Error estimation for the MAP experiment

We report here the first full sky component separation and CMB power spectrum estimation using a Wiener filtering technique on simulated data from the upcoming MAP experiment, set to launch in early 2001. The simulations included contributions from the three dominant astrophysical components expected in the five MAP spectral bands, namely CMB radiation, Galactic dust, and synchrotron emission. We assumed a simple homogeneous and isotropic white noise model and performed our analysis up to a spherical harmonic multipole lmax = 512 on the fraction of the sky defined by |b| > 20 o. We find that the reconstruction errors are reasonably well fitted by a Gaussian with an rms of 24 μK, but with significant deviations in the tails. Our results further support the predictions on the resulting CMB power spectrum of a previous estimate by Bouchet & Gispert (1999), which entailed a number of assumptions this work removes.