{"id":"01f9d161-be7c-45e3-a81f-d1b356fbd5c1","arxiv_id":"2607.27410","paper_version":2,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":13,"one_line_summary":"DESI DR2 Lyman-alpha forest full-shape correlations yield a 1% Alcock-Paczyński measurement at z=2.33 and 0.8% distance ratio constraints.","lead":"DESI's Lyman-alpha forest data constrain the Alcock-Paczyński anisotropy at z=2.33 to 1%, the tightest such measurement at z>1. The result sharpens cosmological distance anchors and slightly reduces the DESI-vs-CMB discrepancy.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Zero systematic error on φs rests solely on FGPA mocks; full-hydro validation needed before claiming 1% AP.","rationale":"The paper is a mature, blinded, mock-validated measurement with a strong internal consistency (auto vs cross, data splits, χ2/ndof≈1.01), and the code and mocks are available. The reader's weakest_assumption identified φs unbiasedness as the key risk; I agree and sharpen it: the unbiasedness is certified only by FGPA mocks, which share a common approximation of small-scale Lyα physics. The new SSM method was introduced precisely because small-scale error leaks into all fitted bins; if the real IGM's small-scale behavior differs from FGPA in a way that changes that leakage, φs could be biased at the 1/3σ level or larger, directly affecting the headline 1% AP constraint and the derived Ωm, H0, and dark-energy parameters. This is not an accusation of error, but a specific missing validation. Because the no-systematic-error decision is load-bearing and the test is feasible with existing hydro simulations, I recommend a CONDITIONAL verdict: accept the measurement only after the full-hydro mock test confirms the bias is below threshold, or after an explicit systematic error is added to φs. If the test passes, the original ACCEPT stands.","tokens_in":50182,"tokens_out":11294,"duration_ms":129155,"concrete_test":"Run the baseline VEGA pipeline on a suite of large-volume full hydrodynamical Lyα mocks with multiple IGM thermal histories and known true cosmology (e.g., Sherwood/Nyx-class light-cones or the simulations used to train ForestFlow in [29]), using the same scale cuts, SSM, and fitting procedure as the DR2 baseline. Require the recovered φs to be within the 0.38% (1/3σ) threshold of truth across all thermal histories; if any variant exceeds this, add an explicit systematic error to φs or revise the model/scale cuts. This directly tests whether the absent φs systematic is justified.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Section IV C states that the approach is expected to hit a 'systematic floor' at higher precision, yet the decision to quote no extra systematic on φs (Sections IV C and V A) is based entirely on two FGPA mock suites (AbacusSummit and CoLoRe 2LPT, Section IV A). These mocks approximate the Lyα optical depth from the matter density via FGPA; they do not model full hydrodynamical IGM effects (thermal broadening, pressure, temperature-density relation). The very effect that motivated the new small-scale marginalization (SSM, §III C)—small-scale model error leaking into all fitted bins through the continuum-fitting distortion matrix—could be quantitatively different in real data than in FGPA mocks, because the real small-scale Lyα fluctuations are not those of the FGPA field. If that leakage is mis-modeled, φs can shift by more than the 0.38% validation threshold (1/3σ) without being detected by the current mock tests, all four Abacus variants sharing the same FGPA approximation. The paper's own warning that future analyses will need EFT/emulator models reinforces that the current linear-plus-empirical model is approximate at percent-level precision. Since φs dominates the combined 1% AP constraint (σ_φs=1.1% vs σ_φp=2.0% before sys), an unrecognized φs bias directly changes the headline DM/DH and the derived Ωm and H0 values.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents Alcock–Paczyński measurements from the full shape of DESI DR2 Lyα‑forest auto‑ and cross‑correlations. Using a template split into a BAO peak and a smooth component, it extracts the broadband AP parameter φs = 1.007 ± 0.011, and combines it with the BAO AP parameter to obtain DM/DH(zeff=2.33) = 4.572 ± 0.046 (1.0%). Combined with the isotropic BAO scale, the analysis yields DM/rd = 39.32 ± 0.33 and DH/rd = 8.600 ± 0.066. Under ΛCDM, LyαFS+BBN gives H0 = 66.5 ± 1.3 km s⁻¹ Mpc⁻¹ and Ωm = 0.325 ± 0.018; combining with DESI BAO reduces the DESI–CMB discrepancy from 2.4σ to 2.2σ and the w0wa preference from 3.2σ to 2.7σ (DESI+CMB). The paper introduces a new small‑scale marginalization (SSM) technique to remove continuum‑fitting leakage and a UVB‑fluctuation model; the UVB bias is measured at bγ = 0.143 ± 0.047, consistent with theoretical expectations.","tokens_in":50756,"tokens_out":7880,"duration_ms":93796,"significance":"If correct, the 1% AP constraint at z>2 is the tightest large‑scale‑structure anchor of the distance ratio at z>1 and a substantial improvement over BAO‑only constraints. The paper’s strengths are the blinded analysis, validation on two independent mock suites (AbacusSummit four variants and CoLoRe 2LPT), multiple data splits, and the good fit quality (PTE = 0.20). The small‑scale marginalization method is an interesting and generally applicable contribution, and the UVB‑bias measurement provides a non‑trivial cross‑check of the model. The cosmological implications—reduced preference for evolving dark energy, a high‑redshift H0 constraint, and updated neutrino‑mass and curvature limits—are important for the current DESI tension landscape.","major_comments":[],"minor_comments":[{"comment":"The decision to quote no extra systematic on φs is the most consequential judgment call in the paper. The SSM argument is strong: marginalizing over every small‑scale bin in the undistorted model space removes sensitivity to the small‑scale model, and §IV B shows that removing the non‑linear correction entirely does not shift the result. Still, the paper should state explicitly that the reported φs uncertainty is statistical‑only and that the decision not to add a systematic is a modeling assumption, and should note what a full‑hydrodynamical validation would add given that both mock suites share the FGPA approximation.","section":"§IV C, §V A"},{"comment":"The BAO systematic Δφp = 0.8% is chosen to match the observed mock bias while the central value is not shifted. This is defensible given that EFT predictions and mocks disagree on the sign, but the text should be more explicit that this is an empirical uncertainty added in quadrature, not a bias correction applied to the measurement. A sentence clarifying this would prevent readers from interpreting the central values as already bias‑corrected.","section":"§V B, Eqs. (22)–(23)"},{"comment":"The symbol S is used for both the full parameter prior covariance and the diagonal prior on the marginalized template amplitudes (later set to S_ii = 100). Please distinguish the two, e.g., S_θ and S_η, to avoid ambiguity in the algebraic marginalization step.","section":"§III C, Eqs. (16)–(17)"},{"comment":"The caption contains a minor redundancy: “the constraints from the cross‑correlations in orange (orange)” repeats “orange”. Please remove the duplicate.","section":"Fig. 4 caption"},{"comment":"In the description of Figure 14, the sentence “By the fifth r⊥ bin shown in brown (0 < r⊥ < 4 h⁻¹ Mpc)” appears to have a typo in the transverse range; the fifth bin should not be the same range as the first. Please correct the transverse interval.","section":"Appendix A"}],"recommendation":"minor_revision","confidential_remarks":"This is a mature DESI collaboration paper with extensive internal review, and the central measurement is well supported. The referee’s main reservation—the absence of an explicit systematic on φs—is mitigated by the SSM design and the accompanying robustness tests, so I regard it as a caveat to be clarified rather than a blocking issue. I recommend acceptance after minor revisions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a well-executed, blinded measurement that delivers the tightest distance constraint at z>1 from large-scale structure — a 1% AP ratio DM/DH at z=2.33 — and I'd send it for peer review. The methodology is an extension of the eBOSS/DR1 full-shape Lyα program rather than a paradigm shift, but the two new pieces (small-scale marginalization, UVB fluctuation model) are real improvements and the analysis is honest about its limits.\n\nWhat stands out: the validation is thorough. Blinded analysis, 1/3σ threshold, four AbacusSummit variants plus 400 CoLoRe mocks, data splits that agree, and a good fit (PTE=0.20). The small-scale marginalization is the right way to handle continuum-fitting leakage; it gives each small-scale bin a free template amplitude, so the exact small-scale physics matters much less than in previous analyses. The UVB measurement b_γ=0.143±0.047 matches the theoretical prediction, which is a nice cross-check. The BAO systematic budget is updated based on mock shifts, which is transparent.\n\nSoft spots: not adding an explicit systematic on φ_s is the main judgment call. The stress-test concern about FGPA mocks is legitimate in principle, but it doesn't fully land — SSM is designed to absorb small-scale model uncertainty, and the mock variants bracket the FGPA freedom. Still, an independent full-hydro test would strengthen the claim at the 1% level; the paper itself says a systematic floor is expected. Minor: the BAO shift systematic is empirically motivated, not derived, and some fixed parameters come from companion papers, but both are tested for impact. The cosmological interpretation is competent and appropriately cautious, including the reduced DESI-CMB tension and the H0=66.5±1.3 from Lyα+BBN.\n\nThis paper deserves a serious referee. My recommendation: accept with the suggestion to add a hydro-based validation or at least an explicit justification for the zero systematic on φ_s. I would cite it.","headline":"Solid 1% AP measurement at z=2.33; SSM absorbs most of the FGPA worry, but an explicit systematic on φ_s would be a cleaner choice.","tokens_in":52066,"tokens_out":3537,"would_cite":true,"duration_ms":40438,"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":"The full shape of Lyman-alpha forest correlations from the survey's second data release measures the Alcock-Paczyński distance ratio at z=2.33 to 1%, the tightest high-redshift expansion anchor from large-scale structure.","keywords":["Lyman-alpha forest","Alcock-Paczyński effect","baryon acoustic oscillations","cosmological distances","dark energy equation of state","Hubble constant","neutrino mass","large-scale structure"],"falsifier":"Replace the linear-theory template with an emulator trained on hydrodynamical simulations and refit the same DR2 correlation functions; if the full-shape D_M/D_H moves by more than about 0.4%—one-third of the quoted uncertainty—the claim that systematics are controlled fails.","tokens_in":50129,"feed_emoji":"🔭","tokens_out":7508,"duration_ms":78659,"temperature":0.7,"pith_summary":"This paper tries to establish that the full shape of the Lyman-alpha forest auto-correlation and its cross-correlation with quasars, from the second data release of the Dark Energy Spectroscopic Instrument, measures the Alcock-Paczyński distance ratio D_M/D_H at effective redshift 2.33 to 1% precision. If correct, this is the tightest large-scale-structure anchor of the expansion history at z>1, about twice as precise as the BAO-only constraint from the same data. It also determines the transverse and line-of-sight distances D_M/r_d and D_H/r_d to 0.84% and 0.77% precision. Assuming a flat ΛCDM background, the measurement yields H0=66.5±1.3 km/s/Mpc when combined with a baryon density prior, and Omega_m=0.325±0.018. These numbers sharpen the high-redshift anchor and modestly shift the DESI-vs-CMB comparison, reducing that discrepancy from 2.4 sigma to 2.2 sigma.","feed_headline":"Lyman-alpha forest full shape pins cosmic distance ratio to 1 percent","feed_subtitle":"Full-shape Lyman-alpha analysis beats BAO-only error by 2x and tightens dark-energy tests.","key_machinery":"The central machinery is a template for the 2D correlation function split into a BAO peak component and a smooth broadband component, each with its own Alcock-Paczyński scale parameters (phi_p, alpha_p and phi_s, alpha_s). The AP parameter phi = q_perp/q_parallel equals the true D_M/D_H divided by the fiducial value, isolating the anisotropic stretch that occurs when the wrong cosmology maps angles and redshifts to distances. Redshift-space distortions—the apparent anisotropy from peculiar velocities—are separated from AP because the Lyman-alpha forest has a large RSD parameter beta_F≈1.2, making the AP response shape distinctly different. A new small-scale marginalization method treats the","core_discovery":"The central claim is that the anisotropy of Lyman-alpha correlations at z_eff=2.33 carries an Alcock-Paczyński signal that can be measured from the broadband, smooth part of the correlation function rather than only from the BAO peak. Combining the Lyman-alpha auto-correlation and the Lyman-alpha-quasar cross-correlation, the paper measures the broadband AP parameter phi_s = 1.007±0.011, corresponding to D_M/D_H = 4.578±0.052; adding the BAO-peak AP constraint gives a full-shape result D_M/D_H = 4.572±0.046, a 1.0% measurement. The joint full-shape and BAO fit gives D_M/r_d = 39.32±0.33 and D_H/r_d = 8.600±0.066 at z_eff=2.33. The paper argues that systematics from small-scale non-linearitie","pith_inferences":["The clean separation between AP and redshift-space distortions in this analysis relies on the large beta_F of the Lyman-alpha forest; below the current 1% precision, a forward-model approach calibrated to hydrodynamical simulations will likely be needed to avoid the systematic floor the paper itself acknowledges.","The new small-scale marginalization method is transferable: any analysis of absorption spectra with anisotropic distortion matrices from continuum fitting could adopt it to prevent small-scale leakage from contaminating large-scale anisotropic signals.","Because the measurement sits deep in the matter-dominated era, it acts as a geometric anchor for models that alter the pre-recombination sound horizon, such as early dark energy or modified gravity, making it a sharper discriminant for those models than low-redshift BAO alone.","If the same pipeline is applied to the next data release with roughly twice the sample, the 1% constraint should shrink toward 0.7%; a deviation from ΛCDM at that level would be far more informative than the current two-sigma tensions."],"forward_implications":["The 1% AP constraint at z=2.33 is the tightest distance-ratio measurement at z>1 from large-scale structure, exceeding the survey's original precision requirements with only three years of data.","Combined with BAO, the distances D_M/r_d and D_H/r_d at z_eff=2.33 are determined at 0.84% and 0.77%, providing a high-redshift anchor for the expansion history.","In flat ΛCDM, the Lyman-alpha forest alone gives Omega_m=0.325±0.018, and with a baryon density prior gives H0=66.5±1.3 km/s/Mpc, independent of CMB anisotropies.","Adding the Lyman-alpha full-shape data reduces the DESI-vs-CMB tension from 2.4 sigma to 2.2 sigma and lowers the preference for evolving dark energy from 3.2 sigma to 2.7 sigma (DESI+CMB) and from about 3.4 sigma to 3.1 sigma (with supernovae).","In ΛCDM, the joint DESI+CMB analysis yields an upper limit on the sum of neutrino masses of <0.0592 eV at 95%, essentially touching the lower bound from neutrino oscillation experiments in the normal ordering."],"fun_headline_variants":["Lyman-alpha forest yields 1% AP measurement, beats BAO by 2x","Lyman-alpha AP measurement doubles precision of DESI BAO","Lyman-alpha forest gives sharpest cosmic ruler yet for dark energy","DESI Lyman-alpha web measures expansion history to 1% precision"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The measurement stands or falls on the assumption that the smooth part of the Lyman-alpha correlations, after the chosen scale cuts and small-scale marginalization, is modeled well enough that any residual confusion between the cosmic-stretch anisotropy and velocity-induced anisotropy is below one-third of the statistical error.","fun_headline_variants_meta":{"raw":{"variants":["Lyman-alpha forest yields 1% AP measurement, beats BAO by 2x","Lyman-alpha AP measurement doubles precision of DESI BAO","Lyman-alpha forest gives sharpest cosmic ruler yet for dark energy","DESI Lyman-alpha web measures expansion history to 1% precision"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000801,"raw_usage":{"total_tokens":3531,"prompt_tokens":1088,"completion_tokens":2443,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":832,"completion_tokens_details":{"reasoning_tokens":2363}},"tokens_in":832,"tokens_out":2443,"duration_ms":19152,"temperature":1.0,"reasoning_tokens":2363,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T01:38:51.956822+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Replace the linear-theory template with an emulator trained on hydrodynamical simulations and refit the same DR2 correlation functions; if the full-shape D_M/D_H moves by more than about 0.4%—one-third of the quoted uncertainty—the claim that systematics are controlled fails.","supporting_citations":[],"review_version":2}