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 clusters with richness class, R 2. We also detect SZ decrements around APM and 2MASS groups at increased significance than previously. 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 also 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 (5.5σ) than predicted by the Chandra X-ray data. These X-ray data cause us to re-interpret our previous result; rather than seeing too much SZ at large scales, we may actually be seeing too little SZ decrement near the cluster centre. 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. This effect is clearly detected in the ACO cluster sample and also by comparing the samples of Lieu et al. (2006) and Bonamente et al. (2006). But 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.