{"id":"c56dc862-05b2-4769-a917-fb754d9221ec","arxiv_id":"2601.07924","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Sahyadri is a new N-body simulation suite with 25x better halo mass resolution than AbacusSummit and seed-matched parameter variations, showing coherent Omega_m signals in power spectra, mass functions, and cosmic-web statistics.","lead":"This paper presents Sahyadri, a new set of dark-matter-only computer simulations with roughly 25 times finer mass resolution than existing parameter-varying suites, resolving halos down to about 3.2 billion solar masses. The authors claim this enables modeling about 1.6 million DESI BGS galaxies at z<0.15 that earlier suites cannot resolve, and demonstrate sensitivity of cosmic-web statistics to the matter density Omega_m.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Seed-matched derivative claim for halo statistics is self-undermined; relies on unpublished stabilization method","rationale":"The numerical specifications check out: 2048^3 particles in 200 h^-1 Mpc yields m_p = 8.1e7 h^-1 M_sun, M_min = 3.2e9 h^-1 M_sun at 40 particles, and a factor ~25 improvement over AbacusSummit. The BGS 40% estimate is indeed dependent on an unpublished galaxy-halo connection and is a legitimate concern, but it is an application rather than a load-bearing pillar of the simulation resource. The more serious gap is in the paper's central methodological claim. Section 2.1.1 asserts that seed-matched ICs enable reliable derivatives of small-scale observables while simultaneously conceding that seed matching does not control small-scale phase-mixing stochasticity for halo statistics, deferring to a companion paper [24]. Since the paper's own showcase statistics (halo power spectrum, VVF, kNN) are halo-based small-scale statistics, and no derivative or validation of [24] is presented, the 'enabling derivatives' headline is unsupported. This warrants a conditional verdict: the suite may be excellent, but the derivative capability must be demonstrated or explicitly scoped. The parameter coverage overstatement (abstract says six, §4 says three completed) is a related but more cosmetic issue that can be fixed by rewording.","tokens_in":22284,"tokens_out":13166,"duration_ms":133554,"concrete_test":"Compute the Ωm derivative of the VVF (or kNN CDF) from the matched-seed fiducial and Ωm± runs, and compare with the derivative from two additional independent-seed realizations of the same cosmologies. If the matched-seed derivative lies within the seed-scatter of the independent-seed estimates, the claim is validated; if not, the paper must either include the stabilization method from [24] or qualify the derivative claim to matter statistics only.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In §2.1.1 the paper claims seed-matched initial conditions ensure differences arise solely from cosmological parameter changes and 'allows Sahyadri to be used reliably for estimating derivatives of small-scale observables.' The same paragraph concedes that seed matching is 'not sufficient to control the gravitationally induced stochasticity due to phase-mixing at small scales, which affects statistical probes involving halos', deferring to a post-processing optimization in the companion paper [24]. All the novel showcase statistics—halo power spectrum (Fig. 8), Voronoi volume function (Fig. 10), and kNN-CDF (Fig. 11)—are small-scale halo statistics, yet no derivative is actually computed; only raw Ωm responses with illustrative jackknife errors are shown. Thus the central capability claim ('enabling cosmological parameter derivatives') is not demonstrated in this paper and depends on an external, unpublished method. If the [24] stabilization fails at Sahyadri resolution/volume, the matched-seed design does not deliver the advertised derivative precision for the very statistics the suite is built to enable.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents Sahyadri, a suite of dark-matter-only N-body simulations with 2048^3 particles in 200 h^-1 Mpc boxes, reaching a particle mass of 8.1e7 h^-1 M_sun and halo masses down to 3.2e9 h^-1 M_sun with 40 particles. The suite is designed around Planck 2018 cosmology with parameter variations, seed-matched initial conditions, and a claimed capability for cosmological parameter derivatives. The paper showcases results for the matter power spectrum, halo mass function, halo power spectrum, Voronoi volume function, kNN statistics, and environment-dependent halo correlations, and it argues that the resolution opens up access to low-mass halos relevant to DESI BGS at z<0.15. Data products are advertised as publicly available with a custom compression scheme.","tokens_in":22585,"tokens_out":5166,"duration_ms":53443,"significance":"If the numerical specifications hold, Sahyadri is a genuinely useful resource: its particle mass is roughly a factor 25 lower than AbacusSummit and more than two orders of magnitude below Quijote/Aemulus, while retaining a non-trivial simulation volume. The paper ships a public pipeline, includes consistency checks against halofit and Tinker, and demonstrates that high-density tracer samples enable measurements of beyond-2-point statistics such as the VVF and kNN-CDFs. These are concrete strengths. However, the two headline claims that would elevate the paper's significance—(i) reliable cosmological parameter derivatives for small-scale halo statistics and (ii) access to ~40% of DESI BGS galaxies—are not demonstrated in the manuscript: the first relies on an unpublished companion paper, and the second relies on an unpublished methodology. The central simulation product itself appears sound, but the advertised capabilities are currently only partly substantiated.","major_comments":[{"comment":"The abstract and §2.1.1 state that seed-matched initial conditions enable reliable estimation of derivatives of small-scale observables involving halos, but the same paragraph concedes that seed matching is not sufficient to control gravitationally induced stochasticity for halo statistics and defers to the unpublished post-processing method of [24]. In the results sections, no derivative is computed for any statistic; Figures 8, 10, and 11 show raw Ωm responses with jackknife errors described as 'for illustrative purpose only.' The central advertised capability is therefore not demonstrated in this paper and depends entirely on an external, unpublished method. Please either include derivative estimates validated at Sahyadri's resolution/volume, or explicitly recast the derivative claim as a design goal to be established in a companion paper.","section":"§2.1.1, §3.2"},{"comment":"The abstract and storage estimates describe the suite as containing 'systematic variations of six cosmological parameters,' and the 117 TB footprint is computed for the full six-parameter set. However, §2 and §4 state that only Ωm, h, and ns variations are complete, with As, w∥, and Ωk 'in progress.' This inconsistency overstates the presently available data products and should be corrected in the abstract, the storage calculations, and the summary section.","section":"Abstract, §2, §4"},{"comment":"The headline estimate that ~40% of DESI BGS galaxies at z<0.15 are accessible to Sahyadri but not AbacusSummit is based on the unpublished 'Alam et al. (in prep)' methodology, applied to the UniverseMachine SHMR and the Bell et al. mass-to-light mapping. The paper gives no details of the BGS selection model, the abundance matching implementation, or validation against observed BGS clustering/luminosity functions. Because this estimate is used to justify the entire survey-motivation of the suite, it needs to be either fully specified in an appendix or released as a public companion document before it can be assessed.","section":"§1, Figure 1"},{"comment":"The summary in §4 states that finite-volume effects on linear halo bias become negligible (≲1%) for halos with M≳10^12 h^-1 M_sun, but Appendix C shows that while the mean halo-by-halo bias is unbiased, the median bias is underestimated by Δb∼0.2 and the scatter σ_b by ∼0.7 in the fiducial 200 h^-1 Mpc box. These effects truncate the tails of the b1 distribution that are used in the environment-correlation analyses (§3.2, Figures 12–13). The text acknowledges that Spearman correlations 'may slightly weaken' the trends, but this is in tension with the 'negligible' summary statement. Please reconcile these statements or quantify the impact on the environment results.","section":"§4 vs Appendix C"}],"minor_comments":[{"comment":"The compression scheme is claimed to maintain 'sub-percent clustering accuracy,' but no direct test of compressed versus uncompressed power spectra or correlation functions is presented. A small validation figure or quantitative statement would substantiate this claim.","section":"§2.1.2"},{"comment":"The text in §3.2 says the matter power spectrum shows 'good agreement' with halofit and deviations begin for k≳4 h/Mpc, but Appendix A reports deviations up to ∼20% at z=0 over a wide range of scales. Please make the characterization consistent and state explicitly whether this level of agreement is expected for the simulation resolution and halofit's calibration.","section":"§3.2, Appendix A"},{"comment":"Typos and minor wording issues: 'reproduceability' (§2.2), 'while while' (Appendix C), and 'spectrua' (§1).","section":"§2.2, Appendix C"},{"comment":"Reference [24] and 'Alam et al. (in prep)' are central to the paper's claims but appear only as unpublished works. If they are not yet on arXiv, please provide versions or at least a detailed public methodological appendix for the referee and readers.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The reader's stress-test concern lands: the derivative capability is the paper's most distinctive promise, yet it is not exercised or validated here, and the BGS fraction estimate is similarly outsourced to an unpublished methodology. The simulation itself appears technically sound, so I would not recommend rejection, but the advertised capabilities must be either demonstrated or appropriately qualified before publication. I would also ask the editor to ensure the authors' companion papers are available or that the manuscript be revised to avoid overstatement."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core numbers check out: Sahyadri is a real step up in mass resolution for a parameter-varying N-body suite. 2048^3 particles in a 200 Mpc/h box with mp = 8.1e7 Msun gives a 40-particle halo limit of 3.2e9 Msun, a genuine ~25x improvement over AbacusSummit. The power spectrum, mass function, VVF, and kNN measurements are a useful first look, and the VVF/kNN results at n ~ 2e-2 Mpc^-3 are genuinely new. The compression scheme is practical, and the GitHub pipeline is a concrete reproducibility plus.\n\nThe soft spots are mostly about packaging and scope. The abstract says six cosmological parameters vary; the body says only Omega_m, h, and ns are complete, with the rest in progress. It also says key data products are publicly available; the data availability statement says they will be released, with only the pipeline public now. That mismatch should be fixed before submission. Appendix A shows 10-20% deviations from halofit and Tinker, so 'excellent agreement' is too strong — ordinary agreement would be honest, and the paper's own appendix says so.\n\nThe bigger caveat is the derivative claim. Section 2.1.1 says seed matching is not sufficient for halo statistics because of gravitational phase-mixing, and defers to a post-processing stabilization method in the companion paper [24]. No derivative is actually computed for the showcase halo statistics; the figures show raw Omega_m responses with jackknife errors explicitly labeled illustrative. So 'enables cosmological parameter derivatives' is an aspiration, not a demonstrated capability in this paper. That is acceptable for a suite paper, but it should be worded as such, and the companion paper's method needs to be public or at least cited with enough detail for a reader to judge it. If that method fails at Sahyadri's resolution or volume, the headline claim erodes even though the simulation itself is fine.\n\nThe 40% BGS estimate is a motivating number, not a measured result. It depends on the UniverseMachine SHMR and Bell et al. mass-to-light mapping implemented in an unpublished Alam et al. in prep. If that model is wrong, the accessibility gain is different. The raw resolution comparison stands, but I would not put the 40% in the abstract without a public companion.\n\nI'd send this to peer review. It is a resource paper with clear value for low-z LSS and cosmic-web studies. The revisions I'd want: fix the abstract/body mismatches, make the data release real with a DOI or portal, soften 'excellent' to match the appendix, and either show one worked derivative or clearly mark derivative-readiness as forthcoming with the companion method.","headline":"The core resolution claim checks out, but the abstract promises six parameters and derivative-ready data that the body does not yet deliver.","tokens_in":23137,"tokens_out":3056,"would_cite":true,"duration_ms":35599,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The Sahyadri simulation suite resolves dark matter halos down to 3.2e9 solar masses, a factor of about 25 better than the AbacusSummit suite, and opens the faint-galaxy regime of the DESI Bright Galaxy Survey to cosmological parameter studi","keywords":["cosmological N-body simulations","cosmic web","DESI Bright Galaxy Survey","halo mass function","Voronoi volume function","k-nearest neighbour statistics","seed-matched initial conditions","Omega_m sensitivity"],"falsifier":"Compare the same halo finder on the Sahyadri snapshots with a version degraded to AbacusSummit particle mass (2e9 h^-1 M_sun) and verify that the 40-particle minimum halo mass shifts by the expected factor of ~25. Separately, generate mock BGS galaxy catalogs from Sahyadri using the paper's assumed stellar-mass-to-halo-mass relation and check whether the predicted BGS stellar mass function at z=0.15 matches observations; a significant mismatch would invalidate the 40% figure.","tokens_in":22169,"feed_emoji":"🕸️","tokens_out":2704,"duration_ms":30113,"temperature":0.7,"pith_summary":"The paper presents Sahyadri, a suite of dark-matter N-body simulations that evolve 2048^3 particles in a 200 h^-1 Mpc box, reaching a particle mass of 8.1e7 h^-1 M_sun. This resolves halos down to 3.2e9 h^-1 M_sun, about 25 times lower mass than the previous best parameter-varying suite, AbacusSummit. The suite varies cosmological parameters with matched initial-condition seeds, so differences between runs isolate the effect of each parameter, enabling derivative and Fisher-matrix calculations. The authors estimate this resolution brings roughly 40% of DESI BGS galaxies at z<0.15 into reach, and they demonstrate clear Omega_m sensitivity in both standard clustering and cosmic-web statistics such as the Voronoi volume function and k-nearest-neighbour distributions.","feed_headline":"Simulation suite resolves halos 25 times smaller than AbacusSummit","feed_subtitle":"Sahyadri's 2048^3-particle runs bring 1.6 million DESI BGS galaxies at z<0.15 into reach for cosmic-web studies.","key_machinery":"The carrying element is the combination of a compact box (200 h^-1 Mpc) with a very large particle count (2048^3), yielding a particle mass of 8.1e7 h^-1 M_sun and resolved halos down to 3.2e9 h^-1 M_sun (40 particles). Seed-matched initial conditions across parameter variations isolate cosmological responses, while a custom compression scheme reduces the data footprint by about a factor of three, keeping the ~117 TB total storage manageable.","core_discovery":"The central claim is that Sahyadri supplies the mass resolution and parameter coverage needed to model the faint galaxy population targeted by low-redshift spectroscopic surveys, especially DESI BGS. By resolving halos at 3.2e9 h^-1 M_sun, the suite captures roughly 1.6 million BGS galaxies at z<0.15 that AbacusSummit cannot reach. The paper shows that the simulations reproduce theoretical matter power spectra and halo mass functions, and that the Voronoi volume function and k-nearest-neighbour statistics respond strongly to Omega_m variations, making these high-density cosmic-web statistics usable for cosmological inference.","pith_inferences":["The 40% BGS accessibility figure depends on assumed stellar-to-halo and mass-to-light relations applied to BGS targets; if those calibrations shift, the percentage changes even though the raw mass-resolution improvement stands.","The small 200 h^-1 Mpc box limits access to large-scale modes; the paper itself shows finite-volume truncation of halo-by-halo bias tails, so users should correct for this when interpreting environment statistics.","Only Omega_m, h, and n_s variations are complete so far; the planned full six-parameter derivative set is needed before the suite can be used as a general emulator across the full parameter space.","The compression scheme's sub-percent clustering accuracy would be worth independent verification with a different power-spectrum estimator, since the claim rests on a single implementation."],"forward_implications":["DESI BGS galaxies at z<0.15 can be modeled at the halo level for roughly 40% more galaxies than with AbacusSummit, enabling mocks and systematics studies for the faint end of the survey.","Seed-matched runs allow cosmological parameter derivatives for beyond-two-point statistics, opening the door to Fisher forecasts that combine VVF, kNN, and the power spectrum.","The demonstrated Omega_m sensitivity of high-density VVF and kNN statistics suggests these observables can contribute constraints from the non-linear regime where two-point statistics are cosmic-variance limited.","The same simulation infrastructure can be extended to 4MOST and other low-redshift surveys, and to higher-order statistics like Minkowski functionals."],"fun_headline_variants":["Sahyadri simulations reach halos 25× finer than AbacusSummit","New simulations bring 1.6M DESI BGS galaxies into reach","Sahyadri resolves halos 25× smaller, capturing 1.6M more galaxies","Halos 25× smaller: Sahyadri's cosmic-web simulations"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The estimate that about 40% of DESI BGS galaxies at z<0.15 become accessible relies on the assumed relation between galaxy stellar mass and halo mass, and between stellar mass and r-band luminosity, calibrated in an unpublished companion methodology; if that galaxy-halo connection is wrong, the headline accessibility gain over AbacusSummit is unsupported even though the mass-resolution claim stands.","fun_headline_variants_meta":{"raw":{"variants":["Sahyadri simulations reach halos 25× finer than AbacusSummit","New simulations bring 1.6M DESI BGS galaxies into reach","Sahyadri resolves halos 25× smaller, capturing 1.6M more galaxies","Halos 25× smaller: Sahyadri's cosmic-web simulations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000627,"raw_usage":{"total_tokens":2803,"prompt_tokens":877,"completion_tokens":1926,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":621,"completion_tokens_details":{"reasoning_tokens":1836}},"tokens_in":621,"tokens_out":1926,"duration_ms":15521,"temperature":1.0,"reasoning_tokens":1836,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T10:59:19.531943+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compare the same halo finder on the Sahyadri snapshots with a version degraded to AbacusSummit particle mass (2e9 h^-1 M_sun) and verify that the 40-particle minimum halo mass shifts by the expected factor of ~25. Separately, generate mock BGS galaxy catalogs from Sahyadri using the paper's assumed stellar-mass-to-halo-mass relation and check whether the predicted BGS stellar mass function at z=0.15 matches observations; a significant mismatch would invalidate the 40% figure.","supporting_citations":[],"review_version":1}