Dark energy model selection with current and future data

Abstract: The main goal of the next generation of weak lensing probes is to constrain cosmological parameters by measuring the mass distribution and geometry of the low redshift Universe and thus to test the concordance model of cosmology. A future all-sky tomographic cosmic shear survey with design properties similar to Euclid has the potential to provide the statistical accuracy required to distinguish between different dark energy models. In order to assess the model selection capability of such a probe, we consider the dark energy equation-of-state parameter $w_0$. We forecast the Bayes factor of future observations, in the light of current information from Planck by computing the predictive posterior odds distribution. We find that Euclid is unlikely to overturn current model selection results, and that the future data are likely to be compatible with a cosmological constant model. This result holds for a wide range of priors.