Dark Energy
Dark energy affects the cosmic history of the Hubble expansion H(z) as well as the cosmic history of mass clustering. If combined, different types of probes of the expansion history and structure history can lead to percent level precision in dark energy parameters. This is because each probe depends on the other cosmological parameters or errors in different weak lensing ways. These probes range from cosmic shear, baryon acoustic oscillations, supernovae, and cluster counting -- all as a function of redshift. Using the CMB as normalization, the combination of these probes will yield the needed precision to distinguish between models of dark energy. What is required is a facility which can undertake all of these probes with deep data over wide area with cross checks to control systematic error.
Due to its unique high etendue, the LSST survey will produce all of these complementary probes of dark energy from the same survey data. Weak gravitational lensing is sensitive to angular diameter distance as a function of redshift and to trends in mass clustering with redshift. Several different probes -- each with its own different sensitivity to these effects - may be undertaken with the same deep weak lens data. These same imaging data (with precision photometric redshifts) will track the baryon acoustic oscillations over cosmic time. Counts of giant clusters of dark matter from these data provide another complementary probe. Finally, the tens of thousands of supernovae out to z=1 provide yet another cross check. When combined with the cosmic microwave background anisotropy data these tests form interlocking checks on cosmological models and the physics of dark energy. These tests are described in the above subpages, along with the program for precision photometric redshifts. Below we describe the resulting precision in 2-parameter dark energy model space enabled by the LSST survey.
More information can be found from the Dark Energy Task Force Whitepaper [PDF, HTML].
LSST precision on Dark Enery Parameters
Top Panel:
Marginalized 1σ errors on the comoving distance (open triangles) and
growth factor (open circles) from a joint analysis of LSST BAO
and WL with a conservative level of systematic uncertainties in the photo-z
error distribution and additive and multiplicative errors in the shear and
galaxy power spectra (see Zhan, Knox, & Tyson 2008 for more details). CMB
priors from Planck are included. The maximum multipole used for WL is
2000, and that for BAO is 3000 [with the additional requirement
Δδ2( l / DA ; z) < 0.4].
The growth parameters, G0 …
D14, are evenly spaced in log(1+z) between
z = 0 and 5, and the distance parameters, D1 …
D14, start at z1 = 0.14. The error
of each distance (growth) parameter is marginalized over all the other
parameters including growth (distance) parameters and other distance
(growth) parameters. The joint constraints on distance are relatively
insensitive to the assumed systematics.
Bottom Panel:
Forecasts of the errors on the dark energy equation-of-state parameters w0
and wa for WL (solid blue line), BAO (dashed magenta line),
cluster counting (dot-dashed green line), supernovae (dotted red line),
joint analysis of WL and BAO (light-green shaded area), and all combined
(gold shaded area). The constraints are marginalized over 9
other cosmological parameters including the curvature and over 120
parameters that model the linear galaxy clustering bias, photometric
redshift bias, and rms photometric redshift error. We take priors
σP(δ z) =
σP(σz)/sqrt(2) = 0.05
σz with σz = 0.05(1+z), which correspond to a
calibration requirement of 400 spectra fairly sampled at 0.1 redshift intervals
if the photo-z errors are Gaussian.
The linear galaxy bias b is allowed to float. We have included a
scale-independent but redshift-bin-dependent additive systematic noise
(power) of 10-8 to BAO angular power spectra and shear noise (power) of
10-9 to WL shear power spectra. The shear multiplicative error of each bin
is assumed to be known within 0.5%. See Zhan (2006, astro-ph/0605696) and
Zhan et al. (2008ab, arXiv:0801.3659 & arXiv:0806.0937) for details. The
LSST cluster counting (dN/dz) results are from Fang & Haiman (2007).