The Halo21 absorption modelling challenge: lessons from ‘observing’ synthetic circumgalactic absorption spectra

In the Halo21 absorption modelling challenge we generated synthetic absorption spectra of the circumgalactic medium (CGM), and attempted to estimate the metallicity, temperature, and density (Z, T, and n_H) of the underlying gas using observational methods. We iteratively generated and analysed three increasingly complex data samples: ion column densities of isolated uniform clouds, mock spectra of 1–3 uniform clouds, and mock spectra of high-resolution turbulent mixing zones. We found that the observational estimates were accurate for both uniform cloud samples, with Z, T, and n_H retrieved within 0.1 dex of the source value for ≳ 90 per cent of absorption systems. In the turbulent-mixing scenario, the mass, temperature, and metallicity of the strongest absorption components were also retrieved with high accuracy. However, the underlying properties of the subdominant components were poorly constrained because the corresponding simulated gas contributed only weakly to the H I absorption profiles. On the other hand, including additional components beyond the dominant ones did improve the fit, consistent with the true existence of complex cloud structures in the source data.