Cathrine Udnes: Gradient Statistics for Cosmic Microwave Background Analysis
There have been many reports on anomalies in the CMB from different research groups. In this work statistical tests using the gradients of the CMB temperature, applied to the seven year WMAP data, have been investigated. These statistical tests could detect anomalies and a preferred direction in the CMB by seeing if there is a particular direction the gradients align with or avoid. This preferred direction statistics was suggested by Bunn and Scott in 2000 for CMB analysis, and has previously been applied to the COBE data. Although no detection of a preferred direction was made an asymmetry in the northern and southern hemispheres were detected in the WMAP data. This asymmetry has previouly been detected by other research groups using other methods to analyse the CMB. These statistical tests were also applied to other templates to test their sensitivity, and how it would be affected by a dipole, foregrounds, and point sources. Also the Bianchi VIIh model and a model with primordial non-Gaussianity were investigated. Along with these statistical tests the power spectra of the gradients were also estimated using the MASTER algorithm previouly used to estimate the temperature and polarization power spectra. It was found that the positive gradient power spectrum has acoustic peaks similarily to the temperature power spectrum, and it has therefore been focused on here. The positive gradient power spectrum was compared to the E-mode polarization power spectrum to be used as a consistency check. This power spectrum was also used to find a power law between the strength of point sources and the frequency of observation.