Cameron Liang | Searcharxiv

arXiv2017

BayesVP: a Bayesian Voigt profile fitting package

We introduce a Bayesian approach for modeling Voigt profiles in absorption spectroscopy and its implementation in the python package, BayesVP, publicly available at https://github.com/cameronliang/BayesVP. The code fits the absorption line profiles within specified wavelength ranges and generates posterior distributions for the column density, Doppler parameter, and redshifts of the corresponding absorbers. The code uses publicly available efficient parallel sampling packages to sample posterior and thus can be run on parallel platforms. BayesVP supports simultaneous fitting for multiple absorption components in high-dimensional parameter space. We provide other useful utilities in the package, such as explicit specification of priors of model parameters, continuum model, Bayesian model comparison criteria, and posterior sampling convergence check.

astro-ph.GA↗

arXiv2017

Observing the circumgalactic medium of simulated galaxies through synthetic absorption spectra

We explore the multiphase structure of the circumgalactic medium (CGM) probed by synthetic spectra through a cosmological zoom-in galaxy formation simulation. We employ a Bayesian method for modelling a combination of absorption lines to derive physical properties of absorbers with a formal treatment of detections, including saturated systems, and non-detections in a uniform manner. We find that in the lines of sight passing through localized density structures, absorption lines of low, intermediate and high ions are present in the spectrum and overlap in velocity space. Low, intermediate and high ions can be combined to derive the mass-weighted properties of a density-varying peak, although the ions are not co-spatial within the structure. By contrast, lines of sight that go through the hot halo only exhibit detectable HI and high ions. In such lines of sight, the absorption lines are typically broad due to the complex velocity fields across the entire halo. We show that the derived gas density, temperature, and metallicity match closely the corresponding HI mass-weighted averages along the LOS. We also show that when the data quality allows, our Bayesian technique allows one to recover the underlying physical properties of LOS by incorporating both detections and non-detections. It is especially useful to include non-detections, of species such as NV or NeVIII, when the number of detections of strong absorbers, such as HI and OVI, is smaller than the number of model parameters (density, temperature, and metallicity).

astro-ph.GA↗