{"id":"ac187268-6569-472b-b6cd-cfccd64cc69c","arxiv_id":"1908.08953","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"kSZ tomography with future CMB and galaxy surveys could constrain correlated compensated isocurvature perturbations to sigma_A = 0.25, more than an order of magnitude better than galaxy surveys alone.","lead":"Combining future galaxy surveys with a reconstructed cosmic microwave background velocity field (kSZ tomography) could detect a class of early-universe density fluctuations, called compensated isocurvature perturbations, at amplitudes comparable to ordinary matter fluctuations. If realized, this would close a roughly two-order-of-magnitude gap in our knowledge of the early universe and test models of the universe's first moments.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The CIP signal itself contaminates the kSZ remote dipole field via baryon-density modulation, so the assumed unbiased tracer and constant optical-depth bias b_v may not absorb the induced scale-dependent bias; the forecast omits this A-dependence.","rationale":"The paper is a careful Fisher forecast with a clear signal model and a thorough treatment of relativistic contributions and degeneracies. Its central claim, σ_A ≈ 0.25, rests on the remote dipole field being a cosmic-variance-free tracer of the total matter density. The weakest link is not the noise level or foregrounds (which the paper partially marginalizes), but the self-consistency of the tracer: correlated CIPs change the baryon/electron density and therefore the optical-depth weighting of the kSZ signal. The quadratic estimator's response (Eq. 16) uses C^{τg}; if this is computed at A=0, the reconstructed dipole field is biased by a scale-dependent factor ∝ A/k² when A≠0. A constant b_v(z) does not absorb this, and the Fisher matrix in Sec. IV does not include the A-dependence of the reconstruction noise or the kSZ-galaxy cross-spectrum. At the low ℓ that dominate the signal, Δ/δ_m is order unity for A~1, so the effect is not small. This is a concrete, checkable gap rather than a fatal flaw; a re-derived forecast that includes the CIP optical-depth modulation could confirm or weaken the headline constraint. Because the paper is a forecast and the issue is a missing term in the analysis, the appropriate disposition remains CONDITIONAL pending this self-consistency check.","tokens_in":12351,"tokens_out":13499,"duration_ms":149165,"concrete_test":"Recompute the Fisher forecast including the CIP dependence of the kSZ reconstruction: set the electron-density fluctuation to δ_e = δ_m + Δ(k) with Δ given by Eq. (5), propagate this through the quadratic estimator (Eqs. 15-17) to obtain the A-dependent reconstruction noise N_ℓ and the A-dependent kSZ-galaxy cross-spectrum, and then re-evaluate σ_A in Table II with both a constant b_v and an ℓ-dependent b_v(ℓ) marginalized. If σ_A degrades by ≳30% or the residual of the sample-variance cancellation exceeds the fiducial CIP signal at ℓ≲30, the central improvement claim is not robust to this self-consistency issue.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section III's central assumption is that the reconstructed remote dipole field is an unbiased tracer of the total matter density, with only a multiplicative 'optical depth bias' b_v(z) to marginalize. But the kSZ signal (Eq. 14) is proportional to the electron density n_e, and correlated CIPs alter the baryon density by δ_b ⊃ Δ(k), with Δ(k) = (5H²Ω_m)/(2ak²) A δ_m (Eqs. 2-5). Thus the optical-depth–galaxy response C^{τg} used in the quadratic estimator (Eqs. 15-17) contains a term ∝ A δ_m/k² when A≠0. An ℓ-independent b_v per redshift bin cannot absorb this scale-dependent contamination; the reconstructed dipole field acquires a spurious component ∝ A/k² correlated with the galaxy density. The Fisher forecast (Sec. IV) treats C^{kSZ,N} and the reconstruction noise N_ℓ as A-independent, so the derivative of the covariance with respect to A omits this back-reaction. Near the fiducial A=0 the effect is a missing O(A) term in the Fisher matrix; for the proposed A~1 detection it is not a small correction at ℓ≲30 where k~ℓ/χ is small and Δ/δ_m ~ 1. If this contamination is included, the sample-variance cancellation that yields σ_A=0.25 (Table II) is weakened because both the galaxy field and the tracer respond to CIPs with different coefficients.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a Fisher-matrix forecast for the amplitude A of correlated compensated isocurvature perturbations (CIPs), using the scale-dependent galaxy bias they induce in combination with kSZ-tomography reconstructions of the remote dipole field. The model includes all linear relativistic contributions to galaxy number counts, treats primordial non-Gaussianity as a competing source of scale-dependent bias, and marginalizes over cosmological parameters and bias functions. For an LSST-like gold sample and a CMB-S4-like experiment, the headline result is sigma_A=0.25 in the fiducial model, degrading to sigma_A=0.49 when f_NL is also marginalized; the authors find an order-of-magnitude improvement over galaxy counts alone and roughly two orders of magnitude over current CMB constraints. The paper also examines a curvaton-motivated model with A=16 and f_NL=6, explicitly showing the sensitivity of that forecast to a prior on the CIP bias b_bc(z).","tokens_in":12630,"tokens_out":21746,"duration_ms":234792,"significance":"If the forecast is robust, the proposed technique would be a genuinely new probe of a poorly constrained early-Universe mode, improving on Planck by two orders of magnitude and offering an independent route to disentangle CIPs from local-type primordial non-Gaussianity. The paper is a clean forward-model forecast rather than a data analysis, and it is careful to include relativistic projection effects, to marginalize over the relevant nuisance parameters, and to test robustness to several experimental choices. The main strengths are the completeness of the galaxy-count model and the explicit treatment of the f_NL degeneracy. The central caveats are that the headline significance is conditional on a halo-model prediction for b_bc(z) and on an idealized kSZ reconstruction with no foreground model; both of these need to be addressed before the headline claim can be taken at face value.","major_comments":[{"comment":"The fiducial result sigma_A=0.25 is computed with b_bc(z) fixed to the halo-model/HOD fit of Eq. (10). The CIP contribution to delta_g in Eq. (4) is proportional to b_bc, so the Fisher sensitivity to A is conditional on this assumed function. The authors' own analysis of the A=16 model shows that a 100% prior on b_bc degrades sigma_A to 5.8, demonstrating a strong degeneracy; the analogous degradation for the fiducial A=0 forecast is not reported. Please add a forecast with a physically motivated prior on b_bc (including the halo-model calibration uncertainty) and show how sigma_A changes, so that the headline improvement is not conditional on an unstated assumption.","section":"Sec. IV, Table II; Sec. II, Eq. (10)"},{"comment":"The reconstruction noise N_l in Eq. (17) is evaluated using a CMB temperature power spectrum C_TT that does not include foreground contamination. With the forecast summing reconstruction modes to l=9000 at 1 microK-arcmin, the small-scale CMB is dominated by tSZ, dust, and radio sources; these increase the effective C_TT and therefore N_l, weakening the sample-variance cancellation that produces sigma_A=0.25. The robustness checks in Fig. 3 vary instrument noise and l_min but not foregrounds. Please add a foreground model to C_TT or demonstrate that foreground cleaning leaves N_l, and hence the forecast, unchanged.","section":"Sec. III, Eqs. (15)-(17); Sec. IV"}],"minor_comments":[{"comment":"The term '5f/3 A b_c b(z) S_psi' appears to have a typo: the CIP contribution should be proportional to b_bc(z), not to a product b_c b(z). Please check the notation.","section":"Sec. II, Eq. (13)"},{"comment":"The paper states that the remote dipole field is an unbiased tracer of the total density, but the estimator's response depends on the optical-depth-galaxy cross-power C^{tau g} in Eq. (16). Correlated CIPs modulate the electron density by f Delta(k); please add a short estimate showing that the resulting A-dependence of C^{tau g} is negligible at the small scales that dominate the reconstruction.","section":"Sec. III, Eq. (16)"},{"comment":"The abstract quotes the two-orders-of-magnitude improvement using the fiducial sigma_A=0.25; because marginalizing over f_NL gives sigma_A=0.49, please state this caveat explicitly in the abstract or at least in the conclusions.","section":"Abstract and Sec. IV, Table II"},{"comment":"The claim that 'a definitive detection of this scenario can be made with future datasets' even with the weakest prior is not supported by the quoted sigma_A=5.8 for A=16, which is only a roughly 2.8-sigma measurement; please either quantify what 'definitive' means or soften the claim.","section":"Sec. IV, A=16 discussion"},{"comment":"The panels in Fig. 3 would be easier to interpret with reference lines at the fiducial parameter values, since the current plots show the dependence without a clear marker for the adopted baseline.","section":"Fig. 3"},{"comment":"The statement that photometric quasar surveys can already improve on current CMB constraints relies on a translation from f_NL to A; please show this mapping explicitly so that the reader can assess the magnitude of the claimed improvement.","section":"Sec. V (Discussion)"}],"recommendation":"major_revision","confidential_remarks":"This is a competent and clearly written forecast. The two main issues are the unquantified dependence on the halo-model b_bc(z) and the absence of foregrounds in the kSZ reconstruction noise; both are addressable with additional plots and discussion and do not appear to require a fundamentally different approach. I would be satisfied with a revised version that adds a prior-marginalized b_bc forecast and a foreground model or a well-justified foreground-cleaning argument."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper is a well-executed forecast, and the new piece is real: combining kSZ-tomography sample-variance cancellation with the scale-dependent galaxy bias from correlated CIPs, and showing that the combination could beat current CMB limits by a large factor. The relativistic treatment of the number counts is thorough, the fNL degeneracy is handled explicitly, and the robustness checks to instrument noise, redshift bins, and magnitude limits are useful. The paper is also honest about the fact that existing quasar data may already improve on CMB constraints, which tempers the headline without undermining it.\n\nThe main soft spot is a genuine back-reaction issue that the forecast does not address. The kSZ signal is proportional to the electron density n_e, and CIPs modulate the baryon density by an amount Δ ∝ A δ_m/k^2. This introduces a scale-dependent term in C^{τg}, the optical-depth–galaxy cross-spectrum that enters the quadratic estimator. The paper marginalizes over a constant optical-depth bias b_v per redshift bin, but a 1/k^2 term cannot be absorbed that way. If the CIP-induced modulation is present, the reconstructed remote dipole field is not exactly an unbiased tracer of the total density, and the sample-variance cancellation that produces σ_A=0.25 is degraded. The Fisher forecast treats the reconstruction noise N_ℓ as A-independent, so the derivative of the covariance with respect to A omits this effect. The missing term is small at the largest scales for A~1, but it is not negligible for the A=16 curvaton scenario, and it could appreciably change the projected constraint even near fiducial A=0. This does not invalidate the concept, but it likely makes the headline number optimistic.\n\nThe A=16 forecast already shows a different, more model-dependent weakness: without a prior on bbc, the degeneracy between A and the bias function is strong, and even a 100% prior yields σ_A=5.8. So the two-orders-of-magnitude claim really comes from the zero-amplitude idealized case.\n\nWho is this for? Anyone thinking about CIP constraints from upcoming surveys, and anyone using kSZ tomography for large-scale structure. The paper is suitable for a serious referee; the reconstruction back-reaction should be addressed in revision, and the forecast would be stronger if the authors either compute N_ℓ consistently at non-zero A or explicitly argue why the effect is small.","headline":"A careful and honest forecast, but the headline sigma_A=0.25 likely misses a back-reaction of the CIPs on the kSZ reconstruction itself, so treat the central number as optimistic until that is modeled.","tokens_in":13187,"tokens_out":6921,"would_cite":true,"duration_ms":73257,"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":"Using galaxy counts with kSZ tomography can expose hidden early-universe perturbations","keywords":["compensated isocurvature perturbations","kinetic Sunyaev-Zeldovich tomography","scale-dependent galaxy bias","sample variance cancellation","remote dipole field","primordial non-Gaussianity","curvaton","cosmic microwave background"],"falsifier":"Measure the kSZ remote-dipole reconstruction noise on a real next-generation CMB data set, treating thermal Sunyaev-Zeldovich emission, dust, and radio sources as they appear in the maps; if the effective noise at multipoles up to 9000 exceeds Eq. (17), the projected CIP sensitivity degrades roughly in proportion, and the claimed improvement over survey-only constraints shrinks.","tokens_in":12137,"feed_emoji":"🔭","tokens_out":8521,"duration_ms":80101,"temperature":0.7,"pith_summary":"Compensated isocurvature perturbations (CIPs) are early-universe fluctuations that shift matter between baryons and dark matter while leaving the total density unchanged, so standard cosmological probes barely see them. Current CMB measurements still allow CIP amplitudes a few hundred times larger than the adiabatic fluctuations. This paper shows that kinetic Sunyaev-Zeldovich (kSZ) tomography can close that gap: reconstructing the remote dipole field and cross-correlating it with galaxy counts cancels cosmic variance, isolating the scale-dependent galaxy bias that correlated CIPs produce. In a forecast with next-generation CMB and galaxy surveys, the correlated CIP amplitude $A$ is constrained to $\\sigma(A)=0.25$, more than an order of magnitude better than using the galaxy survey alone. That sensitivity reaches the amplitudes predicted by curvaton models, making a concrete class of multi-field inflation models testable.","feed_headline":"kSZ tomography can expose hidden early-universe perturbations","feed_subtitle":"Cross-correlating galaxy counts with the remote dipole field cancels cosmic variance, reaching sigma(A)=0.25.","key_machinery":"The load-bearing mechanism is sample-variance cancellation between galaxy number counts and the kSZ-reconstructed remote dipole field, the CMB dipole as seen from each location in the Universe. The reconstruction uses the quadratic estimator of Eq. (15), with noise set by Eq. (17); because the remote dipole field is an unbiased tracer of the total matter density, its cross-correlation with the biased galaxy field isolates the scale-dependent bias generated by correlated CIPs without cosmic variance. The characteristic scale dependence comes from $\\Delta(\\mathbf{k})\\propto A/k^2$, and the CIP bias function $b_{bc}(z)$ is computed from a mass function and halo occupation model. All relativistic and lightcone projection effects in the galaxy number counts are included in the forecast.","core_discovery":"The paper's central claim is that correlated CIPs with amplitude $A$ comparable to the adiabatic curvature fluctuations are detectable through their scale-dependent imprint on galaxy bias. The CIP amplitude enters the galaxy overdensity through the bias term $b_{bc}(z)\\left[\\delta_{bc}+f\\Delta(\\mathbf{k})\\right]$, where $\\Delta(\\mathbf{k})=(5H^2\\Omega_m)/(2a k^2)\\,A\\,\\delta_m$, producing a $k^{-2}$ enhancement at large scales. Combining galaxy number counts with the kSZ-reconstructed remote dipole field and their cross-spectra in a Fisher forecast, and marginalizing over cosmological parameters and all relevant bias functions, the paper obtains $\\sigma(A)=0.25$; marginalizing additionally over the local non-Gaussianity parameter $f_{NL}$ degrades this to $0.49$. The paper concludes that CIPs of order the adiabatic amplitude can be probed, improving on current CMB constraints by over two orders of magnitude, and that curvaton scenarios predicting $A=16$ or $A=-3$ can be confirmed or ruled out.","pith_inferences":["Editorial inference: the projected $\\sigma(A)=0.25$ rests on the remote-dipole reconstruction noise reaching Eq. (17); a direct measurement of that noise, including foregrounds and uncertainties in the free-electron distribution, is the most direct way to test the forecast before the surveys are built.","Editorial inference: because the method uses cross-correlations, it is largely blind to CIP components uncorrelated with the adiabatic mode; such stochastic components would add noise to the measurement, so the quoted sensitivity applies to correlated CIPs only.","Editorial inference: if a future experiment fails to reach the assumed small-scale CMB sensitivity, the constraint degrades, but the paper's parameter scan suggests low multipoles carry most of the signal, so an experiment optimized for large angular scales may retain much of the gain."],"forward_implications":["A future CMB survey with roughly one microKelvin-arcmin noise, together with a deep photometric galaxy survey, could detect correlated CIPs with $A\\sim 1$, turning a CMB-invisible mode into a routine observable.","The same data set must fit $A$ and $f_{NL}$ together: marginalizing over primordial non-Gaussianity weakens the constraint from $0.25$ to $0.49$, so a claimed CIP detection needs a joint analysis.","Curvaton scenarios with $A=16$ and $f_{NL}=6$ would be detected at high significance even with a weak prior on the CIP bias function.","Relativistic and lightcone projection effects cannot be neglected: omitting them biases the inferred CIP amplitude by about $1.5\\sigma$.","Existing photometric quasar samples may already tighten CIP constraints relative to the CMB, because their non-Gaussianity transfer function is comparable in size to the CIP contribution."],"supporting_citations":[{"why":"Current CMB observations that set the existing adiabatic-isocurvature baseline and motivate the CIP gap.","marker":"[1]"},{"why":"Previous CMB-based CIP constraints that the forecast compares against.","marker":"[8]"},{"why":"Derives the scale-dependent galaxy bias from correlated CIPs that the whole method exploits.","marker":"[15]"},{"why":"Introduces the sample-variance cancellation used to beat cosmic variance.","marker":"[17]"},{"why":"Introduces the sample-variance cancellation used to beat cosmic variance.","marker":"[18]"},{"why":"Shows kSZ tomography can constrain local non-Gaussianity, establishing the estimator framework.","marker":"[19]"},{"why":"Provides the forecast prescription, bias modeling, and the galaxy-survey-only comparison baseline.","marker":"[20]"},{"why":"Provides the remote dipole field signal computation including Sachs-Wolfe, ISW, and Doppler contributions.","marker":"[24]"},{"why":"Supplies the mass function, HOD model, and kSZ reconstruction details used for $b_{bc}$ and noise.","marker":"[25]"},{"why":"Supplies the curvaton decay scenarios with specific $A=16$ and $f_{NL}=6$ predictions.","marker":"[32]"}],"fun_headline_variants":["kSZ tomography reveals early-universe isocurvature at adiabatic scale","Correlated CIPs detected via scale-dependent galaxy bias","Cosmic-variance-free probe of compensated isocurvature perturbations","Two-orders-of-magnitude tighter constraints on early-universe CIPs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The forecast assumes that foregrounds and uncertainty in the free-electron distribution do not substantially raise the noise of the remote-dipole reconstruction; if the effective reconstruction noise is larger than Eq. (17) predicts, the sample-variance cancellation weakens and $\\sigma(A)=0.25$ is no longer reached.","fun_headline_variants_meta":{"raw":{"variants":["kSZ tomography reveals early-universe isocurvature at adiabatic scale","Correlated CIPs detected via scale-dependent galaxy bias","Cosmic-variance-free probe of compensated isocurvature perturbations","Two-orders-of-magnitude tighter constraints on early-universe CIPs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000313,"raw_usage":{"total_tokens":1840,"prompt_tokens":1065,"completion_tokens":775,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":681,"completion_tokens_details":{"reasoning_tokens":699}},"tokens_in":681,"tokens_out":775,"duration_ms":7585,"temperature":1.0,"reasoning_tokens":699,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:25:15.912379+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the kSZ remote-dipole reconstruction noise on a real next-generation CMB data set, treating thermal Sunyaev-Zeldovich emission, dust, and radio sources as they appear in the maps; if the effective noise at multipoles up to 9000 exceeds Eq. (17), the projected CIP sensitivity degrades roughly in proportion, and the claimed improvement over survey-only constraints shrinks.","supporting_citations":[{"cited_title":"Baryon-CDM isocurvature galaxy bias with IllustrisTNG","cited_arxiv_id":"1907.04317","evidence_quote":"Derives the scale-dependent galaxy bias from correlated CIPs that the whole method exploits."}],"review_version":1}