What will anisotropies in the clustering pattern in redshifted 21-cm maps tell us?

The clustering pattern in high-redshift H-I maps is expected to be anisotropic for two distinct reasons: the Alcock-Paczynski effect and the peculiar velocities, both of which are sensitive to the cosmological parameters. The signal is also expected to be sensitive to the details of the H-I distribution at the epoch when the radiation originated. We use simple models for the H-I distribution at the epoch of reionization and the post-reionization era to investigate exactly what we hope to learn from future observations of the anisotropy pattern in H-I maps. We find that such observations will probably tell us more about the H-I distribution than about the background cosmological model. Assuming that reionization can be described by spherical, ionized bubbles all of the same size with their centres possibly being biased with respect to the dark matter, we find that the anisotropy pattern at small angles is expected to have a bump at the characteristic angular size of the individual bubbles whereas the large-scale anisotropy pattern will reflect the size and the bias of the bubbles. The anisotropy also depends on the background cosmological parameters, but the dependence is much weaker. Under the assumption that the H-I in the post-reionization era traces the dark matter with a possible bias, we find that changing the bias and changing the background cosmology have similar effects on the anisotropy pattern. Combining observations of the anisotropy with independent estimates of the bias, possibly from the bi-spectrum, may allow these observations to constrain cosmological parameters.