On the evidence for a common-spectrum process in the search for the nanohertz gravitational-wave background with the Parkes Pulsar Timing Array (2107.12112v2)

Abstract: A nanohertz-frequency stochastic gravitational-wave background can potentially be detected through the precise timing of an array of millisecond pulsars. This background produces low-frequency noise in the pulse arrival times that would have a characteristic spectrum common to all pulsars and a well-defined spatial correlation. Recently the North American Nanohertz Observatory for Gravitational Waves collaboration (NANOGrav) found evidence for the common-spectrum component in their 12.5-year data set. Here we report on a search for the background using the second data release of the Parkes Pulsar Timing Array. If we are forced to choose between the two NANOGrav models $\unicode{x2014}$ one with a common-spectrum process and one without $\unicode{x2014}$ we find strong support for the common-spectrum process. However, in this paper, we consider the possibility that the analysis suffers from model misspecification. In particular, we present simulated data sets that contain noise with distinctive spectra but show strong evidence for a common-spectrum process under the standard assumptions. The Parkes data show no significant evidence for, or against, the spatially correlated Hellings-Downs signature of the gravitational-wave background. Assuming we did observe the process underlying the spatially uncorrelated component of the background, we infer its amplitude to be $A = 2.2^{+0.4}_{-0.3} \times 10^{-15}$ in units of gravitational-wave strain at a frequency of $1\, \text{yr}^{-1}$. Extensions and combinations of existing and new data sets will improve the prospects of identifying spatial correlations that are necessary to claim a detection of the gravitational-wave background.

• The paper demonstrates strong evidence (log Bayes Factor >15) for a common-spectrum process in the PPTA dataset using Bayesian analysis.
• It estimates a gravitational-wave strain amplitude of A = 2.2⁺⁰.⁴₋₀.₃ × 10⁻¹⁵ yr⁻¹ and aligns its findings with previous NANOGrav observations.
• The study identifies potential model misspecification issues in pulsar noise analysis, underscoring the need for refined noise-modeling in GW searches.

Analysis of the Common-Spectrum Process in Nanohertz Gravitational-Wave Background Detection

In their paper, Goncharov et al. conduct an investigation into the presence of a common-spectrum process within the context of gravitational-wave detection via the Parkes Pulsar Timing Array (PPTA). This paper builds upon recent findings of the North American Nanohertz Observatory for Gravitational Waves (NANOGrav), which identified evidence for a common-spectrum component within their dataset. The research seeks to determine whether a similar component exists within the PPTA dataset and furthers our understanding of low-frequency gravitational waves.

Methodology and Findings

The authors use the second data release from the PPTA, employing Bayesian statistical methods to scrutinize the dataset for a common-spectrum stochastic process. Their approach tests the null hypothesis, where pulsar timing is subject to independent noise, against an alternative hypothesis, which posits an identical noise process across different pulsars.

Upon analyzing the data, there is strong evidence in favor of the existence of a common-spectrum process (CP) with a logarithmic Bayes Factor greater than 15.0, compelling support for the CP model over the null model where pulsar noise is assumed to be independent. They estimate the gravitational-wave strain amplitude to be $A = 2.2^{+0.4}_{-0.3} \times 10^{-15}$ at a reference frequency of 1~yr$^{-1}$. These findings show consistency with the NANOGrav's earlier reports. However, this evident process lacks the spatial correlations expected from a gravitational-wave background, specifically those predicted by the Hellings-Downs model.

Potential Model Misspecification

One notable issue discussed by Goncharov et al. is potential model misspecification. Simulation studies presented in the paper reveal that models assuming uniform prior distributions for noise parameters may erroneously show evidence for a common process even when one is not truly present. This suggests that the assumed models might be sensitive enough to detect random fluctuations aligned by chance.

Implications

The implications of this paper are multifaceted. Practically, the identified common-spectrum process may provide constraints on the demographic and evolutionary models of supermassive black hole binaries, inferred to be significant sources of nanohertz gravitational waves. The absence of spatial correlations calls into question whether the detected CP is indeed attributable to gravitational waves or if artifacts within current analysis frameworks may influence observations.

Theoretically, the demonstration of a common spectral process challenges existing models of pulsar timing noise and suggests refinements in noise-modeling strategies for pulsar timing arrays. Future work may benefit from testing alternate noise hypotheses that accommodate the variability of noise parameters across different pulsars.

Future Directions

The research heralds the launch of further investigations using combined datasets from international pulsar timing arrays. Improvement in baseline length, heightened reception sensitivity, and broader data aggregation will enhance detection prospects of spatial correlations unique to genuine stochastic gravitational-wave backgrounds. Han et al.’s work lays a critical foundation for refining methodologies that differentiate between astrophysical noise and gravitational-wave-induced signals.

In conclusion, while the existence of a common-spectrum process aligns with theoretical predictions for gravitational-wave backgrounds, the results from the PPTA show that more rigorous analyses and improved data quality are necessary before conclusive claims can be made regarding the origins of this signal.