Binned cosmic microwave background anisotropy power spectra: Peak location

Abstract

We use weighted-mean and median statistics techniques to combine individual cosmic microwave background (CMB) anisotropy detections and determine binned, multipole-space, CMB anisotropy power spectra. The resultant power spectra are peaked. The derived weighted-mean CMB anisotropy power spectrum is not a good representation of the individual measurements in a number of multipole-space bins, if the CMB anisotropy is Gaussian and correlations between individual measurements are small. This could mean that some observational error bars are underestimated, possibly as a consequence of undetected systematic effects. Discarding the most discrepant 5% of the measurements alleviates but does not completely resolve this problem. The median statistics power spectrum of this culled data set is not as constraining as the weighted-mean power spectrum. Nevertheless, it indicates that there is more power at multipoles ℓ ~ 150-250 than is expected in an open cold dark matter (CDM) model, and it is more consistent with a flat CDM model. Unlike the weighted-mean power spectrum, the median statistics power spectrum at ℓ ~ 400-500 does not exclude a second peak in the flat CDM model.