The effect of foreground subtraction on cosmological measurements from intensity mapping

Abstract

We model a 21 cm intensity mapping survey in the redshift range 0.01<z<1.5\ndesigned to simulate the skies as seen by future radio telescopes such as the\nSquare Kilometre Array (SKA), including instrumental noise and Galactic\nforegrounds. In our pipeline, we remove the introduced Galactic foregrounds\nwith a fast independent component analysis (fastica) technique. We present the\npower spectrum of the large-scale matter distribution, C(l), before and after\nthe application of this foreground removal method and calculate the resulting\nsystematic errors. We attempt to reduce systematics in the foreground\nsubtraction by optimally masking the maps to remove high foregrounds in the\nGalactic plane. Our simulations show a certain level of bias remains in the\npower spectrum at all scales l<400. At large-scales l<30 this bias is\nparticularly significant. We measure the impact of these systematic effects in\ntwo different ways: firstly we fit cosmological parameters to the broadband\nshape of the power spectrum and secondly we extract the position of the Baryon\nAcoustic Oscillations (BAO). In the first analysis, we find that the\nsystematics introduce an significant shift in the best fit cosmological\nparameters at the 2 to 3 sigma level which depends on the masking and noise\nlevels. However, cosmic distances can be recovered in an unbiased way after\nforeground removal at all simulated redshifts by fitting the BAOs in the power\nspectrum. We conclude that further advances in foreground removal are needed in\norder to recover unbiased information from the broadband shape of the power\nspectrum, however, intensity mapping experiments will be a powerful tool for\nmapping cosmic distances across a wide redshift range.\n