Multimessenger Probes of the Supermassive Black Hole Binary Population: The Role of Pulsar Timing Arrays

Technical Summary: Multimessenger Probes of the Supermassive Black Hole Binary Population: The Role of Pulsar Timing Arrays

Problem Statement
Pulsar Timing Arrays (PTAs) have detected evidence of a nanohertz gravitational wave background (GWB), likely originating from a population of supermassive black hole binaries (SMBHBs). While PTAs offer a multimessenger approach to astronomy, it remains unclear to what extent PTA data can constrain specific astrophysical parameters of the SMBHB population, particularly those difficult to measure via electromagnetic (EM) observations (e.g., sub-parsec hardening mechanisms). Existing inference efforts often report parameter estimates without rigorously assessing model viability, parameter degeneracies, or the extent to which posterior distributions are shaped by prior choices rather than data. This work addresses the need to quantify the "multimessenger promise" of PTAs by determining which SMBHB population parameters are genuinely informed by GW data versus those that remain prior-dominated.

Methodology
The authors employ a Bayesian inference framework to analyze 200 realistic, simulated PTA data sets. The study utilizes the phenom semi-analytical model for SMBHB demographics and evolution (previously used in NANOGrav 15-year analyses), which is implemented via the holodeck software. The model includes six free parameters:

1. $\tau_f$: Total evolutionary lifetime of the binary (Gyr).
2. $\nu_{inner}$: Hardening rate index at small separations.
3. $\psi_0, m_{\psi,0}$: Normalization and characteristic mass of the Galaxy Stellar Mass Function (GSMF).
4. $\mu, \epsilon_\mu$: Normalization and intrinsic scatter of the MBH–MBulge mass relation.

Two distinct simulated data regimes were constructed:

• A4cast: A realistic data set based on NANOGrav 12.5-year data (45 pulsars, 20-year baseline) with realistic timing noise and cadence.
• SimGWB: An idealized data set with 90 uniformly distributed pulsars, 20-year baselines, regular 14-day cadence, and negligible white noise (10 ns), representing a high-significance future scenario.

The inference pipeline uses the pandora software package with PTMCMC for sampling. To handle the high-dimensional mapping between population parameters ($\theta$) and the GWB power spectral density (PSD) amplitudes ($\rho$), the authors utilize normalizing flows trained on holodeck samples to emulate the likelihood $\pi(\rho|\theta)$. The analysis evaluates the "learnability" of parameters by comparing posterior distributions against priors using Hellinger distances, uncertainty reduction metrics, and posterior-to-prior probability ratios. Additionally, random forest regression is employed to quantify feature importance and parameter sensitivity across frequency bins.

Key Results
The analysis of 200 realizations for both data sets yields the following quantitative findings:

• Constrainable Parameters: PTA data provides significant information gain for parameters governing binary evolution and the GSMF characteristic mass.

  • Hardening Rate ($\nu_{inner}$): There is a $>50\%$ chance of reducing prior uncertainty by more than 50%.
  • Evolutionary Lifetime ($\tau_f$): There is a 25–75% chance of reducing prior uncertainty by 25–75%.
  • GSMF Characteristic Mass ($m_{\psi,0}$): There is an 88% chance of reducing prior uncertainty by 25–50%.
  • In the idealized SimGWB scenario, these constraints are tighter, with uncertainty reductions often exceeding 50% for $\nu_{inner}$ and $\tau_f$.

• Prior-Dominated Parameters: PTA data offers only marginal information gain for parameters already well-constrained by electromagnetic observations.

  • MBH–MBulge Parameters ($\mu, \epsilon_\mu$): Posterior distributions for these parameters remain largely indistinguishable from the priors across all realizations, regardless of data quality. The Hellinger distances between prior and posterior are low, indicating that PTAs cannot further constrain the black hole mass normalization or intrinsic scatter beyond EM constraints.
  • GSMF Normalization ($\psi_0$): For realistic data sets (A4cast), the constraining power on $\psi_0$ is minimal.

• Sensitivity Analysis: Random forest regression confirms that the GWB PSD amplitude is highly sensitive to $\tau_f$, $\nu_{inner}$, and $m_{\psi,0}$, while showing negligible sensitivity to $\mu$ and $\epsilon_\mu$ (feature importance scores $<10\%$).

• Data Realism: Constraints derived from the realistic A4cast data are generally weaker than those from SimGWB, with the A4cast 68% HDI for the GWB spectrum parameter being approximately 7 times larger on average.

Significance and Claims
The paper delineates the specific domains where gravitational-wave data complements electromagnetic observations in the study of SMBHBs. The authors claim that:

1. Multimessenger Synergy: PTAs are uniquely capable of probing dynamical processes (hardening mechanisms, evolutionary lifetimes) that are inaccessible to EM observations, while EM data remains superior for constraining the overall mass function normalization and black hole scaling relations.
2. Model Viability: The study provides a framework to assess model viability, demonstrating that for certain parameters, PTA data does not significantly update astrophysical beliefs beyond existing EM constraints.
3. Future Outlook: The results suggest that future PTA efforts should focus on refining constraints on sub-parsec hardening physics and binary lifetimes, as these represent the primary "information gain" available from the GWB signal. The work cautions that without complementary EM constraints, the MBH–MBulge relation parameters will remain poorly constrained by PTAs alone.

The authors maintain a modest stance, noting that their conclusions are model-dependent (specific to phenom) and that the specific choice of true parameter values in the simulations may influence recovery rates. They do not claim that PTAs will solve all SMBHB population questions, but rather define the boundaries of what can be learned via this specific multimessenger approach.