Abstract

We present results on the clustering of 282 068 galaxies in the Baryon Oscillation Spectroscopic Survey (BOSS) sample of massive galaxies with redshifts 0.4 z 0.7 which is part of the Sloan Digital Sky Survey III project. Our results cover a large range of scales from similar to 500 to similar to 90 h(-1) Mpc. We compare these estimates with the expectations of the flat Lambda cold dark matter (Lambda CDM) standard cosmological model with parameters compatible with Wilkinson Microwave Anisotropy Probe 7 data. We use the MultiDark cosmological simulation, one of the largest N-body runs presently available, together with a simple halo abundance matching technique, to estimate galaxy correlation functions, power spectra, abundance of subhaloes and galaxy biases. We find that the Lambda CDM model gives a reasonable description to the observed correlation functions at z approximate to 0.5, which is remarkably good agreement considering that the model, once matched to the observed abundance of BOSS galaxies, does not have any free parameters. However, we find a greater than or similar to 10 per cent deviation in the correlation functions for scales less than or similar to 1 and similar to 10-40 h(-1) Mpc. A more realistic abundance matching model and better statistics from upcoming observations are needed to clarify the situation. We also estimate that about 12 per cent of the 'galaxies' in the abundance-matched sample are satellites inhabiting central haloes with mass M greater than or similar to 10(14) h(-1) M-circle dot. Using the MultiDark simulation, we also study the real-space halo bias b of the matched catalogue finding that b = 2.00 +/- 0.07 at large scales, consistent with the one obtained using the measured BOSS-projected correlation function. Furthermore, the linear large-scale bias, defined using the extrapolated linear matter power spectrum, depends on the number density n of the abundance-matched sample as b = -0.048 - (0.594 +/- 0.02)log(10)(n/h(3) Mpc(-3)). Extrapolating these results to baryon acoustic oscillation scales, we measure a scale-dependent damping of the acoustic signal produced by non-linear evolution that leads to similar to 2-4 per cent dips at greater than or similar to 3 sigma level for wavenumbers k greater than or similar to 0.1 h Mpc(-1) in the linear large-scale bias.