Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Eccentricity or spin precession? Distinguishing subdominant effects in gravitational-wave data

Isobel M. Romero-Shaw, Davide Gerosa, Nicholas Loutrel

TL;DR

The paper addresses the challenge of distinguishing subdominant effects—orbital eccentricity and spin precession—in gravitational-wave BBH signals. It adopts a Bayesian framework using two waveform models (SEOBNRE for eccentricity with aligned spins and IMRPhenomPv2 for precessing quasi-circular spins) and a likelihood-reweighting strategy to compare competing hypotheses, given the current inability to jointly infer both effects. The key finding is that distinguishability improves with longer signals and favorable geometry (higher eccentricity or larger inclination), while certain configurations (e.g., GW190521-like short signals) remain non-discriminative; the work provides practical thresholds and methods, including a rho-based proxy when reweighting is unreliable, for guiding interpretation of subdominant GW features. Overall, the study clarifies when eccentricity or spin precession can be confidently identified and informs future waveform modeling that captures both effects jointly, with implications for astrophysical inference and tests of relativistic two-body dynamics.

Abstract

Eccentricity and spin precession are key observables in gravitational-wave astronomy, encoding precious information about the astrophysical formation of compact binaries together with fine details of the relativistic two-body problem. However, the two effects can mimic each other in the emitted signals, raising issues around their distinguishability. Since inferring the existence of both eccentricity and spin precession simultaneously is -- at present -- not possible, current state-of-the-art analyses assume that either one of the effects may be present in the data. In such a setup, what are the conditions required for a confident identification of either effect? We present simulated parameter inference studies in realistic LIGO/Virgo noise, studying events consistent with either spin precessing or eccentric binary black hole coalescences and recovering under the assumption that either of the two effects may be at play. We quantify how the distinguishability of eccentricity and spin precession increases with the number of visible orbital cycles, confirming that the signal must be sufficiently long for the two effects to be separable. The threshold depends on the injected source, with inclination, eccentricity, and effective spin playing crucial roles. In particular, for injections similar to GW190521, we find that it is impossible to confidently distinguish eccentricity from spin precession.

Eccentricity or spin precession? Distinguishing subdominant effects in gravitational-wave data

TL;DR

The paper addresses the challenge of distinguishing subdominant effects—orbital eccentricity and spin precession—in gravitational-wave BBH signals. It adopts a Bayesian framework using two waveform models (SEOBNRE for eccentricity with aligned spins and IMRPhenomPv2 for precessing quasi-circular spins) and a likelihood-reweighting strategy to compare competing hypotheses, given the current inability to jointly infer both effects. The key finding is that distinguishability improves with longer signals and favorable geometry (higher eccentricity or larger inclination), while certain configurations (e.g., GW190521-like short signals) remain non-discriminative; the work provides practical thresholds and methods, including a rho-based proxy when reweighting is unreliable, for guiding interpretation of subdominant GW features. Overall, the study clarifies when eccentricity or spin precession can be confidently identified and informs future waveform modeling that captures both effects jointly, with implications for astrophysical inference and tests of relativistic two-body dynamics.

Abstract

Eccentricity and spin precession are key observables in gravitational-wave astronomy, encoding precious information about the astrophysical formation of compact binaries together with fine details of the relativistic two-body problem. However, the two effects can mimic each other in the emitted signals, raising issues around their distinguishability. Since inferring the existence of both eccentricity and spin precession simultaneously is -- at present -- not possible, current state-of-the-art analyses assume that either one of the effects may be present in the data. In such a setup, what are the conditions required for a confident identification of either effect? We present simulated parameter inference studies in realistic LIGO/Virgo noise, studying events consistent with either spin precessing or eccentric binary black hole coalescences and recovering under the assumption that either of the two effects may be at play. We quantify how the distinguishability of eccentricity and spin precession increases with the number of visible orbital cycles, confirming that the signal must be sufficiently long for the two effects to be separable. The threshold depends on the injected source, with inclination, eccentricity, and effective spin playing crucial roles. In particular, for injections similar to GW190521, we find that it is impossible to confidently distinguish eccentricity from spin precession.
Paper Structure (6 sections, 9 equations, 1 figure)

This paper contains 6 sections, 9 equations, 1 figure.

Table of Contents

  1. Introduction
  2. Method
  3. Simulated sources
  4. Model-selection strategies
  5. Results
  6. Discussion & Conclusion

Figures (1)

  • Figure 1: Natural log Bayes factor of eccentricity vs. spin precession as a function of the number of orbital cycles visible in the inspiral. Positive (negative) values of $\ln \mathcal{B}_{\mathrm{E/P}}$ indicate a preference for the aligned-spin eccentric (spin precessing quasi-circular) model. The three panels contain results for each of the injection series described in Sec. \ref{['simsources']}. The top panel shows results for eccentric and non-spinning injections, the middle panel shows results for eccentric and spin-aligned injections, and the bottom panel shows results for quasi-circular and spin precessing injections. For each run, the value of the detector-frame eccentricity $e_\mathrm{10~Hz}$ is indicated by the face colour of the marker and the corresponding source-frame eccentricity $e_{\mathrm{13~cycles}}$ is indicated by the edge colour. Each marker is linked to others in the same injection subset with a grey line. The linestyle indicates the inclination of the source; solid, dot-dashed, and dashed lines are used for $\theta_{JN}=\pi/10$ (i.e. similar to that of GW190521), $\theta_{JN}=\pi/4$, and $\theta_{JN}=\pi/2$ (i.e. edge-on), respectively. The significance region with $|\ln \mathcal{B}_\mathrm{E/P}| < 8$, a conventional value for establishing confidence that one model is preferred over another, is indicated with grey shading. The approximate number of orbital cycles in band for four eccentric-event candidates are indicated with vertical light-grey lines. For injections with eccentricity and aligned spins (middle panel) we pair the Bayes factors $\ln \mathcal{B}_{\mathrm{E/P}}$ (left vertical scale in black) to the approximate criterion based on $\rho$ from Eq. (\ref{['rhoproxy']}) (right vertical scale in green). For injection series with equivalent $\ln\mathcal{B}_\mathrm{E/P}$ values in this panel, we plot $\rho$ with unfilled markers to avoid overcomplicating the plot. We do not plot the fractional change in $\rho$ for edge-on systems since the results are very similar to those already shown for close to face-on injections.