Obtaining the spacetime metric from cosmological observations

TL;DR

This paper proposes a numerical method to derive the universe's spacetime metric from cosmological observations, starting with a spherically symmetric dust model, and demonstrates its viability with artificial data.

Contribution

It introduces a numerical algorithm to determine the spacetime metric from observational data under spherical symmetry and dust conditions, paving the way for more detailed cosmic geometry analysis.

Findings

Successfully reproduces original models with artificial data

Demonstrates the viability of the numerical scheme

Addresses challenges in turning theory into practice

Abstract

Recent galaxy redshift surveys have brought in a large amount of accurate cosmological data out to redshift 0.3, and future surveys are expected to achieve a high degree of completeness out to a redshift exceeding 1. Consequently, a numerical programme for determining the metric of the universe from observational data will soon become practical; and thereby realise the ultimate application of Einstein's equations. Apart from detailing the cosmic geometry, this would allow us to verify and quantify homogeneity, rather than assuming it, as has been necessary up to now, and to do that on a metric level, and not merely at the mass distribution level. This paper is the beginning of a project aimed at such a numerical implementation. The primary observational data from our past light cone consists of galaxy redshifts, apparent luminosities, angular diameters and number densities, together with source evolution functions, absolute luminosities, true diameters and masses of sources. Here we start with the simplest case, that of spherical symmetry and a dust equation of state, and execute an algorithm that determines the unknown metric functions from this data. We discuss the challenges of turning the theoretical algorithm into a workable numerical procedure, particularly addressing the origin and the maximum in the area distance. Our numerical method is tested with several artificial data sets for homogeneous and inhomogeneous models, successfully reproducing the original models. This demonstrates the basic viability of such a scheme. Although current surveys don't have sufficient completeness or accuracy, we expect this situation to change in the near future, and in the meantime there are many refinements and generalisations to be added.