{"id":"efff93c1-1e56-4ccb-9d4a-23785d1fbd6f","arxiv_id":"2412.07615","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A likelihood-based search with particle swarm optimization can detect and estimate supermassive black hole ringdown signals in simulated pulsar timing array data, reaching 99% detection probability at SNR 10.","lead":"This paper shows that a pulsar timing array can detect the gravitational wave ringdown of a supermassive black hole merger at frequencies previously thought out of reach, by combining observations from many pulsars taken at different times. The authors simulate a 100-pulsar array and report 99% detection probability at signal-to-noise ratio 10, opening a possible new way to test black hole physics.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the 99% detection / <0.2% FAP claim is internally consistent for the stated simulation; the dropped Eq. 10 term is not a mismatch because it lies in the template span, and the main limitations are explicitly acknowledged.","rationale":"The reader's conditional verdict is appropriate, but the specific weakest assumption overstates the impact of the dropped term in Eq. 10. Since Eq. 17's four time-dependent templates span all combinations of F+cos, F+sin, F×cos, F×sin with the same ω,τ, and the full Earth-term signal Eq. 10 is another such combination, the GLRT's maximization over extrinsic coefficients can represent it; only the physical reinterpretation of those coefficients would be biased. The detection numbers are therefore not hostage to the 15% statement. The substantial caveats that remain—idealized equal-variance white noise, no red noise, no higher multipoles, and an empirical FAP threshold from only 500 noise realizations—are acknowledged in the text or are small in effect; a 5,000-realization calibration plus a full-waveform injection would confirm this. I therefore would not move the reader's conditional verdict; the proof-of-principle claim is sound as stated.","tokens_in":18785,"tokens_out":22485,"duration_ms":227719,"concrete_test":"Run one extended Monte Carlo: generate 5,000 noise-only realizations to replace the observed-maximum threshold with a calibrated 0.2% false-alarm quantile, and inject the full Eq. 10 Earth-term waveform (exact Q factor, all four basis components) while searching with the Eq. 17 template at SNR 10. If the calibrated threshold stays within a few units of 30.1 and the detection probability remains above about 97%, the present conclusions are unchanged.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is explicitly about a controlled simulation: 100 pulsars, equal 100-ns white Gaussian noise, asynchronous sampling, and a (2,2)-only Earth-term injection generated with the same template as the search. Under those stated conditions, the reported numbers are internally consistent: the SNR-10 fitness distribution N(54.94,10.36) gives 99.2% above the H0 maximum threshold of 30.1, and the empirical count is consistent with that tail probability. I do not find a load-bearing internal error. The most often cited caveat—that Eq. 10 contains a second term and a Q2/(1+Q2) factor omitted in Eq. 17—is less serious than it looks for detection, because the omitted term is a linear combination of the same four basis functions (F+cos, F+sin, F×cos, F×sin) already used in Eq. 17; the F-statistic's free extrinsic coefficients can absorb it. The genuinely unprotected assumptions are the idealized white-noise model and the neglect of higher-order modes, but both are explicitly acknowledged as limitations, and the abstract's 'stringent no-hair tests' goes beyond the Table III precision (j std 0.27, M relative std 15%). This is a framing issue rather than an inconsistency in the detection claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a likelihood-based GLRT-PSO method for detecting and estimating supermassive black hole ringdown signals in pulsar timing array data, exploiting asynchronous sampling of multiple pulsars to overcome the per-pulsar Nyquist limit. The authors simulate a 100-pulsar PTA with 100 ns white Gaussian noise, inject (2,2)-mode Earth-term ringdown signals, and report a 99% detection probability at SNR>10 with false alarm probability <0.2%, along with parameter estimation biases and standard deviations for six sources. The method separates intrinsic and extrinsic parameters, with the latter maximized analytically à la F-statistic and the former searched via particle swarm optimization.","tokens_in":19039,"tokens_out":11957,"duration_ms":112309,"significance":"The central detection result is internally consistent: the reported 99% at SNR=10 matches the fitted Gaussian distribution in Fig. 2, the 500-realization Monte Carlo is adequate, and the PSO convergence checks in Fig. 3 support the optimization. The paper also releases its data and uses machine-checkable simulation consistency checks. The proposed search band above the per-pulsar Nyquist frequency is a novel and timely direction for SKA-era PTAs. However, the abstract's claim that the method enables 'stringent tests of the no-hair theorem' is not supported by the single-mode analysis and by the reported parameter precision (spin standard deviation 0.27, mass relative standard deviation about 15%).","major_comments":[{"comment":"The abstract and Introduction claim that the proposed method will lead to 'stringent tests of the no-hair theorem', but the analysis uses only the dominant (2,2) fundamental mode. A no-hair test requires comparing at least two quasinormal-mode frequencies or damping times (e.g., the (2,2) and (3,3) modes, or an overtone) to check GR-predicted consistency relations; with a single mode one can only estimate M and j under the assumption of GR. The paper's own Sec. V acknowledges that higher-order modes are excluded. Please either remove or substantially qualify the no-hair claim in the abstract, or add a multi-mode analysis that actually performs such a test.","section":"Abstract and Sec. I"},{"comment":"Even if a no-hair test were possible with the (2,2) mode alone, the reported precision does not support 'stringent' constraints: the spin parameter j has a standard deviation of 0.27 and the final mass M has a relative standard deviation of about 15% at SNR 10. These are not stringent bounds. The authors should either quantify the precision required for a meaningful no-hair test and show where the method reaches it (e.g., at higher SNR), or soften the wording throughout the manuscript.","section":"Table III and Sec. IV C"},{"comment":"The expression for the maximum detectable frequency f_h appears to contain a factor error. With ΔT = 1/(2 f_sp), one has τ/ΔT = 2 Q f_sp/(π f_h), which yields f_h = f_sp sqrt(2 Q N_p/π). The text writes f_h = (Q/(π f_h))(f_sp/2) N_p f_sp, which would give a different result. Please check the derivation and correct the formula.","section":"Sec. II, Eq. (30)"}],"minor_comments":[{"comment":"The false alarm probability is estimated from the same 500 noise realizations used to set the threshold (the maximum observed fitness). Please report the binomial uncertainty on this estimate or use an independent set of noise realizations to validate the FAP at the chosen threshold.","section":"Sec. IV B"},{"comment":"The paper does not report estimation results for the extrinsic parameters (ζ, ι, ψ, φ0), even though they are a central part of the analytic F-statistic step. Because Eq. (17) is an approximation to Eq. (10) (dropping the second term and setting Q^2/(1+Q^2)≈1), the extrinsic estimates could be biased. Please report their bias and standard deviation, or discuss why they are not relevant to the claims.","section":"Sec. III B and Table II"},{"comment":"There is a typographical space in 'F AST' in the abstract, and the notation 'NP fsp' after Eq. (30) should be 'N_p f_sp' for consistency.","section":"Abstract and Sec. II"}],"recommendation":"major_revision","confidential_remarks":"The paper's core detection simulation is sound and the Monte Carlo checks are convincing. The main weakness is the gap between the abstract's 'stringent no-hair tests' claim and the actual single-mode, low-precision parameter estimates. If the authors revise the overclaims and correct the noted technical issues, the paper would be a valuable contribution to the PTA gravitational-wave literature. The suitability for the journal depends on whether the authors are willing to frame the results as a proof-of-principle search methodology rather than as a demonstrated no-hair test."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis paper is the first end-to-end search pipeline with a quantitative performance characterization for SMBBH ringdowns in a large asynchronous PTA. The headline claim—99% detection probability at SNR 10 with false alarm below 0.2%—is internally consistent: the fitted Gaussian at SNR 10 (mean 54.94, sigma 10.36) gives 99.2% above the H0 threshold of 30.1, and the 500-realization count matches. The parameter estimation results are also believable: small biases across six sources, and the 13-hour start-time bias is a nice concrete detail.\n\nWhat is actually new here is the performance characterization, not the machinery. The GLRT-plus-PSO method with an analytic F-statistic for extrinsic parameters is inherited from earlier continuous-wave work, including the authors' own papers. Applying it to ringdown signals and actually measuring detection probability and parameter biases in a 100-pulsar asynchronous PTA is a real step forward. The paper is also appropriately honest about its limitations.\n\nThe soft spots: the injected signals use the same approximate template the search uses. The neglected second term in Eq. 10 is less serious for detection than it first looks, because it lies in the span of the same four basis functions and the F-statistic's free coefficients can absorb it. But the Q^2/(1+Q^2) about 1 amplitude factor is about 9% off for Q near 3.2, and no independent waveform is used anywhere. So the 99% number applies to a controlled simulation, not real PTA data. The white equal-variance noise model is also idealized. Both are acknowledged, but the abstract's 'stringent tests of the no-hair theorem' overstates what Table III shows—j std 0.27 and M relative std 15% is not stringent. That framing should be softened. One minor point: data are on Zenodo but code is not, which limits reproducibility for a methods paper.\n\nBottom line: this paper is for PTA data analysts and anyone planning SKA-era searches for resolvable SMBBH sources. The central engineering result is solid and internally consistent. It deserves a serious referee. With the abstract adjusted and, ideally, a robustness test against the full Earth-term expression, it should be publishable. I'd send it to review.","headline":"First quantitative end-to-end PTA ringdown search: internally consistent simulations, but waveform/noise idealizations are untested and the no-hair promise overstates the demonstrated precision.","tokens_in":19583,"tokens_out":4378,"would_cite":true,"duration_ms":41869,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A simulated 100-pulsar array catches supermassive black hole ringdowns 99% of the time.","keywords":["pulsar timing array","supermassive black hole ringdown","gravitational wave detection","generalized likelihood ratio test","particle swarm optimization","staggered sampling","no-hair theorem","quasi-normal modes"],"falsifier":"Take the paper's 100-pulsar, 100 ns simulation and inject ringdowns from the full Earth-term residual (Eq. 10, keeping the secondary term and the $Q^2/(1+Q^2)$ factor) or from a numerical-relativity waveform with higher-order modes, then count detections above the same threshold of 30.1 at SNR 10 with 500 realizations; if the detection fraction falls materially below 99% at a false-alarm probability of 0.2%, the claimed detection performance does not survive realistic waveform systematics.","tokens_in":18530,"feed_emoji":"📡","tokens_out":8301,"duration_ms":72575,"temperature":0.7,"pith_summary":"The paper argues that a pulsar timing array (PTA) can detect the gravitational-wave ringdown of a supermassive black hole merger even when the signal frequency sits above the usual single-pulsar sampling limit, provided the pulsars are observed asynchronously. To show this, it builds a coherent search that combines a generalized likelihood ratio test with an analytic F-statistic for the four extrinsic parameters and particle swarm optimization for the five intrinsic parameters. On a simulated 100-pulsar array with 100-nanosecond white noise and randomized observation times, the search reaches a 99% detection probability at a false-alarm probability below 0.2% for optimal signal-to-noise ratio above 10. The paper also reports accurate recovery of frequency, damping time, sky position, and start time, with enough precision to guide electromagnetic follow-up and to estimate the remnant black hole's spin and mass. If these results carry over to real data, PTAs would open a new mass window for testing the no-hair theorem.","feed_headline":"Simulated pulsar array catches black-hole ringdowns 99% of the time","feed_subtitle":"Staggered pulsar sampling breaks the old frequency limit and opens a new window on black hole tests.","key_machinery":"The mechanism that carries the argument is the combination of asynchronous pulsar sampling with a coherent search statistic. Asynchronous sampling raises the array's frequency reach from the single-pulsar Nyquist rate $f_{sp}$ to about $N_p f_{sp}$, because different pulsars observe the same high-frequency signal at different times; the signal model is the simplified Earth-term timing residual of Eq. 17, which keeps only the dominant (2,2) mode and drops a secondary term of Eq. 10. The extrinsic parameters (amplitude, inclination, polarization, initial phase) enter linearly after a reparametrization and are maximized analytically via the F-statistic, $F = (1/2) N^{T} M^{-1} N$, while particle swarm optimization searches the five-dimensional intrinsic parameter space (sky position, frequency, damping time, start time). PSO's role is to find the global maximum of this multimodal fitness landscape reliably, using eight parallel runs to drive the convergence probability effectively to one.","core_discovery":"The central claim is that a sufficiently large PTA with staggered, asynchronous pulsar sampling can detect and characterize the dominant (2,2) ringdown mode of a supermassive black hole merger using a likelihood-based search. For the paper's simulated array of 100 pulsars, each with 100 ns white timing noise and randomized observation epochs over five years, the generalized likelihood ratio test combined with particle swarm optimization achieves a detection probability of 99% at a false alarm probability below 0.2% once the optimal network SNR exceeds 10; the fitted Gaussian distribution at SNR 10 puts 99.2% of the fitness values above the threshold of 30.1. The same setup yields parameter estimates with negligible bias for frequency, damping time, sky location, and start time, with sky-localization errors around 5 degrees and a start-time bias of about 13 hours. From the estimated frequency and damping time, the paper derives a quality factor with a relative error of about 6.5% and estimates the remnant spin and mass with small biases. These numbers are presented as proof of principle that ringdown astronomy with PTAs is data-analysis-viable, not merely a frequency-argument curiosity.","pith_inferences":["Because the injected signals are generated with the same simplified template the search uses, the quoted 99% detection probability is an upper-bound-like estimate; injecting the full Eq. 10 residual and higher-order modes would test how much of this margin survives realistic waveform systematics.","The same GLRT-plus-PSO architecture should transfer to other short-lived PTA transients, such as memory signals or burst-like emission from eccentric binaries, since the F-statistic separation of parameters does not rely on the ringdown's exponential envelope.","If red noise is handled before the search, for example by subtracting a smooth spline fit, the white-noise results here could still approximately hold; the paper explicitly leaves this step for future work, so a natural next test is to inject red noise at observed levels and re-measure the detection fraction.","The 13-hour start-time bias, though small compared to optical survey cadences, suggests a systematic in the analytic maximization or PSO convergence that could matter for precision follow-up at higher SNR; checking whether the bias scales with SNR would clarify its origin."],"forward_implications":["PTA ringdown searches need not be limited to frequencies below the single-pulsar Nyquist rate; with staggered sampling the reach scales with the number of pulsars, so larger arrays directly buy higher-frequency sensitivity.","An equal-mass, non-spinning supermassive binary with chirp mass $9.52 \\times 10^{9}$ solar masses at 420 Mpc would be detectable at 99% probability with false alarm below 0.2% by a 100-pulsar, 100 ns array at SNR above 10.","Parameter estimation is accurate enough for practical follow-up: sky localization within roughly 5 degrees, frequency to about 7 rad/yr, and start time biased by only about 13 hours.","Ringdown-based estimates of quality factor, spin, and final mass can be recovered with small biases, giving a path toward no-hair theorem tests in the supermassive black hole mass range.","For higher ringdown frequencies the signal duration shortens, fewer pulsars fall inside the window, and parameter-estimation accuracy degrades; this sets a frequency ceiling below the formal array Nyquist limit."],"supporting_citations":[{"why":"Establishes that asynchronous observations extend the PTA frequency reach to about $N_p f_{sp}$ and sketches the SNR prospects for SMBBH ringdowns; this is the premise the search method tests.","marker":"[34]"},{"why":"Supplies the coherent GLRT and F-statistic machinery for continuous-wave PTA searches that this work adapts to ringdown signals.","marker":"[41]"},{"why":"Extends the coherent network analysis and underlies the separation of intrinsic and extrinsic parameters used in the ringdown search.","marker":"[42]"},{"why":"Introduces particle swarm optimization, the numerical global-search algorithm used to maximize the F-statistic over intrinsic parameters.","marker":"[43]"},{"why":"Provides the F-statistic formalism for analytically maximizing over extrinsic parameters, which the paper reparametrizes for the ringdown signal.","marker":"[44]"},{"why":"Gives the phenomenological amplitude formula $A_{2,2}=0.864\\eta$ used to set the injected ringdown signal amplitudes.","marker":"[50]"},{"why":"Provides the fitting formulas for ringdown frequency and quality factor as functions of spin, used for waveform generation and for mapping estimates to remnant mass and spin.","marker":"[51]"},{"why":"Supplies the final-mass formula for black hole mergers, used to connect progenitor parameters to the remnant mass in the simulations.","marker":"[52]"},{"why":"Supplies the final-spin formula used to set the spin parameter $j=0.69$ for equal-mass nonspinning binaries and to interpret spin estimates.","marker":"[53]"},{"why":"Provides the synthetic SKA pulsar catalog from which the 100 nearest pulsars are selected for the simulated array.","marker":"[62]"}],"fun_headline_variants":["Staggered pulsar sampling breaks frequency limit for ringdowns","100-pulsar array detects black-hole ringdowns with 99% success","PTA ringdown search: 99% detection at SNR 10","Staggered pulsars let PTA hear supermassive black-hole ringdowns","Asynchronous pulsar array detects black-hole ringdown signals"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The detection rates are computed on simulated data where the injected ringdown is generated from the same simplified Earth-term template the search uses, and the noise is white, stationary, and identical across pulsars; real waveforms carry extra terms and modes, and real timing noise is red and non-stationary.","fun_headline_variants_meta":{"raw":{"variants":["Staggered pulsar sampling breaks frequency limit for ringdowns","100-pulsar array detects black-hole ringdowns with 99% success","PTA ringdown search: 99% detection at SNR 10","Staggered pulsars let PTA hear supermassive black-hole ringdowns","Asynchronous pulsar array detects black-hole ringdown signals"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000519,"raw_usage":{"total_tokens":2596,"prompt_tokens":1108,"completion_tokens":1488,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":724,"completion_tokens_details":{"reasoning_tokens":1391}},"tokens_in":724,"tokens_out":1488,"duration_ms":9878,"temperature":1.0,"reasoning_tokens":1391,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T18:40:39.564240+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the paper's 100-pulsar, 100 ns simulation and inject ringdowns from the full Earth-term residual (Eq. 10, keeping the secondary term and the $Q^2/(1+Q^2)$ factor) or from a numerical-relativity waveform with higher-order modes, then count detections above the same threshold of 30.1 at SNR 10 with 500 realizations; if the detection fraction falls materially below 99% at a false-alarm probability of 0.2%, the claimed detection performance does not survive realistic waveform systematics.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes that asynchronous observations extend the PTA frequency reach to about $N_p f_{sp}$ and sketches the SNR prospects for SMBBH ringdowns; this is the premise the search method tests."},{"cited_title":"Yunes and X","cited_arxiv_id":null,"evidence_quote":"Supplies the coherent GLRT and F-statistic machinery for continuous-wave PTA searches that this work adapts to ringdown signals."},{"cited_title":"Askar et al., Black holes, gravitational waves and fun- damental physics: a roadmap, Classical and Quantum Gravity 36, 143001 (2019), publisher: IOP Publishing","cited_arxiv_id":null,"evidence_quote":"Extends the coherent network analysis and underlies the separation of intrinsic and extrinsic parameters used in the ringdown search."},{"cited_title":"Coherent network analysis for continuous gravitational wave signals in a pulsar timing array: Pulsar phases as extrinsic parameters","cited_arxiv_id":"1506.01526","evidence_quote":"Provides the F-statistic formalism for analytically maximizing over extrinsic parameters, which the paper reparametrizes for the ringdown signal."},{"cited_title":"Cotesta et al., Analysis of Ringdown Overtones in GW150914, Phys","cited_arxiv_id":null,"evidence_quote":"Gives the phenomenological amplitude formula $A_{2,2}=0.864\\eta$ used to set the injected ringdown signal amplitudes."},{"cited_title":"Kamaretsos, M","cited_arxiv_id":null,"evidence_quote":"Provides the fitting formulas for ringdown frequency and quality factor as functions of spin, used for waveform generation and for mapping estimates to remnant mass and spin."},{"cited_title":"Meidam et al., Testing the no-hair theorem with black hole ringdowns using TIGER, Physical Review D 90, 064009 (2014)","cited_arxiv_id":null,"evidence_quote":"Supplies the final-mass formula for black hole mergers, used to connect progenitor parameters to the remnant mass in the simulations."},{"cited_title":"Berti, V","cited_arxiv_id":null,"evidence_quote":"Supplies the final-spin formula used to set the spin parameter $j=0.69$ for equal-mass nonspinning binaries and to interpret spin estimates."}],"review_version":1}