Cadence of observations is key to variability of radio sources
4 August 2016
Our view of the radio sky changes over time. This is partly because sources change over time: black holes devour stars and gas, stars go supernova, or clouds of fast moving electrons expand and slow. But variability is also partly due to our view through the cosmos: the gas between us and the stars and galaxies is turbulent and distorts our view of stars and galaxies, acting like a set of lenses that focus and defocus the light that we are seeing.
Much work is being done to better understand variability, in particular around a frequency of 1.4GHz. Interestingly, two surveys – both conducted with the Very Large Array (VLA) in the US – differed by a factor of ~4 in the areal density of radio variables on timescales of decades compared to timescales of less than a few years. In a new publication, CAASTRO Affiliate Dr Paul Hancock (ICRAR-Curtin) and colleagues identified a set of archival observations from the Australia Telescope Compact Array (ATCA), suitable for a similarly sensitive survey to explain the discrepancy between the two VLA studies.
The "Phoenix Deep Survey" was conducted over 8 years during which six overlapping areas of sky were observed ("epochs") to eventually combine in a single deep map of the region. In this new study, the researchers instead looked at each of the epochs separately to identify changes in the radio brightness, or flux, of sources in the overlapping areas. Furthermore, sensitivity was similar to the two VLA studies, allowing for an independent measurement of the areal density of variable radio sources.
The team used software solutions to find and classify sources in each of the epochs, cross match sources between epochs, build up light curves for each source, and to identify sources with significant variability. In the end, they had 9 sources that were significantly variable. All but one of these sources was found to be an active galactic nucleus (accreting black hole). The key to the difference in previous areal density estimates, such as the two VLA studies, is in the cadence of the observations: one VLA study was sensitive only to variability on timescales less than a year, whilst the other was sensitive to timescales of ~7 years. Dr Hancock and his colleagues found that the largest amount of variability was seen on timescales of ~5 years.