{"id":"33425fbd-1ee5-4b36-a744-0c29ad64eadf","arxiv_id":"2509.08886","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A zero-lag Pearson correlation between 21-cm and line-intensity maps is predicted to drop sharply at ionized fractions of 1-10%, providing a novel marker for the onset of reionization.","lead":"This paper proposes a new way to detect the very beginning of the Universe's reionization epoch by comparing maps of 21-cm hydrogen emission with maps of star-forming galaxy emission. The cross-correlation between these maps is predicted to dip sharply when only 1-10% of intergalactic hydrogen is ionized, giving an observational anchor for a poorly constrained cosmic transition.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'model-robust' turnover hinges on an unproven exclusion of cold-IGM reionization (L_X,SFR=39.5), where no saturation plateau precedes x_HII~10%.","rationale":"The reader's weakest-assumption identification is exactly the load-bearing issue in the paper. The Pearson coefficient P(z) is proposed as a model-independent boundary condition on the EoR onset, but its interpretability depends on the IGM being fully heated before reionization reaches x_HII~10%. If this is not guaranteed across the observationally allowed parameter space, then a turnover in P(z) cannot be uniquely attributed to the first ionized bubbles. The paper explicitly demonstrates one counterexample (L_X,SFR=39.5) and then argues it is disfavored; however, the exclusion is not demonstrated robustly, especially because the cited observational limits weaken when PopIII mini halos are included. The added smoothing-scale dependence (Fig. 3) reinforces the concern that the turnover is not a universal marker. The paper is otherwise physically plausible, and the simulations support the feature in the standard heating-first regime, but the overclaim of model independence needs to be tempered. Since the reader's verdict already conditional, my stress-test does not change the verdict; it strengthens the specific condition that must be met for the claim to hold.","tokens_in":12510,"tokens_out":4828,"duration_ms":57154,"concrete_test":"Run 21cmFAST v4 with the astrophysical model of Ref. [103], adding PopIII/mini-halo prescriptions as in Ref. [100], and sample L_X,SFR over [39.5, 42] subject to current HERA/EDGES upper limits. Compute P(z) at R=5 Mpc and record whether a saturation plateau (defined, e.g., as |dP/dz| < 0.05 over Δz ≥ 1) exists before x_HII=10%. Report the posterior-weighted fraction of models without such a plateau. Repeat at R=10 Mpc to quantify how often the turnover occurs after x_HII=10%. If either fraction exceeds ~5%, the model-independence claim is falsified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the Pearson drop robustly marks the EoR onset depends on a temporal-ordering premise: P(z) must saturate (21-cm in emission everywhere) before the first ionized bubbles appear, so the subsequent drop is uniquely attributable to reionization. The paper's own Fig. 2 bottom-left violates this premise: for L_X,SFR=39.5, reionization proceeds in a still-cold IGM, P(z) does not reach a stable plateau before x_HII~10%, and the turnover no longer signifies the onset. The authors dismiss this as extreme, citing theoretical and observational constraints, but in the same paragraph concede those limits are 'less robustly' applicable if mini halos hosting PopIII stars are included (Ref. [100]). Thus the claim that the signature appears in 'all models consistent with current observations' is not established; the exception may be observationally allowed. Moreover, Fig. 3 (left) shows the smoothing scale shifts the end of the plateau: at larger R, the plateau extends past x_HII=10%, so the turnover redshift is not a universal function of x_HII. A detected turnover without a preceding plateau—or with a resolution-dependent plateau—cannot alone anchor the EoR onset.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a novel real-space estimator, the zero-lag Pearson cross-correlation coefficient P(z) between 21-cm brightness temperature maps and line-intensity maps of star-forming tracers (e.g., [OIII], CO, [CII]), as a probe of the onset of reionization. Using the Zeus21/oLIMpus simulation pipeline and independent 21cmFASTv4 simulations, the authors show that P(z) first rises from negative to positive as X-ray heating progresses, saturates once the 21-cm field is in emission everywhere, and then drops sharply when the first ionized bubbles form. They argue that this turnover occurs at mean ionized fractions x_HII ~ 1%–10% across a broad range of astrophysical parameters, and therefore provides a model-robust, model-independent observational anchor for the beginning of the EoR, complementing the later zero-crossing of the cross-correlation which signals ongoing reionization. The paper also discusses smoothing-scale dependence and compares P(z) with the Fourier-space cross-correlation coefficient.","tokens_in":12865,"tokens_out":3080,"duration_ms":41254,"significance":"If the central claim holds, the proposed Pearson turnover would fill a genuine observational gap: current probes constrain the end and duration of reionization, not its onset. The estimator is conceptually appealing because it uses only one-point statistics for normalization, potentially mitigating noise propagation relative to power-spectrum estimators. The authors demonstrate the feature with two independent simulation codes, explore variations in several astrophysical parameters, and connect the turnover to a physically motivated temporal ordering (heating before reionization). The use of public, reproducible simulation tools is a strength, and the qualitative prediction—a sharp drop from a saturated plateau in P(z) at early EoR stages—is falsifiable with next-generation 21-cm and line-intensity surveys. However, the robustness of the claim hinges on a premise that the paper itself only partially defends, and the abstract overstates the signal-to-noise analysis and the model independence of the conclusion.","major_comments":[{"comment":"The paper's own counterexample for L_X,SFR = 39.5 directly violates the temporal-ordering premise that saturates P(z) before x_HII~10%. In that case reionization proceeds in a still-cold IGM, no stable plateau is reached before the ionized fraction reaches 10%, and the turnover no longer encodes the EoR onset. The authors dismiss this model as extreme, citing theoretical and observational constraints [96–99], but in the same paragraph they concede these constraints are 'less robustly' applicable if mini halos hosting PopIII stars are included [100]. The concluding sentence 'In all models consistent with current observations' is therefore not established. This is load-bearing for the central claim of a model-robust signature. I ask the authors to quantify the region of parameter space where the plateau precedes x_HII=10% and where it does not, and to test explicitly a PopIII/mini-halo sce","section":"Fig. 2, bottom-left panel and accompanying text"},{"comment":"The smoothing-scale dependence undermines the claim that the turnover is a unique, parameter-free anchor for the EoR onset. The text states that for large R 'the plateau extends beyond the x_HII = 10% point,' meaning the turnover occurs at x_HII > 10% and is therefore not a universal function of the ionized fraction. Since the observable turnover depends on the chosen smoothing scale—and the fiducial R = 5 Mpc is chosen 'unless otherwise stated'—a detected turnover without prior knowledge of R cannot be unambiguously mapped to x_HII~10%. The paper should either demonstrate that R is fixed by the experimental resolution in a way that removes this degeneracy, or define the EoR-onset anchor in an R-independent way (e.g., by using the shape of the full two-point function as a consistency check). Without this, the claim that a turnover in P(z) alone provides an upper bound on x_HII is not ful","section":"Fig. 3, left panel and text on smoothing scale"},{"comment":"The abstract claims 'we provide a preliminary estimate of its signal-to-noise ratio in our fiducial scenario... indicating that it is within reach of next-generation surveys,' but the body explicitly defers all detectability studies to future work [94], which is cited as 'in preparation.' This is an internal inconsistency that overstates the paper's deliverable. Either the SNR estimate should be presented in this paper—with a concrete noise model, integration times, and statistical significance—or the abstract and conclusions should be softened to say that detectability remains to be assessed. As written, the abstract invites a claim the manuscript does not support. Relatedly, the term 'model-independent' is too strong: the estimator depends on the line-luminosity model of Ref. [95], the smoothing scale, and the assumed shot-noise properties; the paper means 'robust across the explored p","section":"Abstract and §4 (detectability statement)"}],"minor_comments":[{"comment":"The notation 'OIII' is inconsistently written as [OIII], OIII, and OIII. Please unify to a single convention, e.g., [OIII] or [O III].","section":"Abstract and text throughout"},{"comment":"The estimator P(z) is called the Pearson coefficient, but its one-point nature and the role of the smoothing scale R should be made explicit in the equation or immediately below it. The current definition sums over voxels, but the voxel size is only mentioned in the text; including R in the definition would clarify the R-dependence discussed in Fig. 3.","section":"Section 2, Eq. (6)"},{"comment":"The third panel is labeled 'unnormalized' while the text says it shows P_un(z), which is not the Pearson coefficient but an unnormalized version. Please clarify in the caption how this quantity is computed and why it is shown instead of the normalized version.","section":"Section 3, Fig. 1 caption"},{"comment":"Reference [94] is listed as 'arXiv:arXiv:2509:XXXX' with a malformed identifier. This placeholder should be corrected or the reference removed if the work is not publicly available yet.","section":"References"},{"comment":"Minor typographical issues: 'per-sedirect' should be 'per se direct'; 'LOF AR' appears with a stray space. These are easy to fix but occur in published-looking text.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper's core idea is interesting and the simulations are reproducible, but the advertised 'model-robust' and 'model-independent' claims are stronger than what the analysis supports. The acknowledged cold-IGM counterexample and the smoothing-scale dependence are not peripheral: they determine whether a detected turnover can be interpreted as an onset-of-reionization anchor. I would like the authors to address these points with quantitative parameter-space tests and a revised abstract before I can recommend acceptance. The missing SNR analysis should also be either removed from the abstract or actually included."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The Pearson drop at x_HII~1-10% is a genuinely new observable feature, and the paper does real work to back it up: two independent simulation codes, a range of astrophysical parameters, and a clear physical mechanism that connects the saturation plateau to the end of X-ray heating and the subsequent drop to the first ionized bubbles. The zero-lag Pearson coefficient is not just a rehash of the Fourier-space r_cross zero crossing; it is designed to catch the smallest bubbles, and the saturation-plateau-drop shape is a new prediction that could genuinely anchor the onset of reionization if it holds up. The paper is also honest about its limitations: it marks the L_X,SFR=39.5 exception in Fig. 2, it shows the smoothing-scale dependence in Fig. 3, and it defers the SNR study to future work.\n\nThe soft spots are real, though. The 'model-independent' framing is stronger than the evidence. The L_X,SFR=39.5 case is waved off with a citation to constraints that the authors themselves concede are 'less robustly' applicable once mini halos hosting PopIII stars are included, so the failure mode may be observationally allowed. And the smoothing-scale dependence means the plateau ends at different x_HII for different R; a detected turnover without a preceding plateau cannot uniquely anchor the onset. The abstract also mentions a preliminary SNR estimate, but the body explicitly defers detectability to future work, so that claim is premature. These are fixable: the paper should present the feature as a promising diagnostic in the standard heating-before-reionization regime, quantify the regime of validity (including the PopIII caveat), and make the SNR a forward-looking statement rather than a headline claim.\n\nWhat the paper does well matters. The physics is transparent, the estimator is easy to compute and interpret, and the independent confirmation with 21cmFASTv4 is a real plus. The citation to prior work on r_cross is fair and the authors correctly distinguish their turnover from the mid-reionization zero crossing.\n\nBottom line: this deserves a serious referee. With the abstract toned down and the caveats made central, it would be a solid methods paper. I'd take it to reading group and cite it, with a caveat in the citing sentence about the regime of validity.","headline":"A genuinely new estimator for the EoR onset, demonstrated in simulations, but the 'model-independent' claim overreaches the paper's own caveats.","tokens_in":13366,"tokens_out":2251,"would_cite":true,"duration_ms":27079,"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 Pearson cross-correlation between 21-cm and line-intensity maps is predicted to drop sharply from saturation at x_HII=1–10%, providing a model-robust anchor for the onset of reionization.","keywords":["epoch of reionization","21-cm cosmology","line intensity mapping","Pearson cross-correlation","reionization onset","X-ray heating","intergalactic medium","star formation tracers"],"falsifier":"A dedicated measurement of P(z) across redshift using co-located 21-cm and line-intensity maps: if the drop occurs when independent probes (e.g., Ly-alpha damping wings or CMB optical depth) estimate the mean ionized fraction well above 10%, the claimed anchor is wrong; conversely, if no saturated plateau and drop appear by x_HII ~10%, the mechanism fails.","tokens_in":12447,"feed_emoji":"📡","tokens_out":6444,"duration_ms":72824,"temperature":0.7,"pith_summary":"This paper proposes that the Pearson coefficient between 21-cm brightness temperature and star-formation line-intensity maps shows a sharp, model-robust drop from saturation when the universe is only 1–10% ionized. This drop marks the appearance of the first ionized bubbles, offering the first direct way to pinpoint the onset of the epoch of reionization—something current probes cannot do. The authors demonstrate the effect across a range of astrophysical parameters and independent simulations, and estimate it is detectable with next-generation 21-cm and line-intensity surveys. If correct, this gives cosmologists a new boundary condition for the last major phase transition of hydrogen.","feed_headline":"Pearson drop signals reionization's onset at 1–10% ionization","feed_subtitle":"A new cross-correlation statistic offers the first direct way to anchor the start of reionization, even without precise ionized-fraction mea","key_machinery":"The Pearson cross-correlation coefficient P(z) between 21-cm brightness temperature and line-intensity maps, evaluated at zero lag. This real-space, one-point statistic is the central object: it rises from negative to positive as X-ray heating completes, saturates, and then drops sharply when the first ionized bubbles appear, separating the onset of reionization from its later evolution. The saturation plateau and subsequent drop encode the transition from an X-ray-heating-dominated correlation to a reionization-dominated anticorrelation.","core_discovery":"The central claim is that the zero-lag Pearson coefficient P(z) between 21-cm and line-intensity tracer maps (e.g., OIII, CO, CII) saturates at a positive value once X-ray heating has driven the 21-cm field into emission everywhere, and then drops rapidly when the first ionized bubbles form, at mean ionized fractions x_HII ≈ 1–10%. Because the drop is driven specifically by the anticorrelation between star-forming regions and neutral hydrogen in ionized bubbles, it cleanly separates the onset of reionization from the preceding epoch of X-ray heating. The authors show that across variations in star-formation efficiency, X-ray luminosity, and escape fraction, the turnover persistently occurs a","pith_inferences":["The same estimator could be applied to lower redshifts to trace the end of X-ray heating, since the transition from negative to positive P(z) may itself carry a sharp feature, effectively providing a second boundary condition for the cosmic dawn.","The shape and width of the drop may encode the size distribution of the first ionized bubbles, since changing the smoothing scale shifts the onset redshift; a full two-point analysis (which the paper defers) could extract this distribution.","If the drop is observed at a redshift significantly different from model predictions, it would indicate that X-ray heating or star formation at high redshift differs from standard assumptions—or that the cold-IGM regime, where the signature fails, is realized.","The method could be extended to other line tracers (e.g., Lyman-alpha) or higher-order statistics, providing cross-checks and improved sensitivity for upcoming surveys."],"forward_implications":["Detection of the predicted drop would fix the redshift (or ionized fraction) at which reionization begins, even without a precise measurement of x_HII.","The turnover provides a new boundary condition for reionization models, complementing constraints from CMB optical depth, quasar absorption spectra, and high-redshift galaxy surveys.","The location of the drop is robust to variations in star-formation efficiency, X-ray luminosity, and escape fraction in standard scenarios, making it a reliable probe of the EoR onset.","The feature is within reach of next-generation 21-cm interferometers and line-intensity mapping surveys, enabling the first direct observation of reionization's earliest stage.","Because the drop occurs earlier than the zero-crossing of the cross-correlation (which marks roughly 30–40% ionization), it extends the observable timeline of the EoR to its very first moments.","The redshift of the drop can be combined with the later zero-crossing to give a two-point timeline of reionization: onset and midpoint."],"fun_headline_variants":["Pearson drop reveals reionization's early onset","New signal pinpoints start of reionization","Cross-correlation marks reionization's first 10%","Sharp drop in cross-correlation signals early reionization"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The signature requires that X-ray heating finishes before reionization reaches about 10% ionization; the paper itself identifies low X-ray luminosity (L_X,SFR=39.5) as a regime where reionization proceeds in a still-cold IGM and the claimed drop vanishes.","fun_headline_variants_meta":{"raw":{"variants":["Pearson drop reveals reionization's early onset","New signal pinpoints start of reionization","Cross-correlation marks reionization's first 10%","Sharp drop in cross-correlation signals early reionization"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000195,"raw_usage":{"total_tokens":1254,"prompt_tokens":864,"completion_tokens":390,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":608,"completion_tokens_details":{"reasoning_tokens":325}},"tokens_in":608,"tokens_out":390,"duration_ms":4489,"temperature":1.0,"reasoning_tokens":325,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T20:03:01.718176+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A dedicated measurement of P(z) across redshift using co-located 21-cm and line-intensity maps: if the drop occurs when independent probes (e.g., Ly-alpha damping wings or CMB optical depth) estimate the mean ionized fraction well above 10%, the claimed anchor is wrong; conversely, if no saturated plateau and drop appear by x_HII ~10%, the mechanism fails.","supporting_citations":[],"review_version":1}