Gravitational-Wave Speed, Dispersion, and Damping
Scrollable formula. Use Left and Right arrow keys to reveal hidden content.
Scrollable table. Use Left and Right arrow keys to reveal additional columns.
Scrollable figure. Use Left and Right arrow keys to inspect hidden content.
Gravitational waves test propagation in several logically distinct ways: an electromagnetic counterpart compares arrival speeds, waveform phasing constrains dispersion, populations of standard sirens probe amplitude damping, and polarization-dependent effects test birefringence. Through the evidence cutoff of 2026-08-20, none of the tested channels requires a departure from general relativity (GR). That compact statement is deliberately bounded: the channels cover different frequencies, redshifts, source assumptions, and phenomenological parameterizations, and GWTC-5 did not update every historical propagation test.
Helpful background. Propagation Speed, Dispersion, and Damping develops the propagation equation and the modified-dispersion and friction parameterizations used here. The present record asks what the dated observations establish about those parameters.
Four propagation questions, not one
Section titled “Four propagation questions, not one”The finite question is whether observations, within their measured frequency and redshift ranges, require any of the following: a gravity-wave speed different from the speed of light; frequency-dependent phase velocity; amplitude evolution different from the GR luminosity-distance law; or polarization-dependent propagation. A coefficient constrained under a constant-power-law ansatz is not a bound on every possible frequency function. Likewise, a speed comparison for one binary-neutron-star event is not a measurement of cosmological damping.
| Assessment at 2026-08-20 | Meaning |
|---|---|
| Question status | Partially resolved: the observed bands and prescribed coefficients are strongly constrained, but frequency, redshift, and polarization coverage remains incomplete |
| Evidence status | Supported: no departure is required in the tested parameterizations |
| Confidence | Moderate overall; high for the GW170817 speed row, moderate for dispersion and damping, and indeterminate for untested birefringent forms |
| Domain | Ground-band compact binaries, GW170817/GRB 170817A, GWTC-3 dispersion tests, and the stated GWTC-5 standard-siren damping model |
| What could change the conclusion | A new joint GW/electromagnetic event, an official dispersion or birefringence update, or a material calibration or population-model revision |
Entry project: expose the source-delay assumption in a speed bound
Section titled “Entry project: expose the source-delay assumption in a speed bound”This is a useful first project for a student with cosmological-distance and uncertainty-propagation skills. It needs only a laptop or a short analytic calculation. A new dispersive-waveform or standard-siren population analysis is a different undertaking requiring waveform inference or cosmological-selection expertise.
Read Abbott et al. 2017, Eqs. (1)–(2) and source-delay discussion, official manuscript P1700308-v8. Extract the observed arrival delay, its sign, the conservative distance input, and the assumed interval between gravitational-wave and gamma-ray emission at the source. Re-derive the quoted speed interval using those inputs, distinguishing an emission delay from a propagation delay and restoring units. Keep the nearby-source approximation and distance uncertainty explicit rather than applying that equation unchanged at arbitrary redshift.
Deliver a derivation and a small plot of the allowed speed difference as the assumed intrinsic-delay interval varies. Mark the original paper’s assumptions as the baseline and other intervals as sensitivity scenarios, not new measurements. The exercise succeeds if another reader can reconstruct how the astrophysical source assumption produces each endpoint. It does not require downloading an entire strain archive and does not establish a frequency-dependent dispersion or amplitude-damping bound.
A candidate extension, with novelty unchecked, is to compare a small number of explicit gamma-ray emission models and determine whether they justify meaningfully different intrinsic-delay distributions. This requires a collaborator who understands merger ejecta and jet emission. Stop if the proposed delay distribution is chosen only because it tightens the gravity constraint, or if the association hypothesis is not secure. A sensible pivot is to publish a reusable sensitivity calculation with no new bound. Before moving to dispersion, read LVK 2023, GWTC-3 tests, §VII and identify the different observable, distance integral, and coefficient convention.
What is established channel by channel
Section titled “What is established channel by channel”For speed, GW170817 and GRB 170817A arrived close enough in time after roughly 40 Mpc of propagation to bound the fractional gravity–light speed difference extremely tightly once an allowed intrinsic emission delay is specified. The original collaboration paper gives the timing construction and both delay brackets in its abstract and Eqs. (1)–(2) on pp. 2–4 Abbott et al. 2017, abstract and pp. 2–4. The result is direct and powerful, but conditional on the association and on bracketing the source delay.
For dispersion, compact-binary phasing constrains prescribed power-law modifications to the group velocity. GWTC-3 found no compelling anomaly in the tested coefficients; the relevant likelihood, exponent basis, and combined bounds are in §VII and Tables IX–X of the catalog test paper LVK 2023, §VII and Tables IX–X. An independent reanalysis extended the negative-power cases and likewise reported no evidence, while showing that parameterization and sampling choices matter Baka et al. 2025, arXiv v2, §§II–IV and Tables I–II.
For damping, the current GWTC-5 standard-siren analysis is consistent with the GR relation between gravitational-wave and electromagnetic luminosity distances. Its inference uses 236 sources, GW170817, DES-Y6 host information, population models, and two implementations; the modified-propagation result and its dependencies are stated in §4.5 and the limitations in §5 LVK 2026a, §§2–5. This is a hierarchical cosmology result, not a waveform-phase test.
Results and open questions at 2026-08-20
Section titled “Results and open questions at 2026-08-20”| Claim or evidence | What it supports | What it does not establish |
|---|---|---|
| GW170817/GRB 170817A timing | Near-luminal propagation for the observed source and frequency band under the stated source-delay bracket | Broadband nondispersion, cosmological damping, or parity symmetry |
| GWTC-3 modified-dispersion posteriors | No required deviation for the tested power-law phase coefficients | Arbitrary frequency dependence or a complete Lorentz-violating theory |
| Independent negative-power reanalysis | Robustness of the null conclusion to an independent implementation in its tested cases | New independent strain data; it reuses GWTC-3 |
| GWTC-5 standard-siren damping analysis | Compatibility of the stated amplitude-distance model with GR | Phase propagation, a population-independent bound, or a unique modified-gravity interpretation |
| GWTC-5 test-suite scope statement | A precise boundary on what the current catalog did and did not test | A new dispersion or anisotropic-birefringence null result |
The high evidentiary weight of the speed row comes from its long propagation baseline and two messengers. The dispersion and damping rows use many events but carry stronger waveform, population, selection, and prior dependence. They agree in not requiring a departure, yet they are neither statistically independent in every respect nor interchangeable.
Competing interpretations and their domains
Section titled “Competing interpretations and their domains”One interpretation says that the combination effectively closes the observational room for modified propagation. That is justified for many simple low-energy theories whose coefficients are approximately constant across the measured domain: a theory predicting an order-one low-redshift tensor-speed shift is untenable, and several smooth dispersion laws are sharply restricted.
The stronger universal version fails. A coefficient can run with frequency or environment, a parity-odd effect can evade an analysis not performed, and damping can be degenerate with source-population and host-catalog assumptions. The current review literature emphasizes that mappings from phenomenological coefficients to covariant theories are conditional and channel dependent Yunes, Siemens, and Yagi 2025, §§3.1–3.5 and §§4.3.6–4.4.3. The evidence therefore supports a bounded null conclusion rather than the claim that propagation has been proved to equal GR at all frequencies and epochs.
Counterexamples, null results, and failed routes
Section titled “Counterexamples, null results, and failed routes”The most instructive failed route is to treat an omitted test as a null measurement. The GWTC-5 tests paper explicitly says that dispersive propagation and anisotropic birefringence were not updated, for methodological and scope reasons LVK 2026b, §1, pp. 2–3 and Table 1. Their absence cannot strengthen or weaken the older numerical bounds.
A second failed route is to quote the arrival-time difference as a source-free speed measurement. Gamma rays need not be emitted at the merger time, so the intrinsic delay is part of the inference. A third is to translate a damping posterior into a universal friction law without propagating the mass spectrum, merger-rate evolution, galaxy completeness, cosmology, and selection function. These limitations narrow the verdict; none supplies positive evidence for non-GR propagation.
What would change the assessment
Section titled “What would change the assessment”A second well-localized joint event with a quantitatively modeled emission delay would test whether the speed result repeats independently. An official catalog analysis that updates dispersion and birefringence with injection-calibrated coverage could change the weaker rows. For damping, the assessment would move if two population/host pipelines, using a public selection function and materially different nuisance models, produced a stable non-GR distance ratio. Any claimed anomaly should survive at least two waveform families and a theory-consistent mapping over the redshift and frequency interval used.
Research checklist
Section titled “Research checklist”- State which propagation channel is being tested and its frequency and redshift support.
- Pin the event set, calibration release, parameter-estimation revision, waveform family, priors, and selection cuts.
- For multimessenger speed, expose the assumed intrinsic-delay interval and association probability.
- For dispersion, state the exponent, coefficient convention, group-versus-phase velocity, and whether coefficients are varied singly or jointly.
- For damping, reproduce the source-population, merger-rate, galaxy-completeness, and cosmology assumptions.
- Calibrate coverage with injections and report how the result moves across waveform or population models.
- Distinguish replotting released posteriors from regenerating them from calibrated strain.
Source and data map
Section titled “Source and data map”The speed result is anchored by the joint GW170817/GRB paper and its official DCC record. Dispersion uses the peer-reviewed GWTC-3 test paper plus the independent 2025 reanalysis. The current damping statement is pinned to LIGO-P2600018-v14. The scope of the current strong-field suite is pinned to LIGO-P2500781-v17, with figure products at LIGO-P2600410-v1.
Underlying strain and posterior products should use GWOSC GWTC-5.0 stable release 9, DOI 10.7935/bk00-6a89. Its 16,384-Hz source strain and 4,096-Hz decimated products are distinct, and data below 10 Hz are not calibrated. No end-to-end rerun was performed for this assessment. A real reproduction must pin the exact strain/posterior revision, calibration, event cuts, power spectral density, waveform, galaxy catalog, software revision, priors, injection selection function, and reported posterior or Bayes factor.
References
Section titled “References”- Abbott, B. P., et al. 2017, “Gravitational Waves and Gamma-Rays from a Binary Neutron Star Merger: GW170817 and GRB 170817A,” Astrophysical Journal Letters 848, L13, DOI 10.3847/2041-8213/aa920c.
- Baka, T., Cirok, B., Haris, K., Noller, J., and Krishnendu, N. V. 2025, “Testing General Relativity with Gravitational Waves—Improving and Extending Modified Dispersion Relation Tests,” arXiv:2511.00497v2, revised 27 February 2026, open preprint.
- LIGO Scientific Collaboration, Virgo Collaboration, and KAGRA Collaboration. 2023, “Tests of General Relativity with GWTC-3,” Physical Review D 107, 022005, DOI 10.1103/PhysRevD.107.022005.
- LIGO Scientific Collaboration, Virgo Collaboration, and KAGRA Collaboration. 2026a, “GWTC-5.0: Constraints on the Cosmic Expansion Rate and Modified Gravitational Wave Propagation,” LIGO-P2600018-v14, official record.
- LIGO Scientific Collaboration, Virgo Collaboration, and KAGRA Collaboration. 2026b, “GWTC-5.0: Tests of General Relativity,” LIGO-P2500781-v17, arXiv:2607.19293v1, official record.
- Yunes, N., Siemens, X., and Yagi, K. 2025, “Gravitational-Wave Tests of General Relativity with Ground-Based Detectors and Pulsar-Timing Arrays,” Living Reviews in Relativity 28, DOI 10.1007/s41114-024-00054-9.