Making maps from Planck LFI 30 GHz data

Abstract

This paper is one of a series describing the performance and accuracy of map-making codes as assessed by the Planck CTP working group. We compare the performance of multiple codes written by different groups for making polarized maps from Planck-sized, all-sky cosmic microwave background (CMB) data. Three of the codes are based on a destriping algorithm, whereas the other three are implementations of a maximum-likelihood algorithm. Previous papers in the series described simulations at 100 GHz (Poutanen et al. 2006(Poutanen et al. , A&A, 449, 1311 and 217 GHz (Ashdown et al. 2007, A&A, 467, 761). In this paper we make maps (temperature and polarisation) from the simulated one-year observations of four 30 GHz detectors of Planck Low Frequency Instrument (LFI). We used Planck Level S simulation pipeline to produce the observed time-ordered-data streams (TOD). Our previous studies considered polarisation observations for the CMB only. For this paper we increased the realism of the simulations and included polarized galactic foregrounds in our sky model, which is based on the version 0.1 of the Planck reference sky. Our simulated TODs comprised dipole, CMB, diffuse galactic emissions, extragalactic radio sources, and detector noise. The strong subpixel signal gradients arising from the foreground signals couple to the output map through the map-making and cause an error (signal error) in the maps. Destriping codes have smaller signal error than the maximum-likelihood codes. We examined a number of schemes to reduce this error. On the other hand, the maximum-likelihood map-making codes can produce maps with lower residual noise than destriping codes.

Residual noise covariance forPlancklow-resolution data analysis

“…On the other hand, if the sky signal varies within the pixels, non-uniform sampling of a pixel can cause a systematic error in the solved map (Ashdown et al 2007a). This is an important aspect, as none of the considered algorithms deal with sub pixel structures by design.…”

“…(16) to estimate its covariance. The CTP papers (Poutanen et al 2006;Ashdown et al 2007aAshdown et al ,b, 2009 have shown that for Planck the two methods produce maps which are very close to one another. Moreover, they have shown that using a generalized destriper, the derived maps eventually become nearly identical to those obtained with the optimal methods, if an appropriate length of the baseline and a number of the baseline offsets is adopted together with a consistently evaluated prior.…”

“…These include robust approaches to producing low-resolution maps and techniques for estimating pixel-pixel correlations from their residual noise. The work is a natural extension of the extensive mapmaking studies conducted within the Planck Working Group 3 (CTP) (Poutanen et al 2006;Ashdown et al 2007aAshdown et al ,b, 2009. The low-resolution maps are expected to compress nearly all the information relevant to the large angular scale in fewer pixels making them more manageable.…”

Planck pre-launch status: Expected LFI polarisation capability

“…For Planck's actual scan strategy, scan circle crossings at substantial angles occur over about ±20 • around the poles, so that a much longer run of the circle is involved in constraining the offsets, giving higher signal to noise, and also reducing the sensitivity to monopole-dipole degeneracy. The actual determination of the offsets will be made in the course of iterative destriping, for instance using the madam algorithm (Keihänen et al 2005;Keihänen et al 2010;Ashdown et al 2007;Kurki-Suonio et al 2009). …”

A deconvolution map-making method for experiments with circular scanning strategies

“…3 together with the results of this analysis. The results of the analysis of more realistic Planck data, used in a previous map-making comparison paper (Ashdown et al 2007a), is described in Sect. 4.…”