J. G., Simon, J. D., &Guhathakurta,P.2015,ApJ,814,L7Kistler,M.D.,Yuksel,H.,Beacom,J.F.,Hopkins,A.M.,&Wyithe,J.S.B.2009,ApJL,705,L104Klypin,A.,etal.,2016,MNRAS,457,4340Klypin,A.,Kravtsov,A.V.,Valenzuela,O.,&Prada,F.1999,ApJ,522,82Klypin,A.,Kravtsov,A.V.,Valenzuela,O.,&Prada,F.1999,ApJ,522,82Knobel,C.,etal.2009,ApJ,697,1842Koch,A.,Kleyna, J. T.,Wilkinson,M. I.,Grebel,E.K.,Gilmore, G. F., Evans, N. W., Wyse, R. F. G., &Harbeck,D.R.2007,AJ,134,566Koch, A., Wilkinson, M. I., Kleyna, J. T., Irwin, M.,Zucker,D.B.,Belokurov,V.,Gilmore,G.F.,Fellhauer,M.,&Evans,N.W.2009,ApJ,690,453Koda,J.,etal.,2014,MNRAS,445,4267KoesterD.,etal.2014,A&A,566,34Koopmans, L.V.E., Treu, T., Bolton,A.S., Burles, S.,&Moustakas,L.A.2006,ApJ,649,599Koposov,S.E.,Belokurov,V.,Torrealba,G.,&Evans,N.W.2015,ApJ,805,130Koposov,S.E.,Gilmore,G.,Walker,M.G.,Belokurov,V.,WynEvans,N.,Fellhauer,M.,Gieren,W.,Geisler,D.,Monaco,L.,Norris,J.E.,Okamoto,S.,Peñarrubia,J.,Wilkinson,M.,Wyse,R.F.G.,&Zucker,D.B.2011,ApJ,736,146Koposov,S.,etal.,2008,ApJ,686,279Korista,K.T.&Goad,M.R.2004,ApJ,606,749Kormendy,J.,Ho,L.,2013,ARA&A,51,511Koss,M.,etal.,2010,ApJ,726,125Kraljic,D.,2015,MNRAS,451,3393Kravtsov2010,AdAst2010,8Krumholz,M.R.,&Dekel,A.2012,ApJ,753,16Kubo,J.,etal.,2007,ApJ,671,466KudritzkiR.P.,etal.2012ApJ,747,15KuziodeNaray,R.,McGaugh,S.S.,&deBlok,W.J.G.2008,ApJ,676,920Kuzmaetal.2015,MNRAS446,3297


206

Laevens,B.P.M.,Martin,N.F.,Bernard,E.J.,Schlafly,E.F.,Sesar,B.,Rix,H.-W.,Bell,E.F.,Ferguson,A.M.N.,Slater,C.T.,Sweeney,W.E.,Wyse,R.F.G.,Huxor,A.P.,Burgett,W. S., Chambers,K.C.,Draper,P.W.,Hodapp,K.A.,Kaiser,N.,Magnier,E.A.,Metcalfe,N.,Tonry, J.L.,Wainscoat,R. J.,&Waters,C.2015,ApJ,813,44Lake,G.1990,Nature,346,39Lallement,R.,etal.,2014,A&A,561,91Lambas,D.,etal.,2003,MNRAS,346,1189Lambas,D.,etal.,2012,A&A,539,45Landecker T., Routledge, D., Reynolds, S.P., Smegal,R.J.,Borkowski,K.J.,etal.1999,ApJ,527,866Lara-Lopez,M.etal.2013a,MNRAS433,35Lara-Lopez,M.etal.2013b,MNRAS434,451Larson, R. B., Tinsley, B.M., & Caldwell, C. N. 1980,ApJ,237,692Larson, R. B., Tinsley, B.M., & Caldwell, C. N. 1980,ApJ,237,692Laureijs,R.,etal.2009,arXiv:0912.0914Laurent,P.,etal.2016,arXiv:1602.09010LeFevre,O.,etal.2005,A&A,439,877LeFevre,O.,etal.2013,A&A,559,14Leauthaud,A.,etal.2012a,ApJ,744,159Leauthaud,A.,etal.2012b,ApJ,746,95LecureurA.,etal.2007,A&A,465,799Leeetal.2014a,ApJ,795,12Lee,H.,etal.2006,ApJ,647,970Lee,KGetal.,2014b,ApJ,788,49Lelli,F.,etal.2014,MNRAS,445,1694Letarte,B.,etal.,2010,A&A,523,17Lewis, G. F., Ibata, R. A., Chapman, S. C.,McConnachie,A.,Irwin,M.J.,Tolstoy,E.,&Tanvir,N.R.2007,MNRAS,375,1364Lewis, I., Balogh, M., De Propris, R., et al. 2002,MNRAS,334,673Li,C.,etal.2006,MNRAS,368,37Lilly,S.J.,etal.2007,ApJS,172,70Lin,L.,etal.2007,ApJL,660,L51Lin,L.,etal.2012,ApJ,756,71Liuetal.2015,ApJ,812,34LiuY.J.,etal.2014ApJ785,94Loeb,A.&Weiner,N.2011,PhysicalReviewLetters,106,171302Loewenstein,M.&Kusenko,A.2010,ApJ,714,652Lotz,J.M.,Primack,J.,&Madau,P.2004,AJ,128,163Lotz,J.,etal.2010a,MNRAS,404,590Lotz,J.,etal.2010b,MNRAS,404,575Lu,T.,Gilbank,D.,McGee,S.,Balogh,M.&Gallagher,S.2012MNRAS420,126Lundgren,B.etal.,2009,ApJ,698,819Madau,P.,etal.1999,ApJ,514,648Madrid, J., & Macchetto, D. 2006, BAAS, vol 38, p.1286Madrid,J.,&Macchetto,D.2009,BAAS,vol49,p.193

MaederA.,MeynetG.,2010,NewAR,54,32Magnelli,B.,etal.2011,A&A,528,A35Mahajan,S.,etal.2011,MNRAS,416,2882Maier,C.,etal.,2014,ApJ,792,3Mandelbaum,R.,Seljak,U.,Kauffmann,G.,Hirata,C.M.,&Brinkmann,J.2006,MNRAS,368,715Marin,F.,etal.2013,MNRAS,432,2654Marshall,D.J.,Robin,A.C.,Reyl\'e,C.,Schultheis,M.,&Picaud,S.2006,A&A,453,635Martig, M., Bournaud, F., Teyssier, R., & Dekel, A.2009,ApJ,707,250Martin,N.F.,Ibata,R.A.,Chapman,S.C.,Irwin,M.,&Lewis,G.F.2007,MNRAS,380,281Martin,N.F., Ibata,R.A., Chapman, S.C., Irwin,M.,&Lewis,G.F.2007,MNRAS,380,281Martinez-Delgado,D.,etal.,2010,AJ,140,962Mateo,M.L.1998,ARA&A,36,435Mateo,M., Olszewski, E.W., &Walker,M. G. 2008,ApJ,675,201Maulbetsch,C.,etal.2007,ApJ,654,53Mayer et al. 2015, Proceedings of AdvancingAstrophysicswiththeSKA,arXiv:1501.01082McCarthy,I.G.,etal.2008,MNRAS,383,593McConnachie,A.W.&Côté,P.2010,ApJ,722,L209McConnachie, A.W., Peñarrubia, J., & Navarro, J. F.2007,MNRAS,380,L75McConnachie,A.W.,2012,AJ,144,4McConnachie,A.W.,etal.2005,MNRAS,356,979McConnachie,A.W.,etal.2009,Nature,461,66McDonald,P.,&Seljak,U.,2009,JCAP,10,7McElroy,Metal.,2015,MNRAS,446,2186McGee,S.L.,Balogh,M.L.,Bower,R.G.,Font,A.S.,&McCarthy,I.G.2009,MNRAS,400,937McGee,S.L.,etal.2011,MNRAS,413,996McGee,S.,Bower,R.&Balogh,M.2014MNRAS442,105McGee, S.,Goto, R.&Balogh,M. 2014MNRAS438,3188McQuinnM.etal.2009,ApJ,694,842McQuinnM.,&WhiteM.2011,MNRAS,415,2257McWilliamA.,etal.2008,AJ,136,367McWilliam,A.,Smecker-Hane,T.,2005,ApJ,622,29Melendez,J.,etal.,2009,ApJ,704,66Menard,B.&Peroux,C.2003,A&A,410,33Meneux,B.,etal.2008,A&A,478,299Meneux,B.,etal.2009,A&A,505,463Merluzzi,P.,etal.,2010,MNRAS,402,753MiceliM.,etal2010,MmSAI,75,282Michel-Dansac,L.,etal.MNRAS,386,82Mihosetal.2015,ApJL,812,34Minniti, D., Lucas, P.W., Emerson, J.P., Saito, R.K.,Hempel,M.,etal.2010,NewAstronomy,15,433Mistani,P.,etal.2016,MNRAS,455,2323Monreal-Ibero & Lallement, Mem. Dell. Soc. Ital.,2015


207

Moore,B.1994,Nature,370,629Moore,B.,Ghigna,S.,Governato,F.,etal.1999,ApJL,524,L19Moore,B.,Ghigna,S.,Governato,F.,Lake,G.,Quinn,T.,Stadel,J.,&Tozzi,P.1999,ApJ,524,L19Mor,R.&Netzer,H.2012,MNRAS,420,526Morales,M.F.2006,ApJ650,L21Mortazavi,S.,etal.2016,MNRAS,455,3058Moster,B,etal.,2011,MNRAS,415,3750Muñozetal.2005,ApJ,631,L137Muñozetal.2006,ApJ,649,201Munozetal.2014,ApJS,210,4Muñoz,R.R.,Geha,M.,&Willman,B.2010,AJ,140,138Murphy,M.etal.,2008,MNRAS,384,1053Murphy,M.,&Liske,J.,2004,MNRAS,386,1192Naab,T.,Oser, L., Emsellem,E., etal. 2014,MNRAS,444,3357Nakamura,F.,&Umemura,M.2001,ApJ,548,19Navarro,Frenk&White,1997,ApJ,490,493Navarro,Frenk&White.1996,ApJ,462,563—.1997,ApJ,490,493Navarro,J.F.,Eke,V.R.,&Frenk,C.S.1996,MNRAS,283,L72Nikolic,B.,etal.2004,MNRAS,355,874Noeske,K.,etal.,2007,ApJ,660,47NomotoK.,etal.2013ARAA51,457Norberg,P.,etal.2002a,MNRAS,332,827Norberg,P.,etal.2002b,MNRAS,332,827Noterdaeme,P.etal.,2008,A&A,481,327Noterdaeme,P.etal.,2009,505,1087Noterdaeme,P.etal.,2010,A&A,523,80Noterdaeme,P.etal.,2011,A&A,526,7Noterdaeme,P.etal.,2015,A&A,577,24Nusser,A.,Davis,M.,1995,ApJ,421,1Oh, S.-H., de Blok,W. J. G., Brinks, E.,Walter, F., &Kennicutt,Jr.,R.C.2011,AJ,141,193Önehagetal.2014,A&A,562,102Ota,K.,2008,ApJ,677,12Ozbek,M.,etal.,2016,MNRAS,456,3610Pacaud, F., et al. 2016, A&A, in press, arXiv:1512.04264Pancoast,A.,Brewer, B.J.,etal.2014,MNRAS,445,3073Pandey,B.,2013,MNRAS,430,3376Parfrey,K.,etal.,2011,PhRvD,83,3511ParisI.,etal.2011,A&A,530,50Park,C.,&Choi,Y.-Y.2009,ApJ,691,1828Park,C.,&Hwang,H.S.2009,ApJ,699,1595Patton,D. R., Ellison, S. L., Simard, L.,McConnachie,A.W.,&Mendel,J.T.2011,MNRAS,412,591Patton,D.R.,etal.2002,ApJ,565,208Patton,D.R.&Atfield,J.E.,2008,ApJ,685,235Patton,D.R.,etal.,2013,MNRAS,433,59Pawlik,M.,etal.,2016,MNRAS,456,3032

Peebles,P.J.E.,1980,TheLarge-ScaleStructureoftheUniverse(Princeton,N.J.,PrincetonUniversityPress)Peñarrubia,J.,Pontzen,A.,Walker,M.G.,&Koposov,S.E.2012,ApJ,759,L42Pengetal.2006,ApJ,639,838Peng,Y.-j.,Lilly,S. J.,Renzini,A.,&Carollo,M.2012,ApJ,757,4Peng,Y.,etal.2010,ApJ,721,193Peng, Y., Lilly, S. J., Renzini, A., & Carollo, M. 2012,ApJ,757,4PennyL.,GiesD.R.,2009,ApJ,700,844Pepe,F.A.,Cristiani,S.,ReboloLopez,R.,Santos,N.C.,Amorim,A.,etal.2010,SPIE,7735,14Percival,W.J.,&White,M.,2009,MNRAS,393,297Perlmutter,S.,Aldering,G.,Goldhaber,G.,Knop,R.A.,Nugent,P.,etal.1999,ApJ,517,565Peroux,C.,etal.,2013,MNRAS,436,2650Peterson,B.M.2011,ArXiv1109.4181PPeth,M.,etal.,2016,MNRAS,458,963Petitjean&Aracil,2004,A&A,422,52Pettini,M.,etal.,2001,ApJ,554,981Phleps,S.,Peacock,J.A.,Meisenheimer,K.,&Wolf,C.2006,A&A,457,145Pichonetal.2001,MNRAS,326,597Pierre, M., et al. 2016, A&A, in press, arXiv:1512.04317Pietrukowicz, P., Udalski, A., Soszy\'nski, I., Nataf,D.M.,Wyrzykowski,L.,etal.2012,ApJ,750,169PineaudesForetsG.,Flower,D.R.,Hartquist,T.W.,&Dalgarno,A.1986,MNRAS,220,801Ploeckinger,S.,etal.,2014,MNRAS,437,3980Plummer,H.C.1911,MNRAS,71,460Poggianti,B.,etal.2006,ApJ,642,188Poggianti,B.,etal.2013,ApJ,762,77PohlenM.,TrujilloI.,2006,A&A,454,759Pontzen,A.&Governato,F.2011,ArXiv:1106.0499—.2014,Nature,506,171Popesso,P.,etal.,2009,A&A,494,443Prescott,M.etal.2011MNRAS,417,1374Pritchet,C.&vandenBergh,S.,1999,AJ,118,883Pritchet, C., et al. 2010, Unveiling the Cosmos: AvisionforCanadianastronomyPrivon,G.,etal.,2013,ApJ,771,120Psychogyios,A.,etal.,2016,arXiv:1602.08500PuspitariniL.,etal.2015,A&A,573,35Querejeta, M., Eliche-Moral, M. C., Tapia, T., et al.2015,A&A,579,L2Rafieferantsoa,M.,etal.,2015,MNRAS,453,3980Raichoor,A.,&Andreon,S.2012,A&A,543,19RamirezI.,etal.2010A&A,521,33Ransom,R.R.,Kothes,R.,Wolleben,M.,&Landecker,T.L.2010,ApJ,724,946Rasmussen, J., Mulchaey, J., Bai, L, Ponman, T.,Raychaudhury,S.,Dariush,A.2012ApJ757,122Rauch,M.,1998,ARA&A,36,267


208

Rauer,H.,etal.2014,arXiv:1310.0696Read,J.I.&Gilmore,G.2005,MNRAS,356,107Reines,A.,Greene,J.&Geha,M.2013ApJ775,116Reyesetal.2010,Nature,464,256Rich, J. A., Kewley, L. J., & Dopita,M. A. 2011, ApJ,734,87Rich, J. A., Torrey, P., Kewley, L. J., Dopita,M. A., &Rupke,D.S.N.2012,ApJ,753,5Ridgwayetal.2014,ApJ,796,53Riess,A.G.,Filippenko,A.V.,Challis,P.,Clocchiatti,A.,Diercks,a.,etal.1998,AJ,116,1009Rines, K., Diaferio, A. &Natarajan, P. 2007 ApJ 657,183Robotham,A.S.G.etal.,2014,MNRAS,444,3986Robotham,A.S.G.,Driver,S.P.,Davies,L.J.M.,etal.2014,MNRAS,444,3986Robotham, A. S. G., Norberg, P., Driver, S. P., et al.2011,MNRAS,416,2640Robotham,A.,etal,2013,MNRAS,431,167Rocha,M., Peter, A. H. G., Bullock, J. S., Kaplinghat,M.,Garrison-Kimmel, S.,Oñorbe, J.,&Moustakas, L.A.2013,MNRAS,430,81Rodighieroetal.,2011,ApJ,739,40Rodrigues,T.,etal.2014,MNRAS,445,2758Roederer,I.,2016,AJ,151,82Rollinde,E.,etal.,2013,MNRAS,428,450RoskarR.,etal.2008,ApJ,675,65Ross, N.P., da Angela, J., Shanks, T., Wake, D.A.,Cannon,R.D.,etal.2007,MNRAS381,573Ruchayskiy, O., Boyarsky, A., Iakubovskyi, D., Bulbul,E.,Eckert,D.,Franse,J.,Malyshev,D.,Markevitch,M.,&Neronov,A.2015,ArXiv:1512.07217Rudie,G.etal.,2012,ApJ,750,67Ryan-Weber,etal.2009,MNRAS,395,1476Sanchez, S. F., Kennicutt, R. C., Gil de Paz, A., et al.2012,A&A,538,A8Sabater,J.,etal.,2013,MNRAS,430,638Sale,S.,etal.,2014,MNRAS,433,2907Salim,etal.,2007,ApJS,173,267Samushia,L,etal.2012,MNRAS,420,2102Sartoris,B.,etal.2016,MNRAS,459,1764Satyapal, S., Secrest, N., McAlpine, W., Ellison, S.,Fischer,J.&Rosenberg,J.2014ApJ784,113Sbordoneetal.2007,A&A,465,815Scannapieco, C., Wadepuhl, M., Parry, O. H., et al.2012,MNRAS,423,1726Scannapieco,E.,etal.2006,MNRAS,365,615Scharf,C.,etal.2000,ApJ,544,49SchlegelD.J.etal.2009,arXiv:0904.0468Schmidt,K.,etal.2013,MNRAS,432,285Schmidt,S.,etal.2014,ApJ,781,24SchneiderR.,etal.2014MNRAS442,1440Schneider, R., Omukai, K., Bianchi, S., & Valiante, R.2012,MNRAS,419,1566Schoenrich,R.,&Binney,J.2009,MNRAS,399,1145

Schorck,T.,etal.,2009,A&A,507,817Scott,C.&Kaviraj,S.,2014,MNRAS,437,2137Scott,N.,Fogarty,L.M.R.,Owers,M.S.,etal.2015,MNRAS,451,2723Scoville, N., Aussel, H., Brusa,M., et al. 2007, ApJS,172,1Scrimgeour,M.,etal.2012,MNRAS,425,116Scudder,J.,etal.,2012,MNRAS,426,549Scudder,J.,etal.,2015,MNRAS,449,3719Searle,L.,&Zinn,R.1978,ApJ,225,357Secrest,N.,etal.,2015,ApJ,798,38Seljak2009,Phys.Rev.L.,vol.102,Issue2,Sethetal.2015,Nature,513,398Seymour,N.,etal.2008,MNRAS,386,1695Shafter,A.W.,Darnley,M.J.,Bode,M.F.,&Ciardullo,R.2012,ApJ,752,156Shakura,N.I.&Sunyaev,R.A.1973,A&Ap,24,337Shapiro,K.L.,Genzel,R.,ForsterSchreiber,N.M.,etal.2008,ApJ,682,231Sharp,R.,&Parkinson,H.2010,MNRAS,408,2495Shen,Y.,etal.2015,ApJS,216,4ShetroneM.D.,etal.2003AJ125,684Siegal-Gaskins&Valluri2008,ApJ681,40Simon,J.D.&Geha,M.2007,ApJ,670,313Simon, J.D.,Geha,M.,Minor,Q.E.,Martinez,G.D.,Kirby,E.N.,Bullock,J.S.,Kaplinghat,M.,Strigari,L.E.,Willman,B.,Choi,P.I.,Tollerud,E.J.,&Wolf,J.2011,ApJ,733,46Simon,J.D.,Jacobson,H.R.,Frebel,A.,Thompson,I.B.,Adams,J.J.,&Shectman,S.A.2015,ApJ,802,93Simon,J.D.,&Geha,M.2007,ApJ,670,313Smith,D.,2015,arXiv:1506.05630Smith,R.2014MNRAS443,69Smith,R. J.,Lucey, J.R.,&Carter,D.2012a,MNRAS,426,2994Smith, R. J., Lucey, J. R., Price, J., Hudson, M. J., &Phillipps,S.2012b,MNRAS,419,3167Sobral,D.,etal.,2014,MNRAS,437,3516Somerville,R.S.,etal.2004,ApJL,600,L171Sommerivaetal.,2012,A&A,539,136SousaS.G.,etal.2011,A&A,533,141Speagleetal.,2014,ApJS,214,15Spergel, D. N. & Steinhardt, P. J. 2000, PhysicalReviewLetters,84,3760SpitoniE.,MatteucciF.,2011,A&A,531,72Springel,V.,DiMatteo,T.,&Hernquist,L.2005,ApJL,620,L79Springel,V.,White,S.D.M., Jenkins,A.,Frenk,C.S.,Yoshida,N.,Gao, L.,Navarro, J., Thacker,R., Croton,D., Helly, J., Peacock, J. A., Cole, S., Thomas, P.,Couchman, H., Evrard, A., Colberg, J., & Pearce, F.2005,Nature,435,629Srianand,R.etal.2008,A&A,482,39Steidel,C.C.,etal.2010,ApJ,717,289


209

Steigman, G., Dasgupta, B., & Beacom, J. F. 2012,Phys.Rev.D,86,023506Stierwalt,S.,etal.2015,ApJ,805,2Stott, J. P., Swinbank, A. M., Johnson, H. L., et al.2016,MNRAS,457,1888SudaT.,HirschiR.,FujimotoM.Y.,2011,ApJ,741,61Sullivan,M.,LeBorgne,D.,Pritchet,C.J.,etal.2006,ApJ,648,868Sweet,S.,etal.,2014,ApJ,782,35Swinbank,A.M.,etal.2006,MNRAS,368,1631Takadaetal.2014,PASJ,66,1Tal,T.,etal.,2009,AJ,138,1417Tayloretal.2003,AJ,125,314Tempel,E.etal.2014A&A566,1Thackeret al. 2016,Unveiling theCosmos:Canadianastronomy2016–2020Ting,Y.-S.,etal.,2012,MNRAS,421,1231Ting,Y.-S.,etal.,2015,ApJ,807,104Tinker, J., Wetzel, A., & Conroy, C. 2011, arXiv:11075046Tollerud,E.J.,Beaton,R.L.,Geha,M.C.,Bullock,J.S.,Guhathakurta,P.,Kalirai, J. S.,Majewski,S.R.,Kirby,E. N., Gilbert, K. M., Yniguez, B., Patterson, R. J.,Ostheimer,J.C.,Cooke,J.,Dorman,C.E.,Choudhury,A.,&Cooper,M.C.2012,ApJ,752,45Tollerud,E.J.,Bullock,J.S.,Strigari,L.E.,&Willman,B.2008,ApJ,688,277Tolstoy,E.,HillV.,TosiM.,2009,ARA&ATolstoy, E., Irwin, M.J., Helmi, A., Battaglia, G.,Jablonka,P.,etal.2004,ApJ,617,L119Tonnesen,S.,Cen,R.,2012,MNRAS,425,2313Torrealba,G.,etal.,2016,MNRAS,459,2370Torrey,P.,etal.2012,ApJ,746,108Tremblay P.E., Bergeron P., Gianninas A., 2011, ApJ,730.1289Tremonti,C.etal.2004ApJ613,898Treu,T.2010,ARA&A,48,87Trujillo,I.etal.2006,ApJ650,18Tully,R.B.,etal.,2014,Nature,513,71Tumlinson,J.,2010,ApJ,708,1398Turon,C.,Primas,F.,Binney,J.,Chiappini,C.,Drew,J,etal.2008,ReportbytheESA-ESOWorkingGrouponGalacticPopulations,ChemistryandDynamicsvanderWel,A.,&vanderMarel,R.P.2008,ApJ,684,260vanDokkum, P. G., & Conroy, C. 2010, Nature, 468,940vanUitert,E.,etal.2016,MNRAS,arXiv:1601.06791Veilleux, S., Cecil, G., & Bland-Hawthorn, J. 2005,ARAA,43,769VennK.A.,etal.2002,ApJ,565,571VennK.A.,etal.2004AJ,128,1177VennK.A.,etal.2014,ApJ791,98Vergely,J.,etal.,2001,A&A,366,1016Verhamme,A.,etal.,2008,A&A,491,89

VielM., Bolton J.S., & HaehneltM.G. 2009,MNRAS,399,L39Villa,E.,etal.,2014,CQGra,31,4005Vogelsberger,M.,Zavala,J.,&Loeb,A.2012,MNRAS,423,3740Vulcani,B.,etal.2010,ApJL,710,L1Wake,D.A.,etal.2011,ApJ,728,46WalkerM.,MateoM.,OlszewskiE.W.,2009a,AJ,137,3100Walker,M.G.&Peñarrubia,J.2011,ApJ,742,20Walker,M.G.,etal.,2016,ApJ,819,53Walker,M. G.,Mateo,M., &Olszewski, E.W. 2008,ApJ,688,L75Walker,M.G.,Mateo,M.,Olszewski,E.W.,Bailey,III,J. I., Koposov, S. E., Belokurov, V., & Evans, N. W.2015a,ApJ,808,108Walker,M.G.,Mateo,M.,Olszewski, E.W.,Gnedin,O.Y.,Wang,X.,Sen,B.,&Woodroofe,M.2007,ApJ,667,L53Walker, M. G., Mateo, M., Olszewski, E. W.,Peñarrubia, J.,Wyn Evans, N., & Gilmore, G. 2009b,ApJ,704,1274Walker,M.G.,Olszewski,E.W.,&Mateo,M.2015b,MNRAS,448,2717Wang, L., Li, C., Kauffmann,G.,&De Lucia,G. 2006,MNRAS,371,537Wareing,C.J.,Zijlstra,A.A.,O'Brien,T.J.,&Seibert,M.2007,ApJ,670,L125Weiner,B.,etal.,2009,ApJ,692,187Weinmann,S.M.,Kauffmann,G.,vonderLinden,A.,&DeLucia,G.2010,MNRAS,406,2249Weinmann, S. M., van den Bosch, F. C., Yang, X., &Mo,H.J.2006,MNRAS,366,2Weinmann,S.M.,etal.2012,MNRAS,426,2797Weinmann, S.M., Neistein, E., & Dekel, A. 2011,MNRAS,1463Wetzel, A. R., Hopkins, P. F., Kim, J.-h., Faucher-Giguere, C.-A., Keres, D., & Quataert, E. 2016,ArXiv:1602.05957Wetzel,A.,Tinker, J.,Conroy,C.&vandenBosch,F.2013,MNRAS432,336Whitakeretal.,2012,ApJ,754,29White,S.D.M.&Rees,M.J.1978,MNRAS,183,341White,S.D.M.,&Frenk,C.S.1991,ApJ,379,52Wijesinghe,D.B.,etal.2012,ApJ,423,7639Willman, B., Dalcanton, J.J., Martinez-Delgado, D.,West,A.A.,Blanton,M.R.,etal.2005,ApJ,626,L85Willman, B., Geha, M., Strader, J., Strigari, L. E.,Simon, J.D.,Kirby,E.,Ho,N.,&Warres,A.2011,AJ,142,128Wolf,C.,etal.2009,MNRAS,393,1302WollebenM.,2007,ApJ,664,349WollebenM.,Fletcher,A.,Landecker,T.L.,Carretti,E.,Dickey,J.M.,etal.2010,ApJ,724,48Wong,K.,etal.,2011,ApJ,728,119


210

Woods,D.&Geller,M.,2007,AJ,134,527Woods,D.,etal.,2010,AJ,139,1857Xu,C.etal.2012,ApJ,760,72Yadav,J.,etal.2010,MNRAS,405,2009Yamanoi,H.,etal.2012,AJ,144,40Yang,X.,Mo,H.J.,&vandenBosch,F.C.2008,ApJ,676,248Yang,X.,Mo,H. J., vandenBosch,F.C.,etal.2007,ApJ,671,153YongD.etal.2013aApJ762,26YongD.etal.2013bApJ762,27Yoonetal.2011,ApJ731,58York,D.etal.,2006,MNRAS,367,945York, D. G., Adelman, J., Anderson, Jr., J. E., et al.2000,AJ,120,1579York, D. G., Adelman, J., Anderson, Jr., J. E., et al.2000,AJ,120,1579

YoungerJ.D.,etal.,2007,ApJ,670,269Zafar,T.,etal.,2015,A&A,584,100Zahid,Hetal.,2013,ApJ,771,19Zahid,Hetal.,2014,ApJ,791,130Zasowski,G.,etal.2015,ApJ,798,35Zehavi,I.,etal.2002,ApJ,571,172Zehavi,I.,etal.2011,ApJ,736,59Zhangetal.2007Zibetti,S.ApJ,2005,ApJ,63L,105Zolotov,A.,Brooks,A.M.,Willman,B.,Governato,F.,Pontzen, A., Christensen, C., Dekel, A., Quinn, T.,Shen,S.,&Wadsley,J.2012,ApJ,761,71Zu, Y., Kochanek,C. S.,&Peterson,B.M.2011,ApJ,735,80Zuckeretal.2006,ApJ,643,L103