GWTC-3: A Gravitational-Wave Transient Catalog

• GWTC-3 is a comprehensive gravitational-wave transient catalog that compiles 90 robust compact binary coalescence events, including BBH, NSBH, and earlier BNS signals.
• It employs advanced multi-pipeline searches and full Bayesian inference to extract detailed source properties such as mass, spin, and redshift distributions.
• The catalog underpins rigorous tests of general relativity and enhances multi-messenger astronomy through improved detection methods and galaxy catalog integrations.

The GW Transient Catalog 3 (GWTC-3) is the third official gravitational-wave (GW) transient catalog produced by the LIGO–Virgo–KAGRA (LVK) collaboration, comprising all events from compact binary coalescences (CBCs) observed in the advanced detector network up to the end of the third observing run (O3) (Collaboration et al., 2021). GWTC-3 catalogs GW signals from binary black holes (BBH), neutron star–black hole binaries (NSBH), and, in earlier observing runs, binary neutron stars (BNS). By nearly doubling the number of statistically significant GW events over its predecessor (GWTC-2.1), GWTC-3 serves as the baseline dataset for statistical analyses of the mass, spin, and redshift distributions of compact objects, multiple tests of general relativity (GR), studies of astrophysical population synthesis, and cosmological parameter inference.

1. Catalog Scope, Detection Criteria, and Event Sample

GWTC-3 contains 90 robust candidates with an astrophysical probability $p_\mathrm{astro} > 0.5$, identified in O3a and O3b, as well as all eligible events from the previous observing runs (O1, O2). Detection is based on multi-pipeline searches (GstLAL, PyCBC, MBTA, and coherent WaveBurst), spanning matched filtering for compact binary waveforms and burst algorithms for unmodeled transients (Collaboration et al., 2021). The significance of each candidate is assessed through thresholds on false alarm rate (FAR; typically $<1~\text{yr}^{-1}$) and astrophysical probability:

$p_\mathrm{astro} > 0.5$

Candidates are parameterized through full Bayesian inference, employing semi-analytical and phenomenological waveform models (e.g., IMRPhenomXPHM, SEOBNRv4PHM) and, for select heavy BBHs, numerical relativity surrogates (NRSur7dq4) (Islam et al., 2023). The resulting sample spans source-frame total masses from $\sim 2.7~M_\odot$ up to $173~M_\odot$, including the first detections of NSBH systems in O3b. No credible BNSs were observed during O3b.

2. Source Property Inference and Population Distributions

Source parameter estimation in GWTC-3 leverages multidetector strain data and advanced waveform models to extract posterior distributions of chirp mass, mass ratio, spins, inclination, luminosity distance, and sky location. Component mass posteriors are central to population and evolutionary inference and are often given by

$$\mathcal{M} = \frac{(m_1 m_2)^{3/5}}{(m_1 + m_2)^{1/5}} \quad M = m_1 + m_2$$

Statistical analyses with these posteriors enabled the identification of features such as:

• A broad BBH mass distribution with no sharp cutoffs, possible “mass gaps,” and a low-mass tail extending toward the classical neutron-star regime (Collaboration et al., 2021).
• The first secure NSBH detections, with some candidates consistent with primaries in the so-called “lower mass gap” ($\sim 3-5~M_\odot$).
• Significant diversity in spin magnitudes and orientations. Comprehensive hierarchical analyses reveal the BBH inclination distribution is fully consistent with random orientations, with skewness $\mathcal{S}_\text{post} = 0.01^{+0.17}_{-0.17}$ and nearly zero Jensen-Shannon divergence ($1.4\times10^{-4}$ bits) when compared to an isotropic expectation (Vitale et al., 2022).
• Emergent evidence for a peak in the BBH mass distribution around $34\,M_\odot$, as inferred via hierarchical mass modeling (Collaboration et al., 2021).
• For a subset of heavy BBHs analyzed with NRSur7dq4, discrepancies between phenomenological and NR-based waveform inferences are significant in $\gtrsim 20\%$ of cases, affecting masses, spins, and especially remnant kick velocities (Islam et al., 2023).

These measurements, now performed on a larger and more diverse event set, enable improved constraints on compact object formation and evolution channels.

3. Astrophysical and Cosmological Applications

The inclusion of high-confidence BBH and NSBH events at luminosity distances reaching $z\sim1$ fundamentally augments the statistical power for population synthesis and cosmology. GWTC-3 has enabled:

• Improved merger rate calculations, with BBH, NSBH, and BNS rates tightly constrained using measured detection efficiencies and sensitive spacetime volumes ($VT$) (Collaboration et al., 2021).
• Cosmological inference using GW “dark sirens”—standard siren distances measured directly from GW signals, with statistical host identification and hierarchical Bayesian modeling. Key formulas include the luminosity distance–redshift relation:

$$H(z) = H_0 \sqrt{\Omega_m (1+z)^3 + \Omega_\Lambda}$$

Combined GWTC-3 dark sirens and GW170817 (the only BNS with electromagnetic counterpart) yield $$H_0=68^{+8}_{-6}~\text{km\,s}^{-1}\,\text{Mpc}^{-1}$$ under model assumptions (Collaboration et al., 2021, Mukherjee et al., 2022, Mancarella et al., 2022). The constraints on $H_0$ now approach classical methods, but remain sensitive to assumptions about the BBH mass distribution and host galaxy catalogs.

• Cross-correlation techniques using galaxy catalogs (e.g., GLADE+) allow for statistical assignment of redshifts and marginalization over host properties. While the dark siren constraint alone is weak, joint analyses with the “bright siren” GW170817 meaningfully improve parameter posteriors.
• The formation time and redshift distribution of BBHs can be reconstructed using GWTC-3's redshift leverage, mapping the delay time distribution $p(\tau) \propto \tau^{\alpha_\tau}$ (with minimum delay time $\tau_\text{min}=10~\text{Myr}$) to infer, for example, that at least one BBH system formed at $z_\text{form}>4.4$ (Fishbach et al., 2023). The BBH progenitor formation yield at $z=4$ is $6.4^{+9.4}_{-5.5}\times 10^{-6}\,M_\odot^{-1}$ and the inferred mean metallicity at that epoch is $$\langle\log_{10}(Z/Z_\odot)\rangle=-0.3^{+0.3}_{-0.4}$$.

4. Tests of General Relativity and Fundamental Physics

GWTC-3 has enabled the most precise and comprehensive strong-field tests of gravity to date. The main GR tests include (Ghosh, 2022, Xie et al., 2 Oct 2025, Santos et al., 26 Mar 2024):

• Inspiral–merger–ringdown (IMR) consistency tests: separate parameter estimation on inspiral and post-merger segments constrains remnant mass and spin deviations, yielding improvements in consistency posteriors by up to a factor of 1.8 relative to previous catalogs.
• Parametrized post-Newtonian (PN) and post-Einsteinian (ppE) analyses: constraints on phase and amplitude corrections at multiple PN orders, including pre-Newtonian $(-1\text{PN})$ dipole radiation (improved bounds especially with long inspiral NSBH events).
• A neural post-Einsteinian (npE) framework is introduced, utilizing a variational autoencoder to construct a continuous latent space for waveform dephasings, covering both traditional PN and non-PN deviations (e.g., massive scalar or dark photon fields). In GWTC-3, hierarchical modeling across events places two-dimensional constraints in the npE space, fully encompassing previous bounds and not revealing significant violations of GR (Xie et al., 2 Oct 2025).
• Black hole remnant ringdown studies and echo searches: combined analyses of the dominant QNM mode (ℓ=2, m=2, n=0) show fractional deviations in frequency (δf$_{220}$) and damping time (δτ$_{220}$) consistent with zero, with a further improvement in bounds by factors of $1.7$–$5.5$ over earlier runs.
• Propagation and dispersion tests: GWTC-3 data exclude anomalous friction and dispersion in the GW propagation equation at high precision—e.g., the friction term parameter $\nu = 0.5^{+3.5}_{-2.6}$ (90% interval) is an order of magnitude improvement over previous GW170817-only results, limiting any time variation of the effective Planck mass (Chen et al., 16 May 2024). Using beyond-Kerr metrics (e.g., Johannsen–Psaltis with parameter $\epsilon_3$) mapped to the ppE framework, GWTC-3 tightens constraints on deviations from Kerr geometry to $|\epsilon_3| \lesssim 0.5$ (95% confidence), an order of magnitude better than X-ray reflection constraints (Santos et al., 26 Mar 2024).

No significant deviations from GR predictions have been found, and GWTC-3 enables stringent bounds on alternatives, including scalar-tensor, dynamical Chern–Simons, Einstein–æther, khronometric, and varying-G theories.

5. Multi-messenger Astronomy and Galaxy Catalog Integration

Electromagnetic (EM) follow-up and multi-messenger studies are heavily informed by GWTC-3 event localizations but complicated by large sky-area uncertainties. The exploitation of rapid galaxy cataloging—specifically, on-the-fly H$\alpha$ surveys using 1–2 m telescopes (e.g., Palomar Transient Factory, MDM), with extended filters to $z \lesssim 0.05$—achieves completeness $\sim 90\%$ with respect to star formation rate in less than a week, covering hundreds of square degrees (Bartos et al., 2014). This sharply reduces the effective search area and false positive rates for transient EM counterparts (e.g., kilonovae).

Further prioritization is achieved by integrating galaxy catalogs (GWGC, GLADE, WISExSCOS) with GW localization maps, using metrics such as blue luminosity, stellar mass, and metallicity (inferred via empirical color–magnitude relations or Bayesian SED fitting) to guide follow-up (Li et al., 2016). Interactive visualization tools (e.g., Skymap Viewer) combine GW posterior probability maps with astrophysical priors, facilitating optimal targeting based on galaxy properties correlated with merger likelihood.

For low-energy neutrino counterparts, limits set by the Borexino experiment for 74 GWTC-3 triggers represent the most stringent constraints to date in the $0.5-5\,\text{MeV}$ range, further informing physical models for matter and radiation in CBCs (Collaboration et al., 2023).

6. Implications for Black Hole and Dark Matter Populations

GWTC-3 has supported sophisticated population modeling efforts, especially for the origin of BBHs:

• Analyses under the primordial black hole (PBH) hypothesis fit event masses to models including lognormal, power-law, critical collapse, and broken power-law mass functions. The GWTC-3 data decisively disfavors a simple power-law PBH mass function (Bayes factor $\log_{10}\mathcal{B}\gtrsim3$), with a lognormal distribution ($M_c \sim 17\,M_\odot$, $\sigma\sim0.7$) strongly preferred (Chen et al., 2022, Liu et al., 2022).
• The fraction of dark matter in stellar-mass PBHs is constrained to $f_\text{PBH}\sim10^{-3}$—PBHs cannot dominate the dark matter density, but can contribute as a minority channel.
• Inclusion of both “astrophysical” BHs and PBHs (the ABH+PBH model), and accounting for first- and multi-merger channels, allows a better fit to the mass/redshift distribution and robust inference on $H_0$.
• Analyses demonstrate that multi-merger effects are negligible ($\lesssim1-2\%$ of the overall BBH merger rate), enabling simpler single-generation population analyses for future constraints (Liu et al., 2022).

7. Future Prospects and Catalog Expansion

GWTC-3 has established the standard for GW transient catalogs, but the landscape is rapidly evolving:

• GWTC-4.0, covering the first half of O4, more than doubles the statistically reliable event count (218 candidates), probes further into the parameter space (higher/low masses, asymmetric mass ratios, higher redshifts, higher SNR), and facilitates even more fine-grained statistical and fundamental-physics studies (Collaboration et al., 25 Aug 2025).
• Event verification and Bayesian parameter estimation have been harmonized and automated, with reproducible workflows (e.g., Asimov blueprints) enabling consistent cross-comparison with independent community catalogs (Williams, 15 Jan 2024). Jensen–Shannon divergence and related metrics are now standard for quantifying agreement between analyses.
• The groundwork is laid for next-generation catalogs and multi-messenger discoveries, with ongoing improvements in detector sensitivity, waveform modeling (including full precession and higher-order modes), galaxy catalog completeness, and multi-modal candidate identification.

GWTC-3 thus represents both a culmination of the first era of gravitational-wave astronomy and a critical stepping stone toward comprehensive population, cosmological, and fundamental-physics investigation with gravitational-wave sources.