Distinguishing a stochastic gravitational-wave signal from correlated noise with joint parameter estimation: Fisher analysis for ground-based detectors

TL;DR

This paper assesses the impact of correlated magnetic noise, such as Schumann resonances, on the detection of stochastic gravitational-wave backgrounds, demonstrating that marginalizing over noise parameters minimally degrades sensitivity while ignoring noise can cause significant bias.

Contribution

It introduces an analytical Fisher analysis model to evaluate the detectability of SGWB amidst correlated magnetic noise, showing robustness of forecasts and potential biases if noise is ignored.

Findings

Marginalizing over noise parameters degrades constraints by at most a factor of 2.

Forecasts are robust against variations in noise parameters, with up to 40% change.

Ignoring correlated noise can lead to serious biases in SGWB parameter estimation.

Abstract

Search sensitivity to a stochastic gravitational-wave background (SGWB) is enhanced by cross-correlating detector signals. However, one of the most serious concerns is the environmental noise correlated between detectors. The global electromagnetic fields on the Earth, known as Schumann resonances, produce the correlated noise through the instrumental magnetic couplings. In this paper, we study the detectability of a SGWB in the presence of the correlated magnetic noise, using the Fisher analysis based on the analytical model of the correlated magnetic noise. We find that there is no significant degeneracy between the SGWB and noise parameters. Marginalizing over the correlated noise parameters degrades the constraint on each SGWB parameter by a factor of $\sim2$ at most in the four-detector case, irrespective of the strength of the magnetic coupling. We also confirm that the forecast results are robust against the variation of correlated noise parameters and can vary up to $40\%$ in the realistic range of the coupling parameters for the second-generation detectors. However, ignoring the correlated noise in parameter estimation generally leads to a biased constraint on the SGWB parameters. If the coupling strength is twice as large as expected, this could result in a serious bias.