{"id":"b952c0ca-16af-4864-827e-4f46fbd6c24f","arxiv_id":"2506.14588","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Rescaling an existing N-body halo catalog can produce 21 cm intensity mapping mocks for WMAP7, Planck18 and CPL cosmologies with 5 to 10 percent agreement in power spectra.","lead":"This paper applies a cosmological remapping technique to turn one supercomputer simulation into halo catalogs for several different cosmologies, then uses them to create mock 21 cm radio maps. The approach matches full simulations to within 5 to 10 percent in clustering statistics, offering a cheaper way to generate test maps for radio surveys like BINGO.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 5-10% 21 cm C_l agreement is demonstrated only for WMAP7; for CPL and Planck18 the claimed match in Section 4.2 is not backed by target C_l curves, leaving the cosmology-independent HOD assumption untested.","rationale":"The reader's weakest assumption correctly identifies the cosmology independence of the HOD relation in Eq. (2.13), but the more direct and load-bearing issue is the absence of target C_l comparisons for CPL and Planck18. The paper's Section 4.2 explicitly claims the same 5-10% residuals for those cosmologies, yet the provided Figure 6 shows only remapped curves, so the claim is not evidenced. This is not a fatal flaw: the WMAP7 comparison validates the full remapping-plus-HOD pipeline for one cosmology, and the CPL halo P(k) comparison gives partial support for the remapping itself. However, the leap to CPL and Planck18 21cm C_l accuracy rests entirely on the unverified HOD transferability, because the HI abundance and clustering depend on the rescaled halo masses and on the HOD parameters. The proposed test would settle the concern directly by applying the identical HOD to the target MiraTitan halo catalog and comparing C_l. This supports the reader's CONDITIONAL verdict without moving it, since the core methodology is credible but the presentation overstates completeness. The Planck18 remapping also uses s=0.8616, outside the Appendix A optimal range (s deviations less than 10%), which further motivates the need for a direct map-level baseline rather than relying on Figure 4's matter P(k) comparison alone.","tokens_in":19097,"tokens_out":7120,"duration_ms":74470,"concrete_test":"Run the same HOD pipeline on the MiraTitan M003 halo catalog and compute C_l at the same frequencies as Figure 6 (980 and 1010 MHz); overlay the target C_l and residual panels against the remapped CPL curves. If the residuals exceed 10%, the claim for CPL fails. Repeat for Planck18 using an appropriate target N-body simulation or an analytic HOD prediction built from the theoretical halo mass function. Additionally, compare the remapped and target halo mass functions to rule out mass-scale-dependent HOD errors that P(k) alone cannot reveal.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that remapped 21 cm C_l match N-body mocks to 5-10% is directly supported only for WMAP7, where Figure 5 compares remapped WMAP7 C_l to OuterRim C_l. For CPL and Planck18, Section 4.2 states 'The same results are found' and Figure 6 shows only the remapped C_l, with no target curves or residual panels. The C_l comparison is the only test that exercises the full chain: remapped halo masses and positions plus the HOD of Eq. (2.13). The halo power spectrum comparisons in Figures 2-4 validate clustering of halos above the catalog threshold, but they do not validate the HI mass assignment, which depends on the rescaled masses M' = s_m M and on the five-parameter HOD. If the HOD is not transferable to CPL or Planck18, or if the remapped halo mass function differs from the target in a way that P(k) comparisons do not reveal, the claimed 5-10% C_l accuracy for those cosmologies is unsupported. Because the paper presents these mocks as suitable for pipeline testing and parameter forecasts, an unvalidated cosmology-dependent HOD bias would propagate directly into the science products.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper applies the cosmological remapping technique of Angulo & White and Mead & Peacock to the Horizon Run 4 WMAP5 halo catalog, generating halo catalogs for WMAP7, Planck18, and a CPL dynamical dark-energy model. It validates the remapped halo and matter power spectra against the OuterRim (WMAP7) and MiraTitan M003 (CPL) simulations and against CLASS theory, reporting roughly 5% agreement in halo P(k) for k greater than about 1e-2 h/Mpc and matter P(k) residuals below 10%. The halos are then populated with HI using a five-parameter HOD from [39], light cones are constructed, and 21 cm brightness-temperature angular power spectra are computed. For WMAP7 the C_l are compared with mocks built from OuterRim, with residuals stated as 5-10%; for CPL and Planck18 the C_l are shown only for the remapped mocks, with no target comparison. Appendix A quantifies remapping accuracy as a function of the rescaling parameters s and s_m and gives empirical thresholds for reliable remapping.","tokens_in":19355,"tokens_out":9018,"duration_ms":88938,"significance":"If the cross-cosmology validation is completed, the paper offers a practical, computationally inexpensive route to 21 cm intensity mapping mocks for a range of cosmologies, which is directly relevant to BINGO and similar experiments. The use of independent N-body simulations for the WMAP7 and CPL halo clustering comparisons is a genuine strength, and the explicit discussion of remapping parameter limits in Appendix A is useful guidance for practitioners. The central end-to-end C_l validation, however, is currently demonstrated only for WMAP7, and the cross-cosmology claim rests on an untested cosmology-independent HOD assumption. With those gaps closed, the method could be a valuable community tool for pipeline testing and forecasts.","major_comments":[{"comment":"The statement that 'the same results are found' for CPL and Planck18 is not supported by the figure: Figure 6 shows only the remapped C_l in four frequency bins, with no target C_l curves and no residual panels. The angular power spectrum is the only quantity that exercises the full chain from remapped halo masses through the HOD to brightness temperature, so the claimed 5-10% C_l agreement for CPL and Planck18 is currently unverified. Please add, for CPL, the corresponding C_l from MiraTitan M003 halo catalogs, and for Planck18 either an N-body-based target if one is available or an explicit statement that no end-to-end C_l validation is performed, with residual panels in both cases.","section":"§4.2, Figure 6"},{"comment":"The HI HOD of Eq. (2.13) is an empirical five-parameter relation calibrated to a single simulation, and the text asserts that it is 'cosmology-independent' without providing a test. For WMAP7, the only C_l comparison, the target simulation is close to the calibration cosmology of the HOD, so that comparison does not test transferability. Since the remapped halo masses M' = s_m M feed directly into Eq. (2.13), any cosmology dependence of the HI-halo mass relation would bias the CPL and Planck18 mocks in a way that the halo and matter power-spectrum checks do not reveal. Please either validate the HOD in at least one additional cosmology (for example with a hydrodynamic simulation) or explicitly quantify this as an unvalidated assumption in the error budget of the mocks.","section":"§2.2, Eq. (2.13)"},{"comment":"For Planck18, Table 3 gives s=0.8616 (13.8% deviation from unity) and s_m=0.7601 (24.0% deviation), while Appendix A.1 defines the accurate-remapping range as s deviations below 10% and s_m deviations below 25%. The Planck18 remap is therefore outside the stated optimal range in s, yet Section 4.1 reports matter power-spectrum residuals below 10%, and the conclusions extend 5-10% accuracy to the Planck18 mocks. The paper should reconcile this, for example by explaining that the Appendix thresholds are conservative or by applying the halo mass and box size thresholds mentioned in Appendix A and reporting the resulting residuals for Planck18.","section":"§4.1 and Appendix A"}],"minor_comments":[{"comment":"The exponential term in Eq. (2.13) is missing its parentheses and minus sign; it should read c exp[-(log10 M - d)^2 / f^2].","section":"§2.2, Eq. (2.13)"},{"comment":"In the bullet list at the end of Section 2.2, the brightness-temperature rescaling step refers to 'Equation (2.14)', but the rescaling is defined in Eq. (2.15).","section":"§2.2, step list"},{"comment":"Reference [33] is cited for the statement that more than 95% of HI lies within the virial radius of dark matter halos in IllustrisTNG, but the reference listed is the COLOSSUS toolkit paper; this citation should be corrected to the appropriate IllustrisTNG HI study.","section":"References"},{"comment":"Tables 4 and 5 list inconsistent values for the WMAP7-like 2 model (s=1.075 versus 1.0659 and s_m=1.271 versus 1.2389); please harmonize the two tables.","section":"Tables 4 and 5"},{"comment":"The abstract and conclusions promise an analysis of halo mass and box size thresholds, but Appendix A only gives thresholds on the rescaling parameters s and s_m; no quantitative halo mass or box size threshold is provided or applied in the main results.","section":"Abstract and Appendix A"},{"comment":"The phrase 'the first time' in Section 5 sits awkwardly with the authors' previous remapped 21 cm mock paper [48] and with the BINGO HOD mocks of [39]; please rephrase to state precisely what is new, such as the first multi-cosmology remapped C_l comparison.","section":"§5, Conclusions"}],"recommendation":"major_revision","confidential_remarks":"The central issue is not the remapping methodology itself, which is sound and competently applied for WMAP7, but the breadth of the validation claims. I would be willing to accept once the C_l comparison for CPL is provided and Planck18 is either validated against an N-body simulation or clearly caveated as unvalidated at the C_l level. The cosmology independence of the HOD is the most important scientific risk and deserves an explicit paragraph in the revised version."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this as a methods paper, not a physics result. What is actually new: they take the Angulo-White/Mead-Peacock remapping machinery, apply it to Horizon Run 4 WMAP5 halo catalogs to produce WMAP7, Planck18, and CPL halo catalogs and 21-cm mocks, push the frequency range beyond their earlier WMAP7 paper, and add a systematic appendix on remapping precision. The halo P(k) validation is the most solid part: remapped WMAP7 and CPL match OuterRim and MiraTitan M003 to about 5% for k ≳ 10^-2 h/Mpc, and the debiased matter P(k) for all three, including Planck18, stays within 10% of CLASS. Those are independent comparisons, not parameters fit to the target simulations. The circularity burden is low.\n\nThe soft spot is exactly where the stress test points. Figure 5 shows WMAP7 21-cm C_l with the OuterRim target and residuals in the 5-10% range. Section 4.2 then says \"the same results are found\" for CPL and Planck18, but Figure 6 shows only the remapped curves, no target, no residual panels. For Planck18 the paper already says no suitable N-body simulation was found, so only theory comparison is possible; that is fine for P(k), but they do not show a theoretical 21-cm C_l either. For CPL they did have MiraTitan M003 and used it for halo P(k), but they do not show its 21-cm C_l as a target. So the map-level 5-10% claim is supported for exactly one cosmology. And the map-level chain depends on the HOD from Zhang et al., applied as a cosmology-independent relation; the P(k) comparisons above the halo threshold do not test the HI mass assignment. Since only WMAP7 is checked end-to-end, the transfer to Planck18 and CPL is an untested assumption. That is a genuine, load-bearing gap in the headline product, not a nitpick.\n\nSmaller issues: the speed advantage is asserted but never quantified; \"first time\" in the conclusions overstates things given their own earlier WMAP7 mock paper; and the Appendix A thresholds are calibrated on matter P(k), so they do not by themselves validate the 21-cm maps. None of these sink the paper. It is a practical, honest piece of work that the 21 cm community can use, and it deserves a real referee. The revision should either add the missing target/residual curves for CPL and Planck18 or explicitly restrict the map-level claim to WMAP7. I would not desk reject.","headline":"A practical remapping paper with solid halo-level validation, but the 21 cm C_l agreement for CPL and Planck18 is asserted in the text without target curves, so the map-level claim currently rests on WMAP7 alone.","tokens_in":19970,"tokens_out":4115,"would_cite":true,"duration_ms":40614,"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":"Cosmological remapping can turn one halo catalog into accurate 21 cm intensity-mapping mocks for WMAP7, Planck18, and CPL cosmologies, with power spectra matching dedicated simulations to within 5–10 percent.","keywords":["cosmological remapping","rescaling","halo catalogs","21 cm intensity mapping","halo occupation distribution","neutral hydrogen","brightness temperature","mock generation"],"falsifier":"Run a dedicated N-body simulation with Planck18 or CPL parameters at comparable volume and resolution, build its 21 cm mocks with the same HOD pipeline, and compare $C_\\ell$ to the remapped Planck18 or CPL mocks; exceeding the 5–10 percent band beyond cosmic variance would falsify the accuracy claim. A second check is to measure the $M_{\\mathrm{HI}}$–$M_h$ relation in two different cosmologies; a shift outside the five-parameter fit of Eq. (2.13) would falsify the cosmology-independence assumption.","tokens_in":18909,"feed_emoji":"📡","tokens_out":6125,"duration_ms":57521,"temperature":0.7,"pith_summary":"The paper argues that cosmological remapping, which rescales the lengths, masses, and redshifts of an existing N-body simulation's halo catalog, can generate halo catalogs for target cosmologies without running new full simulations, and that these catalogs are accurate enough to build 21 cm intensity-mapping mocks. Starting from the Horizon Run 4 WMAP5 simulation, it produces halo catalogs for WMAP7, Planck18, and CPL cosmologies. The remapped halo power spectra match dedicated N-body simulations within about 5 percent for $k\\gtrsim10^{-2}\\,h/\\mathrm{Mpc}$, and the derived 21 cm brightness-temperature angular power spectra match within 5–10 percent. A reader should care because this offers a cheap route to the many mocks needed for survey design, pipeline testing, and cosmological forecasts for experiments like BINGO.","feed_headline":"One simulation, rescaled, yields 21 cm mocks for three cosmologies","feed_subtitle":"Remapped halo catalogs match full simulations to 5–10 percent, enough for BINGO-era forecasts.","key_machinery":"The carrying object is the remapping pair of rescaling factors, $s$ (lengths and positions) and $s_m$ (masses), determined by minimizing the linear-variance mismatch $\\delta^2_{\\mathrm{rms}}$ of Eq. (2.5) between original and target cosmologies. Position shifts are applied through a Zel'dovich displacement field computed from the halo over-density and a bias factor, with a Gaussian filter at the nonlinear scale controlling variance. The second ingredient is the HI halo occupation distribution of Eq. (2.13), a five-parameter relation assigning $M_{\\mathrm{HI}}$ to halo mass, which turns remapped halos into brightness-temperature maps via Eq. (2.14), with 100 random-observer light cones to beat cosmic variance. Appendix A supplies quantitative guardrails: accurate remapping requires $s\\lesssim1.1$ and $s_m\\lesssim1.25$; larger $s_m$ is the stronger warning sign of failure.","core_discovery":"The central claim is that remapping is not just a shortcut for power spectra but carries through to the observable 21 cm signal. Rescaling lengths by $s$ and masses by $s_m$, shifting halo positions with a displacement field derived from the Zel'dovich approximation, and then populating the rescaled halos with neutral hydrogen through the five-parameter HOD relation of Eq. (2.13) reproduces the halo clustering of the target cosmology: WMAP7 and CPL residuals against OuterRim and Mira-Titan M003 remain within 5 percent for $k\\gtrsim10^{-2}\\,h/\\mathrm{Mpc}$, matter power spectra stay below 10 percent for all three targets including Planck18, and 21 cm $C_\\ell$ mocks come within 5–10 percent. The paper is explicit that only WMAP7 receives a direct map-level comparison against a target simulation; Planck18 is checked against the theoretical power spectrum, and the HOD relation itself is assumed cosmology-independent.","pith_inferences":["If the HOD relation is genuinely cosmology-independent, the pipeline extends to arbitrary target cosmologies inside the $s$/$s_m$ guardrails, including dark-energy models beyond CPL, without any new calibration.","The missing direct map-level benchmark for Planck18 and CPL leaves a gap: their claimed 5–10 percent $C_\\ell$ accuracy rests on halo power-spectrum agreement plus the WMAP7 map-level test, and a cosmology-dependent HI-mass mapping would evade all current checks.","A testable extension is to compare remapped Planck18 and CPL mocks against a dedicated simulation's HI maps, or to measure the $M_{\\mathrm{HI}}$–$M_h$ relation in two cosmologies; a shift outside the five-parameter fit would break the cosmology-independence assumption.","The quantitative $s$/$s_m$ thresholds suggest a practical pre-flight check: before remapping to any new cosmology, compute the two factors and reject targets where $s_m$ deviates by more than roughly 25 percent, rather than running the full pipeline."],"forward_implications":["One existing large N-body run, here Horizon Run 4, can serve as the seed for halo catalogs in many target cosmologies, so parameter surveys no longer need one full simulation per cosmology.","The 5–10 percent agreement of the 21 cm $C_\\ell$ means remapped mocks are usable for BINGO-band analyses and similar single-dish intensity-mapping surveys, within the stated scale range.","Remapping accuracy has predictable limits: transitions with $s$ deviations within about 10 percent and $s_m$ deviations within about 25 percent stay reliable, while larger $s_m$ predicts residuals above 50 percent, so the rescaling factors themselves flag when the method should not be trusted.","The same remapped halo catalogs can feed downstream uses beyond 21 cm, such as galaxy formation modeling and galaxy-survey mocks, because the method preserves large-scale clustering."],"supporting_citations":[{"why":"Introduces the original cosmology-rescaling concept that remapping builds on.","marker":"[40]"},{"why":"Provides the extended, self-contained remapping procedure for halo catalogs that this paper applies directly.","marker":"[43]"},{"why":"Supplies the five-parameter HI halo occupation distribution and the light-cone construction used to make 21 cm mocks.","marker":"[39]"},{"why":"Defines the Horizon Run 4 simulation whose WMAP5 halo catalog is the starting point for all remapping.","marker":"[57]"},{"why":"Provides the OuterRim simulation used as the WMAP7 target benchmark for halo power spectra and 21 cm maps.","marker":"[62]"},{"why":"Provides the Mira-Titan M003 simulation used as the CPL target benchmark.","marker":"[61]"},{"why":"Supplies the mass function used to compute the halo bias that connects halo density to matter density for the displacement field.","marker":"[42]"},{"why":"Defines the WMAP7 best-fit parameters that the OuterRim target simulation approximates.","marker":"[63]"}],"fun_headline_variants":["Rescale one simulation, get 21 cm mocks for multiple cosmologies","Remapping yields 21 cm mocks within 5–10%","Fast 21 cm mocks from rescaled N-body runs","One simulation rescaled: 21 cm maps for three cosmologies","Remapped halos yield 21 cm maps within 5–10% of full runs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the HOD relation of Eq. (2.13) maps halo mass to HI mass in a cosmology-independent way, and that this mapping, validated at the map level only for WMAP7, stays valid for Planck18 and CPL; if it does not, the 21 cm mocks inherit a bias that the power-spectrum comparisons do not detect.","fun_headline_variants_meta":{"raw":{"variants":["Rescale one simulation, get 21 cm mocks for multiple cosmologies","Remapping yields 21 cm mocks within 5–10%","Fast 21 cm mocks from rescaled N-body runs","One simulation rescaled: 21 cm maps for three cosmologies","Remapped halos yield 21 cm maps within 5–10% of full runs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001329,"raw_usage":{"total_tokens":5420,"prompt_tokens":971,"completion_tokens":4449,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":587,"completion_tokens_details":{"reasoning_tokens":4349}},"tokens_in":587,"tokens_out":4449,"duration_ms":31311,"temperature":1.0,"reasoning_tokens":4349,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T19:50:43.215236+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run a dedicated N-body simulation with Planck18 or CPL parameters at comparable volume and resolution, build its 21 cm mocks with the same HOD pipeline, and compare $C_\\ell$ to the remapped Planck18 or CPL mocks; exceeding the 5–10 percent band beyond cosmic variance would falsify the accuracy claim. A second check is to measure the $M_{\\mathrm{HI}}$–$M_h$ relation in two different cosmologies; a shift outside the five-parameter fit of Eq. (2.13) would falsify the cosmology-independence assumption.","supporting_citations":[{"cited_title":"Heitmann et al.,THE MIRA–TITAN UNIVERSE: PRECISION PREDICTIONS FOR DARK ENERGY SURVEYS,Astrophys","cited_arxiv_id":null,"evidence_quote":"Provides the Mira-Titan M003 simulation used as the CPL target benchmark."},{"cited_title":"Zhang and others,The BINGO project","cited_arxiv_id":null,"evidence_quote":"Supplies the five-parameter HI halo occupation distribution and the light-cone construction used to make 21 cm mocks."},{"cited_title":"Heitmann et al.,The Outer Rim Simulation: A Path to Many-core Supercomputers, Astrophys","cited_arxiv_id":null,"evidence_quote":"Provides the OuterRim simulation used as the WMAP7 target benchmark for halo power spectra and 21 cm maps."}],"review_version":1}