{"id":"893e402e-35f2-4f31-ac10-498cc9e5b183","arxiv_id":"astro-ph/0601594","paper_version":4,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Weak lensing of the cosmic microwave background is a calculable, observationally important distortion that modifies CMB spectra, generates B modes and non-Gaussianity, and can be mapped back to constrain cosmology.","lead":"This review explains how weak gravitational lensing by large-scale structure distorts the cosmic microwave background, changing its temperature and polarization spectra, creating non-Gaussianity and B-mode polarization, and enabling reconstruction of the lensing mass distribution. It is the standard reference for why future CMB experiments must model lensing and how they might remove it to search for primordial gravitational waves.","discovery_kind":"review","skeptic_critique":{"model":"selfhosted/deepseek-v4.1-flash","headline":"The reader's flagged soft spot is real: Section 3.3's halofit-based non-linear potential omits baryonic feedback, which shifts the arcminute-scale lensed B-mode and trispectrum. But the paper acknowledges the non-linear modelling limit and the central claim survives.","rationale":"The reader's weakest_assumption correctly identifies the halofit/Gaussian-potential assumption as the least secure link in the quantitative chain, and correctly notes that Section 3.3 and Section 5.4 flag it. I agree with that identification and with the ACCEPT verdict. My only refinement is to name the specific missing systematic: baryonic feedback on P(k) at k ~ 1-10 h/Mpc, which the manuscript does not discuss at all. That omission is real but does not undermine the central claim, which is qualitative and holds under any plausible P(k) reshaping. The rest of the review is in good shape: the GR derivation of the deflection (Section 2.1), the Jacobi-map convergence definition in Eq. (2.41) that corrects earlier treatments, the non-perturbative correlation-function method (Section 4.2), and the new full-sky result in Eq. (9.10) are all internally consistent and cross-checked against the flat-sky limit via Eq. (9.11). The sign-convention footnote 10 is a genuine source of cross-code confusion but is disclosed. The delensing forecasts in Section 8 lean on idealized noise assumptions, which the paper acknowledges as foreground-limited. No internal inconsistency, no circular step, no missing proof of a cited result. The verdict should remain ACCEPT; the baryonic caveat is a precision note for the small-scale numbers, not a reason to downgrade a review whose central claim is robust to it.","tokens_in":61196,"tokens_out":2434,"duration_ms":53394,"concrete_test":"Recompute the lensed B-mode spectrum CBl at l ~ 1000, the diagonal of the trispectrum covariance in Eq. (6.8), and the Section 8 delensing residual using a baryon-corrected matter power spectrum (e.g. the Mead et al. halo model with baryons, or an OWLS/Cosmo-OWLS hydrodynamical fit) in place of halofit in Eq. (3.16). If CBl or the delensing residual shifts by more than the quoted few-percent accuracy and by more than cosmic variance at the relevant l, the quantitative claims in Sections 5.4 and 8 need an explicit baryonic systematic. If the shift is below cosmic variance, the concern does not land and the current numbers stand.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim [abstract:s0] is qualitative and robust: lensing changes the spectra, induces non-Gaussianity, and generates a B-mode confusion floor. Nothing in the derivation of the deflection (Eq. 2.13), the lensing potential (Eq. 3.3), or the non-perturbative correlation-function machinery (Eqs. 4.27, 5.44-5.46, 9.10) is at risk. The load-bearing soft spot is narrower and quantitative: the arcminute-scale numbers — CBl ~ 2x10^-6 uK^2 (Section 5.4), the trispectrum covariance in Eq. (6.8), and the delensing forecasts in Section 8 — all inherit their small-scale shape from P_Psi at k ~ 1-10 h/Mpc. Section 3.3 prescribes halofit, which is a dark-matter-only fit. Baryonic feedback (AGN and supernova-driven outflows) suppresses P(k) by 10-30% over exactly that range in hydrodynamical simulations. The manuscript never mentions baryons anywhere in Sections 3.3, 5.4, 6.4, or 8. The paper does acknowledge that halofit is only 'expected to be accurate at the few percent level' and that non-Gaussian corrections matter on small scales, so this is an acknowledged limitation rather than a hidden one. The residual concern is that the quoted few-percent accuracy budget does not include a systematic that is plausibly larger than a few percent, and that the delensing forecasts (factor ~40 reduction, r ~ 10^-6) are stated without a baryonic error bar. This is a precision caveat on the small-scale numbers, not a threat to the review's central thesis.","agreement_with_reader":"agree"},"referee_report":{"model":"selfhosted/deepseek-v4.1-flash","summary":"This is a review of weak gravitational lensing of the cosmic microwave background. It covers the general-relativistic derivation of the deflection angle (§2), the lensing potential and its power spectrum (§3), the lensed temperature and E/B polarization spectra (§4–5), the induced non-Gaussianity in the bispectrum and trispectrum (§6), reconstruction of the lensing potential (§7), delensing (§8), sky-curvature corrections (§9), simulations (§10), and applications to parameter estimation, cluster masses, and moving lenses (§11). The central claim, stated in the abstract, is that lensing changes the CMB power spectra, induces non-Gaussianities, and generates a B-mode polarization signal that is an important source of confusion for primordial gravitational waves, and that all of these effects are calculable from standard general relativity and cosmological perturbation theory. The review is largely expository, but it contains new material in §9, where a full-sky non-perturbative correlation-function result is presented as Eq. (9.10).","tokens_in":61603,"tokens_out":11015,"duration_ms":200335,"significance":"High. The review is rigorous where rigour matters: the GR derivation of the deflection in Eq. (2.13), the Jacobi-map treatment of §2.2, and the non-perturbative correlation-function machinery of §4.2 and §5.3.2 are carefully done, and the manuscript is candid about where approximations enter (the few-percent accuracy of halofit in §3.3, the breakdown of the series expansion at l ~ 2000 in §4.1.3, and the difficulty of cluster lensing in §11.3). It ships reproducible machinery: the lensed spectra are implemented in the public CAMB, CMBFAST, and cmbeasy codes, and the pixel-remapping code lenspix is public. It makes falsifiable predictions — the large-angle lensed B-mode amplitude CBl ~ 2 × 10^-6 µK², the ISW-lensing bispectrum of Eq. (6.2), and the delensing floor — that were subsequently tested against data. The new full-sky result of Eq. (9.10) is a genuine addition to the literature.","major_comments":[{"comment":"The non-linear lensing potential is obtained by rescaling the linear transfer functions with halofit, and §3.3 states that halofit is 'expected to be accurate at the few percent level'. That statement is a dark-matter-only accuracy budget: halofit is a fit to dark-matter simulations, and the quoted few percent does not include baryonic suppression of P(k) over k ~ 1–10 h/Mpc, which hydrodynamical simulations suggest can be comparable to or larger than a few percent. This matters because the arcminute-scale numbers that the review quotes inherit their small-scale shape directly from that range of k: the large-angle lensed B-mode amplitude CBl ~ 2 × 10^-6 µK² in §5.4, the trispectrum covariance of Eq. (6.8) in §6.4, and the delensing forecasts of §8 (the factor ~40 reduction and the r ~ 10^-6 floor). The manuscript should say explicitly in §3.3 that the accuracy budget is dark-matter-only, and add a caveat where the small-scale numbers are quoted, so that readers do not take the few-percent figure as a total error budget. This is a clarification of scope rather than an error in the derivation, and a sentence or two in each place would suffice.","section":"§3.3, Eq. (3.16); §5.4; §8","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"The footnotes dismiss Ref. (31) as 'erroneous' and Ref. (37) as 'incorrect' without giving the counter-argument in the footnote itself. The body text does supply the physics (magnification of a smooth surface is unobservable; only arcminute-coherent magnification affects CMB acoustic scales), but a short pointer in each footnote to the paragraph that makes the argument would help the reader and would soften the tone of what are otherwise bare assertions about published work.","section":"§1.5, footnotes 4 and 5","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The statement that derivations in Refs. (48; 50) 'appear to have neglected that time delay and 1 + ΦN terms required for a consistent definition of the magnification matrix' is a claim of error in published work. Please add a sentence identifying the specific term and its parametric size, or soften the wording to say that the definitions differ, so that the reader can judge the claim without consulting the cited papers.","section":"§2.2.1","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Eq. (9.10) is labelled 'new to this paper', which is appropriate, but a review that contains new material should signpost it in the introduction (or abstract) so that readers know which sections are review and which are original. In addition, the claim that Eq. (9.12) is 'correct to within O(1/l²) factors, giving lensed Cl indistinguishable from the full result' would be stronger with a quantitative residual; Fig. 17 compares flat-sky to full-sky second order, not Eq. (9.12) to Ref. (70).","section":"§9, Eq. (9.10)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The statement that the large-angle white-noise power is 'currently constrained to be CBl = (2.0 ± 0.2) × 10^-6 µK²' should name the data set and the range of l from which the constraint is derived, so that the number can be checked and updated as data improve.","section":"§5.4","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The noise variances are quoted as N Θl = NEl/2 = NBl/2 = 8.5 × 10^-8 µK² and NEl = NBl = 1.7 × 10^-7 µK²; the accompanying √N values (1 µK-arcmin and 1.41 µK-arcmin) are consistent only if the variances are in µK² sr, so the solid-angle unit should be stated. Separately, the Fig. 16 caption gives the beam as 'full-width half-maximum of 4 µK-arcmin', which should read '4 arcmin'.","section":"Figs. 15 and 16 captions","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The PACS code '8.80.Hw' appears malformed; please check it against the PACS scheme (98.80.-k or a similar cosmology code is presumably intended).","section":"PACS line","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"A small number of typographical errors should be corrected: 'the the one' in §2.3, 'the result it not especially enlightening' in §4.2.3, 'cotrolled' in §8, 'paralellizable' in §10, and 'Stoke's parameters' in §11.3.1. Some of the spacing irregularities elsewhere in the text may be typesetting artefacts and are not listed here.","section":"§2.3, §4.2.3, §8, §10, §11.3.1","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The derivation neglects the Θ–ψ correlation on the grounds that 'the correlation only has a tiny effect on the result', but Fig. 5 shows the correlation reaching ~10% at l ~ 60. Since the effect on the lensed Cl is second order in the correlation, the neglect is presumably safe, but a sentence quantifying the resulting error (rather than calling it 'tiny') would make the approximation checkable.","section":"§4.1, Eq. (4.12)","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"minor_revision","confidential_remarks":"This is a long, careful, and already widely used review; the requested changes are local. The one substantive point is the scope of the halofit accuracy statement in §3.3, which is a clarification rather than a correction. The journal may wish to consider the page budget, and to ensure that the new material in §9 is flagged in the front matter so that it is not lost in a review-length article."},"author_rebuttal":null,"desk_editor":{"model":"selfhosted/deepseek-v4.1-flash","letter":"This is the paper everyone cites when they start working on CMB lensing, and reading it again it holds up. Lewis and Challinor take you from the geodesic equation through the lensing potential, the lensed temperature and polarization spectra, the non-Gaussian structure, reconstruction, delensing, and sky curvature, and they do it without hand-waving. If a student asks me where to start, this is what I hand them. The GR section (2.1-2.2) is more careful than most primary papers on the same subject, and the flat-sky correlation-function machinery in 4.2 and 5.3.2 is presented in a way you can actually implement. The gauge and dipole discussion in 2.3 is a clarification the field needed and still doesn't cite enough. There is real new material despite the review label: Eq. (9.10) is a fully non-perturbative full-sky result, Eq. (4.29) is the exact flat-sky series, and the Jacobi-map convergence definition in Eq. (2.41) corrects earlier treatments that dropped the time-delay and (1+Phi_N) terms. That last point is a genuine correction to published work, not a cosmetic one. The citation pattern is thorough and fair; I don't see self-citation padding. On soft spots: the stress-test is right that Section 3.3 prescribes halofit, a dark-matter-only fit, and never mentions baryons. That matters for the arcminute-scale B-mode number in 5.4, the trispectrum covariance in 6.8, and the delensing forecasts in Section 8, which are quoted without an error bar. But the paper does flag that halofit is only good at the few-percent level and that non-Gaussian corrections grow on small scales, so this is an acknowledged limitation with an understated budget, not a hidden one. It's a precision caveat on the small-scale numbers, not a threat to the thesis. The cluster section (11.3) is admirably candid that kSZ confusion makes temperature-based mass reconstruction very hard; more reviews should be that honest about what won't work. Who should read it: anyone doing CMB parameter estimation, inflation searches, or cluster lensing, and anyone who needs the reconstruction estimators in Section 7 in usable form. It deserves a serious referee, though Physics Reports will handle it as a review and that's appropriate.","headline":"The standard reference for CMB lensing, and it earns that status: careful derivations, honest about approximations, with a few genuinely new results buried in a review.","tokens_in":62181,"tokens_out":1099,"would_cite":true,"duration_ms":20729,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"pith_extraction":null,"created_at":"2026-10-29T15:59:19.045539+00:00","model_set":{"reader":"selfhosted/deepseek-v4.1-flash"},"falsifier":null,"supporting_citations":[],"review_version":1}