GALICS – VI . Modelling hierarchical galaxy formation in clusters

High-resolution (HR) N-body resimulations of 15 massive (1014–1015 M ) dark matter (DM) haloes have been combined with the hybrid galaxy formation model GALICS to study the formation and evolution of galaxies in clusters, within the framework of the hierarchical merging scenario. This paper describes the HR resimulation technique used to build the DM halo sample and discusses its reliability. New features incorporated in GALICS include a better description of galaxy positioning after DM halo merger events, a more reliable computation of the temperature of the intergalactic medium as a function of redshift, which also takes into account the reionization history of the Universe, and a semi-analytic description of the ram pressure stripping of cold gas from galactic discs, suffered by galaxies during their motion through the diffuse hot intracluster medium. Within the multitude of available model results, we choose to focus here on the luminosity functions (LFs), morphological fractions and colour distributions of galaxies in clusters and in cluster outskirts, at z = 0. No systematic dependence on cluster richness is found either for the galaxy LFs, morphological mixes, or colour distributions. Moving from higher density (cluster cores) to lower density environments (cluster outskirts), we detect a progressive flattening of the LFs, an increase of the fraction of spirals and a decrease of that of ellipticals and S0s, and the progressive emergence of a bluer tail in the distributions of galaxy colours, especially for spirals. As compared with cluster spirals, early-type galaxies show a flatter LF, and more homogeneous and redder colours. An overall good agreement is found between our results and the observations, particularly in terms of the cluster LFs and morphological mixes. However, some discrepancies are also apparent, with too faint magnitudes of the brightest cluster members, especially in the B band, and galaxy colours tending to be too red (or not blue enough) in the model, with respect to the observations. Finally, ram pressure stripping appears to affect very little our results.