Inferring local quasar IGM damping wing constraints

Abstract

Lyman-$\alpha$ damping wings towards quasars are a highly sensitive probe of the neutral hydrogen (HI) content in the foreground intergalactic medium (IGM), not only constraining the global timing of reionization but also the \textit{local} ionization topology near the quasar. Near-optimal extraction of this information is possible with the help of two recently introduced reionization model-independent summary statistics of the HI distribution in the IGM \textit{before} the quasar started shining, complemented with the quasar's lifetime encoding the effect of its ionizing radiation as a third parameter. We introduce a fully Bayesian JAX-based Hamiltonian Monte Carlo (HMC) inference framework that allows us to jointly reconstruct the quasar's unknown continuum and constrain these local damping wing statistics. We put forward a probabilistic framework that allows us to tie these local constraints to any specific reionization model and obtain model-dependent constraints on the global timing of reionization. We demonstrate that we are able to constrain the (Lorentzian-weighted) HI column density in front of the quasar to a precision of $0.69_{-0.30}^{+0.06}\,\mathrm{dex}$ and its original distance to the first neutral patch before the quasar started shining to $31.4_{-28.1}^{+10.7}\,\mathrm{cMpc}$ (if a noticeable damping wing is present in the spectrum), extracting hitherto unused local information from the IGM damping wing imprint. Once tied to a specific reionization model, we find that the statistical fidelity of our constraints on the global IGM neutral fraction and the lifetime of the quasar improves, while retaining the same precision as achieved by pipelines that infer these parameters directly.