{"id":"ba47b99d-59b3-4551-944e-4cffb148a195","arxiv_id":"1908.05837","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Center-point temperature fluctuations in cylindrical convection cells are universal when scaled by the convective temperature, with a constant value near 0.85, while scaled velocity fluctuations show a weak Ra^{0.07} dependence.","lead":"A study of turbulent convection shows that temperature fluctuations at the cell center collapse to a constant, about 0.85, once scaled by a heat-flux-based temperature, across a huge range of heating rates and for both smooth and rough plates. The result gives engineers and geophysicists a simple way to estimate internal flow fluctuations from global heat-transfer measurements.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The σw,c/w* universality claim rests on a substituted Nusselt correlation for three literature datasets; the paper's own 14% offset in Shang et al. makes DNS with simultaneous Nu the decisive check.","rationale":"The reader's weakest assumption isolates the substituted Nu correlation for the velocity data, and the full text supports this identification. The temperature claim (σT,c/θ*≈0.85) is built on simultaneously measured Nu for all cited sources and is therefore comparatively robust. The velocity claim, however, is not self-contained: for three of the four smooth-cylinder datasets, w* is computed from a global correlation rather than from the experiment's own heat-flux measurement. Since Reσw itself is measured (Fig. 4a) and scales as Ra^0.50, the reported σw,c/w*~Ra^{0.07} is algebraically the residual after dividing by (Ra Nu Pr^{-2})^{1/3}; a bias in the correlation prefactor shifts the amplitude of the collapse, and a bias in its exponent changes the residual slope. The paper's own admission that Shang et al. lie 14% below the others and that this may be due to the substituted Nu is decisive evidence that this is the weak link. Other issues (e.g., the z-offset for rough temperature profiles, the absence of UQ on 0.85) are secondary because they do not affect the central idea of the scale θ*. The one check that would settle the concern is to use DNS data with simultaneous Nu and centre-point σw; if the DNS points reproduce the Ra^0.07 trend, the correlation substitution is exonerated. Until then, the verdict should remain conditional, as the reader concluded.","tokens_in":12904,"tokens_out":12000,"duration_ms":117171,"concrete_test":"Use cylindrical-cell DNS (e.g., Scheel & Schumacher 2016, J. Fluid Mech. 802, 147-173, or a new simulation at Pr≈4.3) to obtain the vertical rms velocity at the cell centre from the time series, compute w* from the simulation's own simultaneously measured Nusselt number, and overlay σw,c/w* versus Ra on Fig. 4(b). If the DNS points do not fall on the same Ra^{0.07} trend and collapse, the velocity universality is an artifact of the substituted correlation rather than a physical scaling.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim has two parts: the temperature universality σT,c/θ*≈0.85 and the velocity scaling σw,c/w*~Ra^{0.07}. The velocity part is the least secure. In Sec. 3.2, for the smooth-cylinder velocity data from Daya & Ecke (2001), Qiu et al. (2004), and Shang et al. (2008), no simultaneous Nusselt number is available, so w* is constructed using the empirical correlation Nu=0.14 Ra^{0.297} Pr^{-0.03} (Xia, Lam & Zhou 2002). The collapse in Fig. 4(b) and the residual exponent 0.07 are therefore not measured independently of this correlation. The paper itself notes the Shang et al. data lie about 14% below the others and suggests the discrepancy 'may be due to the systematic error introduced when calculating w*, which involves Nu that was not measured simultaneously with σw,c'. Since Reσw scales as Ra^0.50, the reported exponent 0.07 is essentially the residual after dividing by (Ra Nu Pr^{-2})^{1/3}; any systematic bias in the substituted Nu among those experiments will directly produce an apparent non-universal offset or a spurious Ra-dependence. The temperature part of the claim does not have this weakness, because for the temperature data Nu was measured simultaneously in each source. But the paper's stated conclusion that w* is the proper velocity scale in the bulk depends on the velocity data, and for a large fraction of those data the normalization is not self-contained.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports new temperature and vertical-velocity fluctuation measurements in an aspect-ratio-one cylindrical Rayleigh–Bénard cell with smooth and rough plates, and combines them with a large set of literature data. The authors introduce the convective temperature and velocity scales θ*=(Q0^2/(αgH))^{1/3} and w*=(αgHQ0)^{1/3}, express them through Ra, Pr, and Nu, and show that the centreline rms temperature scaled by θ*, σT,c/θ*, collapses to about 0.85 for cylindrical cells over Ra≈10^8–10^15 and Pr≈0.7–23.3, independent of plate roughness, while a smaller plateau 0.34 is found in cubic cells. They further report σw,c/w* ~ Ra^{0.07±0.02}, a power-law σT(z)/θ* profile in the mixing zone of cylindrical cells, and a logarithmic σw(z)/w* profile, and suggest that θ* can be used to detect flow-state transitions at ultra-high Ra. The central claimed achievements are that θ* and w* are the proper characteristic scales for bulk temperature and velocity fluctuations in aspect-ratio-unity cylindrical cells.","tokens_in":13214,"tokens_out":9169,"duration_ms":92845,"significance":"If the results hold, the paper provides a valuable empirical simplification: decades of scattered σT,c/ΔT data in cylindrical cells are compressed into a single constant, and the He–Xia result on the functional form of σT(z) is extended to rough plates and to a much wider parameter range. The strength of the paper is its large compiled database and the fact that, for the temperature collapse, Nu was measured simultaneously with σT,c in each source. The velocity part is less secure, because a substantial fraction of the smooth-cell data relies on a substituted global Nusselt correlation rather than simultaneous measurements; a DNS or a new experiment with simultaneous Nu is a concrete, decisive test. The claims are empirical and falsifiable, which is a genuine merit even where the present support is incomplete.","major_comments":[{"comment":"The σw,c/w* collapse is not self-contained for three of the literature cylinder datasets. For Daya & Ecke (2001), Qiu et al. (2004), and Shang et al. (2008), no Nusselt number was measured simultaneously with σw,c, so w* is computed from the empirical correlation Nu=0.14 Ra^{0.297} Pr^{-0.03}. Because Rew*=(Ra Nu Pr^{-2})^{1/3}, this substituted exponent transfers directly to the normalized velocity: with Reσw ~ Ra^{0.50}, the ratio automatically inherits Ra^{0.50-(1+0.297)/3} ≈ Ra^{0.068}, i.e. essentially the quoted 0.07. The paper itself notes that the Shang et al. data are about 14% low and attributes this to the systematic error introduced when calculating w* with a Nu that was not measured simultaneously. Thus the velocity universality claim, and the stated conclusion that w* is the proper bulk velocity scale, currently rest on a fitted heat-transport model for a large fraction of the data. I request either a sensitivity analysis over plausible deviations of Nu for those experiments, or a demonstration using data with simultaneous Nu (DNS or new measurements), before this part of the claim can be accepted.","section":"Sec. 3.2, Eq. (2.1), Fig. 4(b)"},{"comment":"The central temperature result is quoted as σT,c/θ* ≈ 0.85 without any quantitative measure of scatter or uncertainty. Figure 3(b) shows visible spread among sources and no error bars, and the value is obtained by averaging over very different experiments with different reported accuracies. To support the claim of a universal constant, the paper should report the standard deviation or coefficient of variation of the cylindrical data, state the Ra range over which the average is taken, and propagate the uncertainties in Ra, Pr, and Nu into θ*/ΔT for each point. Without this, the 0.85 value cannot be distinguished from a loose trend or from a weak residual Ra dependence.","section":"Sec. 3.1, Fig. 3(b)"},{"comment":"The claim that the σT/θ* profile is universal for smooth and rough plates rests on offsetting the vertical coordinate for rough-cell data by the roughness height h. This coordinate shift is not derived, and no sensitivity test is given; different choices of origin (valley, base, midpoint of the roughness elements) would change the extent and quality of the collapse in the mixing zone. If the physical origin at the roughness tips is intended, that should be stated explicitly and justified; otherwise the apparent universality may be an artifact of the chosen offset.","section":"Sec. 3.1, Fig. 1(b,d)"},{"comment":"The proposed flow-state transition in the ultra-high-Ra data is inferred from only two groups, Ra ≥ 7.90×10^14 and Ra ≤ 1.18×10^13, with no intermediate points and no confidence intervals on the two logarithmic slopes. Since the paper presents this only as a potential application of θ*, the language should be correspondingly cautious, or the analysis should be extended with a quantitative comparison (e.g., overlapping fit ranges and slope uncertainties).","section":"Sec. 3.3, Fig. 6"}],"minor_comments":[{"comment":"The phrase \"the the convective velocity\" in the abstract should read \"the convective velocity.\"","section":"Abstract"},{"comment":"The sentence \"As these data are taken in a cell with Γ=0.7\" appears immediately after a discussion of cube data; a cube does not have Γ=0.7, so the sentence should identify the correct geometry, presumably the cylinder data of Daya & Ecke (2001).","section":"Sec. 3.2"},{"comment":"The text refers to \"the Ra-dependence of θT (z)/θ*\"; this should be σT(z)/θ*.","section":"Sec. 3.3"},{"comment":"The figure should mark, by symbol or legend, which datasets used the substituted Nusselt correlation and which used simultaneous Nu, so that the reader can separate self-contained data from correlation-dependent data.","section":"Fig. 4(b) and Sec. 3.2"},{"comment":"The text describing Fig. 5 says \"σw,c/w* in the range...\" when it is discussing the logarithmic profile plot; the notation should be σw(z)/w* to distinguish the profile from the centre-point values in Fig. 4.","section":"Sec. 3.2, Fig. 5"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe short version: this paper gives the most convincing evidence I've seen that the convective temperature θ* is the right scale for core temperature fluctuations in cylindrical Rayleigh-Bénard cells. The claim that σ_T,c/θ* ≈ 0.85 across Ra from 1e8 to 1e15, Pr from 0.7 to 23, and smooth/rough plates is a genuine empirical result, and it survives contact with the data. The velocity part, σ_w,c/w* ~ Ra^0.07, is not as well supported, and I'd want that fixed before citing it.\n\nWhat's actually new: they assemble a broad set of published and new rough-cell measurements and show that σ_T,c/ΔT scatters wildly but σ_T,c/θ* collapses. That collapse is not forced; for the temperature data, Nu was measured simultaneously in each source. The extension of the He-Xia profile analysis to rough cells is also a real step forward. The rough-cell velocity data are new and show the same Ra^0.5 Reynolds scaling as smooth cells.\n\nSoft spots, in order of real weakness. First, three literature velocity datasets (Daya & Ecke, Qiu et al., Shang et al.) have no simultaneous Nu, so w* is built from a fitted correlation Nu = 0.14 Ra^0.297 Pr^-0.03. That makes the apparent universality of σ_w,c/w* partly a test of that correlation, not purely a measurement. The authors are honest about this—they even flag the Shang et al. 14% offset as possibly systematic error in w*—but it means the velocity universality claim is conditional. Second, the headline constant 0.85 is given as a mean with no scatter or uncertainty; a simple standard deviation or confidence interval would help. Third, offsetting rough-cell profiles by the roughness height h is reasonable physically (origin at valley bottom) but is presented as an ad hoc shift; a sentence on why h is the right offset would close that gap.\n\nThe central temperature argument holds up. I don't think the substituted Nu issue breaks the velocity claim beyond repair, but DNS or experiments with simultaneous Nu and σ_w would be the decisive check. The paper deserves a serious referee; I'd recommend acceptance after a moderate revision addressing the uncertainty on 0.85 and a more guarded statement of the velocity collapse.\n\nWho it's for: anyone working on bulk fluctuation statistics, multi-parameter scaling, or rough-wall convection.","headline":"A careful empirical data-collapse paper that makes a solid case for θ* as the bulk temperature scale; the companion velocity claim is less secure because part of the normalization is borrowed from a global heat-flux correlation.","tokens_in":13793,"tokens_out":2011,"would_cite":true,"duration_ms":19175,"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":"Once rescaled by the convective temperature, the rms temperature at the centre of turbulent convection is a constant ≈0.85 across more than seven decades in Rayleigh number.","keywords":["Rayleigh-Bénard convection","turbulent convection","temperature fluctuations","velocity fluctuations","convective temperature scale","convective velocity scale","rough plates","universality"],"falsifier":"In a $\\Gamma\\approx1$ cylindrical cell, measure $\\sigma_T$ and $\\sigma_w$ at the centre while simultaneously measuring $Nu$, over $10^8 \\leq Ra \\leq 10^{10}$, with at least two different plate roughness geometries; the central claim predicts $\\sigma_{T,c}/\\theta_*$ remains at $0.85$ within scatter and $\\sigma_{w,c}/w_*$ follows $Ra^{0.07\\pm0.02}$ regardless of roughness. A systematic shift of the temperature ratio with roughness height, a drift of either ratio with $Ra$ beyond the quoted uncertainty, or a velocity exponent outside $0.05$–$0.09$ would refute it.","tokens_in":12662,"feed_emoji":"🌡️","tokens_out":9364,"duration_ms":78836,"temperature":0.7,"pith_summary":"Rayleigh–Bénard convection is thermal turbulence driven by heating from below, and its core temperature and velocity fluctuations have resisted a universal description because different experiments report different magnitudes and scaling exponents. This paper argues that the correct measuring stick is the convective temperature $\\theta_*$ and convective velocity $w_*$ constructed from the imposed heat flux, buoyancy, and cell height. Once $\\sigma_T$ is divided by $\\theta_*$, the centreline value collapses to a single constant, about $0.85$, across $10^8 \\leq Ra \\leq 10^{15}$ and $0.7 \\leq Pr \\leq 23.34$, for both smooth and rough plates. The same rescaling nearly collapses the vertical velocity fluctuation, which retains only a weak residual $Ra^{0.07}$ growth, and turns centreline profiles into universal power-law and logarithmic forms. The practical consequence is that the level of bulk thermal fluctuation can be estimated from global quantities alone.","feed_headline":"One constant predicts core temperature jitter in turbulent convection","feed_subtitle":"Across 10^8 to 10^15 Rayleigh and smooth or rough plates, the ratio σ_T,c/θ* holds at 0.85.","key_machinery":"The load-bearing object is a pair of scales from a dimensional argument: $\\theta_* \\equiv Q_0^{2/3}/(\\alpha g H)^{1/3}$ and $w_* \\equiv (\\alpha g H Q_0)^{1/3}$, where $Q_0$ is the specific heat flux, $\\alpha$ the thermal expansion coefficient, $g$ gravity, and $H$ the cell height. These are the only combinations of the supposed bulk-control parameters $\\alpha g$, $Q_0$, and $H$ with dimensions of temperature and velocity. The argument works because the paper rewrites them as $\\theta_*/\\Delta T = Nu^{2/3}/(Ra\\,Pr)^{1/3}$ and $Re_{w_*} = w_* H/\\nu = (Ra\\,Nu\\,Pr^{-2})^{1/3}$, so any dataset that reports global $Ra$, $Pr$, and $Nu$ can be rescaled without new local measurements.","core_discovery":"The central discovery is that in aspect-ratio-unity cylindrical cells the root-mean-square temperature fluctuation at the cell centre, normalized by the convective temperature $\\theta_* \\equiv Q_0^{2/3}/(\\alpha g H)^{1/3}$, is a universal constant $\\sigma_{T,c}/\\theta_* \\approx 0.85$ over $10^8 \\leq Ra \\leq 10^{15}$ and $0.7 \\leq Pr \\leq 23.34$, independent of whether the top and bottom plates are smooth or rough. The vertical rms velocity at the centre, normalized by $w_* \\equiv (\\alpha g H Q_0)^{1/3}$, is not exactly constant but scales as $\\sigma_{w,c}/w_* \\sim Ra^{0.07 \\pm 0.02}$ over the measured range and is likewise independent of plate topography. Outside the thermal boundary layer, the centreline temperature profile $\\sigma_T(z)/\\theta_*$ follows a power law in the distance $z$ from the plate, with exponent about $-0.57$ in cylinders and $-0.74$ in cube and rectangular cells, while the vertical velocity profile $\\sigma_w(z)/w_*$ is logarithmic in $z$. These collapses are presented as evidence that $\\theta_*$ and $w_*$ are the physically appropriate characteristic scales for bulk fluctuations, with geometry-dependent prefactors reflecting the large-scale circulation.","pith_inferences":["If the 0.85 plateau holds beyond the parameter range shown, bulk thermal fluctuation levels in geophysical or industrial convection could be estimated from global heat-flux measurements alone, sidestepping intrusive thermometry.","The cube-cell plateau at $\\sigma_{T,c}/\\theta_* \\approx 0.34$, one-third of the cylinder value, suggests the constant encodes the topology of the large-scale circulation; surveying other aspect ratios would map where universality ends.","The velocity residual $\\sim Ra^{0.07}$ is close to a $1/7$ turbulent boundary-layer exponent, hinting that $w_*$ captures the leading buoyancy balance while a weaker momentum-transport correction leaks in; a Prandtl-number sweep at fixed $Ra$ could test this.","The ultra-high-$Ra$ split in $\\theta_*$-scaled logarithmic profiles could be correlated with independent transition diagnostics, such as changes in heat-transport scaling or spectral signatures, to see whether $\\theta_*$ is a reliable transition detector."],"forward_implications":["The plateau $\\sigma_{T,c}/\\theta_* \\approx 0.85$ means the core temperature fluctuation in a $\\Gamma\\approx 1$ cylinder can be predicted from $Ra$, $Pr$, and $Nu$ alone, without placing a probe in the cell.","Since the plateau is the same for smooth and rough plates, the roughness-induced enhancement of $\\sigma_T/\\Delta T$ is an artefact of the wrong reference scale; $\\theta_*$ absorbs the topography.","The weak residual $\\sigma_{w,c}/w_* \\sim Ra^{0.07\\pm0.02}$ indicates the velocity scale is only nearly universal, leaving a slow dynamical growth that future theories must explain.","The power-law temperature profile with geometry-dependent exponent (about $-0.57$ in cylinders, $-0.74$ in cubes) ties the mixing-zone fluctuation structure to the shape of the large-scale circulation.","At $Ra \\gtrsim 10^{14}$, $\\theta_*$-scaled temperature profiles divide into two distinct logarithmic families, so $\\theta_*$ can serve as a probe of an internal flow-state transition."],"supporting_citations":[{"why":"Introduces the convective temperature and velocity scales that the paper adopts as its normalising units.","marker":"Deardorff (1970)"},{"why":"Supplies the classical hard-turbulence temperature-fluctuation data whose scatter under ΔT normalisation motivates the search for a universal scale.","marker":"Castaing et al. (1989)"},{"why":"Provides rough-cell temperature fluctuation data used to test whether σT,c/θ* is independent of plate topography.","marker":"Du & Tong (2001)"},{"why":"Provides smooth- and rough-plate data with simultaneous heat-transport measurements used to compute θ*.","marker":"Wei et al. (2014)"},{"why":"The present experimental cell and its rough-plate heat-transport regimes are taken from this study.","marker":"Xie & Xia (2017)"},{"why":"Ultra-high-Rayleigh-number temperature profiles in SF6 that the paper rescales with θ* to reveal a possible internal transition.","marker":"Ahlers et al. (2012)"},{"why":"Numerical vertical velocity profiles that provide the σw(z) data for the logarithmic collapse with w*.","marker":"Scheel & Schumacher (2016)"},{"why":"Establishes the power-law versus logarithmic classification of σT(z) that this paper extends to rough plates and wider Ra ranges.","marker":"He & Xia (2019)"},{"why":"Empirical Nu(Ra,Pr) relation used to compute w* for velocity datasets without simultaneous Nusselt-number measurements.","marker":"Xia, Lam & Zhou (2002)"}],"fun_headline_variants":["Convection core temp jitter: universal 0.85","Rough or smooth plates: core temp jitter ratio stays 0.85","0.85: universal core temperature fluctuation in convection","Temperature jitter in convection bulk collapses to one number","Core temp fluctuation ratio independent of Rayleigh, Prandtl, plates"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the bulk region outside the boundary layers is controlled solely by $\\alpha g$, $Q_0$, and $H$, so the convective scales computed from global $Ra$, $Pr$, and $Nu$ correctly represent the core; for the three literature velocity sets without simultaneous $Nu$ measurements, this also depends on the empirical correlation $Nu = 0.14\\,Ra^{0.297}Pr^{-0.03}$ being accurate for those runs.","fun_headline_variants_meta":{"raw":{"variants":["Convection core temp jitter: universal 0.85","Rough or smooth plates: core temp jitter ratio stays 0.85","0.85: universal core temperature fluctuation in convection","Temperature jitter in convection bulk collapses to one number","Core temp fluctuation ratio independent of Rayleigh, Prandtl, plates"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000732,"raw_usage":{"total_tokens":3446,"prompt_tokens":1286,"completion_tokens":2160,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":902,"completion_tokens_details":{"reasoning_tokens":2083}},"tokens_in":902,"tokens_out":2160,"duration_ms":13406,"temperature":1.0,"reasoning_tokens":2083,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:03:39.235928+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In a $\\Gamma\\approx1$ cylindrical cell, measure $\\sigma_T$ and $\\sigma_w$ at the centre while simultaneously measuring $Nu$, over $10^8 \\leq Ra \\leq 10^{10}$, with at least two different plate roughness geometries; the central claim predicts $\\sigma_{T,c}/\\theta_*$ remains at $0.85$ within scatter and $\\sigma_{w,c}/w_*$ follows $Ra^{0.07\\pm0.02}$ regardless of roughness. A systematic shift of the temperature ratio with roughness height, a drift of either ratio with $Ra$ beyond the quoted uncertainty, or a velocity exponent outside $0.05$–$0.09$ would refute it.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the convective temperature and velocity scales that the paper adopts as its normalising units."},{"cited_title":", Gunaratne, G","cited_arxiv_id":null,"evidence_quote":"Supplies the classical hard-turbulence temperature-fluctuation data whose scatter under ΔT normalisation motivates the search for a universal scale."},{"cited_title":", Chan, T.-S","cited_arxiv_id":null,"evidence_quote":"Provides smooth- and rough-plate data with simultaneous heat-transport measurements used to compute θ*."},{"cited_title":"& Xia, K.-Q","cited_arxiv_id":null,"evidence_quote":"The present experimental cell and its rough-plate heat-transport regimes are taken from this study."},{"cited_title":", Bodenschatz, E","cited_arxiv_id":null,"evidence_quote":"Ultra-high-Rayleigh-number temperature profiles in SF6 that the paper rescales with θ* to reveal a possible internal transition."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Numerical vertical velocity profiles that provide the σw(z) data for the logarithmic collapse with w*."},{"cited_title":", Lam, S","cited_arxiv_id":null,"evidence_quote":"Empirical Nu(Ra,Pr) relation used to compute w* for velocity datasets without simultaneous Nusselt-number measurements."}],"review_version":1}