Scaling relations for galaxy clusters in the Millennium-XXL simulation

TL;DR

This paper uses the Millennium-XXL simulation to analyze how galaxy cluster observables scale with mass, revealing correlations and systematic effects crucial for accurate cosmological inferences from cluster surveys.

Contribution

It provides detailed scaling relations for galaxy cluster observables in a large simulation, highlighting the impact of internal structure, orientation, and environment on these relations.

Findings

Scatter in observables is correlated due to internal and environmental factors.

Systematic effects can explain discrepancies in SZ measurements from different surveys.

Proper modeling of observables and selection biases is essential for cosmological conclusions.

Abstract

We present a very large high-resolution cosmological N-body simulation, the Millennium-XXL or MXXL, which uses 303 billion particles to represent the formation of dark matter structures throughout a 4.1Gpc box in a LambdaCDM cosmology. We create sky maps and identify large samples of galaxy clusters using surrogates for four different observables: richness estimated from galaxy surveys, X-ray luminosity, integrated Sunyaev-Zeldovich signal, and lensing mass. The unprecedented combination of volume and resolution allows us to explore in detail how these observables scale with each other and with cluster mass. The scatter correlates between different mass-observable relations because of common sensitivities to the internal structure, orientation and environment of clusters, as well as to line-of-sight superposition of uncorrelated structure. We show that this can account for the apparent discrepancies uncovered recently between the mean thermal SZ signals measured for optically and X-ray selected clusters by stacking data from the Planck satellite. Related systematics can also affect inferences from extreme clusters detected at high redshift. Our results illustrate that cosmological conclusions from galaxy cluster surveys depend critically on proper modelling, not only of the relevant physics, but also of the full distribution of the observables and of the selection biases induced by cluster identification procedures.