2009MNRAS.397..311D

Upcoming ground- and space-based observatories, such as the Dark Energy Survey (DES), the Large Synoptic Survey Telescope, the Joint Dark Energy Mission concepts and the Square Kilometre Array, promise to dramatically increase the size of strong gravitational lens samples. A significant fraction of the systems is expected to be time-delay lenses. Due to the sensitivity of time delays to cosmological parameters, this increase in sample size opens up new avenues of parameter constraint. Many of the existing lensing degeneracies, e.g. between the Hubble parameter H₀ and the slope of the density profile of the lens, become less of an issue with large samples since the distributions of a number of parameters are predictable, and can be incorporated into an analysis, thus helping to lessen the degeneracy. Assuming a mean galaxy density profile that does not evolve with redshift, a Λ cold dark matter cosmology, and Gaussian distributions for bulk parameters describing the lens and source populations, we generate synthetic lens catalogues and examine the relationship between constraints on the Ω_m-Ω_Λ plane and H₀ with increasing lens sample size. We find that, with sample sizes of ∼400 time-delay lenses, useful constraints can be obtained for Ω_m and Ω_Λwith approximately similar levels of precision as from the best of other methods. In addition, sample sizes of ∼100 time-delay systems yield estimates of H₀ with errors of only a couple of per cent, exceeding the level of precision from current best estimates such as the Hubble Space Telescope key project. As such, strong lensing promises to become one of the premier methods for constraining the Hubble constant in the future. We note that insufficient prior knowledge of the lens samples employed in the analysis, via under or overestimates in the mean values of the sample distributions, results in the broadening of constraints. This highlights the need for sound prior knowledge of the sample before useful cosmological constraints can be obtained from large time-delay samples.