Gaussian Approximation of Peak Values in the Integrated Sachs-Wolfe Effect

TL;DR

This paper models the expected peak signals in the cosmic microwave background caused by the Integrated Sachs-Wolfe Effect within the standard cosmological model, comparing simulations with observational data from Planck and WMAP.

Contribution

It introduces a Gaussian simulation framework for the ISW effect that accounts for correlations across redshifts and physical contributions, providing a refined theoretical prediction.

Findings

Simulation results agree with previous large-scale structure estimates.

Observed peaks exceed theoretical mean at 2.5σ significance.

Detected signals differ from null hypothesis at 3.5σ significance.

Abstract

The accelerating expansion of the universe at recent epochs is encoded in the cosmic microwave background: a few percent of the total temperature fluctuations are generated by evolving gravitational potentials which trace the large-scale structures in the universe. This signature of dark energy, the Integrated Sachs-Wolfe Effect, has been detected by averaging temperatures in the WMAP sky maps corresponding to the directions of superstructures in the Sloan Digital Sky Survey data release 6. We model the maximum average peak signal expected in the standard $\Lambda$CDM cosmological model, using Gaussian random realizations of the microwave sky, including correlations between different physical contributions to the temperature fluctuations and between different redshift ranges of the evolving gravitational potentials. We find good agreement with the mean temperature peak amplitude from previous theoretical estimates based on large-scale structure simulations, but with larger statistical uncertainties. We apply our simulation pipeline to four different foreground-cleaned microwave temperature maps from Planck and WMAP data, finding a mean temperature peak signal at previously identified sky locations which exceeds our theoretical mean signal at a statistical significance of about $2.5\sigma$ and which differs from a null signal at $3.5\sigma$.