Cosmological information in Gaussianised weak lensing signals

TL;DR

This paper explores how Gaussianising transformations of weak lensing signals can improve cosmological parameter estimation, finding significant gains in ideal conditions but limited benefits when realistic noise is included.

Contribution

It introduces optimized Box-Cox transformations for Gaussianising weak lensing convergence fields and assesses their impact on cosmological information extraction.

Findings

Optimized Box-Cox transformations outperform offset logarithmic transformations in Gaussianising convergence.

Transformations increase signal-to-noise ratio significantly in noise-free simulations.

Realistic shape noise diminishes the benefits of Gaussianising transformations.

Abstract

We investigate the information on cosmology contained in Gaussianised weak gravitational lensing convergence fields. Employing Box-Cox transformations to determine optimal transformations to Gaussianity, we develop analytical models for the transformed power spectrum, including effects of noise and smoothing. We find that optimised Box-Cox transformations perform substantially better than an offset logarithmic transformation in Gaussianising the convergence, but both yield very similar results for the signal-to-noise and parameter constraints. None of the transformations is capable of eliminating correlations of the power spectra between different angular frequencies, which we demonstrate to have a significant impact on the errors on cosmology. Analytic models of the Gaussianised power spectrum yield good fits to the simulations and produce unbiased parameter estimates in the majority of cases, where the exceptions can be traced back to the limitations in modelling the higher-order correlations of the original convergence. In the idealistic case, without galaxy shape noise, we find an increase in cumulative signal-to-noise by a factor of 2.6 for angular frequencies up to 1500, and a decrease in the area of the confidence region in the Omega_m-sigma_8 plane by a factor of 4.4 in terms of q-values for the best-performing transformation. When adding a realistic level of shape noise, all transformations perform poorly with little decorrelation of angular frequencies, a maximum increase in signal-to-noise of 34%, and even marginally degraded errors on cosmological parameters. We argue that, to find Gaussianising transformations of practical use, one will need to go beyond transformations of the one-point distribution of the convergence, extend the analysis deeper into the non-linear regime, and resort to an exploration of parameter space via simulations. (abridged)