IDENTIFICATION OF MEMBERS IN THE CENTRAL AND OUTER REGIONS OF GALAXY CLUSTERS

The caustic technique measures the mass of galaxy clusters in both their virial and infall regions and, as a byproduct, yields the list of cluster galaxy members. Here we use 100 galaxy clusters with mass M-200 >= 10(14) h(-1) M-circle dot extracted from a cosmological N-body simulation of a Lambda CDM universe to test the ability of the caustic technique to identify the cluster galaxy members. We identify the true three-dimensional members as the gravitationally bound galaxies. The caustic technique uses the caustic location in the redshift diagram to separate the cluster members from the interlopers. We apply the technique to mock catalogs containing 1000 galaxies in the field of view of 12 h(-1) Mpc on a side at the cluster location. On average, this sample size roughly corresponds to 180 real galaxy members within 3r(200), similar to recent redshift surveys of cluster regions. The caustic technique yields a completeness, the fraction of identified true members, f(c) = 0.95 +/- 0.03, within 3r(200). The contamination, the fraction of interlopers in the observed catalog of members, increases from f(i) = 0.020(-0.015)(+0.046) at r(200) to fi = 0.08(-0.05)(+0.11) at 3r(200). No other technique for the identification of the members of a galaxy cluster provides such large completeness and small contamination at these large radii. The caustic technique assumes spherical symmetry and the asphericity of the cluster is responsible for most of the spread of the completeness and the contamination. By applying the technique to an approximately spherical system obtained by stacking the individual clusters, the spreads decrease by at least a factor of two. We finally estimate the cluster mass within 3r200 after removing the interlopers: for individual clusters, the mass estimated with the virial theorem is unbiased and within 30% of the actual mass; this spread decreases to less than 10% for the spherically symmetric stacked cluster.