The first and third year data releases from the WMAP provide evidence of an anomalous Cold Spot (CS) at galactic latitude b=-57deg and longitude l=209deg. We have examined the properties of the CS in some detail in order to assess its cosmological si
gnificance. We have performed a cluster analysis of the local extrema in the CMB signal to show that the CS is actually associated with a large group of extrema rather than just one. In the light of this we have re-examined the properties of the WMAP ILC and co-added cleaned WCM maps, which have previously been used for the analysis of the properties of the signal in the vicinity of the CS. These two maps have remarkably similar properties on equal latitude rings for |b|>30deg, as well as in the vicinity of the CS. We have also checked the idea that the CMB signal has a non-Gaussian tail, localized in the low multipole components of the signal. For each ring we apply a linear filter with characteristic scale R, dividing the CMB signal in two parts: the filtered part, with characteristic scale above that of the filter R, and the difference between the initial and filtered signal. Using the filter scale as a variable, we can maximize the skewness and kurtosis of the smoothed signal and minimize these statistics for the difference between initial and filtered signal. We have discovered that the shape of the CS is formed primarily by the components of the CMB signal represented by multipoles between 10<=L<=20, with a corresponding angular scale about 5-10 degs. This signal leads to modulation of the whole CMB sky, clearly seen at |b|>30deg in both the ILC and WCM maps, rather than a single localized feature. After subtraction of this modulation, the remaining part of the CMB signal appears to be consistent with statistical homogeneity and Gaussianity.
Scale transformations have played an extremely successful role in studies of cosmological large-scale structure by relating the non-linear spectrum of cosmological density fluctuations to the linear primordial power at longer wavelengths. Here we gen
eralize this approach to investigate the usefulness of scale transformations for nonlinear higher-order statistics, specifically the bispectrum. We find that the bispectrum predicted by perturbation theory at tree-level can be rescaled to match the results of full numerical simulations in the weakly and intermediately nonlinear regimes, especially at high redshifts, with an accuracy that is surprising given the simplicity of the procedure used. This discovery not only offers a simple practical way of calculating the matter bispectrum, but also suggests that scale transformations may yet yield even deeper insights into the physics of hierarchical clustering.
