{"id":"805d53f5-926d-465b-940a-13a248f5c479","arxiv_id":"2412.04959","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Forecasted 1-sigma sensitivity to the sum of neutrino masses is 15 meV for Planck+DESI and 7 meV for CMB-S4+MegaMapper, with the one-loop bispectrum providing 10% and 30% of the constraining power.","lead":"Cosmologists forecast that future galaxy surveys combined with cosmic microwave background data will measure the sum of neutrino masses with 15 meV precision using Planck with DESI, and 7 meV using CMB-S4 with MegaMapper. The paper also shows that these bounds are robust to plausible new physics, with shifts in the inferred mass dropping to about 1 meV for the most sensitive combination.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Projected 7 meV sensitivity relies on unvalidated one-loop bispectrum at z=4.3 with EFT parameters extrapolated from BOSS scaling.","rationale":"The reader identified the same weakest assumption (agreement = agree). This assumption is load-bearing because the headline 7 meV number is the main summary statistic, and the bispectrum at z=4.3 is the least empirically grounded input. The paper is transparent about the ~50% systematic uncertainty (Section 4.1) and about the linear-response nature of the new-physics shift estimate (footnote 12), so the correct verdict is a conditional acceptance: the qualitative conclusions (P+B improves constraints, S4+MegaMapper suppresses new-physics shifts to ~O(1-4) meV) are plausible, but the precise projected sensitivities should not be treated as final. No internal inconsistency or unfounded claim beyond this acknowledged extrapolation was found.","tokens_in":36006,"tokens_out":11163,"duration_ms":118836,"concrete_test":"Recompute the S4+MegaMapper Fisher (Table II, baseline row) for the z=4.3 bin using (i) unscaled BOSS EFT parameters, (ii) a HOD-calibrated bias model at z=4.3, and (iii) the current scaling but with the perturbativity prior removed, keeping all other settings fixed. If sigma_Sigma_mnu shifts by more than 30% in any case, the extrapolation is load-bearing. A complementary test would be to run N-body simulations with massive neutrinos at z~4.3, build galaxy mocks matching MegaMapper's shot noise, and fit the one-loop EFT P and B monopole up to kmax=0.76 h/Mpc to check for unbiased neutrino mass recovery.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing concern is the unvalidated reach of the one-loop EFTofLSS bispectrum in the z=4.3 MegaMapper bin. Section 4.1 adopts fiducial EFT parameters by redshift-scaling the BOSS best fit (via Ref. [146]) to z=4.3, and Table I assigns this bin a one-loop reach of kmax=0.76 h/Mpc with bias b1=6.3. The derivatives of the power spectrum and bispectrum with respect to Sigma_mnu, and the Gaussian covariance in Eqs. (9)-(10), are evaluated at these extrapolated values. The one-loop bispectrum has not been validated against simulations or data at such high redshift and high bias, so the ~30% contribution of the bispectrum to the 7 meV sensitivity (Section 5a) could be optimistically estimated if the redshift scaling or the BOSS-calibrated perturbativity prior is inaccurate. Since the projected sensitivity itself is the central claim, and the authors estimate an additional ~50% systematic from Gaussian covariance and neglected AP/window effects, the 7 meV figure should be read as indicative rather than robust.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents Fisher forecasts for the sum of neutrino masses using one-loop EFTofLSS predictions for the galaxy power spectrum and bispectrum, combined with CMB forecasts from Planck and Stage-4 experiments and LSS forecasts for DESI and MegaMapper. The headline results are a projected 1-sigma sensitivity of 15 meV for Planck+DESI and 7 meV for S4+MegaMapper, with the one-loop bispectrum contributing about 10% and 30% of the constraining power, respectively. The paper also considers four theory extensions (extra relativistic species, neutrino self-interactions, curvature, and a time-varying electron mass) and quantifies the shift in the inferred neutrino mass induced by a 1-sigma shift in each new-physics parameter. The analysis pipeline follows Ref. [146], uses public Boltzmann and sampling codes, and includes an analytic Fisher cross-check in Appendix A.","tokens_in":36248,"tokens_out":8192,"duration_ms":89388,"significance":"If taken at face value, the forecasts are timely and useful: they extend earlier power-spectrum-only forecasts by including the one-loop bispectrum, and they provide a quantitative, if approximate, assessment of the robustness of neutrino-mass sensitivities to several motivated new-physics directions. The manuscript is transparent about its main limitations: Section 4.1 states that the Gaussian covariance and neglected Alcock-Paczynski/window effects may make the Fisher constraints about 50% tighter than a full analysis, and footnote 12 explicitly limits the shift estimates. The inclusion of an analytic Fisher estimate in Appendix A and comparisons with Refs. [32-34] help situate the results. However, because the headline 7 meV sensitivity and the 30% bispectrum gain rest on the one-loop bispectrum being reliable at z_eff=4.3 up to k_max=0.76 h/Mpc with EFT parameters extrapolated from BOSS, the central numbers should be presented with a robustness band or an explicit sensitivity test.","major_comments":[{"comment":"The high-redshift MegaMapper bin (z_eff=4.3, b1=6.3, k_1L_max=0.76 h/Mpc) is the bin where the one-loop bispectrum contributes most of its forecasted gain, yet the EFT parameters are obtained by redshift-scaling the BOSS best fit via Ref. [146], and no validation of the one-loop bispectrum in this high-bias, high-redshift regime is presented or cited. Since the Fisher derivatives in Eq. (10) and the Gaussian covariance are evaluated at these extrapolated values, the 30% bispectrum contribution and the 7 meV sensitivity are load-bearing on this assumption. I ask for a robustness test that degrades k_max for this bin (for example to the tree-level reach of 0.28 h/Mpc or to an intermediate 0.5 h/Mpc) and/or widens the EFT priors specifically for this bin, reporting the resulting sigma_Sigma_mnu. The prior-widening check in Appendix B covers only the Planck+DESI power-spectrum case and does not address this concern.","section":"Section 4.1 / Table I"},{"comment":"The covariance is a Gaussian, diagonal approximation with no power-spectrum-bispectrum cross-covariance, and Alcock-Paczynski and window-function effects are neglected; the text estimates that these approximations make the Fisher constraints about 50% tighter than a full analysis. This systematic uncertainty is comparable to the 30% bispectrum gain quoted in the abstract and to the difference between 7 and 10 meV. I recommend that the abstract and Section 5 quote the central forecasts with this systematic explicitly attached, for example as a 50% pipeline uncertainty on the Fisher numbers, or that a shortened-k_max analysis be presented as a more conservative headline.","section":"Section 4.1 / Eqs. (9)-(10)"},{"comment":"The claimed O(1) meV suppression of new-physics shifts for S4+MegaMapper is obtained from the linear-response formula Delta_Sigma = corr * sigma evaluated at the fiducial zero-new-physics point, and footnote 12 states that this is only an estimate requiring a full MCMC with a shifted fiducial for a comprehensive study. The non-Gaussian behavior visible in the Planck+DESI delta_m_e case, where the correlation changes sign between the CMB-only and combined analyses, shows that the linear correlation estimate can be fragile. Because this shift estimate is one of the two central claims in the abstract, I recommend either computing the Fisher matrix at a representative 1-sigma-shifted fiducial for the new-physics parameters or explicitly downgrading the claim to an approximate linear-response estimate.","section":"Section 4.4d / Table V / footnote 12"}],"minor_comments":[{"comment":"The sentence starting 'the best fits of Planck and BOSS analyses are not compatible' contains a broken reference placeholder '[ ? ]' and needs a proper citation.","section":"Section 2b"},{"comment":"In the MegaMapper bullet, the text says 'We use the same redshift binning, kmax, linear bias, and shotnoise specifications for DESI as outlined in Table 3 of [146]'; this should refer to MegaMapper, not DESI.","section":"Section 4.2"},{"comment":"The bispectrum gain is reported as 'about 10% and 30%' in the abstract, '15% and 25%' in Section 4.4b, and '33%' in Section 5a; please define the baseline for each percentage (power-spectrum-only, tree-level bispectrum, or Fisher-information fraction) so the numbers can be compared consistently.","section":"Abstract / Section 4.4b / Section 5a"},{"comment":"The row for 'this work' lists 17 meV for Planck+DESI with P_1-loop, while Table II gives 15 meV for the P+B analysis; this is not contradictory if Table VI is power-spectrum-only, but the table entry should state this explicitly.","section":"Table VI"},{"comment":"The symbol sigma is used both for the log-space standard deviation of the lognormal distribution and for the final 1-sigma bound; please distinguish these two quantities notationally.","section":"Eq. (11)"}],"recommendation":"major_revision","confidential_remarks":"The forecast rests heavily on Ref. [146] by overlapping authors for survey specifications, k_max estimation, EFT parameter priors, and the redshift-scaling prescription. This is not improper, but it means that a reader cannot reproduce the central numbers without that paper, and the referee may wish to confirm that the reliance is adequately disclosed and that the validation status of the high-redshift one-loop bispectrum is clarified."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The useful thing about this paper is that it goes beyond the usual power-spectrum-only forecasts and quantifies what the one-loop bispectrum actually buys for the neutrino mass measurement. The new numbers—15 meV for Planck+DESI, 7 meV for S4+MegaMapper, with the bispectrum contributing ~10% and ~30% respectively—are concrete and cross-checked against a simple analytic estimate. The robustness analysis against new physics (Neff, neutrino self-interactions, curvature, varying electron mass) is also a genuine addition: the correlation-based shift tables give a quick sense of where degeneracies bite and where they do not. The paper is transparent about its approximations, and the internal consistency of the forecast (analytic check, comparison to Refs. [32–34]) gives me confidence the analysis was done carefully.\n\nThe soft spots are real but mostly acknowledged. The stress-test concern about the z=4.3 MegaMapper bin is on point: assuming the one-loop bispectrum holds up to kmax=0.76 h/Mpc with bias b1=6.3, based on EFT parameters redshift-scaled from the BOSS best fit, is a stretch. That bin is precisely where the one-loop bispectrum has not been validated against simulations or data. The authors' own estimate that Gaussian covariance and neglected AP/window effects could tighten the Fisher result by ~50% means the 7 meV should be read as an optimistic lower bound—a realistic sensitivity might be closer to 10 meV. The new-physics shift estimates are also explicitly linear correlation approximations, not full MCMC shifts, so the O(1) meV claim for S4+MegaMapper is indicative rather than settled. None of this is hidden; it is in Section 4.1 and the footnotes, and the conclusions do not oversell.\n\nThe citation pattern is defensible: the framework comes from Ref. [146] and the authors reuse its survey specifications, so the heavy citation is mostly methodological. They do also compare against independent forecasts, which is the right thing to do.\n\nThis paper deserves a serious referee. It is a solid, clearly-written forecast that will be useful to anyone planning LSS neutrino analyses, and the robustness table is a convenient reference. The main referee requests should be a mock-based validation of the high-z bispectrum (or at least a more cautious treatment of that bin) and a fully propagated systematic error budget on the headline numbers. But as a forecast, it is honest and within the norms of the field.\n\nMy recommendation: send it to peer review with a request for those clarifications. I would cite it in my own work on neutrino mass forecasts.","headline":"A careful, honest Fisher forecast that puts real numbers on the one-loop bispectrum's value for neutrino mass, but the headline 7 meV sensitivity should be treated as optimistic given the acknowledged ~50% systematics and unvalidated high-redshift extrapolation.","tokens_in":36778,"tokens_out":1727,"would_cite":true,"duration_ms":21144,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["83F05","85A40"],"pacs":[],"model":"deepseek-v4-flash","headline":"Galaxy surveys plus CMB can weigh neutrinos to 7 meV, forecasts show.","keywords":["sum of neutrino masses","EFTofLSS","one-loop bispectrum","Fisher forecast","DESI","MegaMapper","CMB","new physics"],"falsifier":"Compare the one-loop EFTofLSS power spectrum and bispectrum predictions against N-body simulations with massive neutrinos at z≈2–5 on scales up to the adopted kmax values; if the residuals systematically exceed the survey data errors at k≳0.3 h/Mpc, the projected sensitivities, especially the 7 meV for S4+MegaMapper, would be overestimated. Alternatively, run the same Fisher pipeline on BOSS data and compare the forecast error bars to the actual measured ones to calibrate the roughly 50% systematic from covariance and window approximations.","tokens_in":35799,"feed_emoji":"🔭","tokens_out":6316,"duration_ms":57044,"temperature":0.7,"pith_summary":"This paper forecasts how precisely upcoming cosmological surveys can measure the sum of neutrino masses, Σmν, using effective field theory predictions for galaxy clustering at one loop. It finds that Planck combined with DESI should reach a 1σ sensitivity of 15 meV, while a Stage-4 CMB experiment combined with the proposed MegaMapper survey could reach 7 meV, with the one-loop bispectrum contributing roughly 10% and 30% of the constraining power, respectively. The authors also test how these sensitivities change when the model is extended with new physics—extra relativistic species, neutrino self-interactions, spatial curvature, or a varying electron mass—and find that for S4+MegaMapper a 1σ shift in any of these parameters displaces the inferred Σmν by only O(1) meV. A measurement this precise would complement laboratory neutrino experiments and could discriminate the neutrino mass ordering.","feed_headline":"Future surveys could measure neutrino mass sum to 7 meV","feed_subtitle":"One-loop clustering predictions show Planck+DESI reaching 15 meV and S4+MegaMapper 7 meV, robust against new physics.","key_machinery":"The analysis is built on the Effective Field Theory of Large-Scale Structure (EFTofLSS) at one loop: the galaxy power spectrum monopole and quadrupole and the galaxy bispectrum monopole in redshift space, with bias expansion and counterterms, evaluated with fast loop-integral code, and Fisher matrices for LSS surveys combined with CMB likelihood chains. Neutrino effects enter through the fν corrections—the power spectrum is suppressed by roughly (1−8fν) and the bispectrum by (1−16fν) on scales k≫kFS—so redshift-space distortions help break the degeneracy with galaxy bias. EFT parameters are taken from BOSS best fits and extrapolated in redshift, and the theory reach kmax is set by requiring the estimated two-loop error to stay below the survey data error.","core_discovery":"The central claim is that with the state-of-the-art EFTofLSS, including the one-loop power spectrum and bispectrum, the projected 1σ uncertainty on the sum of neutrino masses is σ(Σmν)=15 meV for Planck+DESI and 7 meV for S4+MegaMapper, assuming normal ordering and a fiducial Σmν=60 meV. The one-loop bispectrum contributes about 10% and 30% of these constraints respectively, and in the S4+MegaMapper combination the shift in Σmν induced by a 1σ shift of new physics parameters (extra relativistic species, neutrino self-interactions, curvature, electron-mass variation) is suppressed to about 1–4 meV, making the measurement robust against those extensions. The robustness comes from the complementarity of CMB and LSS probes breaking parameter degeneracies.","pith_inferences":["The assumed roughly 50% systematic from Gaussian covariance and neglected Alcock-Paczynski and window effects suggests real analyses may yield weaker constraints; testing the pipeline on BOSS-like mocks would quantify this.","The redshift-scaling of EFT parameters from BOSS is an untested extrapolation at z≈2–5; calibrating the one-loop model against N-body simulations with massive neutrinos at those redshifts would directly test the kmax reach behind the 7 meV projection.","The lognormal Fisher treatment of Σmν and the correlation-based shift estimates could be verified by full MCMC runs with shifted fiducial values, which would also probe non-Gaussian posteriors.","The projected gain from the one-loop bispectrum hints that higher-order statistics (two-loop power spectrum, higher-N point functions) could push sensitivities further, though with more nuisance parameters to marginalize."],"forward_implications":["Within about five years, Planck+DESI could reach σ(Σmν)=15 meV, giving a roughly 4σ detection of a non-zero sum and more than 2σ discrimination between normal and inverted ordering for the minimal normal-ordering mass.","With S4+MegaMapper, σ(Σmν)=7 meV would make the mass ordering accessible at about 5σ, a regime where cosmological precision rivals laboratory experiments.","The one-loop bispectrum sharpens the MegaMapper constraints by about 33% over the power spectrum alone, with roughly two-thirds of that gain coming from the one-loop rather than tree-level bispectrum.","Marginalizing over new physics parameters (Neff, self-interacting neutrinos, curvature, varying electron mass) degrades the S4+MegaMapper neutrino constraint by less than about 15%, so the projected sensitivity is nearly model-independent.","The forecast shifts of Σmν from a 1σ shift in new physics parameters stay at about 1–4 meV for S4+MegaMapper, meaning a future measured value would not be significantly biased by these extensions."],"supporting_citations":[{"why":"Supplies the one-loop power spectrum of biased tracers in redshift space in the EFTofLSS, the baseline two-point model.","marker":"[55]"},{"why":"Derives the one-loop bispectrum of galaxies in redshift space in the EFTofLSS, the key three-point prediction that drives part of the forecast gain.","marker":"[56]"},{"why":"Provides the EFTofLSS treatment in the presence of massive neutrinos, justifying the fν counting and renormalization scheme.","marker":"[128]"},{"why":"Enables fast evaluation of the one-loop loop integrals, making the bispectrum Fisher forecast computationally feasible.","marker":"[143]"},{"why":"Supplies the BOSS one-loop bispectrum analysis whose best-fit EFT parameters are the fiducial values for the forecasts.","marker":"[144]"},{"why":"Establishes the forecast framework, survey specifications for DESI and MegaMapper, and the kmax determination and redshift-scaling of EFT parameters.","marker":"[146]"},{"why":"Provides the CLASS Boltzmann solver used to compute the linear power spectrum including massive neutrinos for the fiducial cosmology.","marker":"[159]"},{"why":"Supplies the Planck 2018 best-fit cosmological parameters adopted as the fiducial model.","marker":"[163]"},{"why":"Provides MontePython, used to generate the CMB forecast likelihoods and covariance information for Planck and S4.","marker":"[167]"}],"fun_headline_variants":["Future surveys to weigh neutrinos: 7 meV sensitivity","Neutrino mass sum: 7 meV from future cosmic surveys","Cosmology's neutrino mass precision: 7 meV with S4+MegaMapper","Robust 7 meV neutrino mass forecasts from CMB+galaxy clustering","One-loop bispectrum pushes neutrino mass sensitivity to 7 meV"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The forecasts assume the one-loop EFTofLSS predictions, especially the bispectrum, remain accurate out to the adopted maximum wavenumbers (kmax up to about 0.76 h/Mpc for MegaMapper at z=4.3), with EFT parameters extrapolated from BOSS by a redshift-scaling prescription; if the one-loop model breaks down on those scales, the projected 7 meV sensitivity would be optimistic.","fun_headline_variants_meta":{"raw":{"variants":["Future surveys to weigh neutrinos: 7 meV sensitivity","Neutrino mass sum: 7 meV from future cosmic surveys","Cosmology's neutrino mass precision: 7 meV with S4+MegaMapper","Robust 7 meV neutrino mass forecasts from CMB+galaxy clustering","One-loop bispectrum pushes neutrino mass sensitivity to 7 meV"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000446,"raw_usage":{"total_tokens":2306,"prompt_tokens":1047,"completion_tokens":1259,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":663,"completion_tokens_details":{"reasoning_tokens":1159}},"tokens_in":663,"tokens_out":1259,"duration_ms":12121,"temperature":1.0,"reasoning_tokens":1159,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T21:06:43.896967+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compare the one-loop EFTofLSS power spectrum and bispectrum predictions against N-body simulations with massive neutrinos at z≈2–5 on scales up to the adopted kmax values; if the residuals systematically exceed the survey data errors at k≳0.3 h/Mpc, the projected sensitivities, especially the 7 meV for S4+MegaMapper, would be overestimated. Alternatively, run the same Fisher pipeline on BOSS data and compare the forecast error bars to the actual measured ones to calibrate the roughly 50% systematic from covariance and window approximations.","supporting_citations":[],"review_version":1}