{"id":"d0337b97-3b88-4d2b-93e8-eacb78a196e5","arxiv_id":"1908.03732","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Adding mock SKA 21 cm BAO data to current CMB, BAO, and supernova data would substantially tighten constraints on matter density, the Hubble constant, and dark energy equation of state parameters.","lead":"This paper simulates how the future SKA radio telescope's 21 cm survey would sharpen measurements of the universe's expansion and dark energy. The authors show that adding SKA mock data to existing cosmological observations could shrink error bars on key parameters by roughly half or more.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"SKA mock likelihood uses only diagonal errors and omits the fiducial data vector, so the reported improvement percentages are not reproducible and may be optimistic.","rationale":"The reader's weakest assumption is that the Bull et al. mock errors faithfully represent future SKA constraints; that is exactly the hinge of this forecast. I narrow it to a concrete implementation point: the paper uses diagonal relative errors without a stated covariance or central values. This is not a dispute with the consensus that 21 cm BAO will be powerful; it is a reproducibility and optimism check. If the off-diagonal correlations are non-negligible, the reported improvements are upper bounds rather than forecasts. The proposed test recomputes the likelihood from the primary forecast rather than from eyeballed Fig. 3 errors. Because the qualitative conclusion is well supported by the setup and the quantitative issue is addressable, I keep the reader's CONDITIONAL verdict as unchanged; the condition is that the mock likelihood be fully specified and tested against the full covariance.","tokens_in":9310,"tokens_out":11501,"duration_ms":135380,"concrete_test":"Reconstruct the Bull et al. (2015) forecast as a multivariate Gaussian using the full Fisher matrix for each redshift bin (or recompute it from the survey specifications), including correlations between ln H and ln D_A and between bins. Re-run the CosmoMC chains for CBS+SKA2 with this covariance and an explicit fiducial data vector; compare the marginalized errors in Tables I–III. For instance, if σ(H0) in the ΛCDM CBS+SKA2 case rises from 0.17 km/s/Mpc to >0.19, or σ(w0) in the CPL model rises by more than 10%, the reported improvement percentages are not robust to the likelihood's treatment of correlations.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The forecast likelihood in Section II is built from the relative errors σH/H and σDA/DA 'directly extracted from Fig. 3' of Bull et al. (2015). This supplies only diagonal error bars. Anisotropic BAO forecasts generally yield a correlated 2x2 covariance between ln H and ln D_A within each redshift bin, and the bins may also be correlated; using the diagonal errors as independent Gaussian constraints overstates the information content of SKA and inflates the improvement percentages quoted in Section IV and Tables I–III. Independently, the paper never states the central values of the mock H and D_A data or the fiducial cosmology used to generate them, so the likelihood function is not fully specified and the MCMC results cannot be reproduced or checked. The qualitative claim that SKA improves constraints is probably robust, but the quantitative percentages are load-bearing and rest on these unstated or optimistic choices.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper forecasts the impact of future 21 cm intensity-mapping BAO measurements from the Square Kilometre Array (SKA1-MID and SKA2) and Euclid on cosmological parameter constraints. Using current CMB distance priors from Planck 2018, optical BAO data, and the Pantheon supernova compilation (abbreviated CBS), the authors run CosmoMC fits to the ΛCDM, wCDM, and CPL models, adding mock BAO likelihoods built from relative errors σH/H and σDA/DA digitized from Bull et al. (2015). They report that adding SKA2 mock data improves constraints on Ωm by 34%–70%, on H0 by 52%–73%, and on dark-energy equation-of-state parameters by roughly 50% (Tables I–III, Section IV). The central claim is that future SKA IM BAO data will significantly sharpen cosmological constraints and break parameter degeneracies, especially between Ωm and H0 and between Ωm and w.","tokens_in":9464,"tokens_out":4992,"duration_ms":52690,"significance":"The paper's strength is its systematic comparison across three dark-energy models and five data combinations using standard MCMC methods; the qualitative conclusion that low-redshift BAO data from SKA2 will substantially improve constraints on Ωm, H0, and dark-energy parameters is plausible and consistent with earlier forecasts. The quantitative improvement percentages, however, are the paper's main quantitative output, and they rest on the fidelity of the mock likelihood built from digitized diagonal errors without an explicit data vector or covariance. Because those assumptions are not fully documented, the numerical results should be treated as indicative rather than definitive; the paper would be considerably strengthened by providing the full mock data and error model. The manuscript ships no code or data, so reproducibility rests entirely on the description of the likelihood.","major_comments":[{"comment":"The likelihood is constructed from relative errors σH/H and σDA/DA 'directly extracted from Fig. 3' of Ref. [15], but the paper never specifies the central values of the mock H(z) and DA(z) data points or the fiducial cosmology used to generate them. Without this information the Gaussian likelihood is not fully defined and the MCMC results cannot be reproduced or checked. The authors should provide the full mock data vector (e.g., in a table or appendix) and state the fiducial cosmology, or explicitly present the mock constraints as relative measurements so the forecast is transparent.","section":"Section II"},{"comment":"The likelihood uses only the diagonal components of the BAO error budget. Anisotropic BAO measurements in Bull et al. (2015) generally yield a correlated 2×2 covariance between ln H and ln DA within each redshift bin, and bin-to-bin correlations can also be present. If correlations are neglected, the information content attributed to SKA is inflated, which would bias the improvement percentages in Tables I–III and Section IV upward. The authors should adopt the full covariance from the simulation or justify with a sensitivity test why the diagonal approximation is adequate.","section":"Section II"},{"comment":"The conclusion states that wa is 'promoted by 49.6%' based on relative error, yet Section III explicitly warns that the relative error for wa is unreliable because its central value is near zero and recommends using the absolute error instead. The relative improvement for wa (0.9044 vs. 1.4309, Table III) is not a meaningful metric, and using it in the summary-level claim overstates the improvement. The headline improvements should be based on the absolute error or a model-independent figure of merit.","section":"Section IV"}],"minor_comments":[{"comment":"The name 'Chevalliear-Polarski-Linder' is a typo for 'Chevallier-Polarski-Linder'.","section":"Section I"},{"comment":"The paper should specify which version of the Planck 2018 distance priors is used (Ref. [20] is Chen et al., which is fine) and whether CMB lensing is included, since these choices can affect the baseline constraints.","section":"Section II"},{"comment":"The axis label 'm' should be 'Ωm' with the subscript omega for clarity; the same issue appears in Figure 2.","section":"Figure 1"},{"comment":"The phrase 'the data of Euclid behave much better than SKA1 but worse than SKA2' is vague; it would be clearer to give the quantitative comparative precisions from the tables.","section":"Section III"},{"comment":"The paper should cite the original Planck 2018 likelihood paper (Aghanim et al. 2018) in addition to the distance-priors paper, since it is the source of the CMB data used.","section":"References"},{"comment":"The values of ε(wa) are very large and unstable because wa is near zero; consider reporting only the absolute error for wa in the abstract and conclusion, as the text itself advises.","section":"Table III"}],"recommendation":"major_revision","confidential_remarks":"The paper is competent and the qualitative conclusion is likely robust, but the quantitative claims are not reproducible as written because the mock data vector and covariance are not specified. The diagonal-only approximation and the use of relative error for wa also tend to inflate the headline improvements. If the authors provide the full mock data (or clearly state that they use fiducial-model predictions as the data vector) and address the covariance issue, the paper would be suitable for publication. The improvement over current constraints is a forecast and not circular, but the assumptions should be stated transparently."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things before reading this one. First, the qualitative result is solid: adding SKA 21 cm BAO mock data to CMB, optical BAO, and SN breaks degeneracies and sharpens constraints on Omega_m, H0, and dark-energy EoS parameters. Second, the quantitative improvement percentages are less trustworthy than the text implies, because the mock likelihood is not fully specified and uses only diagonal errors. The stress-test note holds up on closer reading. Section II says the relative errors sigma_H/H and sigma_DA/DA are 'directly extracted from Fig. 3' of Bull et al., but the central values of the mock H and D_A points are never given, and neither is the fiducial cosmology. Without those, the Gaussian likelihood is not reproducible. The paper also treats the H and D_A measurements in each redshift bin as independent diagonal constraints. Anisotropic BAO forecasts generally produce a correlated 2x2 covariance between ln H and ln D_A per bin, and ignoring those correlations overstates the information content. That inflates the improvement fractions in Section IV and Tables I–III. What is actually new: the paper applies the Bull et al. mocks (and Euclid mocks) to three dark-energy models—LambdaCDM, wCDM, CPL—with Planck 2018 distance priors, Pantheon, and current BAO, all in one analysis. It is a routine extension of earlier work, including the same group's Ref. [19], so the novelty is low. But the execution is honest: the MCMC runs are standard CosmoMC, the tables are clear, and the authors do not oversell SKA1 over Euclid. They correctly note that SKA2 is the competitive case. The qualitative conclusion that SKA2 improves constraints is robust and consistent with prior forecasts. The weak spot is the mock likelihood: missing central values, no fiducial cosmology stated, diagonal errors only. This matters because the exact percentages are load-bearing for the survey-planning audience. The paper would be fixable: state the fiducial, provide the data vector and covariance, or at least release the chains. Without that, the numbers are a useful estimate but not a checkable result. The citation pattern is fine; the reliance on Bull et al. is legitimate because that is where the mocks come from. Who gets value: people doing survey forecasting for SKA and Euclid will read the tables and compare with their own projections. Dark-energy phenomenologists will see the degeneracy-breaking plots and take the qualitative message. It deserves a serious referee: a good referee can demand the missing likelihood details and the numbers would then be much more valuable. I would not desk-reject it. My recommendation: send it to review, with the understanding that the reproducibility issue is mandatory to address.","headline":"A competent but modest forecast paper: SKA BAO mocks improve dark-energy parameter constraints, yet the headline percentages rest on an underspecified and possibly optimistic mock likelihood.","tokens_in":9978,"tokens_out":690,"would_cite":false,"duration_ms":10013,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["98.80.-k","95.36.+x","98.70.Dk"],"model":"deepseek-v4-flash","headline":"Adding SKA 21 cm data sharply tightens cosmological constraints, the paper argues.","keywords":["21 cm intensity mapping","SKA","Baryon acoustic oscillations","Cosmological parameter estimation","Dark energy equation of state","CPL parameterization","CosmoMC","Forecast"],"falsifier":"When the real SKA2 data arrive, one could compare the actual measured BAO correlation-function errors in each redshift bin against the assumed σH/H and σDA/DA values; if the real errors are larger or correlated, the predicted improvement fractions of 34–70% on Ωm and 52–73% on H0 would shrink accordingly.","tokens_in":9115,"feed_emoji":"📡","tokens_out":1809,"duration_ms":19526,"temperature":0.7,"pith_summary":"This paper argues that mock 21 cm baryon acoustic oscillation (BAO) measurements from the planned Square Kilometre Array (SKA), especially its second phase (SKA2), will significantly improve constraints on the matter density, the Hubble constant, and the dark energy equation-of-state parameters when combined with current CMB, optical BAO, and supernova data. Using simulated SKA errors and a Markov chain Monte Carlo fit to three dark energy models, the authors find that adding SKA2 data to the current CMB+BAO+SN ('CBS') dataset improves the constraint on the matter density by 34–70% and on the Hubble constant by 52–73%, and also sharpens constraints on the dark energy equation-of-state parameters w, w0, and wa. The main reason is that the 21 cm BAO measurements break parameter degeneracies, particularly between the matter density and the Hubble constant.","feed_headline":"SKA 21 cm data could cut key cosmic errors by half or more","feed_subtitle":"New mock BAO forecasts show sharp gains for dark energy and Hubble constant constraints.","key_machinery":"The central object is a mock BAO dataset constructed from the relative errors in the Hubble expansion rate σH/H and the angular diameter distance σDA/DA extracted from Fig. 3 of Bull et al. (2015), applied to SKA1-MID Band 1, Band 2, and SKA2. These errors are used to build a Gaussian likelihood for the distances, which is combined with the Planck 2018 CMB distance priors, optical BAO measurements (6dFGS, SDSS-MGS, BOSS DR12), and the Pantheon supernova compilation. The fit is performed with the CosmoMC Markov chain Monte Carlo package, and the constraint improvement is quantified by comparing 1σ errors and relative precisions ε(ξ) = σ(ξ)/ξbf across data combinations.","core_discovery":"The central claim is that future SKA 21 cm intensity-mapping BAO mock data, when added to the current CMB+optical BAO+supernova dataset ('CBS'), dramatically improve posterior constraints on the standard cosmological parameters and on dark energy equation-of-state parameters. For the ΛCDM model, adding SKA2 data improves the constraint precision on Ωm from 2.65% to 1.00% and on H0 from 0.93% to 0.25%. For the wCDM model, the dark energy equation-of-state parameter w improves from 5.18% to 2.33%. For the CPL model, w0 improves from 9.66% to 5.16% and the absolute error on wa shrinks from 0.3646 to 0.1836. The paper also finds that SKA2 data break degeneracies between Ωm and H0 in ΛCDM and between Ωm and w in wCDM, and that SKA2 outperforms both SKA1-MID and the future Euclid optical survey in constraining power.","pith_inferences":["The paper does not attempt to forecast the SKA constraints on the neutrino mass sum or on curvature, but the same mock BAO data would likely also sharpen those parameters; a natural extension is to add Σmν and Ωk to the parameter space and rerun the forecast.","The derived improvements assume the mock central values are fixed to a fiducial cosmology, but a more realistic forecast would marginalize over the fiducial model or use a data-driven mean; this could widen the reported errors.","The relative error extraction from Bull et al. (2015) ignores possible redshift-space distortion and foreground contamination effects beyond those already in the simulation, so the real improvement fractions may be lower if systematics correlate between redshift bins.","A testable next step would be to apply the same likelihood to real data from the upcoming HIRAX, CHIME, or Tianlai pathfinder surveys, checking whether the measured BAO scales actually reach the assumed precision."],"forward_implications":["If the SKA2 mock data reflect real future measurements, then a single radio survey could cut the uncertainty on the Hubble constant to roughly a quarter of its current CMB+BAO+SN level, reaching sub-percent precision in ΛCDM.","Combining SKA2 with Euclid would further tighten the dark energy equation-of-state constraints, suggesting that joint radio and optical BAO surveys will be a powerful way to test whether dark energy evolves with redshift.","The demonstrated degeneracy-breaking between Ωm and H0 means SKA BAO data could help arbitrate the current tension between early- and late-universe determinations of H0, if systematics are controlled.","Because the improvement is largest in the more flexible dark energy models (wCDM and CPL), the 21 cm probe is especially suited for testing deviations from a cosmological constant.","The comparison with Euclid indicates that while SKA1 may be less competitive than Euclid, SKA2 would become the most constraining BAO experiment of its era."],"supporting_citations":[{"why":"Provides the simulated SKA1 and SKA2 BAO relative errors σH/H and σDA/DA from which the mock likelihood is constructed.","marker":"[15]"},{"why":"Source of the Planck 2018 CMB distance priors used in the CBS baseline dataset.","marker":"[1]"},{"why":"Source of the Pantheon Type Ia supernova compilation with 1048 data points used in the CBS baseline.","marker":"[22]"},{"why":"Source of the BOSS DR12 BAO measurements at effective redshifts 0.38, 0.51, and 0.61.","marker":"[21]"},{"why":"Source of the SDSS-MGS BAO measurement at effective redshift 0.15.","marker":"[9]"},{"why":"Source of the 6dFGS BAO measurement at effective redshift 0.106.","marker":"[10]"},{"why":"The CosmoMC package used for Markov chain Monte Carlo parameter estimation.","marker":"[23]"},{"why":"Previous work by two of the same authors using the same SKA mock data to forecast neutrino mass constraints, which the paper cites for consistency on degeneracy breaking.","marker":"[19]"},{"why":"Provides the Planck 2018 distance priors in a form the authors adopt for the analysis.","marker":"[20]"},{"why":"Describes the SKA project configuration (SKA1-MID, SKA1-LOW, SKA2) used to define the survey setups.","marker":"[14]"}],"fun_headline_variants":["SKA2 mock data breaks cosmic degeneracies, sharpens dark energy","SKA2 IM BAO forecasts slash errors on H0 and dark energy","SKA 21 cm mock data doubles dark energy and Hubble precision","SKA2 mock data to improve dark energy equation-of-state"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The forecast assumes that the relative errors on the Hubble expansion rate and angular diameter distance taken from the Bull et al. (2015) simulation are accurate predictions for the real SKA measurements, with no unaccounted systematic errors or correlations between redshift bins.","fun_headline_variants_meta":{"raw":{"variants":["SKA2 mock data breaks cosmic degeneracies, sharpens dark energy","SKA2 IM BAO forecasts slash errors on H0 and dark energy","SKA 21 cm mock data doubles dark energy and Hubble precision","SKA2 mock data to improve dark energy equation-of-state"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001392,"raw_usage":{"total_tokens":5629,"prompt_tokens":941,"completion_tokens":4688,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":557,"completion_tokens_details":{"reasoning_tokens":4611}},"tokens_in":557,"tokens_out":4688,"duration_ms":33383,"temperature":1.0,"reasoning_tokens":4611,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:03:18.035881+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"When the real SKA2 data arrive, one could compare the actual measured BAO correlation-function errors in each redshift bin against the assumed σH/H and σDA/DA values; if the real errors are larger or correlated, the predicted improvement fractions of 34–70% on Ωm and 52–73% on H0 would shrink accordingly.","supporting_citations":[],"review_version":1}