Map-making for the Planck 30 Ghz channel with Polar and MADAM destriping codes

H. Kurki-Suonio, V. Heikkilä, E. Keihänen, R. Keskitalo, and T. Poutanen
Department of Physical Sciences, University of Helsinki and Helsinki Institute of Physics

Our two map-making codes are based on the destriping approach, where the correlated noise is represented by a sequence of constant baselines in the time domain. MADAM utilizes information about the noise covariance matrix, whereas Polar does not. This allows MADAM to use shorter baselines (here 1.2 s) than Polar (here 1 min), and thus to model the noise more accurately.

The CTP working group has used simulated data of the four Planck 30 GHz detectors to study different map-making methods. We show here some maps made by Polar and MADAM as a part of this effort. The simulated data contained CMB, foreground, dipole, and noise. The two top rows show the first two of these input components. The columns from left to right are the Stokes parameters I, Q, and U.

The next two rows show the output maps (top) and the output-input difference maps (bottom) from Polar (maps from MADAM look similar). The difference maps are dominated by white noise. The std of these Nside = 512 difference maps (pixel size 7 arcmin) are 42.03 (I), 59.34 (Q), 59.85 (U) microK (antenna temperature) for Polar and 41.85 (I), 59.08 (Q), 59.59 (U) microK for MADAM. These should be compared to the level of white noise which is 41.66 (I), 58.83 (Q), 59.38 (U) microK. Thus MADAM produces maps with slightly lower noise, getting closer to the white noise limit.
In addition to white noise and remaining correlated noise, the difference maps contain a contribution from signal distortion, which is due to the effect of pixelization noise on the map-making method. As a price for the more aggressive noise removal this signal distortion is larger for MADAM (bottom row) than for Polar (top row).

This work is part of an ongoing collaboration in the CTP Working Group of the Planck Consortium. The input data were simulated using the reference sky model being developed by Planck Working Group 2 and software provided by the Planck Level S team. The output data were post-processed using the HEALPix package. This research used resources of the National Energy Research Scientific Computing Center, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098.