Field-level multiprobe analysis of the CMB, integrated Sachs-Wolfe effect, and the galaxy density maps

Extracting information from cosmic surveys is often done in a two-step process, construction of maps and then summary statistics such as two-point functions. We use simulations to demonstrate the advantages of a general Bayesian framework that consistently combines different cosmological experiments on the field level, and reconstructs both the maps and cosmological parameters. We apply our method to jointly reconstruct the primordial CMB, the integrated Sachs-Wolfe effect, and six tomographic galaxy density maps on the full sky on large scales along with several cosmological parameters. While the traditional maximum a posterior estimator has both two-point level and field-level bias, the new approach yields unbiased cosmological constraints and improves the signal-to-noise ratio of the maps.