Using cosmological simulations and synthetic absorption spectra to assess the accuracy of observationally derived CGM metallicities

We used adaptive mesh refinement hydrodynamic cosmological simulations of a z = 1 Milky Way-type galaxy and a z = 0 Dwarf galaxy and generated synthetic quasar absorption-line spectra of their circumgalactic medium (CGM). Our goal is to assess whether standard observational spectroscopic analysis methods accurately reproduce intrinsic column densities, metallicities [Si/H], and hydrogen densities n_H, in simulated absorption-line systems. Without knowledge of the intrinsic simulated properties (blind study), we analysed synthetic COS and HIRES spectra with fixed S/N = 30 to determine the column densities, metallicity, and n_H, using Voigt profile fitting combined with Markov chain Monte Carlo single-phase CLOUDY modelling techniques. To quantify the simulated absorbing gas properties, we objectively determined which gas cells along a line of sight (LOS) contribute to detected absorption in the spectra and adopt the unweighted geometric mean of these properties. For this pilot study, we performed this experiment for five LOS in the two simulated galaxies. We found an average agreement between the 'observed' and intrinsic metallicity overestimated within 0.8σ or 0.2 dex for the 'Milky Way' and overestimated within 1.4σ or 0.2 dex for the Dwarf galaxy. We found that the spectroscopically-derived n_H are underestimated within 0.8σ or 0.4 dex of the intrinsic n_H for the 'Milky Way' and overestimated within 0.3σ or 0.3 dex for the Dwarf galaxy. The overall agreement suggests that, for single-phase ionization modelling of systems where there is substantial spread in gas properties, global metallicity measurements from quasar absorption line studies are capturing the average metallicity and ionization parameters.