Consistency of the Parkes Pulsar Timing Array Signal with a Nanohertz Gravitational-wave Background

• Published in: Astrophysical journal. Letters 2022-06, Vol.932 (2), p.L22
• Main Authors: Goncharov, Boris, Thrane, Eric, Shannon, Ryan M., Harms, Jan, Bhat, N. D. Ramesh, Hobbs, George, Kerr, Matthew, Manchester, Richard N., Reardon, Daniel J., Russell, Christopher J., Zhu, Xing-Jiang, Zic, Andrew
• Format: Article
• Language: English
• Subjects: Amplitudes; Arrays; Astronomical methods; Astronomy data analysis; Background noise; Bayesian statistics; Gravitational wave astronomy; Gravitational waves; Hierarchical models; High energy astrophysics; Importance sampling; Methodology; Millisecond pulsars; Noise; Power spectra; Pulsar timing method; Pulsars; Stochastic processes; Supermassive black holes
• ISSN: 2041-8205; 2041-8213
• DOI: 10.3847/2041-8213/ac76bb

Pulsar timing array experiments have recently reported strong evidence for a common-spectrum stochastic process with a strain spectral index consistent with that expected of a nanohertz-frequency gravitational-wave background, but with negligible yet non-zero evidence for spatial correlations required for a definitive detection. However, it was pointed out by the Parkes Pulsar Timing Array (PPTA) collaboration that the same models used in recent analyses resulted in strong evidence for a common-spectrum process in simulations where none is present. In this work, we introduce a methodology to distinguish pulsar power spectra with the same amplitude from noise power spectra of similar but distinct amplitudes. The former is the signature of a spatially uncorrelated pulsar term of a nanohertz gravitational-wave background, whereas the latter could represent ensemble pulsar noise properties. We test the methodology on simulated data sets. We find that the reported common process in PPTA pulsars is indeed consistent with the spectral feature of a pulsar term. We recommend this methodology as one of the validity tests that the real astrophysical and cosmological backgrounds should pass, as well as for inferences about the spatially uncorrelated component of the background.