We analyze the small-scale clustering in "MegaZ-LRG", a large photometric-redshift catalogue of Luminous Red Galaxies extracted from the imaging dataset of the Sloan Digital Sky Survey. MegaZ-LRG, presented in a companion paper, spans the redshift range 0.4 < z < 0.7 with an r.m.s. redshift error dz ~ 0.03(1+z), covering 5,914 deg^2 to map out a total cosmic volume 2.5 h^-3 Gpc^3. In this study we use 380,000 photometric redshifts to measure significant deviations from the canonical power-law fit to the angular correlation function in a series of narrow redshift slices, in which we construct volume-limited samples. These deviations are direct signatures of the manner in which these galaxies populate the underlying network of dark matter haloes. We cleanly delineate the separate contributions of the "1-halo" and "2-halo" clustering terms and fit our measurements by parameterizing the halo occupation distribution N(M) of the galaxies. Our results are successfully fit by a "central" galaxy contribution with a "soft" transition from zero to one galaxies, combined with a power-law "satellite" galaxy component, the slope of which is a strong function of galaxy luminosity. The large majority of galaxies are classified as central objects of their host dark matter haloes rather than satellites in more massive systems. The effective halo mass of MegaZ-LRG galaxies lies in the range log_10 (M_eff / h^-1 M_sol) = 13.61 - 13.8 (increasing with redshift, assuming large-scale normalization sigma_8 = 0.8) for corresponding number densities in the range n_g = 5.03 - 0.56 x 10^-4 h^3 Mpc^-3. Our results confirm the usefulness of the halo model for gaining physical insight into the patterns of galaxy clustering.

We describe the construction of MegaZ-LRG, a photometric redshift catalogue of over one million luminous red galaxies (LRGs) in the redshift range 0.4 < z < 0.7 with limiting magnitude i < 20. The catalogue is selected from the imaging data of the Sloan Digital Sky Survey Data Release 4. The 2dF-SDSS LRG and Quasar (2SLAQ) spectroscopic redshift catalogue of 13,000 intermediate-redshift LRGs provides a photometric redshift training set, allowing use of ANNz, a neural network-based photometric-redshift estimator. The rms photometric redshift accuracy obtained for an evaluation set selected from the 2SLAQ sample is sigma_z = 0.049 averaged over all galaxies, and sigma_z = 0.040 for a brighter subsample (i < 19.0). The catalogue is expected to contain ~5 per cent stellar contamination. The ANNz code is used to compute a refined star/galaxy probability based on a range of photometric parameters; this allows the contamination fraction to be reduced to 2 per cent with negligible loss of genuine galaxies. The MegaZ-LRG catalogue is publicly available on the World Wide Web from http://www.2slaq.info .

We analyze MegaZ-LRG, a photometric-redshift catalogue of Luminous Red Galaxies (LRGs) based on the imaging data of the Sloan Digital Sky Survey (SDSS) 4th Data Release. MegaZ-LRG, presented in a companion paper, contains 10^6 photometric redshifts derived with ANNz, an Artificial Neural Network method, constrained by a spectroscopic sub-sample of 13,000 galaxies obtained by the 2dF-SDSS LRG and Quasar (2SLAQ) survey. The catalogue spans the redshift range 0.4 < z < 0.7 with an r.m.s. redshift error ~ 0.03(1+z), covering 5,914 deg^2 to map out a total cosmic volume 2.5 h^-3 Gpc^3. In this study we use the most reliable 600,000 photometric redshifts to present the first cosmological parameter fits to galaxy angular power spectra from a photometric redshift survey. Combining the redshift slices with appropriate covariances, we determine best-fitting values for the matter and baryon densities of Omega_m h = 0.195 +/- 0.023 and Omega_b/Omega_m = 0.16 +/- 0.036 (with the Hubble parameter h = 0.75 and scalar index of primordial fluctuations n = 1 held fixed). These results are in agreement with and independent of the latest studies of the Cosmic Microwave Background radiation, and their precision is comparable to analyses of contemporary spectroscopic-redshift surveys. We perform an extensive series of tests which conclude that our power spectrum measurements are robust against potential systematic photometric errors in the catalogue. We conclude that photometric-redshift surveys are competitive with spectroscopic surveys for measuring cosmological parameters in the simplest vanilla models. Future deep imaging surveys have great potential for further improvement, provided that systematic errors can be controlled.

The popular halo model predicts that the power spectrum of the galaxy fluctuations is simply the sum of the large scale linear halo-halo power spectrum and the weighted power spectrum of the halo profile. Previous studies have derived halo parameters from the observed galaxy correlation function. Here we test the halo model directly for self-consistency with a minimal set of theoretical assumptions by utilising the 2dF Galaxy Redshift Survey (2dFGRS). We derive empirically the halo occupation and galaxy radial distributions in the haloes of the 2dF Percolation-Inferred Galaxy Group (2PIGG) catalogue. The mean halo occupation number is found to be well-fitted by a power-law, <N|M> ~ M^b, at high masses, with b = 1.05, 0.88, 0.99 for red, blue and all galaxies respectively (with 1-sigma errors of 15-19%). We find that the truncated NFW profile provides a good fit to the galaxy radial distributions, with concentration parameters c=3.9, 1.3, 2.4 for red, blue and all galaxies respectively (with 1-sigma errors of 8-15%). Adding the observed linear power spectrum to these results, we compare these empirical predictions of the halo model with the observed correlation functions for these same 2dF galaxy populations. We conclude that subject to some fine tuning it is an acceptable model for the two-point correlations. Our analysis also explains why the correlation function slope of the red galaxies is steeper than that of the blue galaxies. It is mainly due to the number of red and blue galaxies per halo, rather than the radial distribution within the haloes of the two galaxy species.

We introduce ANNz, a freely available software package for photometric redshift estimation using Artificial Neural Networks. ANNz learns the relation between photometry and redshift from an appropriate training set of galaxies for which the redshift is already known. Where a large and representative training set is available ANNz is a highly competitive tool when compared with traditional template-fitting methods. The ANNz package is demonstrated on the Sloan Digital Sky Survey Data Release 1, and for this particular data set the r.m.s. redshift error in the range 0 < z < 0.7 is 0.023. Non-ideal conditions (spectroscopic sets which are small, or which are brighter than the photometric set for which redshifts are required) are simulated and the impact on the photometric redshift accuracy assessed.