Detection of a non-Gaussian Spot in WMAP

Abstract

An extremely cold and big spot in the WMAP 1-year data is analyzed. Our work is a continuation of a previous paper (Vielva et al. 2004) where non-Gaussianity was detected, with a method based on the Spherical Mexican Hat Wavelet (SMHW) technique. We study the spots at different thresholds on the SMHW coefficient maps, considering six estimators, namely number of maxima, number of minima, number of hot and cold spots, and number of pixels of the spots. At SMHW scales around 4 degrees (10 degrees on the sky), the data deviate from Gaussianity. The analysis is performed on all sky, the northern and southern hemispheres, and on four regions covering all the sky. A cold spot at (b = -57, l = 209) is found to be the source of this non-Gaussian signature. We compare the spots of our data with 10000 Gaussian simulations, and conclude that only around 0.2% of them present such a cold spot. Excluding this spot, the remaining map is compatible with Gaussianity and even the excess of kurtosis in Vielva et al. 2004, is found to be due exclusively to this spot. Finally, we study whether the spot causing the observed deviation from Gaussianity could be generated by systematics or foregrounds. None of them seem to be responsible for the non-Gaussian detection.