{"id":"a30bf457-a3a2-4a99-b1e9-c76b61e055f2","arxiv_id":"2607.27149","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":6,"one_line_summary":"Shear–kSZ correlates kSZ, tomographic line-of-sight velocity, and lensing convergence to measure P_me(k) and thereby the baryonic matter-power suppression S(k) at high forecast significance.","lead":"A new three-field estimator (kSZ × reconstructed velocity × weak-lensing shear) measures how ionized gas traces all matter, not just galaxies. It targets the baryonic suppression that limits Stage-IV cosmic shear, with a forecast ~16σ detection using SO, DESI, and LSST-like data.","discovery_kind":"new_method","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"The ≲5% validation and SNR≈16 rest on an in-realization calibration that cannot be performed on real data: A_ij is measured against the same realization's true halo velocities, which both absorbs sample variance and sidesteps the actual reconstruction response.","rationale":"The reader's weakest_assumption identified the same locus — the smoothed-true-velocity proxy and the absorption of reconstruction effects into a simulation-measured A_ij. My pass sharpens it: the concern is not merely that the proxy differs from reconstruction, but that the specific calibration used for validation (A_ij from the same realization) provides a sample-variance cancellation that is structurally unavailable in a real analysis, so the ≲5% agreement overstates the demonstrated end-to-end accuracy even though the estimator algebra itself is correct. This does not overturn the paper's central methodological claim — the factorization, parity arguments, and the P_me→S(k) identity (validated on MillenniumTNG in App. C) are solid — so it supports keeping CONDITIONAL rather than moving to REJECT, and is not grounds for ACCEPT because the fix (end-to-end reconstruction on the light cone with external calibration) is exactly the kind of addressable item the conditional verdict already contemplates. The proposed test is directly executable with public tools and the same simulation products the paper already uses.","tokens_in":41594,"tokens_out":1582,"duration_ms":74924,"concrete_test":"Run standard continuity-equation reconstruction (e.g., pyrecon at R_s=10–15 h⁻¹ Mpc) on the AbacusSummit LRG-like halo light-cone catalog used in the paper; rebuild the templates ṽ_i from the reconstructed velocities; calibrate A_ij from an *independent* set of mocks (different phases) or the linear-theory Eq. (17) rather than the same realization; then recompute Ĉ^{TV}_ℓ and the Fisher SNR. If the meas/model ratio stays within ~10% for ℓ≳500 and the SNR stays within ~30% of 16, the forecast stands; a larger drift quantifies exactly how much of the headline number was borrowed from the in-realization calibration.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The master formula (Eq. 19) is validated with the band matrix A_ij ≡ ⟨ṽ_i v̄_j⟩ measured by cross-correlating the smoothed true-halo-velocity template with the deposited true velocities of the same AbacusSummit realization (Sec. III E, II D). The paper itself notes this calibration \"captures the sample variance of the long-wavelength radial velocity modes\" and that the coherent 10–20% per-bin amplitude modulation \"cancels against the calibrated A_ij within a single realization.\" That cancellation is doing real work in the quoted agreement: the ≲5% residuals in Fig. 3/Fig. 9 are residuals after a calibration that has access to information unavailable on the sky. In a data analysis, the true velocity field is unknown; A_ij must come from mocks or linear theory, which (a) do not share the data realization's large-scale modes, so the sample-variance cancellation disappears and the modulation shows up as scatter/bias in the recovered P_me, and (b) must model the full response of continuity-equation reconstruction — smoothing, bias, shot noise, redshift-space distortions, fiber incompleteness — rather than a Gaussian filter on true velocities. The paper asserts the smoothed template performs \"very similar\" to reconstruction (r≈0.6–0.7 vs ⟨ṽv⟩/σ²_v≈0.5) but never runs the reconstruction pipeline, so this comparison is plausibility, not evidence. The estimator's algebra is sound and the velocity-parity foreground argument is clean; the soft spot is entirely that the end-to-end chain from reconstructed velocities to calibrated kernel to P_me is validated only in a configuration where the calibration is given the answer. The 16σ and \"1% on S(k)\" numbers inherit this untested transfer.","agreement_with_reader":"agree"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The manuscript proposes \"shear-kSZ,\" a three-field cross-correlation between the CMB temperature (kSZ) map, a tomographically binned line-of-sight velocity template from spectroscopic tracers, and the weak-lensing convergence. The expectation value factorizes (Eqs. 12–19) into a per-bin velocity kernel — encoded in a band matrix A_ij = ⟨ṽ_i v̄_j⟩ — multiplying the projected matter–electron cross-power P_me(k). Because the density leg is κ rather than a galaxy overdensity, the estimator accesses P_me for the full matter field, and an exact species decomposition (Eq. 22) converts P_me into the baryonic suppression S(k) up to a bounded electron auto-spectrum term and sub-percent dark-matter backreaction. The analytic model is validated against AbacusSummit at ≲5% for ℓ≳500, unbiasedness is tested with independent noise seeds, the P_me→S(k) closure is verified on MillenniumTNG to 0.2–0.5%, and a Fisher forecast gives SNR≈16 for SO-like CMB × DESI-like LRGs × LSST-like high-z sources over f_sky=0.2.","tokens_in":42008,"tokens_out":7725,"duration_ms":246677,"significance":"If the forecast holds, this is a valuable addition to the kSZ toolkit: a direct, tracer-independent probe of P_me(k) with a concrete, near-term path to constraining the baryonic suppression that dominates Stage-IV shear systematics. The paper ships several genuine strengths: a first-principles derivation including an exact line-of-sight velocity sum rule (Eq. 18) that dictates the tomographic design; a parameter-free end-to-end comparison to the simulation (Fig. 3); an unbiasedness test with independent noise seeds plus a useful pipeline caution (App. B1); a velocity-parity argument for foreground cancellation; and a hydro-simulation verification of the P_me→S(k) closure (App. C). The forecast is for a concrete, falsifiable measurement on current/near-term data. The soft spot is that both the quoted validation precision and the SNR rest on an in-realization calibration and a smoothed-true-velocity stand-in for reconstruction, neither of which is available on real data.","major_comments":[{"comment":"The ≲5% validation (Figs. 3, 9) and the App. D precision claim (σ_x/x = 1/16 → ~1% on S) use A_ij measured against the same realization's velocities, which absorbs the coherent 10–20% long-mode modulation (App. B3). On real data A_ij must come from mocks or linear theory, so this modulation reappears as multiplicative scatter on the recovered P_me, coherent over ~3–4 bins and all ℓ. A back-of-envelope estimate (15% per coherence patch, ~7 independent patches across 24 bins) gives ~6% scatter on the overall amplitude — comparable to the 6.25% statistical error at SNR=16 — potentially doubling the S(k) error budget. The test needed is feasible in-scope: redo Figs. 3/9 with A_ij from linear theory (Fig. 2 shows the normalized shape matches) or from a different AbacusSummit phase, and propagate the result into App. D.","section":"§II D, §V A / Fig. 3, Fig. 9"},{"comment":"The forecast SNR=16 (Table I) uses a Gaussian-smoothed true-halo-velocity template (R_s=15 h⁻¹Mpc) as a stand-in for continuity-equation reconstruction. Since A_ij absorbs only the amplitude, the SNR depends on the template's noise through C^{ViVj}_ℓ in Eq. (29), and the stand-in contains no reconstruction noise, RSD, bias, or fiber incompleteness. The fidelity comparison offered (reconstruction r≈0.6–0.7 vs ⟨ṽv⟩/σ²_v≈0.5) contrasts a correlation coefficient with a regression slope — these are not the same statistic and do not establish equivalent noise properties. Given that reconstruction on AbacusSummit LRG mocks is standard (and the smoothing suppression 0.51 vs 0.70 transverse-linear prediction already shows the proxy is imperfectly understood), the headline '16σ' and '~10σ at LSST Y1' claims need either an actual reconstruction run or an explicit, quantified sensitivity statement.","section":"§III E, §V E / Table I"},{"comment":"The P_me→S(k) closure (Eq. 22) uses the Cauchy–Schwarz bound with r_me≈1 and a small dark-matter backreaction, verified only on MillenniumTNG at z=0.5 (App. C, Figs. 10–11). The paper itself cites observational evidence for feedback stronger than MTNG, where r_me and the backreaction could deviate more. Since the sub-percent S(k) accuracy is a headline claim, the closure test should be repeated on at least one strong-feedback hydrodynamical model (e.g., a CAMELS/FLAMINGO variant) at the LRG redshift range, or the claim should be explicitly scoped to MTNG-like physics.","section":"§II F / App. C"}],"minor_comments":[{"comment":"Caption gives σ_z=0.25, n_eff=26 arcmin⁻², and 'top/bottom rows', whereas §III C 2 uses σ_z=0.05(1+z), n_eff≈4.6 arcmin⁻², and the figure shows two panels arranged left/right. Please reconcile — this looks like a leftover from an earlier draft.","section":"Fig. 5 caption vs §III C 2"},{"comment":"Two consecutive paragraphs state the SNR impact of the diagonal approximation as '<1%' and '15–20% at low ℓ', with the inversion sentence repeated nearly verbatim. Presumably total vs low-ℓ SNR, but as written it reads as contradictory.","section":"§V D"},{"comment":"The Gaussian covariance of Eq. (29) omits connected four-point terms in the V_i–V_j covariance; κ is substantially non-Gaussian at ℓ∼2000–4000 where the SNR peaks. A brief quantification (or a statement that C^{ViVj} measured from the realization partially captures this) would strengthen the forecast.","section":"§IV A / Eq. (29)"},{"comment":"Please clarify whether the adopted SO ILC Deproj-0 noise curve includes astrophysical foreground residuals (tSZ, CIB). Velocity parity removes their bias but not their variance, and C^{TT}_ℓ dominates the noise budget.","section":"§III D"},{"comment":"The edge-bin correction of Eq. (A1) relies on linear-theory Ψ∥ and an assumed τ̄′ beyond the tracer volume; please state the expected accuracy of this correction versus the ~8% cost of excising edge bins.","section":"App. A"},{"comment":"Typos: 'to mimic the effective of smoothing' (§III E); 'in very agreement' (§V B); 'with gaining very little independent information' (§II C). Notation: Eq. (2) writes P_me where Eq. (14) defines C^{me}_ℓ; please harmonize.","section":"various"},{"comment":"The Fisher sum starts at ℓ_min=2 although the model is validated only for ℓ≳500 (§II E). For the observed cases the low-ℓ contribution is negligible, but for the ideal cases it inflates the SNR; a one-line statement of the ℓ<500 contribution would help.","section":"§IV B"}],"recommendation":"major_revision","confidential_remarks":"Single-author methods paper, technically careful and unusually transparent about its own approximations. The citation pattern leans heavily on the author's prior stacked-kSZ work, but that reflects the field's authorship rather than padding. The core estimator algebra and simulation validation are sound; the requested revisions are quantitative tightening of the calibration and forecast realism, all feasible within the manuscript's existing simulation infrastructure. Well suited to the journal's scope."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"Punchline: this is a real methods step past stacked kSZ. Swap δ_g for κ, keep the velocity weighting, and the density leg is total matter—so you get P_me and a direct route to baryonic S(k) without HOD mass extrapolation. That is the point of the paper, and it lands.\n\nWhat is new is the three-field factorization (Eq. 19), the LOS velocity sum rule that forces tomography, the parity argument that kills velocity-independent foregrounds, and the high-z source cut that removes IA and boost by construction. The Wick/Limber derivation is standard in spirit but carefully written. End-to-end Abacus checks at ≲5% on signal multipoles, the unbiasedness test with independent noise seeds, the τ–matter consistency check, and the MillenniumTNG S(k) closure (sub-percent) are all honest work. Citation pattern is appropriate; she knows the stacked-kSZ and feedback literature.\n\nSoft spot, in proportion: the validation uses A_ij measured against the same realization’s true (smoothed) halo velocities. That absorbs long-mode sample variance and sidesteps a full continuity-equation reconstruction on the light cone. On data, A_ij comes from mocks or linear theory, so the cancellation is weaker and the transfer function must carry real reconstruction response. The paper says the smoothed proxy is “very similar” (r ~ 0.5–0.7) and treats A_ij like the usual stacked-kSZ velocity transfer—fair analogy, not a free lunch. Edge-bin leakage and gas painting from MTNG are smaller, flagged issues. None of this breaks the estimator algebra or the foreground argument; it means the 16σ / “1% on S(k)” numbers are survey-shaped forecasts, not yet pipeline-ready.\n\nWho it is for: people doing kSZ tomography, Stage-IV shear systematics, and DESI×SO×LSST cross-correlations. Worth a serious referee. I would engage, cite the estimator definition and the S(k) map, and push for an end-to-end reconstruction mock before treating the SNR as locked.\n\nSend it to peer review.","headline":"Clean three-field estimator that actually targets P_me (and thus S(k)); the algebra and sim checks hold, the 16σ number is a forecast with a standard reconstruction-proxy caveat.","tokens_in":42225,"tokens_out":564,"would_cite":true,"duration_ms":19250,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"A new shear–kSZ estimator turns kSZ, reconstructed velocities, and weak lensing into a direct measurement of the matter–electron power spectrum and the baryonic suppression of cosmic structure.","keywords":["kinematic Sunyaev-Zel'dovich","weak lensing","matter-electron power spectrum","baryonic suppression","kSZ tomography","velocity reconstruction","cosmic shear systematics","Stage-IV surveys"],"falsifier":"Measure the three-field cross-spectrum on real SO-like temperature maps, DESI LRG velocity reconstructions, and LSST (or Euclid) high-z shear; if the multipole shape and multi-bin amplitude match the beam-convolved model with an A_ij kernel calibrated on the same reconstruction mocks, and if a direct τ–matter cross-check in hydro simulations still recovers S(k) to ≲0.5%, the central claim holds.","tokens_in":41904,"feed_emoji":"🌌","tokens_out":1187,"duration_ms":24847,"temperature":0.7,"pith_summary":"Stage-IV weak-lensing surveys will be limited by how baryons rearrange gas and suppress the matter power spectrum on small scales. This paper proposes shear–kSZ: cross-correlate the kinematic Sunyaev–Zel’dovich temperature map with a tomographic line-of-sight velocity template and the weak-lensing convergence. The expectation value factorizes into a calibratable velocity kernel times the projected matter–electron cross-power P_me(k), so the density leg is the full matter field rather than a biased galaxy tracer. From P_me one recovers the baryonic suppression S(k) up to sub-percent corrections, without halo-mass extrapolations. Validated on simulations at the few-percent level and forecast at about 16σ with Simons Observatory–like CMB, DESI-like LRGs, and high-redshift LSST sources over 20% of the sky, the estimator is presented as a near-term observational target that can calibrate the leading astrophysical systematic in cosmic shear.","feed_headline":"New shear–kSZ probe pins baryon drag on structure to ~1%","feed_subtitle":"Cross-correlating kSZ, galaxy velocities, and lensing yields the matter–electron spectrum Stage-IV shear needs.","key_machinery":"The shear–kSZ estimator and its factorization: after pairing the coherent large-scale velocities, ⟨Ĉ^{T V_i}_ℓ⟩ collapses to a single velocity kernel V_i (or band matrix A_ij) multiplying W_κ P_me/χ². All reconstruction fidelity, smoothing, and sampling live in the calibratable kernel; scale dependence is carried only by P_me.","core_discovery":"The shear–kSZ estimator Ĉ^{T V_i}_ℓ ≡ ⟨T, (ṽ_i/c) κ⟩ has expectation value equal to a calibratable velocity kernel V_i times the lensing kernel and the projected matter–electron spectrum P_me(k_ℓ, z_i)/χ_i². Because κ traces total matter, P_me maps to the baryonic suppression S(k)=P_mm/P_DMO_mm via an exact cold-dark-matter plus baryon decomposition, up to a sub-percent dark-matter backreaction and a small bounded electron auto-spectrum term. The analytic model matches AbacusSummit measurements at ≲5% for ℓ≳500; CMB foregrounds cancel by velocity parity; and a realistic SO+DESI+LSST forecast yields SNR≈16 (f_sky=0.2), corresponding to roughly 1% on S(k).","pith_inferences":["If the method works on data, stacked kSZ (P_ge) and shear–kSZ (P_me) together become a differential test of how electrons occupy biased tracers versus the full matter field.","The same velocity-parity cancellation that kills tSZ/CIB bias could be reused for other odd-parity momentum–density–density estimators beyond kSZ.","Edge-bin leakage from gas outside the reconstruction volume is a generic survey-boundary systematic that any tomographic kSZ analysis will need to model or cut.","A polarization-only CMB-lensing κ channel would push the usable multipole range deeper into the beam-limited regime where much of the SNR lives."],"forward_implications":["P_me from shear–kSZ can be turned into a sub-percent constraint on the baryonic suppression S(k) used in Stage-IV cosmic shear.","Joint shear–kSZ plus cosmic shear can constrain baryonic feedback internally instead of only marginalizing flexible correction models.","A ~10σ detection is already expected with early LSST releases because CMB noise, not lensing depth, dominates.","Advanced Simons Observatory is projected to roughly double the signal-to-noise; lower-redshift tracers (e.g. BGS) open more background source samples.","CMB-lensing κ can replace galaxy shear for higher-redshift foregrounds and maximal overlap with the temperature map."],"fun_headline_variants":["Shear-kSZ estimator maps matter-electron spectrum for Stage-IV shear","kSZ plus lensing pins baryonic suppression S(k) to ~1%","New kSZ-velocity-lensing probe yields full P_me(k)","Shear-kSZ forecasts 16σ on ionized gas-matter power","Cross-correlating kSZ, velocities, lensing constrains baryon drag"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The forecast treats a Gaussian-smoothed true halo velocity field as a faithful stand-in for real continuity-equation velocity reconstruction on spectroscopic galaxies, with all losses absorbed into a simulation-measured calibration matrix.","fun_headline_variants_meta":{"raw":{"variants":["Shear-kSZ estimator maps matter-electron spectrum for Stage-IV shear","kSZ plus lensing pins baryonic suppression S(k) to ~1%","New kSZ-velocity-lensing probe yields full P_me(k)","Shear-kSZ forecasts 16σ on ionized gas-matter power","Cross-correlating kSZ, velocities, lensing constrains baryon drag"]},"model":"grok-4.5","effort":"low","cost_usd":0.004928,"raw_usage":{"total_tokens":1569,"prompt_tokens":1067,"num_sources_used":0,"completion_tokens":108,"cost_in_usd_ticks":49284000,"prompt_tokens_details":{"text_tokens":1067,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":394,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":1067,"tokens_out":108,"duration_ms":8580,"temperature":1.0,"reasoning_tokens":394,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-30T11:12:11.067193+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Measure the three-field cross-spectrum on real SO-like temperature maps, DESI LRG velocity reconstructions, and LSST (or Euclid) high-z shear; if the multipole shape and multi-bin amplitude match the beam-convolved model with an A_ij kernel calibrated on the same reconstruction mocks, and if a direct τ–matter cross-check in hydro simulations still recovers S(k) to ≲0.5%, the central claim holds.","supporting_citations":[],"review_version":1}