A Joint Analysis for Cosmology and Photometric Redshift Calculation Using Cross Correlations

We present a method of calibrating the properties of photometric redshift bins as part of a larger Markov Chain Monte Carlo (MCMC) analysis for the inference of cosmological parameters. The redshift bins are characterised by their mean and variance, which are varied as free parameters and marginalised over when obtaining the cosmological parameters. We demonstrate that the likelihood function for cross-correlations in an angular power spectrum framework tightly constrains the properties of bins such that they may be well determined, reducing their influence on cosmological parameters and avoiding the bias from poorly estimated redshift distributions. We demonstrate that even with only three photometric and three spectroscopic bins, we can recover accurate estimates of the mean redshift of a bin to within $Δμ\approx 3-4 \times10^{-3}$ and the width of the bin to $Δσ\approx 1\times10^{-3}$ for galaxies near $z = 1$. This indicates that we may be able to bring down the photometric redshift errors to a level which is in line with the requirements for the next generation of cosmological experiments.