{"id":"1a41f79b-6627-469c-9381-d2669a0db559","arxiv_id":"2501.04809","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A review with new simulations showing that thermal boundary conditions set the scale of large-scale convection patterns, and that compressible convection shows enhanced dissipation intermittency.","lead":"This review summarizes recent direct numerical simulations of mesoscale convection, the intermediate-scale ordered patterns seen in turbulent convection in stars and the atmosphere. It also reports new simulations on rectangular domains and compressible convection, focusing on how boundary conditions set the pattern scale and how kinetic energy dissipation is distributed.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The Dirichlet/Neumann dichotomy is not yet fully load-bearing: the Neumann 'final' scale is set by the box by construction, and the paper does not rule out that in larger domains or longer runs an intrinsic scale would emerge.","rationale":"The reader's weakest assumption identifies the same load-bearing concern: the structure-formation dichotomy rests on domain-size and simulation-time convergence, especially for the Neumann case where the final supergranule size equals the domain width by construction. My stress-test confirms that this is the most serious gap. The paper is honest about its limits, and the existing evidence—Γ = 60 Neumann runs, the λcrit = ∞ linear-stability argument, and prior Dirichlet TSS studies—supports a provisional version of the claim. But the strongest phrasing, that thermal boundary conditions control the pattern scale with an intrinsic Dirichlet scale of about 5H and an unbounded Neumann scale, requires either a larger-domain Neumann run showing continued aggregation past 60H or a demonstration that the 5H scale is stable under a doubling of Γ and run time. Because the new quantitative universality claims (Acoh ≈ 40%, ΛTSS ≈ 5H saturation) are based on limited parameter coverage and hand-chosen analysis thresholds, the conditional verdict is appropriate. I do not see an internal inconsistency or a reason to reject; the concern is about missing evidence at the parameter and domain extremes, not about the internal logic of the simulations. The proposed spectral-peak tracking test would directly separate an unbounded aggregation process from a finite-domain artifact, and the Acoh reanalysis would check whether the claimed universality survives a more neutral choice of diagnostic height and box count.","tokens_in":34467,"tokens_out":6991,"duration_ms":77904,"concrete_test":"Reanalyse the Γ = 60 Neumann data of refs. [13,67] and, if feasible, run one Γ = 120 case at the same Ra ≈ 10^5, Pr = 1: track the peak of the horizontal temperature-variance spectrum k_peak(t). If k_peak decays as a self-similar power law all the way to k_min until the end of the run, the supergranule scale is genuinely unbounded; if it plateaus near k ≈ 2π/(5H) before a late finite-domain takeover, the Dirichlet/Neumann distinction is partly a transient or finite-domain artifact. As a cheaper complement, recompute Acoh for the existing runs at a common normalized height, e.g. x3 = δU,rms, and with Nbox changed by a factor of 4; if the 40% value shifts by more than a few percent, that universality claim is threshold-sensitive.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim's strongest form—that thermal boundary conditions alone set the large-scale structure scale, with Dirichlet saturating near 5H and Neumann growing to the domain size—is underdetermined exactly where the two cases differ most. In Sec. 4.1, ΛSG = ΓH is a consequence of the finite box: the reported 'final' state is the largest scale the domain permits. The linear-stability argument (λcrit = ∞, ref. [76]) shows the Neumann base state has no intrinsic wavelength, so any finite box will eventually imprint its own size. What must be demonstrated is that in a larger box the aggregation continues to the new domain scale on accessible times, and that the Dirichlet 5H scale is not likewise creeping with Γ or run time. The paper's own limitations—Sec. 7 calls the work 'a starting point only,' and Sec. 4.1 concedes 'nor is there evidence that the supergranules vanish beyond certain critical Ra or Pr'—mean the dichotomy is established only for Ra ≲ 10^8, Pr roughly 0.001–100, and Γ ≤ 60, with supergranule transients of order 10^4 τf. In that regime domain size and aggregation time are controlling experimental factors, so the claimed universal control by thermal boundary conditions is not yet a converged physical law. The Acoh ≈ 40% universality is also threshold-defined at different wall-normal positions for different Pr (x3 = δT,rms versus δU,rms), so it does not independently support the structure-formation claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript is a review-style article, with some new DNS results, on turbulent mesoscale convection in plane layers. It derives the fully compressible equations and the Boussinesq limit, reports low-Prandtl-number simulations at Pr=0.001, compares Dirichlet and Neumann thermal boundary conditions and different mechanical boundary conditions, examines the effect of rectangular domain shape on turbulent superstructures, and analyzes compressible convection through stratification regimes, temperature-dependent material properties, and the statistics and multifractal properties of the kinetic energy dissipation rate. The central claimed result is that thermal boundary conditions control the large-scale structure formation: constant-temperature plates produce turbulent superstructures with a nearly fixed scale of about 5H, while constant-heat-flux plates produce supergranules that grow to the domain size, with mechanical boundary conditions playing a subdominant role.","tokens_in":34848,"tokens_out":8210,"duration_ms":83879,"significance":"If the central claim holds, the paper identifies a physically important control parameter for the scale of organized convection in planetary and stellar settings, with direct consequences for parametrizations of unresolved mesoscale convection. The paper's strengths include a consistent derivation of the governing equations, a clear framing of compressible convection regimes in the (ε,D) plane, the use of genuinely demanding DNS (for example, more than half a trillion grid points at Pr=0.001), and the connection of the Neumann-case self-organization to linear stability arguments and independent spectral-transfer analyses. The authors are also explicit about several limitations. However, the quantitative support for the central dichotomy and for several accessory universal claims is thin; the Neumann supergranule scale is limited by the domain size by construction, and the Dirichlet saturation scale is not demonstrated by a systematic convergence test. These issues affect the strongest form of the paper's main message.","major_comments":[{"comment":"The central claim that Neumann boundary conditions produce supergranules growing to the domain size, whereas Dirichlet conditions produce a converged scale near 5H, is underdetermined exactly where the two cases differ. Because the reported final supergranule scale is ΛSG=ΓH, the final state is by construction the largest scale the finite box permits; the λcrit=∞ linear-stability argument (ref. [76]) establishes that the Neumann base state has no intrinsic wavelength, but it does not demonstrate that in a larger domain the aggregation would continue to the new domain scale on accessible times rather than saturating at some intrinsic intermediate scale. The Dirichlet leg also lacks a convergence test: Sec. 4.1 quotes ΛTSS≈5H for Pr∼1 without a table of ΛTSS versus Γ and run time, and Sec. 3 reports ΛTSS≈3H at Pr≤0.02, so the quoted scale is not Pr-independent. Given the paper's own caveats in Sec. 4.1 ('nor is there evidence that the supergranules vanish beyond certain critical Ra or Pr') and Sec. 7 ('should be considered as a starting point only'), the thermal-boundary-condition dichotomy should be presented as an observation valid for the simulated range Γ≤60, Ra≲10^8, and run times of order 10^4 τf, not as a converged universal law.","section":"Sec. 4.1, Figs. 6-8"},{"comment":"The claim that the coherent near-wall area fraction Acoh is always approximately 40% and independent of Prandtl number is not supported as a controlled comparison. The analysis plane is set to x3=δT,rms for Pr=0.7 and to x3=δU,rms for Pr=0.001, i.e., at different wall-normal positions, and the coherent/incoherent classification uses the threshold |u_h|≷u_rms(x3); no sensitivity to this threshold or to the coarse-graining box size is reported. With eleven table entries and no uncertainty estimates on Acoh, the apparent universality could be an artifact of the threshold definition rather than a physical invariant, and the statement in Sec. 7 that Acoh is 'independent of Prandtl number' overreaches the data.","section":"Sec. 3, Table 1 and Fig. 5"},{"comment":"The scaling laws Nu∼Ra^0.29 and Re∼Ra^0.5 are fitted to only three Rayleigh numbers (Ra=10^5, 10^6, 10^7) at Pr=0.001, a fact the text itself notes in the preceding sentence. With three points, no assessment of curvature or of a possible regime change is possible, and the low values Nu=1.21, 2.48, 4.57 make a pure power-law fit particularly fragile. These expressions should be labeled as preliminary fits to three points, or supplemented by additional Rayleigh numbers, before they are presented as scaling laws in the summary of the regime.","section":"Sec. 3, eq. (39) and Table 1"},{"comment":"The conclusion that fully compressible convection is more intermittent than Oberbeck-Boussinesq convection rests on two FCC runs and two OB runs without uncertainty estimates for the generalized dimensions D(q) obtained from power-law fits in Fig. 15, and the bulk volume for FCC1 differs from that used for FCC2, OB1, and OB2 (Vb=L^2×[0.2,0.6]H versus L^2×[0.2,0.8]H). The differences between the D(q) curves in Fig. 16 are modest and could be within fit error. Since this intermittency claim is one of the four highlighted 'aspects' of the paper, it should either be presented explicitly as a preliminary trend or be supported by a fit-range and bootstrap sensitivity analysis.","section":"Sec. 6.3, Tables 2-4 and Fig. 16"}],"minor_comments":[{"comment":"The abstract contains two idiomatic errors: 'are partly not anymore accessible' should be 'are no longer accessible', and 'Beside these experiments' should be 'Besides these experiments'.","section":"Abstract and Sec. 1"},{"comment":"The derivative notation in Eq. (20) mixes d and ∂: the first term should be written with ∂T~/∂x3 consistently.","section":"Eq. (20)"},{"comment":"The grid entries such as '5122' and '1002' should be typeset as 512^2 and 100^2; the current notation is confusing, especially when Nbox is defined in the same table.","section":"Table 1"},{"comment":"Reference [29] is incomplete: it contains the placeholders 'Flow XXX (2024) XXX. doi:XXX' and needs the actual journal volume, article number, and DOI.","section":"Reference [29]"},{"comment":"The sentence 'the critical wavelength is λcrit=2√2 and≈2.02 for free-slip [73] and no-slip conditions [74,75], respectively' is grammatically awkward and should list the two values in separate clauses for clarity.","section":"Sec. 4.1"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The two things you should know: this is largely a review of the group's own prior DNS with a few genuinely new results, and the central claim about thermal boundary conditions setting the pattern scale is credible in the Dirichlet case but under-supported in the Neumann case.\n\nThe new pieces are worth taking seriously. The rectangular-cuboid aspect-ratio scans in Sec. 4.2 (Figs. 9–10) are a legitimate extension, and the FCC/OB comparison of dissipation-rate PDFs and multifractal spectra (Sec. 6, Figs. 13, 16) adds something new to the compressible-convection literature. The equation framework is standard and internally consistent, and the simulations are carefully documented with parameter tables. As a review of the authors' recent work, it is thorough and the references are appropriate.\n\nNow the soft spots. The paper's strongest claim—that Dirichlet plates give TSS saturating near 5H while Neumann plates give supergranules growing to the domain size—is not yet a converged physical law. The stress-test concern lands: ΛSG = ΓH is the largest scale the box permits, and the linear-stability argument (λcrit = ∞) means there is no intrinsic wavelength for the Neumann base state. So observing that the aggregate reaches the box size does not tell you what would happen in a larger domain. The paper does concede the lack of evidence at higher Ra/Pr and calls itself a starting point, but the abstract and Sec. 4.1 state the dichotomy more flatly than the evidence supports. The Acoh ≈ 40% universality is also threshold-defined at different wall-normal positions (δT,rms vs. δU,rms), so it does not independently reinforce the structure-formation story.\n\nThese are real limitations, but they are not fatal. The Dirichlet 5H scale is supported by multiple DNS at several aspect ratios, and the qualitative difference between Dirichlet and Neumann is backed by linear stability and several independent groups. The paper would benefit from tempering the claims and explicitly framing the Neumann result as a finite-domain observation that needs larger-domain confirmation.\n\nWho is this for? Convection researchers working on large-scale structure, low-Pr flows, or compressible dissipation statistics will find it useful, both as a review and for the new simulation data. It deserves a serious referee: the work is substantial, honest, and the new results are worth publishing, but I would push for revision that scales back the universality language.\n\nWrite it up as a conditional accept and ask for the overclaims to be tamed.","headline":"A useful review with some genuinely new DNS, but the headline claim that thermal boundary conditions alone set the pattern scale is not yet as load-bearing as the text suggests; the Neumann 'domain-size' scale is partly a finite-box artifact.","tokens_in":35355,"tokens_out":1601,"would_cite":false,"duration_ms":18848,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["47.27.te","44.25.+f"],"model":"deepseek-v4-flash","headline":"The thermal boundary condition, not the mechanical one, sets the pattern scale in turbulent mesoscale convection.","keywords":["mesoscale convection","turbulent superstructures","supergranules","thermal boundary conditions","constant heat flux","low Prandtl number","compressible convection","kinetic energy dissipation"],"falsifier":"Run a direct numerical simulation of constant-heat-flux convection at aspect ratio 120, matching the Rayleigh and Prandtl numbers of the Γ=60 runs, and track the largest connected coherent temperature pattern over time; if it stops growing before the domain width, the supergranule aggregation is a finite-domain effect and the universal picture fails.","tokens_in":34324,"feed_emoji":"🌀","tokens_out":5938,"duration_ms":52538,"temperature":0.7,"pith_summary":"This review-plus-extension paper consolidates direct numerical simulations of mesoscale convection in a plane layer heated from below and cooled from above. It argues that the thermal boundary condition controls the formation of long-lived large-scale flow structures: constant-temperature plates give turbulent superstructures with a fixed horizontal scale of about five layer heights, whereas constant-heat-flux plates produce supergranules that aggregate until they fill the whole domain. The paper also reports that about 40% of the near-wall area is coherent shear-dominated motion regardless of aspect ratio, Prandtl number, or Rayleigh number, and it extends the analysis to low Prandtl numbers, compressibility, and the statistics of kinetic energy dissipation. The motivation is that mesoscale convection in the atmosphere and Sun sits between small-scale turbulence and global circulation and must be parametrized in any global model.","feed_headline":"Heat-flux plates grow convection cells to domain size","feed_subtitle":"Fixed-temperature plates cap patterns near 5 layer depths; fixed-heat-flux plates produce supergranules that keep growing.","key_machinery":"The organizing object is the pair of thermal boundary conditions: the Dirichlet condition of fixed plate temperature versus the Neumann condition of fixed heat flux. In the linear regime the Neumann condition has an infinite critical wavelength $\\lambda_{\\mathrm{crit}}=\\infty$ while the Dirichlet condition has a finite one; the paper shows that this distinction persists far into the turbulent regime and produces the two contrasting classes of long-living large-scale flow structures (TSS vs supergranules). Supporting machinery includes very large horizontally extended DNS domains (aspect ratios up to 60), the decomposition of the dissipation field into solenoidal, dilatational, and inhomogeneous components, and the multifractal analysis of dissipation statistics.","core_discovery":"The central claim is that thermal boundary conditions determine the large-scale structure formation in turbulent mesoscale convection. For fixed-temperature (Dirichlet) plates the flow organizes into turbulent superstructures (TSS) whose characteristic horizontal extent saturates at about $\\Lambda_{\\mathrm{TSS}}\\approx 5H$ for aspect ratios up to at least 16, while for fixed-heat-flux (Neumann) plates the same flow evolves through a slow aggregation into a supergranule whose final size is the domain width, $\\Lambda_{\\mathrm{SG}}=\\Gamma H$. The mechanical boundary conditions (no-slip vs free-slip) are subdominant for this self-organization. The paper further claims that near-wall coherent regions occupy a universal fraction of roughly 40% independent of $\\mathrm{Ra}$, $\\mathrm{Pr}$, and $\\Gamma$; that low-Prandtl-number convection exhibits a Kolmogorov $k^{-5/3}$ kinetic energy spectrum with superstructure scale approaching about $3H$; and that in compressible convection the kinetic energy dissipation rate is multifractal and more intermittent than in the Oberbeck–Boussinesq limit, with solenoidal and dilatational components concentrated around pre-shock regions.","pith_inferences":["If pattern scale is governed by thermal boundary condition, then observations of pattern scales in natural convection (granules, cloud streets) could be inverted to infer effective boundary conditions at the top and bottom of convective layers.","Partially conducting plates interpolate between Dirichlet and Neumann; a testable prediction is that as the plate-to-fluid conductivity ratio decreases, the asymptotic superstructure scale should increase from $\\sim 5H$ toward the domain size.","The universal 40% coherent fraction suggests that near-wall shear organization is governed by a local instability mechanism independent of the large-scale pattern; if true, the mechanism should be reproducible in a much smaller domain with the same local boundary-layer dynamics.","Because supergranule growth is stopped by weak rotation, rotation can serve as a control knob to test the inverse-cascade interpretation; measuring the aggregate scale versus the Rossby number would sharpen the theory."],"forward_implications":["If thermal boundary conditions set the pattern scale, then global climate and stellar models can use the known boundary type to estimate the size of unresolved mesoscale convection structures without resolving them.","Because supergranules aggregate to domain size under constant heat flux, flux-driven convection in nature will have a scale set by the domain, not by fluid properties.","Low-Prandtl-number convection (relevant to stars and liquid metals) produces Kolmogorov-type inertial ranges and superstructures of about $3H$, implying structure size varies weakly with $\\mathrm{Pr}$ and $\\mathrm{Ra}$.","The universal ~40% near-wall coherent area fraction offers a simple, robust statistic for wall models and for interpreting experimental boundary-layer measurements.","Compressible convection shows enhanced small-scale intermittency relative to the Boussinesq limit even at lower Reynolds numbers, so parametrizations based on Boussinesq turbulence statistics will underestimate extreme dissipation events."],"supporting_citations":[{"why":"Establishes the supergranule aggregation under constant heat flux boundary conditions, which is the central contrast to the Dirichlet case.","marker":"[13]"},{"why":"Reports turbulent thermal superstructures with a fixed characteristic scale in Dirichlet convection.","marker":"[11]"},{"why":"Identifies turbulent superstructures in Rayleigh–Bénard convection and their dependence on aspect ratio.","marker":"[10]"},{"why":"Shows inverse cascades of kinetic energy and thermal variance that drive supergranule aggregation.","marker":"[64]"},{"why":"Shows supergranule aggregation is independent of Prandtl number, extending the claim across the accessible parameter range.","marker":"[67]"},{"why":"Provides the coherent/incoherent near-wall decomposition and the observed 40% coherent area fraction.","marker":"[50]"},{"why":"Supplies the low-Prandtl-number DNS data and the superstructure scale around 3H.","marker":"[45]"},{"why":"Justifies the claim that domains with aspect ratio at least 16 are wide enough to avoid finite-size effects.","marker":"[53]"},{"why":"Gives the infinite critical wavelength for the Neumann boundary condition, linking the linear instability to the turbulent pattern scale.","marker":"[76]"}],"fun_headline_variants":["Heat-flux plates unlock supergranules that span the whole domain","Fixed-temperature caps convection patterns; fixed heat flux grows supergranules","Convection pattern size set by thermal boundaries, not Rayleigh","Supergranules fill the box: thermal boundary conditions rule mesoscale convection","Fixed heat flux grows supergranules to box size; fixed temp caps at 5H"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusions assume that the simulated domains are large enough and the runs long enough that the observed patterns (the 5H superstructure scale and the domain-filling supergranule) reflect intrinsic dynamics rather than finite-size or transient artifacts, even though the supergranule's final size is the domain width by construction.","fun_headline_variants_meta":{"raw":{"variants":["Heat-flux plates unlock supergranules that span the whole domain","Fixed-temperature caps convection patterns; fixed heat flux grows supergranules","Convection pattern size set by thermal boundaries, not Rayleigh","Supergranules fill the box: thermal boundary conditions rule mesoscale convection","Fixed heat flux grows supergranules to box size; fixed temp caps at 5H"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001189,"raw_usage":{"total_tokens":4937,"prompt_tokens":1008,"completion_tokens":3929,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":624,"completion_tokens_details":{"reasoning_tokens":3834}},"tokens_in":624,"tokens_out":3929,"duration_ms":26045,"temperature":1.0,"reasoning_tokens":3834,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:25:10.455389+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run a direct numerical simulation of constant-heat-flux convection at aspect ratio 120, matching the Rayleigh and Prandtl numbers of the Γ=60 runs, and track the largest connected coherent temperature pattern over time; if it stops growing before the domain width, the supergranule aggregation is a finite-domain effect and the universal picture fails.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the supergranule aggregation under constant heat flux boundary conditions, which is the central contrast to the Dirichlet case."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports turbulent thermal superstructures with a fixed characteristic scale in Dirichlet convection."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows inverse cascades of kinetic energy and thermal variance that drive supergranule aggregation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows supergranule aggregation is independent of Prandtl number, extending the claim across the accessible parameter range."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the coherent/incoherent near-wall decomposition and the observed 40% coherent area fraction."},{"cited_title":"Pandey, D","cited_arxiv_id":null,"evidence_quote":"Supplies the low-Prandtl-number DNS data and the superstructure scale around 3H."}],"review_version":1}