Reconstructing Dark Energy

TL;DR

This paper surveys model-independent reconstruction of the cosmological expansion history and the nature of dark energy, framing the problem through both physical and geometrical DE. It divides reconstruction into parametric and non-parametric approaches, detailing how observables such as $D_L(z)$ and $H(z)$ relate to the DE density and equation of state, and how growth data can provide cross-checks. It introduces robust diagnostics—the $w$-probe and the statefinder pair $(r,s)$—to distinguish DE models beyond the EOS, while highlighting the sensitivity of $w(z)$ to priors like $Ω_m$ and the pitfalls of overfitting. The review finds that current data favor a cosmological constant but permit modest evolution of dark energy, and it outlines future directions where improved measurements and model-independent techniques could uncover new physics governing cosmic acceleration.

Abstract

This review summarizes recent attempts to reconstruct the expansion history of the Universe and to probe the nature of dark energy. Reconstruction methods can be broadly classified into parametric and non-parametric approaches. It is encouraging that, even with the limited observational data currently available, different approaches give consistent results for the reconstruction of the Hubble parameter $H(z)$ and the effective equation of state $w(z)$ of dark energy. Model independent reconstruction using current data allows for modest evolution of dark energy density with redshift. However, a cosmological constant (= dark energy with a constant energy density) remains an excellent fit to the data. Some pitfalls to be guarded against during cosmological reconstruction are summarized and future directions for the model independent reconstruction of dark energy are explored.

Figures (7)

• Figure 1: An early reconstruction of the supernova data (left panel) using the parametric fit (\ref{['eq:taylor']}) shows an evolving equation of state to be marginally preferred over the cosmological constant alam04b. The same data were independently analyzed by means of a non-parametric ansatz (right panel) huterer05. It is encouraging that both reconstructions appear to give similar results for $0 < z \hbox{$\buildrel < \over \sim$} ~ 1$ where most of the data points lie. The crossing of the so-called 'phantom divide' at $w=-1$ led to much theoretical interest and considerable model building activity. More recent SNe results, together with constraints from other measurements such as the CMB, LSS and Baryon acoustic oscillations imply less pronounced evolution of $w(z)$ with redshift as shown in figure \ref{['fig:EOS']}, see also leandrostirth05gongwangteg05daly05huterer05paddy06zhao06otto06hao_wei. The left panel is from Alam, Sahni and Starobinsky alam04b while the right panel is from Huterer and Coorey huterer05.
• Figure 2: The equation of state of dark energy $w(z)$ reconstructed using the WMAP 3 year data + 157 "gold" SNIa data + SDSS. Median (central line), 68%(inner, dark grey) and 95%(outer, light grey) intervals. The two parameter fit (\ref{['eq:cpl']}) has been used in this exercise. From Zhao et al.zhao06.
• Figure 3: Dark energy density $\rho_X(z)$ reconstructed using SN Ia data goldsnls, combined with the WMAP 3 year data, the SDSS baryon acoustic oscillation data, and the 2dF linear growth data for density perturbations. The 68% (shaded) and 95% confidence contours are shown. Beyond $z_{\rm cut}$=1.4 (upper panel) and 1.01 (lower panel), $\rho_X(z)$ is parametrized by a power law $(1+z)^\alpha$. The horizontal dashed line shows the unevolving density associated with a cosmological constant. From Wang and Mukherjee wangm06.
• Figure 4: Pitfalls in cosmological reconstruction 1. This figure from Maor et al.maor02 shows the results of a reconstruction exercise performed by assuming: (i) $w_{\rm Q}$ = constant as a prior, which gives the larger lower contour with $w_{\rm Q} < -1$; (ii) the additional constraint $w_{\rm Q} \geq -1$, results in the smaller upper contour with $w_{\rm Q} = -1$ as the best fit. Both (i) and (ii) give confidence contours and best fit values of $w_{\rm Q}$ and $\Omega_m$ which are widely off the mark since they differ from the fiducial Quintessence model which has $w_{\rm Q}(z) = -0.7 + 0.8 z$ and $\Omega_m = 0.3$.
• Figure 5: Pitfalls in cosmological reconstruction 2. The reconstructed equation of state $w(z)$ is shown for $1000$ realizations of an $\Omega_m=0.3, w=-1$, $\Lambda$CDM model assuming SNAP quality data. An incorrect value for the matter density, $\Omega_{\rm m 0}=0.2$, is assumed in the reconstruction exercise which uses the polynomial ansatz (\ref{['eq:taylor']}). The dashed line represents the fiducial $\Lambda$CDM model with $w=-1$ while the solid lines show the mean value of the (incorrectly) reconstructed $w(z)$ and $1\sigma$ confidence levels around the mean. Note that the reconstructed EOS excludes the fiducial $\Lambda$CDM model to a high degree of confidence. From Shafieloo et al.arman.