CROSS-CORRELATION IN THE HIGH-Z SKY FEDERICO BIANCHINI THE QUEST - - PowerPoint PPT Presentation

CROSS-CORRELATION IN THE HIGH-Z SKY

FEDERICO BIANCHINI

2

THE QUEST FOR Λ

Energy Density

z

z ∼ 1

Dark Energy is

• Learn about the Dark Energy/Gravity sector

3

THE QUEST FOR Λ

Energy Density

z

z ∼ 1

Dark Energy is

• Learn about the Dark Energy/Gravity sector

TOMOGRAPHY

4

THE QUEST FOR Λ

Energy Density

z

z ∼ 1

Dark Energy is

• Learn about the Dark Energy/Gravity sector
• Investigate Astrophysics

TOMOGRAPHY

5

THE QUEST FOR Λ

Energy Density

z

z ∼ 1

Dark Energy is

• Learn about the Dark Energy/Gravity sector
• Investigate Astrophysics
• Isolate Systematics

TOMOGRAPHY

6

A BRIEF HISTORY OF TIME

7

A BRIEF HISTORY OF TIME

• Initial conditions
• Particle content
• Reionization
• Cosmic acceleration

8

CMB LENSING IN A NUTSHELL

CMB photons are weakly gravitationally deflected by the intervening matter distribution during their cosmic journey

credits: ESA and Planck team

φ(ˆ n) = −2 Z χ∗ dχ fK(χ∗ − χ) fK(χ∗)fK(χ)Ψ(χˆ n; η0 − χ)

Growth of 
 structures Geometry

˜ X(ˆ n) = X(ˆ n + rφ(ˆ n))

B(ˆ n) (±2.5µK) T(ˆ n) (±350µK) E(ˆ n) (±25µK)

*no primordial B-modes

B(ˆ n) (±2.5µK) T(ˆ n) (±350µK) E(ˆ n) (±25µK)

Lensing convolves the unlensed CMB power spectra with CMB lensing power spectrum

11

LENSING IS SMOOTH

`

`(` + 1)CTT

`

/2⇡ [µK2]

AL

Lensing convolves the unlensed CMB power spectra with CMB lensing power spectrum

12

LENSING IS SMOOTH

`

`(` + 1)CTT

`

/2⇡ [µK2]

AL

IDEA

Lensing introduces statistical anisotropy, i.e. correlates previously uncorrelated multipoles

13

CMB LENSING RECONSTRUCTION

hX(l)Y ∗(l L)i = 0

Hu&Okamoto02
 Seljak&Zaldarriaga97

hX(l)Y ∗(l L)i / φ(L)

We can extract lensing by looking at the off-diagonal correlations between X and Y

METHOD

Lensing Potential Normalization Optimally-chosen weight function Filtered (data) maps
 X,Y ϵ [T,E,B] 


14

APPLICATIONS OF CMB LENSING

φCMB

ICγB

ytSZ

x

φCMB

κgal

δg

TCIB

{

• Neutrino masses
• Geometrical degeneracy
• Dark energy
• Reionization Ase-2τ

Azabajian+13 PlanckXIII(2015)

15

APPLICATIONS OF CMB LENSING

φCMB

ICγB

ytSZ

x

φCMB

κgal

δg

TCIB

{

• Neutrino masses
• Geometrical degeneracy
• Dark energy
• Reionization Ase-2τ
• Tracers bias
• Photo-z calibration
• Primordial non-Gaussianities
• Multiplicative bias calibration
• Intrinsic alignments
• Star formation history
• Tracers bias
• Redshift distribution estimation
• Warm-hot intergalactic medium
• Clusters environment
• Dark Matter studies
• Leverage for high-z astrophysical 


contribution to gamma sky

16

TWO SIDES OF THE SAME COIN

MATTER DISTRIBUTION Hosts galaxies Lens CMB photons

Light is a (biased) tracer of matter Lensing is insensitive to matter’s nature Bias

δg = b δ

Galaxy fluctuations Matter fluctuations

Galaxy
 fluctuations Matter
 fluctuations Light is a biased tracer of matter Lensing is insensitive to nature of matter

MATTER 
 DISTRIBUTION Lens CMB 
 photons Host 
 Galaxies

IDEA

17

SIGNAL MODELING

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

z

0.2 0.4 0.6 0.8

W κ

0.2 0.4 0.6 0.8

dN/dz

Observables trace DM fluctuations with different weightings
 Overlap between kernels means that there is a xc signal

Cg

⇥

= Z z∗ dz c H(z) χ2(z)W (z)b(z)dN dz P ⇣ k, z ⌘ ∝

δdata

κdata

=

δdata

δdata

= Cgg

• =

Z z∗ dz c H(z) χ2(z) h b(z)dN dz i2 P

• k, z
• ∝ b2σ2

8

bσ2

8

Lensing of background galaxies by foreground LSS induces an apparent clustering in the sky

• Modifies observed area
• (De-)magnify observed fluxes

18

MAGNIFICATION BIAS

log S250 [mJy]

log N(> S)

N(> S) ∝ S−α

α ' 3

δobs

g

(ˆ n) = δclust

g

(ˆ n) + δµ

g (ˆ

n)

WL limit

− − − − − − → δµ

g (ˆ

n) ∝ (α − 1)

19

THE INGREDIENTS

PlanckTM HerschelTM

• 9 bands in [30,857] GHz
• Beams FWHM in [30’ - 5’]
• Noise ~30/40 µK-arcmin
• SPIRE 250, 350, 500 µm
• Beams FWHM 18”, 26”, 36”
• Noise 5.8, 6.3, 6.8 mJy

CMB LSS

∗2009 − †2013

2015 Planck ˆ φWF

• 7e-05

6e-05

20

PLANCK CMB LENSING MAPS

`

T : φ = COBE : Planck

To test the robustness of results we make use of CMB lensing maps from both 2013 and 2015 Planck data releases

PlanckXVII (2013)
 PlanckXV (2015)

21

H-ATLAS GALAXY SAMPLE

• Selection criteria
• Baseline (flux-based)
• Gonzalez-Nuevo+12 


(flux + color)

• Tomographic photo-z bins
• k

S250µm > 35 mJy

S350µm > 3σ

S350µm/S250µm > 0.6 S500µm/S350µm > 0.4

zph ≥ 1.5 1.5 ≤ zph < 2.1 zph ≥ 2.1

fsky ' 0.01

Ng '90000

22

SUB-MM MAGIC

The sub-mm flux remains approximately constant for z > 1

Lapi+11

Photo-zs are estimated through template fitting assuming the 
 SED of SMM J1235-0102

Swinbank+10

Convergence NGP

• 0.411

Galaxies NGP

• 0.737

Convergence SGP

• 0.411

Galaxies SGP

• 0.737

Convergence G09

• 0.411

Galaxies G09

• 0.737

Convergence G12

• 0.411

Galaxies G12

• 0.737

Convergence G15

• 0.411

Galaxies G15

• 0.737

23

THE PATCHES

Sub-mm galaxies trace the peaks of matter density field and are denser in regions where the CMB convergence is enhanced

24

CORRELATION BY EYE

hδi

h i

z > 1.5 1.5 < z < 2.1 z > 2.1

Maps smoothed to ~ 1 degree scale (~ 30 Mpc at z = 2)

AND NOW DATA!

26

2013 VS 2015 DATA CROSS-SPECTRA

100 200 300 400 500 600 700 800

`

0.0 2.5 5.0 7.5

Cg

` (×10−7)

ˆ zph > 1.5 Theory b = 2.79, A = 1.65 Theory b = 2.80, A = 1.62 2013 × gGN12 M2013 2013 × gGN12 M2015 2015 × gGN12 M2015 2015 × g35mJy M2015

∼ 6 Mpc ∼ 50 Mpc

z ∼ 2

FB +16

ˆ CXY

L

= X

L0

( X PLM0B2

0Q0L0)−1PL0 ˜

CXY

• Hivon+01

h

i =

27

ERRORS & NULL-TESTS

100 200 300 400 500 600 700 800

ℓ

0.0 0.2 0.4 0.6 0.8 1.0 1.2

∆Cκg

ℓ (×10−7)

Planck Sims Real Herschel Real Planck Analytical

Cκg

ℓ

Cκg

• We use two sets of sims:
• 500 correlated Gaussian galaxy and CMB lensing maps
• 100 sims of CMB lensing released by Planck team
• Null test to validate power spectrum extraction pipeline
• Covariance matrices evaluation

χ2/d.o.f. = 7.2/7 → P.T.E. = 41% χ2/d.o.f. = 5.9/7 → P.T.E. = 55% FB+15

28

DATA TOMOGRAPHY

100 200 300 400 500 600 700 800

`

0.0 2.5 5.0 7.5

Cg

` (×10−7)

2015 × g35 mJy Theory b = 3.54, A = 1.45 Theory b = 2.89, A = 1.48 Theory b = 4.75, A = 1.37 ˆ zph ≥ 1.5 1.5 ≤ ˆ zph < 2.1 ˆ zph ≥ 2.1 100 200 300 400 500 600 700 800

`

0.0 0.8 1.6 2.4

Cgg

` (×10−6)

g35 mJy Theory b = 3.54 Theory b = 2.89 Theory b = 4.75 ˆ zph ≥ 1.5 1.5 ≤ ˆ zph < 2.1 ˆ zph ≥ 2.1

Null hypothesis (= no correlation between fields) is rejected at a significance between 10𝜏 to 22𝜏

FB+16

29

CONSTRAINTS FROM JOINT ANALYSIS

We introduce an amplitude parameter A that rescales theoretical cross- power spectrum and combine observed cross- and auto-power spectra

2.4 3.0 3.6 4.2 4.8 5.4

b

0.9 1.2 1.5 1.8 2.1

A

z > 1.5 1.5 < z < 2.1 z > 2.1

?

FB+16, Aversa+15

A > 1 @ 2-3 s ˆ Cκg

L = ACκg,th L

∝ Ab

30

TESTING ΛCDM

CMB lensing and galaxy clustering measurements can be combined to provide tests of GR

ˆ EG(`, z) = Γ Cg

`

Cgg

`

Pullen+16 Zhang+07 Giannantonio+16

Planck X SDSS SPT-SZ X DES Gravitational Slip Growth of structure

31

PHOTO-Z TEST

To test robustness against changes in photo-z templates, we redo the analysis adopting the SED of Pearson+13

2.4 3.2 4.0 4.8 5.6 6.4

b

0.6 0.9 1.2 1.5 1.8 2.1 2.4

A

z > 1.5 1.5 < z < 2.1 z > 2.1

A ~1 for high-z bin, however bias is higher than found by studies with similar galaxy samples (Xia+12; Viero+13; Hildebrandt+13)

32

THE POLARBEAR EXPERIMENT

• Dedicated CMB polarization

experiment

• Targeting large & small scales
• Located on Cerro Toco @ 5190 m 


in Atacama desert (Chile)

• High & dry
• Observable fsky up to 80%
• Off-axis Gregorian-Dragone design

(3.5 m primary mirror)

• 3.5’ FWHM beam @ 150 GHz
• 1274 superconducting TES

bolometers

ABS

33

PB-2 AND SIMONS ARRAY

• PB-2a: deploying 2017
• 95 GHz + 150 GHz
• 7588 detectors
• Simons Array
• 3 receivers (22762 detectors)
• 95 GHz (50%) + 150 GHz (33%) + 220 GHz (16%)

*picture taken in March 2017 Suzuki+15

34

PB X H-ATLAS: OVERLAP

Overlap in two patches
 RA12 and RA23 (SGP)
 fsky~0.0005

35

PB X H-ATLAS: OVERLAP

Overlap in two patches
 RA12 and RA23 (SGP)
 fsky~0.0005

POLARBEAR collaboration+14

36

PB X H-ATLAS: FORECASTS

101 102 103

Multipole `

N

500 1000 1500 2000

Multipole `

Cg

500 1000 1500 2000

Multipole `min

Expected S/N

• Cumulative S/N obtained by fixing lmax = 2000 and lowering lmin
• Expected overall S/N ~ 4 (~17 for SA) if lmin ~ 100/200

WRAP-UP

• XC science from detection regime to standard

cosmological probe

• Sub-mm galaxies trace DM fluctuations
• Ongoing effort with sub-orbital experiments

THANKS!

BACKUP SLIDES

40

H-ATLAS REDSHIFT DISTRIBUTIONS

Budavari+03; Pearson+13; FB+16

∼ N(0, [σ(1 + z)]2)

p(z|W) = p(z) Z dzphW(zph)p(zph|z)

σ ' 0.26

0.25 0.50 0.75

1.5 ≤ z < 2.1

p(z) SMM p(z) Pearson

0.2 0.4 0.6

z ≥ 2.1

p(z|W) SMM p(z|W) Pearson

1 2 3 4 5

z

0.00 0.25 0.50

z ≥ 1.5

W κ W(zph)

Arbitrary units

41

PHOTO-Z PDF

42

FURTHER TESTS

`

Cg1g2

(×10−6)

2.5 3.0 3.5 4.0 4.5 5.0 5.5

b

0.9 1.2 1.5 1.8 2.1 2.4

A

z > 1.5 1.5 < z < 2.1 z > 2.1

Cross-correlation of galaxy positions in the two redshift bins

Flux threshold at 350 μm: 3 —> 5σ

43

CORRELATION MATRICES

Corrkg