0227-(WMAP) Observations: Foreground Emission arxiv:0302208

PAPER

First Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Foreground Emission arxiv:0302208

The purpose of a foreground mask is to exclude map pixels that contain “too much” foreground signal from use in CMB analyses. Some pixels often need to be masked even after a foreground reduction technique has been applied to a map. What constitutes “too much” foreground signal depends on the particular analysis. It is therefore useful to have masks available with different flux cut-off levels. For example, these masks can be used to demonstrate (in)sensitivity of scientific results to the level of foreground sky cut

A series of masks are made using multiples of these step sizes. A cut at the peak value of the remainder histogram is referred to as “p0”

the mask defined by the peak of the remainder histogram using the K-band map is “Kp0”. A mask with a temperature cut-off two steps smaller in temperature (more severe) is called “Km2”, where “m” means “minus” and (for a less severe cut) “p” means “plus”.

In addition to diffuse Galactic emission, point sources also contaminate the WMAP maps. We have constructed a point source source mask that includes: (1) all of the sources from Stickel et al. (1994) (which is the updated K ̈ uhr et al. (1981) 5 GHz 1 Jy compilation with attached optical source identifications); (2) sources with 22 GHz fluxes ≥ 0.5 Jy from the Hirabayashi et al. (2000) comprehensive compilation of flat spectrum AGN; (3) Additional flat spectrum objects from Ter ̈asranta et al. (2001); and (4) sources from the X-ray/radio blazar survey of Perlman et al. (1998) and Landt et al. (2001). The mask excludes a radius of 0.6◦ about each of almost 700 source positions. Table 2 gives the percentages of the sky that are cut by each of the analysis masks with and without masking point sources.

For both Galactic and extragalactic contamination, the most affected WMAP pixels should be masked and not used for cosmological purposes.

Bias-Hardened CMB Lensing with Polarization arxiv:1310.2372

For realistic CMB observations, there are so-called mean-field biases for the standard minimum-variance quadratic lensing estimators because of non-lensing sources of statistical anisotropy, such as masking, inhomogeneous map noise, beam asymmetry, or spatially varying errors in the detector polarization angles. With a perfect statistical understanding of the unlensed CMB and the instrument used to observe it, these biases can be corrected for, but given the imperfections in our understanding of these quantities it can be useful to design estimators that are less sensitive to them. Approaches have been proposed in the literature to mitigate some of the mean-field biases. For example, the mean fields from masking in temperature have been studied by several authors, with approaches including simply avoiding mask boundaries (Hirata et al. 2008; Carvalho & Tereno 2011), or using inpainting/apodization (Perotto et al. 2010; Plaszczynski et al. 2012; Benoit-Levy et al. 2013) to smooth them. These techniques could also be utilized for polarization, in conjunction with pure estimators for E and B modes (Smith 2006; Smith & Zaldarriaga 2007). For temperature, the mean-field bias from an inhomogeneous map noise has also been studied by Hanson, Rocha & Gorski (2009).