The Baryonic Acoustic Feature and Large-Scale Clustering in the SDSS LRG Sample

TL;DR

This study measures the large-scale two-point correlation function $\xi(s)$ of SDSS LRGs in DR7 to test ΛCDM predictions and characterize the baryonic acoustic feature. Using 160 LasDamas mock catalogs to derive a robust covariance and test systematics, the authors find DR7 clustering is broadly consistent with ΛCDM/HOD expectations, with a BAO-like peak persisting across subsamples. They quantify the BAO peak position $s_p$ and relate it to a cosmic distance ratio via $D_V$ and the sound horizon $r_s$, obtaining $r_s/D_V(0.278)=0.1394\pm0.0049$ and $D_V(0.278)=1099\pm38\,h^{-1}$Mpc, in agreement with prior analyses. The brighter DR7 subsample shows an enhanced large-scale signal that is not readily explained by the fiducial mocks, but calibration systematics cannot account for it. Overall, the DR7 LRG sample supports ΛCDM with a robust BAO signature, while highlighting sample variance and subsample dependence as important considerations for future, larger surveys.

Abstract

We examine the correlation function ξof the Sloan Digital Sky Survey (SDSS) Luminous Red Galaxy sample (LRG) at large scales (60<s<400 Mpc/h) using the final data release (DR7; 105,831 LRGs between 0.16<z<0.47). Using mock catalogs, we demonstrate that the observed baryonic acoustic peak and larger scale signal are consistent with LCDM at the 1.5σlevel. The signal at 155<s<200 Mpc/h tends to be high relative to theoretical expectations; this slight deviation can be attributed to a bright subsample of the LRGs. Fitting data to a non-linear, redshift-space, template based-model, we constrain the peak position at s_p=103.6+3.6-2.4 Mpc/h when fitting the range 60<s<150 Mpc/h (1σuncertainties measured from the mocks. This redshift-space distance s_p is related to the comoving sound horizon scale r_s after taking into account matter clustering non-linearities, redshift distortions and galaxy clustering bias. Mock catalogs show that the probability that a DR7-sized sample would not have an identifiable peak is at least 10%. As a consistency check of a fiducial cosmology, we use the observed s_p to obtain the distance D_V=[(1+z)^2D_A^2cz/H(z)]^(1/3) relative to the acoustic scale. We find r_s/D_V(z=0.278)=0.1394+-0.0049. This result is in excellent agreement with Percival et. al (2009), who examine roughly the same data set, but using the power spectrum. Comparison with other determinations in the literature are also in very good agreement. We have tested our results against a battery of possible systematic effects, finding all effects are smaller than our estimated sample variance.

Figures (19)

• Figure 1: Angular Selection Function : SDSS DR7 LRG sky coverage. For plotting purposes we present one tenth of the $105,831$ DR7-Full galaxies. The solid blue line is the Galactic plane.
• Figure 2: Radial selection function: Comoving number density $n(z)$ of the full DR7 (DR7-Full; black) and its subsamples DR7-Dim (green), DR7-Bright (blue), DR7-Bright2 (red) and DR3 (cyan).
• Figure 3: DR3 and DR7 $\xi(s)$: Our DR3 results (green diamonds) show excellent agreement on the scales investigated here with those published by eisenstein05b (red crosses and uncertainty bars). The remaining discrepancies are consistent with shot-noise in the random catalogs. The dashed lines are our results for DR7-Full which shows a stronger clustering signal at $135<s<180$$h^{-1}$Mpc . In the thick blue dashed line we used the same $\Omega_{M0}=0.3$ flat cosmology as the DR3 results, and thin black dashed line $\Omega_{M0}=0.25$. Both cosmologies agree very well at scales discussed here $80<s<200$$h^{-1}$Mpc. In our DR3 result we use number ratio of random to data points $r \sim 50$ and for DR7 $r \sim 15.6$.
• Figure 4: DR7 Subsample $\xi(s)$: Comparing DR7-Full (thick dashed line) to subsamples DR7-Dim (green diamonds), DR7-Bright (blue crosses) and DR7-Bright2 (red squares). DR7-Bright shows a stronger signal than the other samples on most scales. The peak position appears consistent for all subsamples. For the uncertainties of DR7-Dim and DR7-Bright please refer to Figures \ref{['dr7qvl']} and \ref{['dr7vl']}, respectively.
• Figure 5: DR7-Dim $\xi(s)$: Results from SDSS (black diamonds) and from the LasDamas mock catalogs. The mock mean $\overline{\xi}_{mock}$ is the green solid line and the uncertainties in the mean are small vertical green lines. The variance for one realization is presented by the gray bands: $68.2\%$ light gray ($1\sigma_{mock}$), and $95.4 \%$ by dark ($2\sigma_{mock}$). The blue dotted lines are the outermost result of all mocks in each separation bin (not one realization in particular). Top Inset: Same format as main figure extending the results with wider bins to larger scales. Bottom Inset: Significance of large-scale clustering - we average $\xi$ in bins $[130,400]$$h^{-1}$Mpc . The observed $\langle \xi \rangle$ (red thick dashed line) is clearly within $2\sigma$ of the mock realizations (black histogram), with a $\chi^2$ fitting on $10$ d.o.f yielding $\chi^2/d.o.f=0.72$. The mock mean result is the thin green dot-dashed line.