Statistics of neutral regions during hydrogen reionization

Abstract

We present predictions for two statistical measures of the hydrogen reionization process at high redshift. The first statistic is the number of neutral segments identified in spectra of high-redshift QSOs as a function of their length. The second is the cross-correlation of neutral regions with possible sources of ionizing radiation. These independent probes are sensitive to the topology of the ionized regions. If reionization proceeded from high- to low-density regions, then the cross-correlation will be negative, while if voids were ionized first we expect a positive correlation and a relatively small number of long neutral segments. We test the sensitivity of these statistics for reionization by stars in high-redshift galaxies. The flux of ionizing radiation emitted from stars is estimated by identifying galaxies in an N-body simulation using a semianalytic galaxy formation model. The spatial distribution of ionized gas is traced in various models for the propagation of the ionization fronts. A model with ionization proceeding from high- to low-density regions is consistent with the observations of Becker et al., while models in which ionization begins in the lowest density regions appear to be inconsistent with the present data. Further data are required to confirm this result.