Finding clusters in SZ surveys

We use simulated maps to investigate the ability of high resolution, low noise (roughly 10 μK-arcmin beam) surveys of the CMB to create catalogues of clusters of galaxies by detecting the characteristic signature imprinted by the Sunyaev Zel'dovich effect. We compute the completeness and efficiency of the catalogues in our simulations for several survey strategies, and find that while the primary CMB anisotropy does not degrade this effort, point source confusion is a more significant problem, which may necessitate the use of either multi-frequency observations or a single frequency at which the effect is likely to be relatively small. We also evaluate the relative merit of some Fourier and wavelet based filtering techniques, and find that simple frequency-space filtering is at least as good as wavelet techniques for the purpose of creating complete and efficient catalogues of clusters.