Precision Cosmology with

1
Precision Cosmology with
2
Another classic high-z QSO with Lyman-a and metal
lines
Rauch, Sargent Barlow
3
Mean Flux and Continuum Fitting of the z gt 2
Lyman-a Forest

• Probability of Transmission ltFgt ltObserved
  Flux / QSO Continuumgt
• Probability of Scattering from Lyman-a DA 1 -
  ltFgt between 1070-1150
• Long history of measurements, first direct
  observable of the forest
• Modern measurements use 2 techniques
• Direct Continuum Fitting (DCF) to high
  resolution, high S/N echelle spectra.
• e.g. Kirkman et al. 2005, Kim et al. 2001,
  McDonald et al. 2000
• Requires exhaustive treatment of individual
  spectra
• Currently limited to 10s of lines of sight per
  redshift bin
• Errors in DA at the 5 level
• Continuum Extrapolation (CE) to large samples of
  lower resolution samples
• E.g. Press, Rybiki, Schneider Bernardi et al.
  2003
• Requires well fluxed QSO samples, and a motivated
  model of continuum.
• Up to 100s of lines of sight, potential precision
  of 1

4
Bernardi et al 2003
5
(No Transcript)
6

• All measurements of ltFgt, DA depend on comparison
  to simulations
• So lets use the distribution of flux (PDF) as
  well as the mean,
• as suggested by Lidz. Et al 2005
• Advantages
• Does not require absolute continuum determination
• Trend removal (Croft et al. 98, Hui et al. 01)
  can be done on QSO by QSO
• basis. One does not require composite, nor
  complicated PCA procedures with large training
  samples.
• Disadvantages
• The observed PDF depends sensitvely to resolution
  and noise
• The PDF bins are highly correlated when
  normalized by trend removal.
• A correct model of the forest should agree with
  both the mean flux as well as the PDF

7
(No Transcript)
8
(No Transcript)
9
Lidz et al. 2005
10
K
K
11

• Some possible questions for discussion
• Can QSO lines of sight yield unbiased estimates
  of ltFgt?
• Do high column systems contribute excess metals
  beyond the average dN/dz?
• Is it possible to reconcile the current set of
  available measurements?
• Metals and high column systems are significant
  contributions
• What is the ultimate goal in precision for DA is
  1 attainable?
• How should we combine constraints from high and
  low resolution samples?

12
(No Transcript)
13
Correlation function of SDSS LRGs
Eisenstein et al. 2005
14
Dv
z
15
(No Transcript)
16
First attempt with DR3 QSOs, z gt 2.7
17
S2 x
Comoving distance (Mpc/h)
18
Jackknife estimates of 1-sigma error
19
From Ruperts simulation, 1000 LOS
20
2d image of flux correlation in velocity/redshift
z
100 - 5 Mpc/h
4.0
3.0
2.3
0
2200
10,000
17,000
Velocity separation (km/s)