Phleps

Uploaded from authorPOINT
Views:
 
     
 

Presentation Description

No description available.

Comments

Presentation Transcript

SISCO meeting September 2005:New Science Report from the Edinburgh node: 

SISCO meeting September 2005: New Science Report from the Edinburgh node Stefanie Phleps* (IfA Edinburgh), John Peacock, Klaus Meisenheimer and Christian Wolf Edinburgh, September 15th 2005 * sp@roe.ac.uk

Latest results: 

Latest results 'Galaxy clustering from COMBO-17: The halo occupation distribution at andlt;zandgt;=0.6 ' (Phleps, Peacock, Meisenheimer and Wolf 2005) Paper (astro-ph/0506320) submitted to Aandamp;A June 2005, still waiting for referee report

The simplest description of structure: the two-point correlation function: 

The simplest description of structure: the two-point correlation function The two-point correlation function x(r) describes the excess probability of finding a galaxy at distance r from a galaxy selected at random over that expected in a uniform, random distribution usually parametrised by a power law with correlation length r0:

The galaxy samples in the present analysis: 

The galaxy samples in the present analysis 10360 galaxies with 0.4andlt;zandlt;0.8 Iandlt;23 MBandlt;-18 and Volume limited galaxy sample complete to MBandlt;-18

How to measure x(r): 

How to measure x(r) Simplest estimator: (andlt;DDandgt; is the (normalised) number of galaxy pairs per distance bin, andlt;RRandgt; is the number of random pairs) We use: Peebles 1980 Landy andamp; Szalay 1993

Slide6: 

Use COMBO17 data (multicolor redshifts) Large redshift errors, x(r) breaks down Calculate the projected correlation function w(rp) instead

The projected correlation function: 

The projected correlation function Definition: rp: projected distance between pairs of galaxies, p: distance parallel to the line of sight But: in practice impossible to integrate to infinity (finite redshift bin)! Not much of a problem if redshift errors are small (correlation function approximately a power law), but…

The influence of redshift errors: 

The influence of redshift errors Model x(rp,p), including peculiar velocities and coherent infall The same model convolved with the pairwise error distribution of the COMBO17 galaxies

And this is how it looks like for COMBO17:: 

And this is how it looks like for COMBO17: Impossible to integrate over pandgt;150 h-1 Mpc Want to minimize random noise in x(rp,p) at a given rp

Possible solution: 

Possible solution Integrating a fitted model instead Advantage: decreasing random noise and capturing entire signal Disadvantage: we don’t know the form of the correlation function a priori and want to make as much use of the available data as possible Find a compromise!

How to make use of the model: 

How to make use of the model Convolve it with pairwise redshift error distribution Fit the amplitude of each x(rp=const,p) to the data Integrate the data out to p=100 h-1 Mpc Then integrate the model

How to make use of the model: 

How to make use of the model Convolve it with pairwise redshift error distribution Fit the amplitude of each x(rp=const,p) to the data Integrate the data out to p=100 h-1 Mpc Then integrate the model

How to make use of the model: 

How to make use of the model Convolve it with pairwise redshift error distribution Fit the amplitude of each x(rp=const,p) to the data Integrate the data out to p=100 h-1 Mpc Then integrate the model

How to make use of the model: 

How to make use of the model Convolve it with pairwise redshift error distribution Fit the amplitude of each x(rp=const,p) to the data Integrate the data out to p=100 h-1 Mpc Then integrate the model

How to integrate x(rp,p): 

How to integrate x(rp,p)

How to integrate x(rp,p): 

How to integrate x(rp,p)

How to integrate x(rp,p): 

How to integrate x(rp,p)

How to integrate x(rp,p): 

How to integrate x(rp,p)

How to integrate x(rp,p): 

How to integrate x(rp,p)

The integral constraint: 

The integral constraint Mean galaxy density is determined from observed galaxy counts in each field Groth and Peebles 1977: estimator will be biased low with respect to the true correlation by a constant I: I can be estimated by Monte-Carlo integration and added to the measured w(rp)

What do we expect for w(rp)?: 

What do we expect for w(rp)? In the local universe the correlation function is featureless only on intermediate scales On large (andgt;20 h-1Mpc) scales a cutoff has been reported (Maddox et al.1990, Collins et al. 1992, Conolly et al. 2002) On very small scales a change of slope is expected (and in local samples also measured) (Zehavi et al. 2004a,b, Abazajian et al. 2004, Zengh et al. 2004)

Not too surprising…: 

Not too surprising… Two different regimes: On small scales, the correlation function is dominated by pairs of galaxies within one single dark matter halo On large scales, the clustering of the haloes themselves dominate (linear power spectrum) There must be a transition region, the shape and scale of which depends on the typical halo masses the galaxies occupy

Dark-matter haloes and bias: 

Dark-matter haloes and bias

The halo model of galaxy clustering: 

The halo model of galaxy clustering Mass function and density profiles of the haloes known from N-body simulations Clustering properties can be expressed analytically (for Gaussian density fields) Unknown: How the galaxies populate the haloes -andgt; subject of current investigation

The halo model of galaxy clustering: 

The halo model of galaxy clustering Mass function and density profiles of the haloes known from N-body simulations Clustering properties can be expressed analytically (for Gaussian density fields) Unknown: How the galaxies populate the haloes -andgt; subject of current investigation

The halo model of galaxy clustering: 

The halo model of galaxy clustering Mass function and density profiles of the haloes known from N-body simulations Clustering properties can be expressed analytically (for Gaussian density fields) Unknown: How the galaxies populate the haloes -andgt; subject of current investigation

A model of the correlation function: 

A model of the correlation function Gravitational clustering of dark matter determines the population of virialized dark matter halos, galaxy formation physics determines the Halo Occupation Distribution (HOD) Given cosmological parameters (Wm, WL, H0, s8) and a specified HOD, the correlation function can be calculated analytically (see Peacock and Smith 2000, Seljak 2000, 2002, Scoccimarro et al. 2001, Berlind et al. 2002, 2003, Takada andamp; Jain 2003, Yang, van den Bosch, Mo et al. 2002a, 2002b, 2003, Abazaijan 2004)

The Halo Occupation Distribution (HOD): 

The Halo Occupation Distribution (HOD) The Halo Occupation Number tells you the probability of finding N galaxies in a halo of mass M: N Ma for Mandgt;Mmin In a model where light traces mass exactly, Mmin=0 and a=1

Slide29: 

Simplest model: N = (M/Mmin)a only one free parameter if number density fixed logN M Not a bad approx to more complex models (e.g. Zheng et al. 2005) Mmin

Locally: Prediction matches correlation data: 

Locally: Prediction matches correlation data Zehavi et al. (astro-ph/0301280) Luminous SDSS galaxies need weight M0.11 for M andgt; Mmin= 1013.6

Our model (John A. Peacock): 

Our model (John A. Peacock) Original method of calculation as in Peacock and Smith 2000+simple HOD model (Flat LCDM with Wm=0.25, Wb=0.045,h=0.73 (H0=100 h km s-1Mpc-1) Use observed number density of galaxies as further constraint: only free parameter in the HOD is a (for a given value of a the observed number density determines the cutoff mass) Can leave s8 as free parameter!

Red and blue: 

Red and blue Divide the sample into Red sequence and blue cloud galaxies, following the definition of Bell et al. 2004, ApJ, 608, 752: Red sequence galaxies have (U-B) colours redder than

COMBO17: 

COMBO17 Blue galaxies Red galaxies

COMBO17: 

COMBO17 Model plotted for Red galaxies: a=0.45,0.5,0.55 Blue galaxies: a=0.15,0.2,0.25 For each a Mmin is chosen so as to match the observed comoving densities

The ratio of HOD model and best-fitting power-law: 

The ratio of HOD model and best-fitting power-law

Clustering fits at <z>=0.6: 

Clustering fits at andlt;zandgt;=0.6 0 0 2 2 1 1 1 8 a 0 red blue

Results from the fit (COMBO17):: 

Results from the fit (COMBO17): Red galaxies: a=0.528+-0.030 s8=0.853+-0.075 Mmin=1012.1 h-1 M0 Blue galaxies: a=0.077+-0.025 s8=1.197+-0.078 Mmin=1011.5 h-1 M0 Mean s8 = 1.01+-0.03

Local sample: 2dF and SDSS: 

Local sample: 2dF and SDSS Red galaxies Blue galaxies

Clustering fits at z=0: 

Clustering fits at z=0 0 0 2 2 1 1 1 8 a 0 red blue

Results from the fit (2dF and SDSS):: 

Results from the fit (2dF and SDSS): Red galaxies: a=0.528+-0.030 s8=0.853+-0.075 Mmin=1012.1 h-1 M0 Blue galaxies: a=0.077+-0.025 s8=1.197+-0.078 Mmin=1011.5 h-1 M0 Red galaxies: a=0.433+-0.034 s8=0.971+-0.084 Mmin=1012.1 h-1 M0 Blue galaxies: a=0.080+-0.012 s8=1.046+-0.030 Mmin=1011.4 h-1 M0 Mean s8 = 1.01+-0.03 Mean s8 = 1.02+-0.17 COMBO-17

The projected correlation function of red sequence galaxies: 

The projected correlation function of red sequence galaxies

The projected correlation function of blue cloud galaxies: 

The projected correlation function of blue cloud galaxies

Results: 

Results The correlation functions of both red and blue galaxies at z=0.6 display deviations from the pure power law, as observed at low redshift (Hawkins et al 2003; Zehavi et al. 2004) We fit the correlation function using a two-parameter HOD, leaving s8 and a as free parameters Measurement of local value of s8 ~1 from COMBO-17 is consistent with value measured from local data Confirms standard hierarchical growth

Future prospects: 

Future prospects Going deeper: COMBO17+4(Pi: K. Meisenheimer), a NIR extension of COMBO17: will include approximately 4200 galaxies with sz=0.02 up to z=2 wider area, more galaxies – VST16 (PIs: R. Bender; K. Meisenheimer, G. Busarello) : area approximately 20 times larger than COMBO17 (20 ), Expect approximately 600 000 galaxies between z=0.2 and z=1.2, with sz=0.02 Comparison with simulations (-andgt; Eelco)