{"id":"46076c11-a435-4e4d-8d0c-a897d32201a1","arxiv_id":"1908.08951","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Cross-correlating gravitational wave lensing with galaxy surveys is forecast to detect weak lensing of gravitational waves at z<0.5 within about 10 years, with black hole-neutron star mergers as the most promising source.","lead":"Gravitational waves passing through clumpy matter get slightly magnified, and this paper forecasts when we can see that effect by comparing gravitational wave events with galaxy maps. It predicts that within about a decade, LIGO-class detectors plus galaxy surveys should detect this weak lensing signal at low redshift, especially from black hole-neutron star mergers.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The BH-NS forecast's claimed gravity test rests on the unproven assumption that ϵ_s, the EM-derived distance error, is uncorrelated with δ and κ_gw; if ϵ_s correlates with environment or cosmology, the cross-correlation signal is biased at its own amplitude.","rationale":"The reader's weakest_assumption is exactly the load-bearing point: Eq. (17) requires ⟨ϵ_s κ_g⟩ and ⟨ϵ_s δ⟩ to vanish, and this is asserted, not shown. The issue is not merely a disagreement with consensus; it is an internal statistical gap in the estimator's derivation. The signal is small (C^{κgwδ}_l at the level of a few×10^{-7} in the figures' units, or a few 10^{-3} per-source fractional lensing), while ϵ_s from a 3% photo-z is ~5-10% per source, so even a small residual correlation (a few percent of ϵ_s) can shift the measured cross-correlation by order unity relative to the forecast. The paper's own noise model in Eqs. (20)-(21) includes a σ_b term from 'uncertain values of cosmological parameters,' acknowledging that cosmology dependence enters the distance estimate, but it does not propagate this dependence through the estimator or allow for a modified-gravity d_L(z). The conclusions also state that 'The detailed potential for constraints on the theoretical models... remain to be determined,' which confirms the gravity-test claim is not yet backed by a concrete modified-gravity forecast. This is addressable: a simulation or an analytic derivation of the residual correlators would settle it, and a single worked modified-gravity example (e.g., a non-zero μ or η in Eq. (9)) would show whether the claimed probe actually separates the lensing signal from the distance-redshift signal. Verdict should remain CONDITIONAL, with the condition being that the unbiasedness in Sec. 4.2.1 be demonstrated under realistic photometric-redshift and magnification conditions and that at least one modified-gravity model be propagated through the forecast to show the claimed test has discriminating power.","tokens_in":17859,"tokens_out":8986,"duration_ms":71035,"concrete_test":"Run a small end-to-end injection study: (a) generate a ΛCDM matter field with lensing convergence fields κ_gw, κ_g, and δ; (b) populate host galaxies for NS-NS/BH-NS events with photometric-redshift errors σ_z/(1+z)=0.03 that are correlated with galaxy properties (e.g., magnitude or local density, mimicking magnification bias); (c) apply the estimator of Eq. (17) exactly as written, computing d_L^es from the assigned redshift and fiducial cosmology; (d) compare the reconstructed C^{κgwδ}_l to the true input C^{κgwδ}_l. If the bias exceeds the forecast 1σ errors, the unbiasedness assumption fails. A second run with d_L(z) shifted by 1% (mimicking a modified-gravity change in the distance-redshift relation, with the same matter power spectrum) would test whether the claimed gravity probe is degenerate with the fiducial-distance assumption.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper's most concrete claim is the BH-NS galaxy-overdensity cross-correlation C^{κgwδ}_l reaching >3σ at z>0.35 within ~5 years of Advanced-LIGO data. The estimator (Eqs. 15-17) is unbiased only if ϵ_s — the fractional error in the 'true' luminosity distance computed from the EM redshift and best-fit ΛCDM parameters — is uncorrelated with the lensing and galaxy density fields, so that cross terms such as ⟨ϵ_s δ⟩ and ⟨ϵ_s κ_gw⟩ vanish after averaging. This is load-bearing because the lensing signal is at the ~10^{-3} per-source level, while σ_s from a 3% photometric redshift (C=0.03 in Sec. 5) is several percent per source and is suppressed only by 1/√N_gw. Three concrete failure modes are left unaddressed: (1) photometric-redshift errors correlate with environment through source-lens clustering and magnification bias, so ⟨ϵ_s δ⟩ need not vanish; (2) the 'true' distance is defined with a fiducial ΛCDM relation d_L(z), so any modified-gravity deviation in the distance-redshift relation — precisely the signal the paper claims to test — is absorbed into ϵ_s and biases the measured lensing cross-correlation, making the gravity test circular; (3) the noise model in Eq. (21) includes σ_b from 'uncertain values of cosmological parameters' but treats it as a variance term rather than a source of correlated bias. The paper treats the vanishing of these correlators by assertion in Sec. 4.2.1, not by derivation or simulation, and gives no bound on the size of residual ⟨ϵ_s δ⟩ relative to ⟨κ_gw δ⟩.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a new observational probe: measuring the weak lensing convergence of gravitational waves (GWs) by cross-correlating GW-derived luminosity distances with galaxy density and galaxy-lensing fields. The authors derive analytic estimators for the cross-spectra C^{κgwκg}_l and C^{κgwδ}_l, including separate treatments for GW sources with and without electromagnetic counterparts, and forecast signal-to-noise ratios for Advanced LIGO/Virgo, LISA, and various galaxy surveys. Their central forecast is that black-hole--neutron-star mergers with EM counterparts, cross-correlated with galaxy overdensity, can yield a >3σ detection at z>0.35 within about five years, and that the method tests general relativity and modified-gravity theories by probing GW propagation and the relation between metric perturbations and matter.","tokens_in":18333,"tokens_out":3490,"duration_ms":41339,"significance":"If the estimator is unbiased as claimed, this is a genuinely novel and timely multi-messenger cosmological probe. The analytic signal equations are standard and internally consistent, and the forecasts are concrete, falsifiable predictions rather than post-hoc fits; the BH-NS result is a specific, testable claim for the LIGO/Virgo era. The paper also usefully lays out the combined covariance of GW lensing, galaxy clustering, and galaxy lensing. However, the central gravity-testing framing is not backed by quantitative forecasts for any modified-gravity model, and the unbiasedness of the estimator rests on an unverified correlation assumption that is load-bearing because the signal is orders of magnitude smaller than per-source distance errors. The paper's value as a forecast is solid, but its stronger claim to 'probe the theory of gravity' is not yet established.","major_comments":[{"comment":"The unbiasedness of the EM-counterpart estimator is asserted rather than demonstrated. The estimator is unbiased only if the correlators ⟨ϵ_s κ_g⟩ and ⟨ϵ_s δ⟩ vanish, where ϵ_s is the fractional error in the 'true' luminosity distance computed from the EM redshift and best-fit cosmological parameters. This is load-bearing because the lensing signal is at the few×10^-3 level at z~0.5, while ϵ_s from a 3% photometric redshift is an order of magnitude larger per source and is suppressed only by 1/√N_gw. Realistic photo-z errors can correlate with environment through magnification bias or source-lens clustering, and modified-gravity deviations in the distance--redshift relation would appear in ϵ_s. The authors provide no derivation, simulation, or bound on residual ⟨ϵ_s δ⟩ and ⟨ϵ_s κ_g⟩; without such a bound, the claimed unbiasedness and the resulting SNR forecasts are not fully supported.","section":"Sec. 4.2.1, Eq. (17)"},{"comment":"The paper's central claim to 'probe the theory of gravity' is not operationalized. The 'true' luminosity distance in Eq. (16) is defined using the fiducial ΛCDM distance–redshift relation, so any modified-gravity effect on d_L(z) is absorbed by construction into ϵ_s and removed from the estimator. The cross-correlation signal is therefore sensitive only to modifications in the lensing kernels and the Poisson/GR relation, not to the distance-redshift modifications listed in Sec. 3. The forecasts in Sec. 5 are computed entirely from GR + ΛCDM inputs with no modified-gravity model, no predicted signal difference, and no distinguishability criterion. The title and abstract oversell what is currently demonstrated; the paper needs either a quantitative modified-gravity forecast or a careful statement of which gravity modifications the proposed estimator can actually constrain.","section":"Sec. 3 and Eq. (16)"},{"comment":"The noise model in Eq. (21) treats σ_b, the error from 'uncertain values of cosmological parameters,' as an additional variance term, but a common shift in cosmological parameters induces a correlated error in d_L^est across all sources. Such a correlated error enters ϵ_s and will not average down like independent noise; depending on how the source redshift distribution overlaps the galaxy survey, it can bias C^{κgwδ}_l coherently rather than merely inflating its variance. The same issue affects the no-EM-counterpart estimator in Eq. (18), where the multiplicative factor ⟨1/(1+ϵ_s)⟩ is assumed removable once the GW source redshift distribution N_gw(z) is known from clustering, but clustering redshifts do not remove a cosmology-dependent or environment-correlated ϵ_s. The paper should quantify this correlated-bias contribution to the SNR or explicitly justify why it is subdominant.","section":"Sec. 5, Eq. (21)"}],"minor_comments":[{"comment":"The sentence 'The cross-correlation ... probe theories of gravity' has a subject-verb agreement issue; also, the abstract says 'within 10 years' while the BH-NS forecast in Sec. 5 is quoted for five years, so the time-line wording should be made consistent.","section":"Abstract"},{"comment":"The notation ⟨·⟩ in Eq. (17) is used for an average over source–galaxy pairs, but the distinction between ensemble average and survey average is not defined; please specify how the estimator is computed in practice over a finite sky area and finite N_gw.","section":"Sec. 4.2.1, Eq. (17)"},{"comment":"The caption says the detection threshold for NS-NS at z>0.2 'is going to be less than 10−σ'; this phrase appears garbled and should be rewritten to state whether the events are below the detection threshold or the SNR is below 10.","section":"Fig. 4 caption"},{"comment":"The parameter θ_min is described as the sky localization area, but it appears in the exponential as an angle; please clarify whether θ_min is an angular radius or a solid angle and give its units consistently.","section":"Eq. (21)"},{"comment":"The text states C=0 for spectroscopic redshifts for some EM-counterpart cases but later uses photometric redshifts for LISA sources; please state explicitly which source classes are assigned C=0 and which use C=0.03, and discuss the implication for the claimed unbiasedness.","section":"Sec. 5"}],"recommendation":"major_revision","confidential_remarks":"The manuscript fits the journal's scope and the forecast direction is timely. In my view the main barrier is not novelty or internal inconsistency but the unverified uncorrelated-error assumption in Eq. (17), which is central to the claimed SNR and to the 'testing gravity' narrative. If the authors can provide a simulation or analytic bound showing that residual correlated ϵ_s is below the signal level, or alternatively reframe the claims to avoid the circularity with the ΛCDM distance relation, the paper would be suitable for publication. I did not find evidence of citation manipulation or scope problems."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear [Colleague],\n\nThe short version: this is a legitimate forecast paper for a genuinely new observable—lensing of GWs measured through luminosity-distance ratios and cross-correlated with galaxy surveys—but the \"testing gravity\" framing oversells what the calculation actually does. The paper predicts a GR signal and quotes SNRs; it does not forecast sensitivity to any specific modified-gravity parameter.\n\nWhat's new: the estimator in Eqs. 15–17, which turns per-event DL residuals into a κ_gw map and cross-correlates with δ_g and κ_g, is a real construction. Earlier work treated GW lensing as noise for standard sirens or considered auto-correlations. The signal kernels are standard, but applying them to BH-NS and LISA sources with this estimator is a reasonable step forward. The forecast that BH-NS cross-correlation with galaxy overdensity beats 3σ at z>0.35 within five years of Advanced-LIGO is internally consistent, given optimistic but not crazy event rates.\n\nThe soft spots are concentrated in one place, and it is load-bearing. Eq. 17 assumes ϵ_s—the fractional error in the \"true\" distance from EM redshift and fiducial cosmology—is uncorrelated with δ and κ_gw, so those cross terms vanish. That is asserted, not derived or simulated. The lensing signal is ~10^-3 per source while ϵ_s from a 3% photo-z is orders of magnitude larger per source; only the ensemble average saves you. If photo-z errors correlate with environment (magnification bias, source-lens clustering) or if modified gravity changes d_L(z) relative to ΛCDM, the measured cross-correlation is biased at the level of the signal. The second failure mode is acute for the paper's stated purpose: the \"true\" distance is defined with ΛCDM, so an MG deviation is absorbed into ϵ_s and the gravity test becomes circular. The authors should at least bound ⟨ϵ_s δ⟩, or run a simulation with realistic photo-z scatter, before claiming unbiasedness.\n\nMinor: the numerical forecasts are not reproducible from the text alone—event rates, sky localizations, and survey parameters are given but no code or data release. And the paper cites its own companion papers for clustering redshifts; that's fine, but it doesn't replace the missing validation.\n\nOverall: the paper deserves a serious referee. It's a plausible new probe, the math is standard, and the main flaw is addressable. The right outcome is likely major revision, not rejection. I'd want the uncorrelated-error assumption tested before I'd bet on the BH-NS detection.","headline":"A plausible forecast paper for a genuinely new observable—GW lensing cross-correlations—whose central unbiasedness assumption is asserted, not proven, and whose 'testing gravity' claim is overstated.","tokens_in":18809,"tokens_out":1851,"would_cite":false,"duration_ms":19670,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper shows that weak gravitational lensing of gravitational waves can be measured by cross-correlating LIGO/Virgo sources with galaxy surveys, forecasting detection at z<0.5 within a decade.","keywords":["gravitational lensing","gravitational waves","weak lensing","galaxy surveys","cross-correlation","luminosity distance","modified gravity","LISA"],"falsifier":"Run the proposed estimator on a forward simulation that assigns photometric redshifts to gravitational-wave sources with realistic environmental correlations and source--lens clustering: if $\\langle\\epsilon_s\\delta\\rangle$ or $\\langle\\epsilon_s\\kappa\\rangle$ does not vanish, the recovered $\\hat{C}^{\\kappa_{\\rm gw}\\delta}_\\ell$ will be biased at the level of the signal. On real data, split the galaxy catalog by redshift-error properties and check that the measured cross-correlation is stable between subsamples.","tokens_in":17692,"feed_emoji":"🌀","tokens_out":15216,"duration_ms":126640,"temperature":0.7,"pith_summary":"The paper proposes that weak gravitational lensing of gravitational waves can be measured statistically by cross-correlating gravitational-wave sources with galaxy catalogs. In general relativity, gravitational waves and light follow the same perturbed spacetime geodesics, so the strain of a gravitational-wave signal should be magnified or demagnified by foreground matter in the same way as electromagnetic radiation. Using an estimator that compares the luminosity distance inferred from the gravitational-wave signal with the distance expected from an electromagnetic redshift, the authors forecast that Advanced LIGO and Virgo combined with planned galaxy surveys should detect this lensing at redshifts below $z<0.5$ within about ten years. Their most specific prediction is that black-hole--neutron-star mergers with electromagnetic counterparts, cross-correlated with galaxy overdensity, will reach a greater-than-$3\\sigma$ detection at $z>0.35$ within roughly five years. A successful measurement would test how gravitational waves propagate through spacetime and probe the gravitational imprint of dark matter.","feed_headline":"Weak lensing of gravitational waves becomes detectable within a decade","feed_subtitle":"Cross-correlating LIGO-Virgo events with galaxy maps can reveal how matter bends gravitational waves at redshifts below 0.5.","key_machinery":"The load-bearing object is the estimator $\\hat{D}_L(\\hat{n})$, the fractional deviation of the gravitational-wave-inferred luminosity distance from the electromagnetic-based expected distance, along with its cross-correlation against galaxy overdensity $\\delta_g$ and galaxy lensing convergence $\\kappa_g$. The lensing kernel $W_{\\kappa_{\\rm gw}}(\\chi(z)) = (1/H(z))\\int_z^\\infty dz'\\, (dn_{\\rm gw}/dz')(\\chi(z')-\\chi(z))/\\chi(z')$ projects the matter power spectrum along the line of sight, so the theory spectra $C^{\\kappa_{\\rm gw}\\kappa_g}_\\ell$ and $C^{\\kappa_{\\rm gw}\\delta}_\\ell$ follow from the same Limber-approximation machinery used in galaxy weak lensing. The estimator converts a per-event multiplicative distortion $(1+\\kappa_{\\rm gw})$ into an additive field whose mean can be extracted: detector noise, redshift error, and sky-localization uncertainty enter as additive noise terms in the covariance, and averaging over many sources suppresses them. This same comparison of gravitational-wave lensing with light lensing is what tests gravity, because modified theories alter either the gravitational-wave propagation equation or the relation between metric potentials and matter density.","core_discovery":"The paper's central claim is that the weak-lensing convergence of gravitational waves, $\\kappa_{\\rm gw}$, is observable as a statistical signal rather than only as a source of noise. The observed strain is written as $\\tilde{h} = h(f_z)[1 + \\kappa_{\\rm gw}(\\hat{n})]$, which makes the apparent luminosity distance $D_L = d_L/(1 + \\kappa_{\\rm gw}) + \\epsilon_{\\rm gw}$. For sources with electromagnetic counterparts, the authors build $\\hat{D}_L \\equiv 1 - D_L/d_L^{\\rm es}$, with $d_L^{\\rm es}$ the luminosity distance computed from the electromagnetic redshift and best-fit cosmological parameters, and show that to first order $\\hat{D}_L \\approx \\kappa_{\\rm gw} - \\epsilon_{\\rm gw}/d_L + \\epsilon_s - \\epsilon_s \\kappa_{\\rm gw}$. Cross-correlating $\\hat{D}_L$ with the galaxy convergence $\\hat{C}^{\\kappa_{\\rm gw}\\kappa_g}_\\ell$ and with the galaxy overdensity $\\hat{C}^{\\kappa_{\\rm gw}\\delta}_\\ell$ recovers the lensing signal, provided the redshift-error term $\\epsilon_s$ averages to zero. The forecasts place the strongest LIGO-era signal in the black-hole--neutron-star channel: $\\hat{C}^{\\kappa_{\\rm gw}\\delta}_\\ell$ from these mergers exceeds $3\\sigma$ at $z>0.35$ within five years, making it the most promising avenue for Advanced LIGO, and the multi-detector network plus galaxy surveys detects gravitational-wave lensing at $z<0.5$ within a decade. For LISA, the same estimator yields high-signal-to-noise measurements for supermassive black hole binaries of $10^4$--$10^7\\,M_\\odot$, extending the probe to high redshift.","pith_inferences":["The paper does not quantify what happens if photometric redshift errors correlate with environment; a simulation with realistic magnification and source--lens clustering could test whether the claimed unbiased estimator survives, since the lensing signal is only about $10^{-3}$ while per-source redshift error is an order of magnitude larger.","If modified gravity changes the distance--redshift relation, the 'true' distance used in $\\hat{D}_L$ is itself theory-dependent; recovering $\\kappa_{\\rm gw}$ and testing gravity would then require fitting cosmological parameters jointly with the lensing signal rather than fixing them to the standard model.","The same cross-correlation approach can be pointed at CMB lensing or 21-cm intensity maps at $z>3$, where galaxy surveys lose sensitivity; the paper gestures toward CMB lensing, and this would be a direct extension of the forecast."],"forward_implications":["Within about ten years, Advanced LIGO, Virgo, and planned galaxy surveys should detect weak lensing of gravitational waves at $z<0.5$, giving a new multi-messenger view of cosmic structure.","Black-hole--neutron-star mergers with electromagnetic counterparts are the most promising Advanced-LIGO source: their cross-correlation with galaxy overdensity should exceed $3\\sigma$ at $z>0.35$ after roughly five years.","For LISA, the same cross-correlation estimator gives high signal-to-noise measurements across supermassive black hole masses $10^4$--$10^7\\,M_\\odot$, extending gravity tests to large redshift.","A measured $\\kappa_{\\rm gw}$ would probe the growth of the gravitational potential and the gravitational distribution of dark matter, and comparing the gravitational-wave and galaxy lensing signals tests whether gravitational waves propagate exactly as general relativity predicts."],"supporting_citations":[{"why":"Establishes the geometric-optics lensing magnification of gravitational-wave strain used to write the observed strain as h(1+kappa_gw).","marker":"Takahashi 2006"},{"why":"Shows how lensing perturbs luminosity-distance estimates from gravitational-wave events, the effect the estimator is designed to isolate.","marker":"Cutler & Holz 2009"},{"why":"Supplies the electromagnetic-counterpart route to gravitational-wave source redshifts used to compute the expected luminosity distance for EM-bright events.","marker":"Nissanke et al. 2013"},{"why":"Shows that clustering of gravitational-wave sources with galaxies can recover their redshift distribution, used for events without electromagnetic counterparts.","marker":"Oguri 2016"},{"why":"Develops the clustering-redshift technique for gravitational-wave source catalogs, used to assign redshifts for black-hole--black-hole events.","marker":"Mukherjee & Wandelt 2018"},{"why":"Provides the fitting form for lensing-induced noise in luminosity distance that enters the variance of the estimator.","marker":"Hirata et al. 2010"},{"why":"Derives the chirp-based relation that lets the gravitational-wave luminosity distance be measured independently of the source chirp mass.","marker":"Schutz 1986"},{"why":"Supplies the nonlinear matter power spectrum used to compute the predicted cross-correlation spectra.","marker":"Blas et al. 2011"}],"fun_headline_variants":["GW lensing detectable within a decade via galaxy cross-correlation","Lensing of gravitational waves emerges as real signal, not noise","Galaxy maps turn gravitational wave lensing into a gravity test","BH-NS mergers lead path to first gravitational wave lensing detection","Gravitational wave lensing seen in LIGO-Virgo data within 10 years"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire measurement rests on one assumption: the fractional error in the luminosity distance calculated from an electromagnetic redshift and the assumed cosmology is uncorrelated with the galaxy density or lensing field, so that it averages to zero. If redshift errors are tied to environment, the measured cross-correlation is biased at the level of the lensing signal itself.","fun_headline_variants_meta":{"raw":{"variants":["GW lensing detectable within a decade via galaxy cross-correlation","Lensing of gravitational waves emerges as real signal, not noise","Galaxy maps turn gravitational wave lensing into a gravity test","BH-NS mergers lead path to first gravitational wave lensing detection","Gravitational wave lensing seen in LIGO-Virgo data within 10 years"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000274,"raw_usage":{"total_tokens":1696,"prompt_tokens":1056,"completion_tokens":640,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":672,"completion_tokens_details":{"reasoning_tokens":546}},"tokens_in":672,"tokens_out":640,"duration_ms":6933,"temperature":1.0,"reasoning_tokens":546,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:26:44.841274+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the proposed estimator on a forward simulation that assigns photometric redshifts to gravitational-wave sources with realistic environmental correlations and source--lens clustering: if $\\langle\\epsilon_s\\delta\\rangle$ or $\\langle\\epsilon_s\\kappa\\rangle$ does not vanish, the recovered $\\hat{C}^{\\kappa_{\\rm gw}\\delta}_\\ell$ will be biased at the level of the signal. On real data, split the galaxy catalog by redshift-error properties and check that the measured cross-correlation is stable between subsamples.","supporting_citations":[],"review_version":1}