extragalactic science with decam galaxy evolution at z

27
Extragalactic Science with DECam Galaxy Evolution at z<1 Jennifer Lotz STScI

Upload: others

Post on 19-Oct-2020

1 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Extragalactic Science with DECam Galaxy Evolution at z

Extragalactic Science with DECamGalaxy Evolution at z<1

Jennifer LotzSTScI

Page 2: Extragalactic Science with DECam Galaxy Evolution at z

• what do we know about z<1 universe?

• unsolved problems

• DECam advantage

• the low surface brightness universe

Page 3: Extragalactic Science with DECam Galaxy Evolution at z

What have we learned at z<1?

• blue disks (red spheroids) were more (less) common in past

• most z<1 star-formation occurs in disk galaxies

• galaxy merger rate evolves modestly

• typical AGN are in massive, bulge-dominated galaxies

Page 4: Extragalactic Science with DECam Galaxy Evolution at z

Morphology

Assembly History

Star-formationHistory

Hogg et al. 2004

disks

spheroids

SAURON, ATLAS3d

Page 5: Extragalactic Science with DECam Galaxy Evolution at z

Evolution in SFR density

Bouwens et al. 2009

tota

l st

ar-

form

atio

n /

co-m

ovin

g vo

lum

e

redshift

Page 6: Extragalactic Science with DECam Galaxy Evolution at z

and our results are tabulated in Table 3. Note that we haveused the rest-frame B-band primarily as an accountingdevice here, calculating M=LBh i and using a Schechter fit tothe luminosity function to correct for sample incomplete-ness at the faint end. Our stellar masses are actually derivedusing NICMOS and ground-based near-IR observationsthat reach significantly redder rest-frame wavelengths(0.54–1.44 lm for the Ks-band data) over the entire redshiftrange being considered here.

Regardless of the model adopted for the mass estimates,there is good evidence that the global stellar mass densityincreased with cosmic time. For example, for galaxiesbrighter than present-day L!B , which are abundant in theHDF-N at all redshifts z < 3, the minimum estimate for !!at z " 1 is #7 times larger than that from the maximum-M=Lmodels at z " 2:7.

Figure 7 places the HDF-N mass density estimates into aglobal context, comparing them with other data at

TABLE 3

Stellar Mass Densities (log !!)

Z = 1.0Z$ Z = 0.2Z$

z range!(t) = 1 Component

(h70M$Mpc%3)!(t) = 2 Components

(h70M$Mpc%3)!(t) = 1 Component

(h70M$Mpc%3)!(t) = 2 Components

(h70M$Mpc%3)

0.5–1.4 ....... 8.46 & 0.07 8.61 & 0.07 8.52'0:07%0:08 8:62'0:07

%0:07

1.4–2.0 ....... 8.06'0:17%0:13 8.22'0:16

%0:12 7.97 & 0.17 8:12'0:16%0:13

2.0–2.5 ....... 7.58'0:11%0:07 8.01'0:09

%0:08 7.36'0:11%0:08 7:79'0:10

%0:09

2.5–3.0 ....... 7.52'0:23%0:14 7.89'0:20

%0:15 7.27'0:27%0:18 7:65'0:22

%0:15

Note.—"M ( 0:3,"# ( 0:7, h ( 0:7.

Fig. 7.—Redshift evolution of the comoving stellar mass density,"!. The horizontal axis scaling and error bars are as in Fig. 5. Open symbols show resultsfrom the literature at 0 < z < 1 (circle, Cole et al. 2001; triangles, Brinchmann & Ellis 2000; squares, Cohen 2002). The HDF-N points ( filled squares) showresults for the Z$, one-component ! t) * model fits. The vertical error bars show the 68% random errors for this choice of model, combining the samplingvariance and the mass fitting uncertainties. The vertical extent of the boxes shows the range of systematic uncertainty introduced by varying the metallicityand the star formation histories of the mass-fitting model. The blue curves show the result of integrating the cosmic star formation history, SFR(z), traced byrest-frame UV light, with and without corrections for dust extinction (solid and dotted curves, respectively). The dashed green curve shows the integrated starformation history from Pei et al. (1999), with their 95% confidence range indicated by the shaded region.

No. 1, 2003 EVOLUTION OF GLOBAL STELLAR MASS DENSITY 33

Evolution in stellar mass density

Dickinson et al. 2003

tota

l st

ell

ar

mas

s /

co-m

ovin

g vo

lum

e

redshift

Page 7: Extragalactic Science with DECam Galaxy Evolution at z

Evolution in the Galaxy Merger Rate

(1+z)2.1± 0.2(1+z)0.1± 0.4

Major Mergers

Major+Minor Mergers

merger rate evolves modestly

0.5-1 major merger per L* galaxy since z~1

1-3 minor mergers per L* galaxy since z~1

Lotz et al. 2011

Page 8: Extragalactic Science with DECam Galaxy Evolution at z

How Do Galaxies Assemble?

Ceverino et al. 2009Lacey & Cole 1993

galaxy mergers vs. gas accretion

time

Page 9: Extragalactic Science with DECam Galaxy Evolution at z

theory: mergers v. gas accretion

Keres et al. 2009redshift

mas

s ac

cre

tio

n r

ate

p

er

co-m

ovin

g vo

lum

e

gas accretion

Page 10: Extragalactic Science with DECam Galaxy Evolution at z

Unsolved problems in galaxy evolution

• inside-out growth of disks and spheroids: gas accretion and minor mergers

• quenching + spheroid formation

• assembly of the most massive galaxies

• survival of the smallest galaxies

Page 11: Extragalactic Science with DECam Galaxy Evolution at z

LSST Science Book

Deep/Wide Extragalactic Surveys

Chapter 9: Galaxies

will also undoubtedly be progress in smaller-area deep surveys following on from the Hubble DeepFields, GOODS, COSMOS, and the Subaru Deep Fields.

LSST will be a unique tool to study the universe of galaxies. The database will provide photometryfor 1010 galaxies, from the local group to redshifts z > 6. It will provide useful shape measurementsand six-band photometry for about 4 × 109 galaxies (§ 3.7.2). Figure 9.1 and Figure 9.2 provideindications of the grasp of LSST relative to other existing or planned surveys.

Figure 9.1: Comparison of survey depth and solid angle coverage. The height of the bar shows the solid anglecovered by the survey. The color of the bar is set to indicate a combination of resolution, area, and depth withrgb values set to r = V/V (HUDF), g = (mlim − 15/)/16, and b = θ/2��, where V is the volume within which thesurvey can detect a typical L∗ galaxy with a Lyman-break spectrum in the r band, mlim is the limiting magnitude,and θ is the resolution in arcseconds. The surveys compared in the figure are as follows: SDSS: Sloan Digital SkySurvey; MGC: Millennium Galaxy Catalog (Isaac Newton Telescope); PS1: PanSTARRS-1 wide survey, starting in2009 in Hawaii; DES: Dark Energy Survey (Cerro-Tololo Blanco telescope starting 2009); EIS: ESO Imaging survey(complete); CADIS: Calar Alto Deep Imaging Survey; CFHTLS: Canada France Hawaii Telescope Legacy Survey;NOAODWS: NOAO Deep Wide Survey; COSMOS: HST 2 deg2 survey with support from many other facilities;PS1MD: PanSTARRS-1 Medium-Deep Survey covering 84 deg2; GOODS: Great Observatories Origins Deep Survey(HST, Spitzer, Chandra, and many other facilities); WHTDF: William Herschel Telescope Deep Field; HDF, HUDF:Hubble Deep Field and Ultra Deep Field.

A key to testing our understanding of galaxy formation and evolution will be to examine the fullmulti-dimensional distributions of galaxy properties. Tools in use today include the luminosityfunction of galaxies, the color-luminosity relation, size-luminosity relation, quantitative morphol-ogy, and the variation of these distributions with environment (local density or halo mass). Asdata sets and techniques evolve, models will be tested not just by their ability to reproduce themean trends but by their ability to reproduce the full distribution in multiple dimensions. Studiesof the tails of these distributions – e.g., galaxies of unusual surface brightness or morphology – giveus the leverage to understand short-lived phases of galaxy evolution and to probe star formationin a wide range of environments.

310

Page 12: Extragalactic Science with DECam Galaxy Evolution at z

LSST Science Book

Deep/Wide Extragalactic Surveys

Chapter 9: Galaxies

will also undoubtedly be progress in smaller-area deep surveys following on from the Hubble DeepFields, GOODS, COSMOS, and the Subaru Deep Fields.

LSST will be a unique tool to study the universe of galaxies. The database will provide photometryfor 1010 galaxies, from the local group to redshifts z > 6. It will provide useful shape measurementsand six-band photometry for about 4 × 109 galaxies (§ 3.7.2). Figure 9.1 and Figure 9.2 provideindications of the grasp of LSST relative to other existing or planned surveys.

Figure 9.1: Comparison of survey depth and solid angle coverage. The height of the bar shows the solid anglecovered by the survey. The color of the bar is set to indicate a combination of resolution, area, and depth withrgb values set to r = V/V (HUDF), g = (mlim − 15/)/16, and b = θ/2��, where V is the volume within which thesurvey can detect a typical L∗ galaxy with a Lyman-break spectrum in the r band, mlim is the limiting magnitude,and θ is the resolution in arcseconds. The surveys compared in the figure are as follows: SDSS: Sloan Digital SkySurvey; MGC: Millennium Galaxy Catalog (Isaac Newton Telescope); PS1: PanSTARRS-1 wide survey, starting in2009 in Hawaii; DES: Dark Energy Survey (Cerro-Tololo Blanco telescope starting 2009); EIS: ESO Imaging survey(complete); CADIS: Calar Alto Deep Imaging Survey; CFHTLS: Canada France Hawaii Telescope Legacy Survey;NOAODWS: NOAO Deep Wide Survey; COSMOS: HST 2 deg2 survey with support from many other facilities;PS1MD: PanSTARRS-1 Medium-Deep Survey covering 84 deg2; GOODS: Great Observatories Origins Deep Survey(HST, Spitzer, Chandra, and many other facilities); WHTDF: William Herschel Telescope Deep Field; HDF, HUDF:Hubble Deep Field and Ultra Deep Field.

A key to testing our understanding of galaxy formation and evolution will be to examine the fullmulti-dimensional distributions of galaxy properties. Tools in use today include the luminosityfunction of galaxies, the color-luminosity relation, size-luminosity relation, quantitative morphol-ogy, and the variation of these distributions with environment (local density or halo mass). Asdata sets and techniques evolve, models will be tested not just by their ability to reproduce themean trends but by their ability to reproduce the full distribution in multiple dimensions. Studiesof the tails of these distributions – e.g., galaxies of unusual surface brightness or morphology – giveus the leverage to understand short-lived phases of galaxy evolution and to probe star formationin a wide range of environments.

310

DES SNe

Page 13: Extragalactic Science with DECam Galaxy Evolution at z

DECam advantages

field of view: statistics, large objects

depth: distant objects, low-surface brightness

y-band: better photz, stellar masses

what DES will not do:

• very deep fields (except for SNe fields but no y)

• special places (local galaxies, nearby clusters)

• special filters (u, narrow-band)

Page 14: Extragalactic Science with DECam Galaxy Evolution at z

Unsolved problems in galaxy evolution

• inside-out growth of disks and spheroids: gas accretion and minor mergers

• quenching + spheroid formation

• assembly of the most massive galaxies

• formation + survival of the smallest galaxies

⇒ low-surface brightness universe

μ ~ 27-29 mag per sq arcsec

Page 15: Extragalactic Science with DECam Galaxy Evolution at z

Steidel et al. 2011

Ly-alpha halos: scattering or accretion?

μg ~ 27.1

The Astrophysical Journal, 736:160 (18pp), 2011 August 1 Steidel et al.

Figure 4. Left: scaled far-UV continuum image produced (as described in the text) from the average of 92 continuum-selected LBGs, drawn from three independentfields. The regions shown are 20!! ("160 physical kpc at z = 2.65) on a side, with a grid spacing of 2!!. Right: the continuum-subtracted, stacked Ly! image for the samesample of galaxies. In both panels, the contours are logarithmically spaced in surface brightness with the lowest contour shown at "2.5#10$19 erg s$1 cm$2 arcsec$2.(A color version of this figure is available in the online journal.)

Figure 5. Observed average surface brightness profile for the 1220 Å continuumlight (blue) and the Ly! line (red) for the full sample of 92 continuum-selectedgalaxies, evaluated over the same rest-frame bandwidth sampled by the Ly!image (24.3 Å). Note that these profiles are simply the azimuthal averages of thestacked images shown in Figure 4. The light-shaded region indicates the rangeof typical Ly! surface brightness threshold reached by deep Ly! surveys for thedetection of individual objects. The dashed lines show the surface brightnessprofile assuming that S(b) = Clexp($b/bl) with parameters given in Table 2.The corresponding angular scale at %z& = 2.65 is given along the top axis.For the purpose of comparison, we also show the Ly! profile expected forthe same sources under the assumption of the “Case B” Ly! to CB ratio, nodestruction of Ly! by dust, and no spatial diffusion of Ly! photons due toresonant scattering (i.e., the Ly! and CB profiles would be identical in shape,and since W0(Ly!) = 100 Å, the Ly! line image would be a factor of "4.1brighter than the continuum in the effective rest-frame bandwidth of 24.3 Å).(A color version of this figure is available in the online journal.)

(i.e., Case B), and Ly! and CB light had the same spatialdistribution on average.

The surface brightness profiles for both the continuum andthe Ly! line images are reasonably well-fit by an exponential

of the form S(b) = Ciexp($b/bi) for projected radii beyondthe central arcsec; the parameters of the best-fit values for thenormalization Ci and scale length bi are given in Table 2 for eachsub-sample as well. In the full image stack, the effective surfacebrightness detection thresholds are a factor of '10 lower thanfor individual galaxies; it is clear that the distribution of Ly!emission is very different from that of the continuum for everysub-sample, with best-fit Ly! scale lengths of bl " 20–30 kpc,compared with the corresponding continuum emission whichhas bc " 3–4 kpc. It is important to note that the true differ-ence in scale length is larger, since we have made no attemptto deconvolve the profiles from the seeing disk, which wasFWHM " 0.!!86, 1.!!20, and 0.!!80 for HS1549, HS1700, andSSA22a, respectively. The continuum profiles of the stackedcomposite CB images have FWHM " 1.!!2–1.!!4, indicating aver-age (seeing deconvolved) galaxy continuum sizes of FWHM "0.!!80 (" " 0.!!35). These deconvolved angular sizes are also con-sistent with measurements of similar galaxies in deep HubbleSpace Telescope/Advanced Camera for Surveys images (e.g.,Peter et al. 2007; Law et al. 2007).

The stacked Ly! and CB images as in Figures 5 and 6represent unweighted averages of all galaxies in the sample (withmasking as described above). This choice was motivated by thedesire to preserve the photometric integrity of the stacks so thatfluxes could be measured directly using aperture photometry, butalso because any scaling or weighting would require decidingwhat the relevant figure of merit should be. Medians are oftenused to suppress outliers in stacked data sets, but they havethe disadvantage for the present application of not preservingflux in two-dimensional images, of working best when scalinghas been applied to individual images going into the stack,and of suppressing real signal as it approaches the noise level.Nevertheless, in Figure 7 we show a comparison of the surfacebrightness profiles for median-combined stacks as compared tomean-combined for both line and continuum.

We have argued that our sample of galaxies with %z& = 2.65has emission line and continuum properties characteristic ofthose in the full LBG spectroscopic surveys at these redshifts.Figure 8 shows that the diffuse Ly! emission is also consistentamong the three survey fields taken individually. This is im-portant, since each field samples galaxies at a different redshift

7

The Astrophysical Journal, 736:160 (18pp), 2011 August 1 Steidel et al.

Figure 4. Left: scaled far-UV continuum image produced (as described in the text) from the average of 92 continuum-selected LBGs, drawn from three independentfields. The regions shown are 20!! ("160 physical kpc at z = 2.65) on a side, with a grid spacing of 2!!. Right: the continuum-subtracted, stacked Ly! image for the samesample of galaxies. In both panels, the contours are logarithmically spaced in surface brightness with the lowest contour shown at "2.5#10$19 erg s$1 cm$2 arcsec$2.(A color version of this figure is available in the online journal.)

Figure 5. Observed average surface brightness profile for the 1220 Å continuumlight (blue) and the Ly! line (red) for the full sample of 92 continuum-selectedgalaxies, evaluated over the same rest-frame bandwidth sampled by the Ly!image (24.3 Å). Note that these profiles are simply the azimuthal averages of thestacked images shown in Figure 4. The light-shaded region indicates the rangeof typical Ly! surface brightness threshold reached by deep Ly! surveys for thedetection of individual objects. The dashed lines show the surface brightnessprofile assuming that S(b) = Clexp($b/bl) with parameters given in Table 2.The corresponding angular scale at %z& = 2.65 is given along the top axis.For the purpose of comparison, we also show the Ly! profile expected forthe same sources under the assumption of the “Case B” Ly! to CB ratio, nodestruction of Ly! by dust, and no spatial diffusion of Ly! photons due toresonant scattering (i.e., the Ly! and CB profiles would be identical in shape,and since W0(Ly!) = 100 Å, the Ly! line image would be a factor of "4.1brighter than the continuum in the effective rest-frame bandwidth of 24.3 Å).(A color version of this figure is available in the online journal.)

(i.e., Case B), and Ly! and CB light had the same spatialdistribution on average.

The surface brightness profiles for both the continuum andthe Ly! line images are reasonably well-fit by an exponential

of the form S(b) = Ciexp($b/bi) for projected radii beyondthe central arcsec; the parameters of the best-fit values for thenormalization Ci and scale length bi are given in Table 2 for eachsub-sample as well. In the full image stack, the effective surfacebrightness detection thresholds are a factor of '10 lower thanfor individual galaxies; it is clear that the distribution of Ly!emission is very different from that of the continuum for everysub-sample, with best-fit Ly! scale lengths of bl " 20–30 kpc,compared with the corresponding continuum emission whichhas bc " 3–4 kpc. It is important to note that the true differ-ence in scale length is larger, since we have made no attemptto deconvolve the profiles from the seeing disk, which wasFWHM " 0.!!86, 1.!!20, and 0.!!80 for HS1549, HS1700, andSSA22a, respectively. The continuum profiles of the stackedcomposite CB images have FWHM " 1.!!2–1.!!4, indicating aver-age (seeing deconvolved) galaxy continuum sizes of FWHM "0.!!80 (" " 0.!!35). These deconvolved angular sizes are also con-sistent with measurements of similar galaxies in deep HubbleSpace Telescope/Advanced Camera for Surveys images (e.g.,Peter et al. 2007; Law et al. 2007).

The stacked Ly! and CB images as in Figures 5 and 6represent unweighted averages of all galaxies in the sample (withmasking as described above). This choice was motivated by thedesire to preserve the photometric integrity of the stacks so thatfluxes could be measured directly using aperture photometry, butalso because any scaling or weighting would require decidingwhat the relevant figure of merit should be. Medians are oftenused to suppress outliers in stacked data sets, but they havethe disadvantage for the present application of not preservingflux in two-dimensional images, of working best when scalinghas been applied to individual images going into the stack,and of suppressing real signal as it approaches the noise level.Nevertheless, in Figure 7 we show a comparison of the surfacebrightness profiles for median-combined stacks as compared tomean-combined for both line and continuum.

We have argued that our sample of galaxies with %z& = 2.65has emission line and continuum properties characteristic ofthose in the full LBG spectroscopic surveys at these redshifts.Figure 8 shows that the diffuse Ly! emission is also consistentamong the three survey fields taken individually. This is im-portant, since each field samples galaxies at a different redshift

7

stack of 92 z~3 LBGs LRIS NB over 0.04 sq degrees

~3 fields x 10 hr exptime on Keckneed to go 10x deeper to see

fluorescence?

DECam (+u, NB) for 4 sq degrees, 100x more objects in stack?

Page 16: Extragalactic Science with DECam Galaxy Evolution at z

964

MA

RT

INE

Z-D

EL

GA

DO

ET

AL

.V

ol.1

40

(a)

(c)

(e)

(g)

(h)

(f)

(d)

(b)

Fig

ure

1.L

umin

ance

filte

rim

ages

ofne

arby

gala

xies

from

ourp

ilots

urve

y(s

eeSe

ctio

n3

ford

iscu

ssio

n)sh

owin

gla

rge,

diff

use

light

subs

truc

ture

sin

thei

rout

skir

ts:

(a)

apo

ssib

leSg

r-lik

est

ream

inM

essi

er63

;(b)

gian

tplu

mes

arou

ndN

GC

1084

;(c)

part

ialt

idal

lydi

srup

ted

sate

llite

sin

NG

C42

16;(

d)an

umbr

ella

-sha

ped

tidal

debr

isst

ruct

ure

inN

GC

4651

;(e)

anen

orm

ous

stel

larc

loud

inN

GC

7531

;(f)

diff

use,

larg

e-sc

ale

and

mor

eco

here

ntfe

atur

esar

ound

NG

C35

21;(

g)a

prom

inen

tspi

kean

dgi

antw

edge

-sha

ped

stru

ctur

ese

enem

anat

ing

from

NG

C58

66(B

BO

0.5

m);

(h)

ast

rang

ein

ner

halo

inN

GC

1055

,spr

inkl

edw

ithse

vera

lspi

kes

ofde

bris

(RdS

0.5

m).

Eac

hpa

neld

ispl

ays

a(l

inea

r)su

per-

stre

tche

dco

ntra

stve

rsio

nof

the

tota

lim

age.

Aco

lor

inse

tof

the

disk

ofea

chga

laxy

(obt

aine

dfr

omda

tafr

omth

esa

me

tele

scop

eas

the

lum

inan

ceim

ages

)ha

sbe

enov

erpl

otte

dfo

rre

fere

nce

purp

oses

.In

addi

tion,

som

eof

the

orig

inal

imag

esw

ere

also

crop

ped

tobe

tter

show

the

mos

tin

tere

stin

gre

gion

sar

ound

each

targ

et.

with

3056

!30

56pi

xels

with

12!

12µ

m.

The

Moo

rook

(MrK

)ob

serv

ator

yus

esa

sim

ilar

CC

Dca

mer

a,th

eA

poge

eA

lta-K

AF1

6803

with

4096

!40

96pi

xels

and

asm

alle

rpi

xel

size

(9!

m).

The

tidal

stre

amde

tect

ion

stra

tegy

and

the

proc

edur

esus

edfo

rda

taac

quis

ition

and

redu

ctio

nof

the

data

are

the

sam

eas

thos

ede

scri

bed

inou

rpr

evio

uspa

pers

from

this

proj

ect

(Mar

tınez

-Del

gado

etal

.200

8,20

09).

Insu

mm

ary,

this

stra

tegy

striv

esfo

rm

ultip

lede

epex

posu

res

ofea

chta

rget

usin

ga

wid

eba

ndpa

ss,

high

thro

ughp

ut,

clea

rop

tical

filte

rw

itha

near

-IR

cuto

ff,a

lso

know

nas

alu

min

ance

filte

r,L.

The

typi

calc

umul

a-tiv

eex

posu

retim

esar

ein

the

rang

eof

6–11

hr.D

ata

redu

ctio

n

follo

wed

stan

dard

tech

niqu

esde

scri

bed

inth

eaf

ore-

men

tione

dpa

pers

.The

listo

ftar

gets

,tog

ethe

rwith

the

tele

scop

es,a

ndto

tal

expo

sure

time

used

inea

chca

sear

egi

ven

inTa

ble

2.Ph

otom

etri

cca

libra

tion

ofth

elu

min

ance

filte

r(L

)im

ages

isno

tcur

rent

lyav

aila

ble.

So,t

oas

sess

thei

rde

pth

and

typi

cal

qual

ityin

term

sof

back

grou

ndan

dfla

t-fie

ldin

g,w

ere

lied

onim

ages

ofsi

xof

our

gala

xies

—N

GC

1055

,N

GC

1084

,N

GC

3521

,NG

C42

16,N

GC

4651

,and

NG

C58

66—

obta

ined

byth

eSl

oan

Dig

italS

kySu

rvey

(SD

SS;Y

ork

etal

.200

0;D

ata

Rel

ease

7;A

baza

jian

etal

.200

9).

Bas

edon

SDSS

phot

omet

ry,

we

deriv

edph

otom

etri

ceq

ua-

tions

toco

nver

tth

eL-

band

coun

tsin

tog-

band

mag

nitu

des.

Martinez-Delgado et al. 2010

Outer halos of local disk galaxies

μg ~ 27.1

Page 17: Extragalactic Science with DECam Galaxy Evolution at z

964

MA

RT

INE

Z-D

EL

GA

DO

ET

AL

.V

ol.1

40

(a)

(c)

(e)

(g)

(h)

(f)

(d)

(b)

Fig

ure

1.L

umin

ance

filte

rim

ages

ofne

arby

gala

xies

from

ourp

ilots

urve

y(s

eeSe

ctio

n3

ford

iscu

ssio

n)sh

owin

gla

rge,

diff

use

light

subs

truc

ture

sin

thei

rout

skir

ts:

(a)

apo

ssib

leSg

r-lik

est

ream

inM

essi

er63

;(b)

gian

tplu

mes

arou

ndN

GC

1084

;(c)

part

ialt

idal

lydi

srup

ted

sate

llite

sin

NG

C42

16;(

d)an

umbr

ella

-sha

ped

tidal

debr

isst

ruct

ure

inN

GC

4651

;(e)

anen

orm

ous

stel

larc

loud

inN

GC

7531

;(f)

diff

use,

larg

e-sc

ale

and

mor

eco

here

ntfe

atur

esar

ound

NG

C35

21;(

g)a

prom

inen

tspi

kean

dgi

antw

edge

-sha

ped

stru

ctur

ese

enem

anat

ing

from

NG

C58

66(B

BO

0.5

m);

(h)

ast

rang

ein

ner

halo

inN

GC

1055

,spr

inkl

edw

ithse

vera

lspi

kes

ofde

bris

(RdS

0.5

m).

Eac

hpa

neld

ispl

ays

a(l

inea

r)su

per-

stre

tche

dco

ntra

stve

rsio

nof

the

tota

lim

age.

Aco

lor

inse

tof

the

disk

ofea

chga

laxy

(obt

aine

dfr

omda

tafr

omth

esa

me

tele

scop

eas

the

lum

inan

ceim

ages

)ha

sbe

enov

erpl

otte

dfo

rre

fere

nce

purp

oses

.In

addi

tion,

som

eof

the

orig

inal

imag

esw

ere

also

crop

ped

tobe

tter

show

the

mos

tin

tere

stin

gre

gion

sar

ound

each

targ

et.

with

3056

!30

56pi

xels

with

12!

12µ

m.

The

Moo

rook

(MrK

)ob

serv

ator

yus

esa

sim

ilar

CC

Dca

mer

a,th

eA

poge

eA

lta-K

AF1

6803

with

4096

!40

96pi

xels

and

asm

alle

rpi

xel

size

(9!

m).

The

tidal

stre

amde

tect

ion

stra

tegy

and

the

proc

edur

esus

edfo

rda

taac

quis

ition

and

redu

ctio

nof

the

data

are

the

sam

eas

thos

ede

scri

bed

inou

rpr

evio

uspa

pers

from

this

proj

ect

(Mar

tınez

-Del

gado

etal

.200

8,20

09).

Insu

mm

ary,

this

stra

tegy

striv

esfo

rm

ultip

lede

epex

posu

res

ofea

chta

rget

usin

ga

wid

eba

ndpa

ss,

high

thro

ughp

ut,

clea

rop

tical

filte

rw

itha

near

-IR

cuto

ff,a

lso

know

nas

alu

min

ance

filte

r,L.

The

typi

calc

umul

a-tiv

eex

posu

retim

esar

ein

the

rang

eof

6–11

hr.D

ata

redu

ctio

n

follo

wed

stan

dard

tech

niqu

esde

scri

bed

inth

eaf

ore-

men

tione

dpa

pers

.The

listo

ftar

gets

,tog

ethe

rwith

the

tele

scop

es,a

ndto

tal

expo

sure

time

used

inea

chca

sear

egi

ven

inTa

ble

2.Ph

otom

etri

cca

libra

tion

ofth

elu

min

ance

filte

r(L

)im

ages

isno

tcur

rent

lyav

aila

ble.

So,t

oas

sess

thei

rde

pth

and

typi

cal

qual

ityin

term

sof

back

grou

ndan

dfla

t-fie

ldin

g,w

ere

lied

onim

ages

ofsi

xof

our

gala

xies

—N

GC

1055

,N

GC

1084

,N

GC

3521

,NG

C42

16,N

GC

4651

,and

NG

C58

66—

obta

ined

byth

eSl

oan

Dig

italS

kySu

rvey

(SD

SS;Y

ork

etal

.200

0;D

ata

Rel

ease

7;A

baza

jian

etal

.200

9).

Bas

edon

SDSS

phot

omet

ry,

we

deriv

edph

otom

etri

ceq

ua-

tions

toco

nver

tth

eL-

band

coun

tsin

tog-

band

mag

nitu

des.

No.

4,20

10ST

EL

LA

RT

IDA

LST

RE

AM

SIN

TH

EL

OC

AL

VO

LU

ME

967

–22

Fig

ure

2.Il

lust

rativ

eco

mpa

riso

nof

som

eof

the

diff

eren

tfea

ture

sde

tect

edin

ourp

ilots

urve

yto

the

surv

ivin

gst

ruct

ures

visi

ble

inco

smol

ogic

alsi

mul

atio

nsof

Milk

yW

ay-l

ike

gala

xies

(see

Sect

ion

4).T

heth

ree

cent

ralp

anel

spr

ovid

ean

exte

rnal

pers

pect

ive

real

ized

thro

ugh

asi

mul

atio

n(f

rom

left

tori

ght,

halo

mod

els

num

bere

d9,

10,a

nd11

from

John

ston

etal

.200

8)w

ithin

the

hier

arch

ical

fram

ewor

kan

dsh

owlu

min

ous

stre

ams

resu

lting

from

tidal

lydi

srup

ted

sate

llite

s.E

ach

snap

shot

is30

0kp

con

asi

de.T

hetid

alfe

atur

esla

bele

din

the

snap

shot

sid

entif

yst

ruct

ures

sim

ilart

oth

ose

obse

rved

inou

rdat

a:A

,gre

atci

rcle

sfe

atur

es(M

essi

er63

);Sp

,spi

kes

(NG

C58

66);

U,u

mbr

ella

-sha

ped

stru

ctur

es(N

GC

4651

);P

D,p

artia

llydi

srup

ted

sate

llite

s(N

GC

4216

);G

P,g

iant

plum

es(N

GC

1084

);M

x,po

ssib

lym

ixed

-typ

est

ream

s(N

GC

1055

).(A

colo

rver

sion

ofth

isfig

ure

isav

aila

ble

inth

eon

line

jour

nal.)

and

grav

itatio

nalp

oten

tialo

fth

em

assi

veda

rkm

atte

rha

los

inw

hich

they

resi

de.T

his

will

ther

eby

offe

ra

uniq

ueop

port

unity

tost

udy

the

appa

rent

lyst

illdr

amat

icla

stst

ages

ofga

lact

icas

sem

bly

inth

eL

ocal

Uni

vers

e.In

this

rega

rd,t

hesu

rvey

will

beco

mpl

emen

tary

to(a

nddi

rect

lyin

form

the

inte

rpre

tatio

nof

)lo

calg

alac

tic“a

rcha

eolo

gica

l”da

tafr

omre

solv

edga

laxi

essu

chas

M31

and

the

Milk

yW

ay.

We

than

kK

.V

.Jo

hnst

onfo

rpr

ovid

ing

the

mod

els

used

inth

ispa

per

and

for

usef

uldi

scus

sion

.W

eal

soth

ank

G.

van

deV

enn,

Dav

idV

alls

-Gab

aud,

And

rew

Coo

per,

and

M.

A.

Gom

ez-F

lech

oso

foru

sefu

ldis

cuss

ions

.D.M

.-D.a

ckno

wle

dges

fund

ing

from

the

Span

ish

Min

istr

yof

Edu

catio

n(R

amon

yC

ajal

fello

wsh

ipan

dre

sear

chpr

ojec

tA

YA

2007

-650

90)

and

the

Inst

ituto

deA

stro

fisic

ade

Can

aria

s(p

roye

ct3I

).S.

R.M

.ap

prec

iate

ssu

ppor

tfro

mN

SFgr

antA

ST-0

8079

45.

RE

FER

EN

CE

S

Aba

zajia

n,K

.N.,

etal

.200

9,A

pJS,

182,

543

Bel

l,E

.,et

al.2

008,

ApJ

,680

,295

Bel

okur

ov,V

.,et

al.2

006,

ApJ

,642

,L13

7B

elok

urov

,V.,

etal

.200

7a,A

pJ,6

57,8

9B

elok

urov

,V.,

etal

.200

7b,A

pJ,6

58,3

37B

ullo

ck,J

.S.,

&Jo

hnst

on,K

.V.2

005,

ApJ

,635

,931

But

a,R

.198

7,A

pJS,

64,1

Col

e,S.

,Lac

ey,C

.G.,

Bau

gh,C

.M.,

&Fr

enk,

C.S

.200

0,M

NR

AS,

319,

168

Coo

per,

A.,

etal

.201

0,M

NR

AS,

406,

744

Forb

es,

D.

A.,

Bea

sley

,M

.A

.,B

ekki

,K

.,B

rodi

e,J.

P.,

&St

rade

r,J.

2003

,Sc

ienc

e,30

1,12

17G

rillm

air,

C.J

.200

6,A

pJ,6

51,L

29

Ibat

a,R

.,Ir

win

,M.,

Lew

is,G

.,Fe

rgus

on,A

.M.N

.,&

Tanv

ir,N

.200

1,N

atur

e,41

2,49

John

ston

,K.V

.,B

ullo

ck,J

.S.,

Shar

ma,

S.,F

ont,

A.,

Rob

erts

on,B

.E.,

&L

eitn

er,

S.N

.200

8,A

pJ,6

89,9

36Ju

ric,

M.,

etal

.200

8,A

pJ,6

73,8

64K

auff

man

n,G

.,W

hite

,S.D

.M.,

&G

uide

rdon

i,B

.199

3,M

NR

AS,

264,

201

Mal

in,D

.,&

Had

ley,

B.1

997,

PASA

,14,

52M

ajew

ski,

S.R

.,M

unn,

J.A

.,&

Haw

ley,

S.L

.199

6,A

pJ,4

59,L

73M

ajew

ski,

S.R

.,Sk

ruts

kie,

M.,

Wei

nber

g,M

.,&

Ost

heim

er,J

.200

3,A

pJ,5

99,

1082

Mar

tınez

-Del

gado

,D.,

Pena

rrub

ia,J

.,G

aban

y,R

.J.,

Truj

illo,

I.,M

ajew

ski,

S.R

.,&

Pohl

en,M

.200

8,A

pJ,6

89,1

84M

artın

ez-D

elga

do,

D.,

Pena

rrub

ia,

J.,

Juri

c,M

.,A

lfar

o,E

.J.

,&

Ivez

ic,

Z.

2007

,ApJ

,660

,126

4M

artın

ez-D

elga

do,D

.,Po

hlen

,M.,

Gab

any,

R.J

.,M

ajew

ski,

S.R

.,&

Palm

a,C

.20

09,A

pJ,6

92,9

55M

cCon

nach

ie,A

.W.,

etal

.200

9,N

atur

e,46

1,66

Mou

hcin

e,M

.,Ib

ata,

R.,

&R

ejku

ba,M

.201

0,A

pJ,7

14,1

2R

obai

na,

A.,

Bel

l,E

.F.

,va

nde

rW

el,

A.,

Som

ervi

lle,

R.

S.,

Skel

ton,

R.

E.,

McI

ntos

h,D

.H.,

Mei

senh

eim

er,K

.,&

Wol

f,C

.201

0,A

pJ,7

04,3

24R

ober

tson

,B.,

etal

.200

6,A

pJ,6

45,9

86R

ocha

-Pin

to,H

.J.,

Maj

ewsk

i,S.

R.,

Skru

tski

e,M

.F.,

Cra

ne,J

.D.,

&Pa

tters

on,

R.J

.200

4,A

pJ,6

15,7

32Sh

ang,

Z.,

etal

.199

8,A

pJ,5

04,L

23Sm

ith,A

.J.,

etal

.200

2,A

J,12

3,21

21To

omre

,A.,

&To

omre

,J.1

972,

ApJ

,178

,623

Tum

linso

n,J.

2010

,ApJ

,708

,139

8va

nde

rKru

it,P.

C.1

979,

A&

AS,

38,1

5V

oron

tsov

-Vel

yam

inov

,B.A

.195

9,A

tlas

and

Cat

alog

ofIn

tera

ctin

gG

alax

ies

(Mos

cow

:Ste

rnbe

rgIn

stitu

te,M

osco

wSt

ate

Uni

v.)

Wal

ter,

F.,B

rink

s,E

.,de

Blo

k,W

.J.G

.,B

igie

l,F.

,Ken

nicu

tt,R

.C.,

Tho

rnle

y,M

.D.,

&L

eroy

,A.2

008,

AJ,

136,

2563

Whi

te,S

.D.M

.,&

Fren

k,C

.S.1

991,

ApJ

,379

,52

Yor

k,D

.G.,

etal

.200

0,A

J,12

0,15

79Z

ibet

ti,S.

2009

,MN

RA

S,su

bmitt

ed(a

rXiv

:091

1.49

56)

Zib

etti,

S.,C

harl

ot,S

.,&

Rix

,H.-W

.200

9,M

NR

AS,

400,

1181

Martinez-Delgado et al. 2010

Outer halos of local disk galaxies

μg ~ 27.1

Page 18: Extragalactic Science with DECam Galaxy Evolution at z

Quenching - environment, mass, AGN

!"#$!%&!'()!*)+!,*-./#01!#1!2322!-4!,51./#014!0,!4/)66-*!7-44!-1+!)18#*017)1/9!!!:#/(!/()!8-65)4!!;!/0!!<!$#8)1!#1!'-=6)!>?!@60//)+!-/!#1/)*8-64!0,!A9>!+)B!#1!"!-1+! 9! ! !'()! 4)@-*-/#01! 0,! /()! ),,)./4! 0,! 7-44! -1+! )18#*017)1/! #4! !

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

/()!*)+E=65)!7#B!0,!$-6-B#)4!#1!2322!#4!#1+)@)1+)1/!0,!$-6-./#.!4/)66-*!7-44?!-1+!8#.)!8)*4-9!'(#4!$00+!)7@#*#.-6!4)@-*-=#6#/C!0,!7-44!-1+!)18#*017)1/!7)-14!/(-/!:)!.-1!:*#/)!/()!*)+!,*-./#01!#1!/)*74!0,! "!-1+! ?!=C!)#/()*!0,!/()!,#*4/!/:0!)F5-/#014?!:(#.(!*)+5.)!/0!/()!/(#*+&!! ! !

! ! ! ! H%I!!

:#/(! "! #1+)@)1+)1/!0,! ! -1+!:#/(! ! #1+)@)1+)1/!0,!"9!'(#4!#7@6#)4! -! 4#7@6)! 4C77)/*C! /0! /()! #$%&H ?!"I! 45*,-.)?! :(#.(! #4! !#6654/*-/)+!#1!"#$!%9! ! ! ! !2#1.)! !#4!D)*0!#1!/()!60:)4/!+)14#/C!*)$#014!H#9)9!/()!80#+4I?!

/(#4!4)@-*-=#6#/C!7)-14!/(-/! "H"I!#4!)-4#6C!#1/)*@*)/)+!-4!/()!*)+!,*-./#01!#1!/()4)!60:)4/!+)14#/C!*)$#0149!J#K):#4)?! H I!#4!/()! !

!

!"#$!L&!M4!,0*!"#$5*)!N?!=5/!,0*!/()!DOP2QP2!4-7@6)!-/!A9R!S!D!S!A9%9!

!!*)+!,*-./#01!,0*!8)*C!60:!7-44!$-6-B#)4?!,0*!:(#.(! 7!#4!=C!.01E4/*5./#01!D)*09!TC!#14)*/#1$!/()!/:0!,#//)+!*)6-/#014!HNI!#1/0!H%I?!:)!*).08)*! !!

! ! ! ! ! HLI!!

:(#.(!:-4!@*)8#0546C!@*0@04)+!=C!T-6+*C!)/!-69!H>AA%I!-4!01)!0,! /:0! )7@#*#.-6! ,#//#1$! ,51./#014! ,0*! /()! #$%&H ?!"I! 45*,-.)! #1!23229!'()!.6)-*!4)@-*-=#6#/C!0,!/()!),,)./4!0,!)18#*017)1/!-1+!7-44?!

:()1! @-*-7)/)*#D)+! #1! /(#4! :-C?! 45$$)4/4! /(-/! /()*)! -*)! /:0! !+#4/#1./!@*0.)44)4!-/!:0*K9!U)!:#66!()1.),0*/(!*),)*!/0!/()4)!-4!V)18#*017)1/EF5)1.(#1$V!-1+!V7-44EF5)1.(#1$V!/0!*),6)./!/()#*!H#1+)@)1+)1/I!),,)./4!01!#*)+!-.*044!/()!H ?!"I!@6-1)9!'()4)!/:0!F5)1.(#1$!@*0.)44)4!:#66!=)!$08)*1)+!=C! *-/)4! H#9)9! /()!@*0=-E=#6#/C!0,!=)#1$!F5)1.()+!@)*!$-6-BC!@)*!51#/!/#7)I!0,! !-1+! "!*)4@)./#8)6C9! ! !'()! +#4/#1./#01! =)/:))1! /()! /:0! ),,)./4! :#66! =)! )8)1! 70*)!

.6)-*6C!4))1!:()1!:)!.014#+)*!(0:?!0=4)*8-/#01-66C?! "!-1+! !+)@)1+! 01! .047#.! )@0.(9! "0*! /(#4! :)! /5*1! /0! 05*! DOP2QP2!4-7@6)!#1!/()!1)B/!2)./#019!!!

'()*+,-*./%*%012$,0"%0.345%06/207*,!%$8.%9*!

'()(:*;/%*%"!2$268<*92708.5$%*,#*%012$,0"%0.345%06/207*!"#$!L!4(0:4!/()!)F5#8-6)1/!@60/4!0,! "!-1+! !,*07!/()! !

#$%& H ?"I " H"?"AI H ? AI ; " H"?"AI

H ? AI " H"?"AI ; H ? AI

" "

#$%& H ?"I ; )B@ W !;!> " W !R

!<

Peng et al. 2010

slope 0:06 ! 0:09 dex per unit redshift ( jB rising with time).Taking the 1 ! minimum slope of "0.03 dex per unit redshift(falling with time) and assuming a minimum change in mass-to-light ratio M #/LB of 1.0 mag gives a rise in total red stellarmass of a factor of 2.3. The geometric mean of these two num-bers gives a minimum increase of a factor of 2 in the numberof red L# galaxies from z $ 1 to now, the same as claimed byB04.

To obtain the maximum rise, we take the formal differencein "# estimated in x 4.2 between z $ 0 and 1 using method 1 andall data (0:56 ! 0:09 dex) and add the 1 ! error bar. This yieldsa maximum rise of 0.65 dex, or a factor of 4.5. However, thismaximum still does not contain a correction for the likely grow-ing contamination by dusty galaxies at z % 1, which could addanother 30% or more, for a factor of about 6.

6. DISCUSSION

6.1. A ‘‘Mixed’’ Scenario for the Formationof Spheroidal Galaxies

The most arresting conclusion to emerge so far from thestudy of luminosity functions is the growth in the number of redL# galaxies since z $ 1, as shown in Figure 7. Barring mergersamong a large and undiscovered population, which would haveto be tiny and/or highly obscured to avoid detection in presentsurveys, this rise means that the ancestors of a large fraction oftoday’s red L# galaxies must be visible in existing blue samplesat z $ 1 and later. The implications of this were discussed byB04. We build on their arguments by adding data on the blueluminosity function (measured here by DEEP2 and COMBO-17for the first time) and the properties of local E galaxies, which weargue also have strong implications for formation scenarios. Ourdiscussion focuses on typical field galaxies at high redshift, sinceDEEP2 and COMBO-17 sample all galaxies regardless of loca-tion. The red sequence in distant clusters has also been exten-sively studied, but we do not try to fold these data into the pictureat this time.

It is well established that residence on the red sequence (inthe absence of dust) requires that star formation be quenched orat least strongly reduced. Stellar populations become red enoughto join the red sequence just 1Y2 Gyr after star formation isstopped (e.g., Newberry et al. 1990; Barger et al. 1996; Bower

et al. 1998; Poggianti et al. 1999), but in order for them to staythere, the star formation rate must remain low. Gebhardt et al.(2003) explored a ‘‘frosting model’’ with an early high rate ofstar formation followed by a slowly decaying # component.Based on colors, they found that only 7% of the total stellar masscould be formed in the # component; similar limits on present-day star formation rates are set byGALEX observations (Yi et al.2005; Salim et al. 2005). In short, the large buildup in red stellarmass after z $ 1 could not have arisen from star formation withinred galaxies themselves (see also B04). Rather, the stellar massnear L# on the red sequence must have migrated there via oneof three processes: (1) the quenching of star formation in bluegalaxies, (2) the merging of less luminous already quenched redgalaxies, or (3) some combination of the two. In the followingdiscussion, we focus on galaxies arriving on the red sequencenear L# because the data are complete there and photometricerrors are not as serious as they are for brighter spheroids. Gal-axies may of course also be migrating to the lower end of the redsequence, and the data reviewed in x 5 (point 8) suggest that thisis also happening.

It is helpful to visualize this mass migration as the move-ment of progenitor galaxies through the CMD or, more funda-mentally, the color-mass diagram. Sample tracks are shown inFigure 10. Two parent regions are illustrated, a narrow red locuscorresponding to the red sequence and a broader blue clump,which we call the ‘‘blue cloud.’’ The rather constant morphol-ogy of the CMD since z $ 1 (B04; Weiner et al. 2005; Paper I )suggests that these parent regions are relatively stable in sizeand location. In reality, they are also moving as galaxies evolve,but this will not be too important if individual galaxies movethrough them more rapidly. With that assumption, we show theclumps as fixed and the galaxies as moving through them withtime.

Each final galaxy today is represented by its most massiveprogenitor at any epoch. Stellar mass is migrating toward theupper left corner, where luminous red galaxies reside. For a gal-axy to get there, two things must happen: the mass composingthe final galaxy must be assembled via gravitational collapse,and star formation must be quenched. A key question in theformation of red sequence galaxies therefore is, did quench-ing occur early in the process of mass buildup, midway, or late?If extremely early, the pieces that would become the final galaxy

Fig. 10.—Schematic arrows showing galaxies migrating to the red sequence under different versions of the merging hypothesis. Evolutionary tracks are plotted in thecolor-mass diagram. Here it is assumed that red galaxies arise from blue galaxies when star formation is quenched during amajor merger, causing the galaxy to double inmass, but the exact nature of the quenching mechanism is not crucial. Quenching tracks are shown by the nearly vertical black arrows. The mergers would be gas-rich (or‘‘wet’’) because the progenitor galaxies are blue objects making stars and hence contain gas. Once a galaxy arrives on the red sequence, it may evolve more slowly alongit through a series of gas-poor, or ‘‘dry,’’ mergers. These are shown as the white arrows. They are tilted upward to reflect the aging of the stellar populations during themore gradual dry merging. A major variable is the time of mass assembly vs. the time of quenching. Three possibilities are shown. Track A represents very earlyquenching while the fragments of the galaxy are still small. In that case, most mass assembly occurs in dry mergers along the red sequence. Track B is the other extreme,having maximally late quenching. In that case, galaxies assemble most of their mass while still blue and then merge once to become red with no further dry merging.Track C is intermediate, with contributions from both mechanisms. This ‘‘mixed’’ scenario best matches the properties of both distant and local ellipticals. In addition tothe merging scenario illustrated here, the gas supply of some disks may simply be choked off or stripped out without mergers, to produce disky S0s. Such tracks wouldbe vertical, but aside from this their histories are similar. S0s dominate on the red sequence below L#, ellipticals above (Marinoni et al. 1999).

GALAXY LFs TO z%1 FROM DEEP2 AND COMBO-17 285No. 1, 2007

Faber et al. 2007

– 23 –

Fig. 4.— The distributions of the total volume-weighted [OIII] luminosity in Type 2 AGN

over log M!, log µ!, C, and Dn(4000). The description of the line types is the same as in Fig.1.

Page 19: Extragalactic Science with DECam Galaxy Evolution at z

Quenching - environment, mass, AGN

!"#$!%&!'()!*)+!,*-./#01!#1!2322!-4!,51./#014!0,!4/)66-*!7-44!-1+!)18#*017)1/9!!!:#/(!/()!8-65)4!!;!/0!!<!$#8)1!#1!'-=6)!>?!@60//)+!-/!#1/)*8-64!0,!A9>!+)B!#1!"!-1+! 9! ! !'()! 4)@-*-/#01! 0,! /()! ),,)./4! 0,! 7-44! -1+! )18#*017)1/! #4! !

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

/()!*)+E=65)!7#B!0,!$-6-B#)4!#1!2322!#4!#1+)@)1+)1/!0,!$-6-./#.!4/)66-*!7-44?!-1+!8#.)!8)*4-9!'(#4!$00+!)7@#*#.-6!4)@-*-=#6#/C!0,!7-44!-1+!)18#*017)1/!7)-14!/(-/!:)!.-1!:*#/)!/()!*)+!,*-./#01!#1!/)*74!0,! "!-1+! ?!=C!)#/()*!0,!/()!,#*4/!/:0!)F5-/#014?!:(#.(!*)+5.)!/0!/()!/(#*+&!! ! !

! ! ! ! H%I!!

:#/(! "! #1+)@)1+)1/!0,! ! -1+!:#/(! ! #1+)@)1+)1/!0,!"9!'(#4!#7@6#)4! -! 4#7@6)! 4C77)/*C! /0! /()! #$%&H ?!"I! 45*,-.)?! :(#.(! #4! !#6654/*-/)+!#1!"#$!%9! ! ! ! !2#1.)! !#4!D)*0!#1!/()!60:)4/!+)14#/C!*)$#014!H#9)9!/()!80#+4I?!

/(#4!4)@-*-=#6#/C!7)-14!/(-/! "H"I!#4!)-4#6C!#1/)*@*)/)+!-4!/()!*)+!,*-./#01!#1!/()4)!60:)4/!+)14#/C!*)$#0149!J#K):#4)?! H I!#4!/()! !

!

!"#$!L&!M4!,0*!"#$5*)!N?!=5/!,0*!/()!DOP2QP2!4-7@6)!-/!A9R!S!D!S!A9%9!

!!*)+!,*-./#01!,0*!8)*C!60:!7-44!$-6-B#)4?!,0*!:(#.(! 7!#4!=C!.01E4/*5./#01!D)*09!TC!#14)*/#1$!/()!/:0!,#//)+!*)6-/#014!HNI!#1/0!H%I?!:)!*).08)*! !!

! ! ! ! ! HLI!!

:(#.(!:-4!@*)8#0546C!@*0@04)+!=C!T-6+*C!)/!-69!H>AA%I!-4!01)!0,! /:0! )7@#*#.-6! ,#//#1$! ,51./#014! ,0*! /()! #$%&H ?!"I! 45*,-.)! #1!23229!'()!.6)-*!4)@-*-=#6#/C!0,!/()!),,)./4!0,!)18#*017)1/!-1+!7-44?!

:()1! @-*-7)/)*#D)+! #1! /(#4! :-C?! 45$$)4/4! /(-/! /()*)! -*)! /:0! !+#4/#1./!@*0.)44)4!-/!:0*K9!U)!:#66!()1.),0*/(!*),)*!/0!/()4)!-4!V)18#*017)1/EF5)1.(#1$V!-1+!V7-44EF5)1.(#1$V!/0!*),6)./!/()#*!H#1+)@)1+)1/I!),,)./4!01!#*)+!-.*044!/()!H ?!"I!@6-1)9!'()4)!/:0!F5)1.(#1$!@*0.)44)4!:#66!=)!$08)*1)+!=C! *-/)4! H#9)9! /()!@*0=-E=#6#/C!0,!=)#1$!F5)1.()+!@)*!$-6-BC!@)*!51#/!/#7)I!0,! !-1+! "!*)4@)./#8)6C9! ! !'()! +#4/#1./#01! =)/:))1! /()! /:0! ),,)./4! :#66! =)! )8)1! 70*)!

.6)-*6C!4))1!:()1!:)!.014#+)*!(0:?!0=4)*8-/#01-66C?! "!-1+! !+)@)1+! 01! .047#.! )@0.(9! "0*! /(#4! :)! /5*1! /0! 05*! DOP2QP2!4-7@6)!#1!/()!1)B/!2)./#019!!!

'()*+,-*./%*%012$,0"%0.345%06/207*,!%$8.%9*!

'()(:*;/%*%"!2$268<*92708.5$%*,#*%012$,0"%0.345%06/207*!"#$!L!4(0:4!/()!)F5#8-6)1/!@60/4!0,! "!-1+! !,*07!/()! !

#$%& H ?"I " H"?"AI H ? AI ; " H"?"AI

H ? AI " H"?"AI ; H ? AI

" "

#$%& H ?"I ; )B@ W !;!> " W !R

!<

Peng et al. 2010

slope 0:06 ! 0:09 dex per unit redshift ( jB rising with time).Taking the 1 ! minimum slope of "0.03 dex per unit redshift(falling with time) and assuming a minimum change in mass-to-light ratio M #/LB of 1.0 mag gives a rise in total red stellarmass of a factor of 2.3. The geometric mean of these two num-bers gives a minimum increase of a factor of 2 in the numberof red L# galaxies from z $ 1 to now, the same as claimed byB04.

To obtain the maximum rise, we take the formal differencein "# estimated in x 4.2 between z $ 0 and 1 using method 1 andall data (0:56 ! 0:09 dex) and add the 1 ! error bar. This yieldsa maximum rise of 0.65 dex, or a factor of 4.5. However, thismaximum still does not contain a correction for the likely grow-ing contamination by dusty galaxies at z % 1, which could addanother 30% or more, for a factor of about 6.

6. DISCUSSION

6.1. A ‘‘Mixed’’ Scenario for the Formationof Spheroidal Galaxies

The most arresting conclusion to emerge so far from thestudy of luminosity functions is the growth in the number of redL# galaxies since z $ 1, as shown in Figure 7. Barring mergersamong a large and undiscovered population, which would haveto be tiny and/or highly obscured to avoid detection in presentsurveys, this rise means that the ancestors of a large fraction oftoday’s red L# galaxies must be visible in existing blue samplesat z $ 1 and later. The implications of this were discussed byB04. We build on their arguments by adding data on the blueluminosity function (measured here by DEEP2 and COMBO-17for the first time) and the properties of local E galaxies, which weargue also have strong implications for formation scenarios. Ourdiscussion focuses on typical field galaxies at high redshift, sinceDEEP2 and COMBO-17 sample all galaxies regardless of loca-tion. The red sequence in distant clusters has also been exten-sively studied, but we do not try to fold these data into the pictureat this time.

It is well established that residence on the red sequence (inthe absence of dust) requires that star formation be quenched orat least strongly reduced. Stellar populations become red enoughto join the red sequence just 1Y2 Gyr after star formation isstopped (e.g., Newberry et al. 1990; Barger et al. 1996; Bower

et al. 1998; Poggianti et al. 1999), but in order for them to staythere, the star formation rate must remain low. Gebhardt et al.(2003) explored a ‘‘frosting model’’ with an early high rate ofstar formation followed by a slowly decaying # component.Based on colors, they found that only 7% of the total stellar masscould be formed in the # component; similar limits on present-day star formation rates are set byGALEX observations (Yi et al.2005; Salim et al. 2005). In short, the large buildup in red stellarmass after z $ 1 could not have arisen from star formation withinred galaxies themselves (see also B04). Rather, the stellar massnear L# on the red sequence must have migrated there via oneof three processes: (1) the quenching of star formation in bluegalaxies, (2) the merging of less luminous already quenched redgalaxies, or (3) some combination of the two. In the followingdiscussion, we focus on galaxies arriving on the red sequencenear L# because the data are complete there and photometricerrors are not as serious as they are for brighter spheroids. Gal-axies may of course also be migrating to the lower end of the redsequence, and the data reviewed in x 5 (point 8) suggest that thisis also happening.

It is helpful to visualize this mass migration as the move-ment of progenitor galaxies through the CMD or, more funda-mentally, the color-mass diagram. Sample tracks are shown inFigure 10. Two parent regions are illustrated, a narrow red locuscorresponding to the red sequence and a broader blue clump,which we call the ‘‘blue cloud.’’ The rather constant morphol-ogy of the CMD since z $ 1 (B04; Weiner et al. 2005; Paper I )suggests that these parent regions are relatively stable in sizeand location. In reality, they are also moving as galaxies evolve,but this will not be too important if individual galaxies movethrough them more rapidly. With that assumption, we show theclumps as fixed and the galaxies as moving through them withtime.

Each final galaxy today is represented by its most massiveprogenitor at any epoch. Stellar mass is migrating toward theupper left corner, where luminous red galaxies reside. For a gal-axy to get there, two things must happen: the mass composingthe final galaxy must be assembled via gravitational collapse,and star formation must be quenched. A key question in theformation of red sequence galaxies therefore is, did quench-ing occur early in the process of mass buildup, midway, or late?If extremely early, the pieces that would become the final galaxy

Fig. 10.—Schematic arrows showing galaxies migrating to the red sequence under different versions of the merging hypothesis. Evolutionary tracks are plotted in thecolor-mass diagram. Here it is assumed that red galaxies arise from blue galaxies when star formation is quenched during amajor merger, causing the galaxy to double inmass, but the exact nature of the quenching mechanism is not crucial. Quenching tracks are shown by the nearly vertical black arrows. The mergers would be gas-rich (or‘‘wet’’) because the progenitor galaxies are blue objects making stars and hence contain gas. Once a galaxy arrives on the red sequence, it may evolve more slowly alongit through a series of gas-poor, or ‘‘dry,’’ mergers. These are shown as the white arrows. They are tilted upward to reflect the aging of the stellar populations during themore gradual dry merging. A major variable is the time of mass assembly vs. the time of quenching. Three possibilities are shown. Track A represents very earlyquenching while the fragments of the galaxy are still small. In that case, most mass assembly occurs in dry mergers along the red sequence. Track B is the other extreme,having maximally late quenching. In that case, galaxies assemble most of their mass while still blue and then merge once to become red with no further dry merging.Track C is intermediate, with contributions from both mechanisms. This ‘‘mixed’’ scenario best matches the properties of both distant and local ellipticals. In addition tothe merging scenario illustrated here, the gas supply of some disks may simply be choked off or stripped out without mergers, to produce disky S0s. Such tracks wouldbe vertical, but aside from this their histories are similar. S0s dominate on the red sequence below L#, ellipticals above (Marinoni et al. 1999).

GALAXY LFs TO z%1 FROM DEEP2 AND COMBO-17 285No. 1, 2007

Faber et al. 2007

– 23 –

Fig. 4.— The distributions of the total volume-weighted [OIII] luminosity in Type 2 AGN

over log M!, log µ!, C, and Dn(4000). The description of the line types is the same as in Fig.1.

Heckman et al. 2004

SDSS Type2-AGN

Page 20: Extragalactic Science with DECam Galaxy Evolution at z

Quenching - environment, mass, AGN

– 23 –

Fig. 4.— The distributions of the total volume-weighted [OIII] luminosity in Type 2 AGN

over log M!, log µ!, C, and Dn(4000). The description of the line types is the same as in Fig.1.

Chapter 10: Active Galactic Nuclei

Figure 10.2: The probability of detecting an AGN as variable as a function of redshift and absolute magnitude. Left:two epochs separated by 30 days. Right: 12 epochs spanning a total of 360 days. Nearly all of the AGN between thelimiting apparent magnitudes would be detected as variable after one year.

all of the AGN to a limiting apparent magnitude of 24 will be detected as variable. The detectionfraction will increase as the number of epochs increases, and the use of all six bands will improvethe detection fraction even further. Ultimately, LSST will provide ∼ 200 epochs for each AGNcandidate in each band, thus increasing the detection fraction as well as increasing the limitingmagnitude.

The LSST temporal information will be especially useful for selecting low-luminosity AGN whichwould otherwise be swamped by their hosts, as well as radio-loud AGN, which have larger variabilityamplitudes and shorter variability timescales (e.g., Giveon et al. 1999). Variability will also allowselection of AGN which are confused with stars of similar color, particularly at z ∼ 2.7 where SDSSis highly incomplete. Variability timescales, coupled with LSST’s 6-band photometry will allowclean separation of AGNs from variable stars. For example, RR Lyrae stars have similar colors toAGNs but have very different variability timescales (Ivezic et al. 2003), and can thus efficiently beidentified as contaminants.

Selection by Combination with Multiwavelength Surveys

Cross-correlation of LSST imaging with multi-mission, multiwavelength surveys will also contributeto the AGN census by allowing selection of sources, such as optically obscured quasars, that cannoteasily be identified as AGN by color selection, lack of proper motion, or variability. LSST’s “deep-wide” nature will allow it to be combined both with shallower all-sky surveys at other wavelengthsin addition to having both the areal coverage and depth to be paired with the growing numberof multi-wavelength pencil beam surveys. For example, cross-correlations of LSST images withChandra or XMM-Newton observations can reveal obscured AGN that are not easily identifiablevia standard optical techniques; X-ray sources that have no LSST counterparts in any band may be

350

LSST Science Book

AGN variability is detectablefor i < 24 objects in 2-epoch, dt =30 days

what is correlation of AGN with stellar mass,

environment, color?

Page 21: Extragalactic Science with DECam Galaxy Evolution at z

van Dokkum 2005, 2010; LSST Science Book

Spheroid galaxy assembly

Chap

ter

9:G

alax

ies

Fig

ure

9.14

:Lef

t:M

USY

CU

VR

imag

eof

az

=0.

1ga

laxy

with

red

low

surf

ace

bri

ghtn

ess

feat

ure

sre

veal

ing

are

cent

inte

ract

ion

withou

tac

tive

star

form

atio

n(v

anD

okkum

2005

).T

his

imag

ere

aches

a1σ

surf

ace

bri

ghtn

ess

lim

itof

mag

29.5

arcs

ec−

2,a

good

mat

chto

the

expec

ted

LSST

dep

th.

Rig

ht:

SD

SS

gri

imag

eof

the

sam

e3�×

2�

pie

ceof

sky,

show

ing

that

thes

efe

ature

sar

enot

acce

ssib

leto

the

curr

ent

gener

atio

nof

full-s

ky

surv

eys

bey

ond

the

very

loca

lU

niv

erse

.

num

bers

ofgr

oups

afte

rm

aski

ngth

ega

laxi

esw

illen

able

the

mea

nha

lolig

htpr

ofile

tobe

trac

edto

very

low

surf

ace

brig

htne

ss.

Nov

aew

illpr

ovid

ea

uniq

uew

ayto

prob

edi

ffuse

light

.Sh

ara

(200

6)es

tim

ates

that

LSST

will

obta

ingo

odlig

htcu

rves

and

henc

edi

stan

cees

tim

ates

for∼

50tram

pno

vae

per

year

wit

hin

40M

pcif

the

diffu

sest

ella

rm

ass

is10

%of

the

stel

lar

mas

sin

gala

xies

.W

em

ight

cons

ider

putt

ing

one

ofth

eLS

STde

epdr

illin

gfie

lds

(§2.

1)on

ane

arby

clus

ter

ofga

laxi

essu

chas

Forn

ax.

If10

%of

the

tota

lst

ella

rm

ass

ofFo

rnax

(∼2.

1011

M⊙)

inin

trac

lust

erlig

ht,

and

we

obse

rved

it9

mon

ths

ofev

ery

year

,we

wou

lddi

scov

erro

ughl

y17

0in

tra-

clus

ter

nova

e.

9.7

Wid

eA

rea,M

ult

iband

Searches

for

Hig

h-R

edshift

Gala

xie

s

Dee

p,na

rrow

surv

eys

wit

hsp

ace-

born

ete

lesc

opes

have

iden

tifie

dne

wpo

pula

tion

sof

high

-red

shift

gala

xies

atre

dshi

fts

z>

5th

roug

hph

otom

etri

cdr

opou

tte

chni

ques

.W

hile

thes

eob

serv

atio

nal

effor

tsha

vere

volu

tion

ized

our

view

ofth

ehi

gh-r

edsh

iftU

nive

rse,

the

smal

lfiel

dsof

such

surv

eys

seve

rely

limit

thei

rco

nstr

aini

ngpo

wer

for

unde

rsta

ndin

gth

ebr

ight

end

ofth

ehi

gh-r

edsh

iftga

laxy

lum

inos

ityfu

ncti

onan

dfo

rid

enti

fyin

got

her

rare

obje

cts,

incl

udin

gth

em

ost

mas

sive

,old

est,

and

dust

iest

gala

xies

atea

chep

och.

By

com

bini

ngth

epo

wer

ofm

ulti

-ban

dph

otom

etry

for

drop

out

sele

ctio

nan

dth

eun

prec

eden

ted

com

bina

tion

ofw

ide

area

and

deep

imag

ing,

LSST

will

unco

verth

era

rest

high

-red

shift

gala

xies

(Fig

ure

9.5)

.T

hedi

scov

ery

and

char

acte

riza

tion

ofth

em

ost

mas

sive

gala

xies

athi

ghre

dshi

ftw

illpr

ovid

ene

wco

nstr

aint

son

earl

yhi

erar

chic

alst

ruct

ure

form

atio

nan

dw

illre

veal

the

gala

xyfo

rmat

ion

proc

ess

asso

ciat

edw

ith

high

-red

shift

quas

ars

(§10

.1.1

).

Obs

erva

tion

sof

i-dr

opou

tan

dz-

drop

out

gala

xies

inth

eH

ubbl

eU

ltra

Dee

pan

dG

OO

DS

field

sha

veen

able

da

dete

rmin

atio

nof

the

rest

-UV

lum

inos

ityfu

ncti

onof

z∼

6ga

laxi

es(e

.g.,

Dic

kins

onet

al.

2004

;B

unke

ret

al.

2004

;Y

an&

Win

dhor

st20

04;

Mal

hotr

aet

al.

2005

;Y

anet

al.

2006

;B

ouw

ens

etal

.200

4,20

06).

The

reis

asc

atte

rof

1-2

orde

rsof

mag

nitu

dein

dete

rmin

atio

nsof

the

brig

hten

dof

the

LF.F

igur

e9.

15sh

ows

the

gala

xyso

urce

coun

tsu

rfac

ede

nsit

ies

for

5.5

�z

�7

gala

xies

inth

ez-

band

calc

ulat

edfr

oma

rang

eof

z∼

6re

st-fr

ame

UV

lum

inos

ityfu

ncti

ons

take

nfr

omth

elit

erat

ure.

LSST

will

incr

ease

the

coun

tsof

gala

xyca

ndid

ates

atz

>5.

5by∼

5or

ders

ofm

agni

tude

.T

heLS

STsu

rvey

will

prob

eal

mos

tth

een

tire

lum

inos

ityra

nge

inth

isfig

ure

and

334

No. 2, 2010 THE GROWTH OF MASSIVE GALAXIES SINCE z = 2 1023

Figure 5. Top panels: stacked images of galaxies with constant number density in four redshift bins. Each image is 24!! " 24!!. The images reach surface brightnesslevels of #28.5 AB mag arcsec$2, and correspond to #3000 hr of total exposure time on a 4 m class telescope. Middle panels: deconvolved stacks, highlighting thefact that the radial extent of the low surface brightness emission decreases with redshift. Broken (solid) contours show the radii where the flux is 5% (0.5%) of thepeak flux. The 5% contour is similar at all redshifts, but the 0.5% contour evolves rapidly. Bottom panels: radial surface brightness profiles, normalized to the peakflux in the original stacks. Observed profiles are shown in blue, deconvolved profiles in red. The black curve is for stacked images of stars. The galaxies are resolvedat all redshifts, and are progressively smaller at higher redshifts.(A color version of this figure is available in the online journal.)

tens of effective radii for high-redshift galaxies. In practice, theaverage value of pixels with r > 75 kpc is subtracted from eachof the stacks. This procedure is very robust; bootstrapping thestacks (see Section 3.3) shows that the uncertainty in the back-ground correction is only a few percent. Finally, the images aredivided by the total flux in the image. The final stacks thereforehave a total flux of 1 within a 75 kpc radius aperture and a meanflux of zero outside of this aperture.

3.2. Surface Brightness Profiles

The observed stacks are shown in the top panels of Figure 5.There are no obvious residuals in the background, thanks tothe aggressive masking. The images are very deep: the surfacebrightness profiles can be traced to levels of #28.5 AB magarcsec$2 in the observed frame. For the z = 0.6 stack, theselevels are reached at radii of #70 kpc (#10!!); as we show laterthis corresponds to #10 effective radii. The depth is slightlylarger for the z = 0.6 and z = 1.1 stacks than for the higherredshift stacks: the Jx-band images are deeper than the Hx- andK-band data when expressed in AB magnitudes, and the ellipsefitting routine averages over more pixels for the low-redshiftgalaxies as they are more extended (as we show later).

The stellar PSF is fairly broad in this study, with a full widthat half-maximum (FWHM) of %1.!!1, and we first investigatewhether the observed stacks are resolved at this resolution.Radial surface brightness profiles of the stacked images areshown in blue in the bottom panels of Figure 5. Black curvesshow the profiles of stacked images of stars, derived from thesame data. The stars were identified based on their colors (seeK. Whitaker et al. 2010, in preparation) in a narrow magnituderange similar to the galaxies in the sample. They were shifted,masked, visually inspected, averaged, and normalized in thesame way as the galaxy images. The galaxy profiles and thestellar profiles were normalized to a peak flux of 1. Theblue curves are broader than the black curves at all redshifts,demonstrating that the galaxies are resolved.

To investigate the behavior of the galaxy profiles with red-shift the stacks were deconvolved using carefully constructedPSFs. The PSFs were created by averaging images of brightunsaturated stars, masking companion objects. The COSMOSand AEGIS fields have slightly different PSFs; for each stack aseparate PSF was constructed using the appropriate filters andappropriately weighting the PSFs of the two fields. As a test, werepeated the analysis using the stacked stellar images describedabove. Differences were small and not systematic; the differ-

No. 2, 2010 THE GROWTH OF MASSIVE GALAXIES SINCE z = 2 1023

Figure 5. Top panels: stacked images of galaxies with constant number density in four redshift bins. Each image is 24!! " 24!!. The images reach surface brightnesslevels of #28.5 AB mag arcsec$2, and correspond to #3000 hr of total exposure time on a 4 m class telescope. Middle panels: deconvolved stacks, highlighting thefact that the radial extent of the low surface brightness emission decreases with redshift. Broken (solid) contours show the radii where the flux is 5% (0.5%) of thepeak flux. The 5% contour is similar at all redshifts, but the 0.5% contour evolves rapidly. Bottom panels: radial surface brightness profiles, normalized to the peakflux in the original stacks. Observed profiles are shown in blue, deconvolved profiles in red. The black curve is for stacked images of stars. The galaxies are resolvedat all redshifts, and are progressively smaller at higher redshifts.(A color version of this figure is available in the online journal.)

tens of effective radii for high-redshift galaxies. In practice, theaverage value of pixels with r > 75 kpc is subtracted from eachof the stacks. This procedure is very robust; bootstrapping thestacks (see Section 3.3) shows that the uncertainty in the back-ground correction is only a few percent. Finally, the images aredivided by the total flux in the image. The final stacks thereforehave a total flux of 1 within a 75 kpc radius aperture and a meanflux of zero outside of this aperture.

3.2. Surface Brightness Profiles

The observed stacks are shown in the top panels of Figure 5.There are no obvious residuals in the background, thanks tothe aggressive masking. The images are very deep: the surfacebrightness profiles can be traced to levels of #28.5 AB magarcsec$2 in the observed frame. For the z = 0.6 stack, theselevels are reached at radii of #70 kpc (#10!!); as we show laterthis corresponds to #10 effective radii. The depth is slightlylarger for the z = 0.6 and z = 1.1 stacks than for the higherredshift stacks: the Jx-band images are deeper than the Hx- andK-band data when expressed in AB magnitudes, and the ellipsefitting routine averages over more pixels for the low-redshiftgalaxies as they are more extended (as we show later).

The stellar PSF is fairly broad in this study, with a full widthat half-maximum (FWHM) of %1.!!1, and we first investigatewhether the observed stacks are resolved at this resolution.Radial surface brightness profiles of the stacked images areshown in blue in the bottom panels of Figure 5. Black curvesshow the profiles of stacked images of stars, derived from thesame data. The stars were identified based on their colors (seeK. Whitaker et al. 2010, in preparation) in a narrow magnituderange similar to the galaxies in the sample. They were shifted,masked, visually inspected, averaged, and normalized in thesame way as the galaxy images. The galaxy profiles and thestellar profiles were normalized to a peak flux of 1. Theblue curves are broader than the black curves at all redshifts,demonstrating that the galaxies are resolved.

To investigate the behavior of the galaxy profiles with red-shift the stacks were deconvolved using carefully constructedPSFs. The PSFs were created by averaging images of brightunsaturated stars, masking companion objects. The COSMOSand AEGIS fields have slightly different PSFs; for each stack aseparate PSF was constructed using the appropriate filters andappropriately weighting the PSFs of the two fields. As a test, werepeated the analysis using the stacked stellar images describedabove. Differences were small and not systematic; the differ-

z=0.1 μB+V+I ~ 29

MUYSC

SDSS

spheroidal galaxies grow outer envelopes from z~2via ‘dry’ or minor mergers?

Page 22: Extragalactic Science with DECam Galaxy Evolution at z

CFHTLS Deep i-band exptime ~ 30 hours

μi ~ 29 mag/sq arcsec

~1500 interactions to z~1found in 2 sq degrees

Bridge et al. 2010

Distant major mergers

No. 2, 2010 CFHTLS-DEEP: GALAXY INTERACTION FRACTION AT z < 1.2 1071

Figure 1. Examples of galaxies classified to have tidal tails or bridges from the CFHTLS-Deep Catalog of Interacting Galaxies. The i!-band images (left) and colorcomposites (g!, r !, i!) range in class from galaxies with tidal tails (various lengths), close pairs with tidal bridges to double nuclei with tidal tails. Each stamp is 100 h"1

kpc on a side. The white circle marks the galaxy that has been classified. The XY axes are in arcseconds.

sensitivities of our survey. The results that follow are thereforesecure lower limits.

As an additional test, C.B. re-classified a set of 500 galaxies(350 interacting, 150 non-interacting) randomly selected. Thisblind self-test addresses the reproducibility of the authors own

classifications. The classifications remained the same 97% ofthe time. The 3% variation was primarily a result of galaxiesbeing classed into morphologically similar types, for example,an intermediate tidal tailed galaxy being classed as having ashort tidal tail, and not into grossly different types.

No. 2, 2010 CFHTLS-DEEP: GALAXY INTERACTION FRACTION AT z < 1.2 1071

Figure 1. Examples of galaxies classified to have tidal tails or bridges from the CFHTLS-Deep Catalog of Interacting Galaxies. The i!-band images (left) and colorcomposites (g!, r !, i!) range in class from galaxies with tidal tails (various lengths), close pairs with tidal bridges to double nuclei with tidal tails. Each stamp is 100 h"1

kpc on a side. The white circle marks the galaxy that has been classified. The XY axes are in arcseconds.

sensitivities of our survey. The results that follow are thereforesecure lower limits.

As an additional test, C.B. re-classified a set of 500 galaxies(350 interacting, 150 non-interacting) randomly selected. Thisblind self-test addresses the reproducibility of the authors own

classifications. The classifications remained the same 97% ofthe time. The 3% variation was primarily a result of galaxiesbeing classed into morphologically similar types, for example,an intermediate tidal tailed galaxy being classed as having ashort tidal tail, and not into grossly different types.

Page 23: Extragalactic Science with DECam Galaxy Evolution at z

Brightest Cluster Galaxy Formation

de Lucia & Blaziot 2007 Brown et al. 2007

why don’t very massive galaxies grow at z<0.7 ?

Page 24: Extragalactic Science with DECam Galaxy Evolution at z

No. 1, 2005 DIFFUSE LIGHT IN VIRGO CLUSTER L43

Fig. 2.—Diffuse light in the Virgo Cluster, color-coded by surface bright-ness: blue, ; yellow, ; green, ; and red,m p 25–26 m p 26–27 m p 27–28V V V

. North is up; east is to the left.m p 28–29V

Fig. 3.—Schematic diagram showing location of diffuse features discussedin the text.

tends beyond, toward a group of galaxies to the north. Thereis a surface brightness gradient along its length; near M87 ithas a surface brightness of , dropping tom p 27 m p 27.8V V

near the pair. From the point at which it emerges from M87’sstellar halo to where it fades into the background, the streamerhas total length of 38! (178 kpc), a characteristic width of 3!.5(16 kpc), and a total luminosity of !109 L,. Because NGC4458 and NGC 4461 have a very high velocity relative to oneanother ( km s!1), it is unlikely the streamer comesDv p 1296from any strong interaction between the pair; instead, strippingfrom one of the galaxies is more likely.The second streamer (B) is perhaps even more remarkable.

Also appearing to emanate from M87 to the northwest, it mea-sures approximately 28! (130 kpc) long and only 1!.2 (6 kpc)wide, with a peak surface brightness of . The lowm p 27.0V

surface brightness dwarf VCC 1149 is projected along thestream, just before the position at which the stream blends intoM87’s halo; we may well be seeing the destruction of thisdwarf in the tidal field of the cluster—a process that may con-tribute to the variation in the faint-end slope of the luminosityfunction in clusters. The thinness of the stream supports thisconjecture, as the low velocity dispersion of dwarf galaxiesensures that stripped tidal streams will be dynamically cold andnot diffuse significantly in the tangential direction. The thinnessand linearity of the streamer also argue that the dwarf must beon a highly radial orbit or one that is seen projected along itsorbital plane.Not all detected streams are linear, however. A broad, curved

stream (C) can be seen surrounding a low surface brightnessdwarf; the curvature of this stream, arcing away from M87,suggests that the galaxy is orbiting within or around the M86group (or perhaps the NGC 4435/4438 pair) rather than aroundM87 itself.Aside from long ICL streams, many galaxies show tidal

features on smaller scales. Perhaps the most obvious is the tidaltail to the north of the interacting pair NGC 4435/4438 (D).Originally identified by Malin (1994) in photographic imaging,our imaging traces it to fainter surface brightnesses ( ),m 1 28V

where it takes a sudden 90! bend to the west. This kind of tidal“dogleg” is expected during close and slow encounters in agalaxy cluster, where the tidal forces of both the individualgalaxies and the cluster itself conspire to create rather complex

tidal morphologies (J. C. Mihos 2005, in preparation). Othernotable features include NGC 4388’s tidal plumes to the westand northeast (E), a southern tail in NGC 4425 (F; both notedby Malin [1994]) that connects to VCC 987, another tail ex-tending from IC 3349 (G) and passing through VCC 942, anda loop to the east of IC 3475 (H).At low surface brightness, the giant elliptical M87 shows a

very extended and irregular stellar envelope. From photo-graphic images, Weil et al. (1997) and Katsiyannis et al. (1998)show M87’s luminous halo extending to !100 kpc scales; onour image it can be traced further out to 175 kpc (I). Here theouter isophotes are not simply the radial extension of M87’sinner regular luminosity profile; traced to the northwest, theisophotes become boxier and show irregular loops and shellsthat may be remnants from earlier stripping from, or canni-balization of, smaller galaxies. The lopsidedness and irregu-larity of M87’s halo is again indicative of a diffuse envelopethat is not fully relaxed.Because of the close proximity of M86, M84, and a number

of smaller galaxies, uniquely tracing out any individual haloin that complex of galaxies proves difficult. The isophotes ofboth M86 and M84 appear fairly regular out to a surface bright-ness of (J), as noted by Malin (1994). At fainterm ! 26.5V

surface brightnesses, the isophotes blend together and becomequite irregular, and we see a “crown” of diffuse light (at

) to the north of the M84/M86 system (K), and am p 27–28V

diffuse extension to the south of the pair, blending into thenetwork of galaxies to the south. Whether this extended en-velope is truly physical or simply a projection effect remainsunclear. Several low surface brightness filaments (L1–L3) arealso visible around M86: one to the northeast ( ), onem ! 25.5V

to the west ( ), and one to the north ( ).m ! 26.5 m ! 27.5V V

We also see a common envelope of light (M) surroundingthe complex of galaxies to the south of M86, including NGC4413, IC 3363, and IC 3349, along with a tidal bridge con-necting NGC 4413 and IC 3363, and perhaps extending to NGC4388 to the east. This envelope has fairly irregular isophotes,arguing it is more than just simply the extended profiles ofgalaxies overlapping in projection. This group of galaxies maybe a physically coherent subgroup within the Virgo Cluster.Finally, Oosterloo & van Gorkom (2005) have recently iden-

tified an extended ( kpc) H i cloud streaming110 kpc# 25from NGC 4388, which they attribute to ram pressure stripping,

IntraCluster Light in Virgo

Mihos et al. 2005

~2 degreesμV ~ 25–26 μV ~ 26–27μV ~ 27–28 μV ~ 28-29

Page 25: Extragalactic Science with DECam Galaxy Evolution at z

Tollerud et al. 2008 / Willman et al 2009/ LSST Science Book

Ultra-faint dwarf galaxy detectionChapter 7: Milky Way and Local Volume Structure

Figure 7.10: Far left: Map of all stars with r < 21.5 in the field around the Ursa Major I dwarf satellite, MV = −5.5,d =100 kpc. Middle: Map of stars passing a color-magnitude filter projected m −M = 20.0 Far right: Spatiallysmoothed number density map of the stars in the middle panel. The galaxy has a central V -band surface brightnessof only 27.5 mag arcsec−2 (Martin et al. 2008). Figure and caption from Willman (2009) with permission. Datafrom SDSS Data Release 7.

7.10 Stellar Tracers of Low-Surface Brightness Structure in theLocal Volume

Beth Willman, Jay Strader, Roelof de Jong, Rok Roskar

The unprecedented sensitivity to point sources and large sky coverage of LSST will for the first

time enable the use of resolved stellar populations to uniformly trace structure in and around a

complete sample of galaxies within the Local Volume. This section focuses on the science that can

be done by using resolved stars observed by LSST to discover and study low-surface brightness

stellar structures beyond the virial radius of the Milky Way. Chapter 6 of this Science Book

focuses on the science that can be done by using resolved stars observed by LSST to study stellar

populations. § 9.3 and § 9.6 of this Science Book focus on the study of low surface brightness

structures using diffuse light.

7.10.1 The Landscape of the Local Volume

The groups, galaxies, and voids that compose the Local Volume are the landscape that resolved

stars in LSST will be used to map. In this subsection, we provide an overview of this landscape.

We then detail specific studies in the remainder of this Section.

Karachentsev et al. (2004) (KK04) cataloged 451 galaxies within the Local Volume (d�10 Mpc),

hereafter referred to as the KK04 catalog. In a conference proceeding, Karachentsev et al. (2007)

report that the KK04 catalog has been updated to include 550 galaxies, half of which have been

imaged with HST and thus have distances measured with an accuracy of ∼ 8%. The searchable

volume reachable by LSST is expected to include over an order of magnitude more galaxies than

currently known in that volume (see also § 7.9 and § 9.6).

228

Ursa Major I dwarf satellite, MV!= −5.5, d =100 kpc

μV ~ 27.5

Page 26: Extragalactic Science with DECam Galaxy Evolution at z

Tollerud et al. 2008 / Willman et al 2009/ LSST Science Book

Ultra-faint dwarf galaxy detectionChapter 7: Milky Way and Local Volume Structure

Figure 7.10: Far left: Map of all stars with r < 21.5 in the field around the Ursa Major I dwarf satellite, MV = −5.5,d =100 kpc. Middle: Map of stars passing a color-magnitude filter projected m −M = 20.0 Far right: Spatiallysmoothed number density map of the stars in the middle panel. The galaxy has a central V -band surface brightnessof only 27.5 mag arcsec−2 (Martin et al. 2008). Figure and caption from Willman (2009) with permission. Datafrom SDSS Data Release 7.

7.10 Stellar Tracers of Low-Surface Brightness Structure in theLocal Volume

Beth Willman, Jay Strader, Roelof de Jong, Rok Roskar

The unprecedented sensitivity to point sources and large sky coverage of LSST will for the first

time enable the use of resolved stellar populations to uniformly trace structure in and around a

complete sample of galaxies within the Local Volume. This section focuses on the science that can

be done by using resolved stars observed by LSST to discover and study low-surface brightness

stellar structures beyond the virial radius of the Milky Way. Chapter 6 of this Science Book

focuses on the science that can be done by using resolved stars observed by LSST to study stellar

populations. § 9.3 and § 9.6 of this Science Book focus on the study of low surface brightness

structures using diffuse light.

7.10.1 The Landscape of the Local Volume

The groups, galaxies, and voids that compose the Local Volume are the landscape that resolved

stars in LSST will be used to map. In this subsection, we provide an overview of this landscape.

We then detail specific studies in the remainder of this Section.

Karachentsev et al. (2004) (KK04) cataloged 451 galaxies within the Local Volume (d�10 Mpc),

hereafter referred to as the KK04 catalog. In a conference proceeding, Karachentsev et al. (2007)

report that the KK04 catalog has been updated to include 550 galaxies, half of which have been

imaged with HST and thus have distances measured with an accuracy of ∼ 8%. The searchable

volume reachable by LSST is expected to include over an order of magnitude more galaxies than

currently known in that volume (see also § 7.9 and § 9.6).

228

Ursa Major I dwarf satellite, MV!= −5.5, d =100 kpc

μV ~ 27.5

Chapter 7: Milky Way and Local Volume Structure

Figure 7.9: Left panel: Maximum detection distance of dwarf galaxies in the SDSS Data Release 5 stellar catalogs,

and projected for future surveys. The dwarf galaxies known within 1 Mpc are overplotted in red. Right panel: The

predicted luminosity function of dwarf galaxies within 400 kpc of the Milky Way (blue line) over 4π steradians.

Overplotted are the expected number of these dwarfs that may be discovered over the entire sky with survey data

similar to the upcoming SkyMapper Southern Sky Survey (Keller et al. 2007) and LSST. The green line for LSST is

hiding behind the blue line. Both figures are from Tollerud et al. (2008), with permission.

predictions of such models (Wang & White 2007). Moreover, dwarf galaxy kinematic studies will

be useful in placing limits on (or measuring the existence of) a phase-space limited core in their

dark matter halos. This will provide an important constraint on the nature of dark matter. The

ability for dark matter to pack in phase space is limited by its intrinsic properties such as mass

and formation mechanism. CDM particles have negligible velocity dispersion and very large central

phase-space density, resulting in cuspy density profiles over observable scales. Warm Dark Matter

(WDM), in contrast, has smaller central phase-space density, so that density profiles saturate to

form constant central cores.

• Low-luminosity Threshold of Galaxy Formation: The discovery of dark-matter dominated galax-

ies that are less luminous than a star cluster raises several basic questions, including the possibility

of discovering a threshold luminosity for galaxy formation. LSST will enable discovering, enumer-

ating, and characterizing these objects, and in doing so provide a testing ground for the extreme

limits of galaxy formation.

• The Underlying Spatial Distribution of the Milky Way’s Dwarf Galaxy Population: The epoch

of reionization and its effect on the formation of stars in low mass dark matter halos also leaves

an imprint on both the spatial distribution (Willman et al. 2004; Busha et al. 2009) and mass

function of MW satellites (Strigari et al. 2007; Simon & Geha 2007). Other studies have claimed

that the spatial distribution of MW satellites is inconsistent with that expected in a Cold Dark

Matter-dominated model (Kroupa et al. 2005; Metz et al. 2008). The interpretation of such results

hinges critically on the uniformity of the MW census with direction and with distance.

226

search for ultra-faint dwarfs in southern sky

Page 27: Extragalactic Science with DECam Galaxy Evolution at z

Summary

‘sweet spot’ for low-surface brightness universe μ ~ 27-28 mag per sq arcsec (~10-20 hours of exptime)

distant galaxies -- stacking to examine outer halos

nearby galaxies/clusters -- deep, targeted observations

ultra-faint dwarfs -- blind search, deep follow-up