Anomalous SZ Contribution to 3 Year WMAP Data

We first show that the new WMAP 3 year data confirm the detection by Myers et al 2004 of an extended SZ signal centred on 606 Abell (ACO) clusters with richness class, R>1. Our results also show SZ decrements around APM and 2MASS groups at increased significance than previously detected. We then follow the approach of Lieu et al 2006 and compare the stacked WMAP results for the decrement in 31 clusters with ROSAT X-ray profiles where Lieu et al found on average less SZ decrement in the WMAP 1 year data than predicted. We confirm that in the 3 year data these same clusters again show less SZ decrement than the X-ray data predicts. We then analysed the WMAP results for the 38 X-ray clusters with OVRO/BIMA measured SZ decrements as presented by Bonamente et al 2006. We again find that the average decrement is measured to be significantly less (5.5 sigma) than predicted by the Chandra X-ray data. Thus while we confirm the original detection of an extended SZ effect by Myers et al 2004, these X-ray comparisons may now suggest that the central SZ amplitudes detected by WMAP may actually be lower than expected. One possible explanation is that there is contamination of the WMAP SZ signal by radio sources in the clusters but we argue that this appears implausible. We then consider the possibility that the SZ decrement has been lensed away by foreground galaxy groups. Such a model predicts that the SZ decrement should depend on cluster redshift. A reduction in the SZ decrement with redshift is suggested from the ACO cluster sample and also from comparing the samples of Lieu et al 2006 and Bonamente et al 06. However, the mass power-spectrum would require a far higher amplitude than currently expected if lensing was to explain the SZ deficit in high redshift clusters.