Ringdown Tests and Strong-Field Spectroscopy

Ringdown observations turn the remnant of a compact-binary merger into a strong-field consistency experiment. The chapter’s central discipline is to keep three layers distinct: black-hole perturbation theory predicts a spectrum, radiation modeling predicts how that spectrum enters a waveform, and statistical inference asks which parts of that waveform the detector data resolve. Confusing those layers can turn a useful fit into an overstated physical claim.

Throughout the chapter the time dependence is $e^{-i\omega t}$, with

$$\omega=\omega^R-\frac{i}{\tau}, \qquad \tau>0,$$

so a stable quasinormal contribution decays as $e^{-t/\tau}$. Observed time scales use detector-frame remnant mass $M_f^{\rm det}=(1+z)M_f^{\rm src}$. The global spacetime and Fourier conventions of the volume otherwise apply.

Helpful background. Binary Parameters and Observable Combinations supplies the inspiral–merger parameter combinations propagated into the remnant. For the underlying theory, consult Quasinormal Boundary Conditions and Resonances; for waveform construction, consult Quasinormal-Mode Waveform Expansion.

Enter this chapter

The pages form one inference chain.

1. Predicting the Remnant from the Full Signal pushes the inspiral–merger posterior through calibrated remnant relations and then through the Kerr spectral map. It keeps source-frame and detector-frame masses separate and makes the later ringdown comparison genuinely predictive.
2. Ringdown Likelihoods and Start-Time Dependence builds a finite-window likelihood. It shows why truncation changes covariance and why amplitude priors must transform when the reference time moves.
3. Overtones and Mode Identifiability distinguishes waveform reconstruction, predictive necessity, and physical mode identification. A hierarchy of diagnostics shows what extra information is needed to move from an improved fit to a physically identified mode.
4. Multimode Black-Hole Spectroscopy asks whether independently informative complex frequencies lie on the spectrum of one Kerr remnant. Mode separation, excitation, angular mixing, and detector response are all part of the answer.
5. No-Hair and Parametrized Ringdown Tests compares remnant-consistency coordinates with free spectral deformations, exposing prior-volume, labeling, trials, and catalog-combination assumptions.
6. Echo and Near-Horizon Searches treats delayed-response searches as a contested frontier. It derives the cavity time scale while requiring a full background, injection recovery, and model-bounded interpretation.

This order matters. Starting with a free-frequency fit before defining the predicted remnant, window, and mode-resolution criterion makes the meaning of “agreement with Kerr” ambiguous.

For a first quantitative pass, solve the redshifted-frequency exercise, transport one amplitude between reference times, and calculate the two-template overlap before reading the observational examples. These three short calculations explain why remnant prediction, start time, and mode identifiability cannot be separated.

The common likelihood and its assumptions

Every empirical claim in this chapter should identify:

• the event and detector network;
• the strain, calibration, and data-quality products;
• the detector-frame peak and start-time convention;
• the analysis window, conditioning, and noise covariance;
• the angular basis, mode labels, and finite waveform content;
• the amplitude, frequency, remnant, and deviation priors;
• the injection, off-source, and posterior-predictive checks.

The likelihood is conditional on these choices. A result that survives one start-time scan but fails under a plausible noise treatment is not robust; a Bayes factor that changes under an equally defensible amplitude prior is not a prior-independent fact. Conversely, agreement across calibrated variations is positive evidence and should be reported as such.

How to read the frontier evidence

The observational assessment is pinned to 2026-08-18. GW250114 and the official GWTC-5.0 tests-of-general-relativity products are the selected high-information event and catalog examples at that cutoff LVK 2025, LVK 2026. They support stringent Kerr-consistency and spectroscopy analyses under their released assumptions. They do not make start time, waveform incompleteness, calibration, or mode identifiability disappear.

Each frontier page therefore uses the same reading rule:

Observation	Legitimate conclusion	Illegitimate shortcut
A larger mode set fits better	The added flexibility or constrained physics improves that comparison.	Every component has been independently measured.
A deviation posterior includes zero	The tested model is compatible with Kerr at the reported sensitivity.	Kerr has been proved or all alternatives excluded.
A local echo statistic is large	The candidate merits a globally calibrated follow-up.	The single-template tail probability is the search false-alarm probability.
Several events combine tightly	The stated population model gains precision.	All events share a universal deviation without testing that assumption.

Refresh these statements when official strain or calibration changes, the GWTC TGR release is superseded, or independent analyses materially change recovery or false-positive rates.

Review the chapter

1. Why must a ringdown amplitude always be accompanied by a reference time?

Solution

For a damped mode,

$$A(t_a)=A(t_0)e^{-i\omega(t_a-t_0)}.$$

Its magnitude and phase therefore change with the chosen reference. Comparing amplitudes quoted at different times without transporting them can manufacture apparent disagreement, especially for short-lived overtones.

2. Give one test that separates a resolved second mode from a flexible correction to an inadequate one-mode waveform.

Solution

Fit a model in which the second complex frequency is free, repeat across start times and plausible noise treatments, and test recovery on two-mode injections together with false identification on one-mode injections. Localization near a predicted spectral branch that is stable across those checks supplies independent spectral information; a constrained amplitude that appears only for one window does not.

3. Explain why a null echo search does not exclude every horizon-scale modification.

Solution

The search has sensitivity only to its specified delays, frequency band, reflectivity, morphology, event set, and data quality. Other boundary conditions can yield weak, dispersive, irregular, or absent echoes. The null result therefore constrains the searched model family and parameter region.

References

• LIGO Scientific Collaboration, Virgo Collaboration, and KAGRA Collaboration (2025). “Black Hole Spectroscopy and Tests of General Relativity with GW250114.” arXiv:2509.08099.
• LIGO Scientific Collaboration, Virgo Collaboration, and KAGRA Collaboration (2026). “GWTC-5.0: Tests of General Relativity.” arXiv:2607.19293.

Further reading

• Isi, M., and Farr, W. M. (2021). “Analyzing black-hole ringdowns.” arXiv:2107.05609.