Gaussianity and statistical isotropy of the Universe are modern cosmology’s minimal set of hypotheses. In this work we introduce a new statistical test to detect observational deviations from this minimal set. By defining the temperature correlation function over the whole celestial sphere, we are able to independently quantify both angular and planar dependence (modulations) of the CMB temperature power spectrum over different slices of this sphere. Given that planar dependence leads to further modulations of the usual angular power spectrum $C_{\ell }$, this test can potentially reveal richer structures in the morphology of the primordial temperature field. We have also constructed an unbiased estimator for this angular-planar power spectrum which naturally generalizes the estimator for the usual $C_{\ell }$’s. With the help of a chi-square analysis, we have used this estimator to search for observational deviations of statistical isotropy in WMAP’s 5 year release data set (ILC5), where we found only slight anomalies on the angular scales $\ell =7$ and $\ell =8$. Since this angular-planar statistic is model-independent, it is ideal to employ in searches of statistical anisotropy (e.g., contaminations from the galactic plane) and to characterize non-Gaussianities.

• Received 15 July 2009

DOI:https://doi.org/10.1103/PhysRevD.80.063525