|Title:||Estimating Redshifts from Distant HI Galaxies with Bayesian Statistics|
|Date:||Thursday, 27 October 2016|
|Time:||11:00 - 12:00|
The upcoming Square Kilometre Array (SKA) telescope will be the most sensitive radio telescope ever built, detecting the neutral hydrogen (HI) emission from up to a billion galaxies. However, HI emission from distant galaxies is intrinsically faint and easily lost in noise. Significant detection of the HI emission line is only expected for a small fraction of galaxies, making redshift estimation extremely difficult.
In this talk, I will discuss a new approach to redshift estimation from radio galaxies using Bayesian statistics. We perform parameter inference using the well-known HI line profile as a model. Using Bayesian model selection, we are able to robustly distinguish between an emission line and noise or RFI. This method would dramatically increase the number of well-characterised galaxies in the SKA catalogue, as well as providing the full probability distribution for each galaxy’s redshift. Preliminary results from a realistic SKA1-MID simulation indicate that we can recover the line profile parameters and a competitive redshift estimate for very low signal-to-noise lines, showing the promise of this new technique.