Drift scans yield lower uncertainty in EoR observations with MWA
8 June 2015
The Epoch of Reionisation (EoR) marks a period in the first billion years of the Universe when the first stars, galaxies and black holes produced light and changed the nature of the Intergalactic Medium, the gas between the galaxies. In this period, neutral hydrogen gas between galaxies was ionised by the radiation from these stars, shifting the Universe from a dark and neutral place, to one that more closely resembles the Universe we observe locally. The initial conditions of the Universe, and the evolution of the growth of the first structures, are imprinted in the neutral hydrogen gas, and we can probe the evolution of this early period by measuring emission from hydrogen gas. This can be achieved using low-frequency radio telescopes, such as the Murchison Widefield Array (MWA), which can detect the redshifted emission line from the neutral hydrogen nucleus and trace the gas in exquisite detail. The signal is exceptionally weak compared with the measurement noise of our experiments and contamination by other, substantially brighter, astrophysical sources (e.g. radio galaxies and Galactic emission), making this a challenging experiment to undertake and requiring 1000s of hours of data. At this point, no detection of this signal has been made, but there are many teams and instruments pursuing this goal.
The relative weakness of the signal, relative to other signals and noise, demands a careful analysis of the design of the experiment to maximise information available for science. In her recent paper, CAASTRO Associate Investigator Dr Cath Trott (ICRAR – Curtin University) studied the trade-offs for estimation of the high redshift hydrogen signal by observing the sky in different modes. The primary comparison made was ‘tracked scans’, whereby the instrument concentrates on a small region of sky, and drift scans, where a larger area of sky is measured but with less time per region. These modes yield a different balance of reducing the measurement noise efficiently and of reducing the 'sample variance' – an error which reflects that having more measurements of a quantity leads to a more precise measurement. She found that both of these modes produce comparable results, but drift scans showed an added advantage in providing an instrument that is more easily calibrated. This work demonstrated that current and future instruments should consider drift scan modes for performing EoR experiments.