{"id":"308b2977-3619-440b-85a3-c241c1d48a79","arxiv_id":"1908.05463","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"A wind tunnel study of stable and convective boundary layers over a regular urban array quantifies how stratification alters friction velocity, roughness length, displacement height, and the Monin-Obukhov stability length.","lead":"Wind tunnel tests over a model city show that stable air suppresses turbulence, while hot unstable air strengthens it, and the building roughness changes how the stability is felt. The new dataset is one of the few laboratory records of such conditions and can be used to validate computer models of pollution dispersion in cities.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"CBL Monin-Obukhov doubling is confounded by unheated model buildings; roughness-only mechanism is not established.","rationale":"The reader's verdict is CONDITIONAL, and I agree with that disposition. The strongest quantitative support for the headline claim is the CBL L doubling; the unheated-building confound is identified explicitly in the manuscript (Section 4.1) but is not quantified. This is more load-bearing than the linear-extrapolation concern because it can bias the result even if the flux profiles are perfectly linear and the extrapolation is unbiased. It is also a design asymmetry between the two states being compared (approach vs array), not a statistical uncertainty that would shrink with longer averaging. The SBL 80% L increase is less obviously affected, but the same floor-only temperature control raises the same question. I therefore do not move the verdict: CONDITIONAL is the right outcome, with the additional condition that the authors either heat the building models, measure their surface temperatures, or area-weight the heat budget to separate roughness effects from thermal-boundary effects. The paper otherwise has real value: it is a rare stratified urban-array dataset, the data are openly deposited, and the directional turbulence changes are consistent with prior work.","tokens_in":15944,"tokens_out":7796,"duration_ms":83075,"concrete_test":"Compute a corrected CBL surface heat flux for the array from a heat budget: measure (or add instrumented) building surface temperatures during a Riapp=-1.5 run, estimate the convective heat flux from the building surfaces with a standard correlation and the known plan-area fraction of the H x 2H blocks, add it to the floor-panel heat flux, and recompute θ*, L, and L_array/L_approach. If the corrected ratio drops from ~2 toward ~1, the unheated buildings explain the claimed stratification reduction; if it remains ~2, the roughness mechanism survives.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central CBL claim, that the urban array's increased roughness reduces surface stratification and doubles the Monin-Obukhov length relative to the approaching flow, rests on a comparison in which the thermal boundary condition is not the same between the two configurations. In the CBL runs the floor is heated by electrical mats, but the wooden building blocks are unheated (Section 4.1: 'The vertical heat flux over the array appears reduced, also as consequence of the wooden buildings not being heated'). The approaching-flow reference has a uniformly heated floor and no buildings. The array therefore replaces part of the heated floor with cold roughness elements, reducing the effective surface heat flux regardless of any roughness-induced change in turbulence. Since L is proportional to (w'θ')_0/u*^3, a reduction in surface heat flux and an increase in u* both act to increase L; the reported doubling is thus mechanically expected even if roughness had no stability-modifying effect. The paper does not quantify the building heat-flux contribution or provide building surface temperatures, so the claim 'increased roughness causes a reduction in surface stratification' is not uniquely supported. This concern is distinct from, and prior to, the linear-extrapolation assumption identified by the reader: even a perfect extrapolation of the measured w'θ' profile cannot separate the roughness effect from the unheated-building effect. The same concern may affect the SBL runs, where only the floor is cooled and the building models are not described as temperature-controlled.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports wind-tunnel measurements of stable and convective boundary layers over a regular array of rectangular building models, with approaching-flow comparisons intended to isolate roughness effects. Surface parameters (u*, z0, d, θ*, z0h, dh, L) are derived by fitting Monin-Obukhov similarity profiles to measured velocity and temperature profiles and by linearly extrapolating Reynolds shear stress and vertical heat flux profiles to the surface. The main findings are that in stable stratification the array increases the Monin-Obukhov length by up to 80%, reduces z0, increases d, and suppresses in-canopy turbulence; in convective stratification the array roughly doubles L, increases z0 by about 55%, reduces d, and increases u*. The paper also reports internal boundary-layer heights and velocity integral length scales, and makes the dataset openly available.","tokens_in":16215,"tokens_out":9602,"duration_ms":92291,"significance":"If the causal interpretation can be secured, the dataset is a valuable experimental resource for validating urban canopy models under non-neutral conditions, and the directional findings are broadly consistent with prior work. Strengths include the realistic building geometry, comparisons with field observations and earlier experiments, and open data availability. The main weakness is that the central claim that increased roughness reduces surface stratification is not cleanly identified, because the array configuration also changes the thermal boundary conditions (unheated buildings in the CBL cases and, most likely, uncooled buildings in the SBL cases). The quantitative percentages also lack uncertainty estimates and the abstract disagrees with the full text. These issues must be resolved before the quantitative conclusions can be accepted.","major_comments":[{"comment":"The CBL claim that 'the increased roughness causes a reduction in the surface stratification' (L doubled over the array) is confounded by the thermal boundary condition. Section 4.1 states that 'the vertical heat flux over the array appears reduced, also as consequence of the wooden buildings not being heated.' In the approaching-flow reference the heated floor is uniform, while in the array case only the floor between the buildings is heated and the buildings themselves are unheated. The array therefore reduces the total upward heat flux simply by replacing part of the heated floor with unheated solid blocks; since |L| is proportional to u*^3/(w'θ')_0, both the increase in u* and the decrease in surface heat flux act to increase L, so the reported doubling can occur even if the roughness has no stability-modifying effect. The paper does not report building surface temperatures or a heat-budget estimate that would separate the roughness effect from the surface-temperature effect. Please quantify the building heat-flux contribution (or provide a bounding estimate) and revise the causal wording in the abstract and Section 5 accordingly.","section":"Section 4.1, Table 2"},{"comment":"The same identification problem applies to the stable cases. The lower-roughness reference is generated with uniformly cooled floor panels, while in the array case only the floor is cooled and the wooden building blocks are not described as actively cooled. If the blocks are warmer than the cooled floor, they will locally reduce the magnitude of the negative surface heat flux, which increases L and partially mimics a roughness-induced weakening of stability. The paper should state the thermal state of the building surfaces in the SBL runs and quantify its contribution to the reported L increase of up to 80%, or explicitly justify that the effect is negligible.","section":"Section 3.1, Table 1"},{"comment":"The abstract supplied with this version reports SBL z0 reduction up to 35% and d increase up to 12%, and CBL z0 increase up to 50%; the full-text abstract and Section 5 report 27%, 5%, and 55%, respectively. These are different quantitative claims for the same experiments. The authors must correct the inconsistency and ensure that the abstract, main text, tables, and conclusion report identical numbers.","section":"Abstract vs. full text and Conclusion"},{"comment":"The headline ratios (L doubling, 80% increase, z0 and d changes) are reported as exact numbers without uncertainty estimates. Section 2.2 reports standard errors of 10-25% on the covariances that feed the flux extrapolation, so the propagated uncertainties in u*, θ*, and L are likely material to the claimed differences. Please propagate the measurement uncertainties into the derived surface parameters and report confidence intervals, or at least provide a sensitivity analysis of L to the chosen linear-extrapolation interval.","section":"Tables 1 and 2, Section 2.3"},{"comment":"Equation (3) as written has a sign error relative to standard Monin-Obukhov theory. With θ* = -(w'θ')_0/u*, the correct expression is ζ = (g/Θ0)θ*/(u*^2/kz), not the negative of that quantity. The tabulated δ/L values have the conventional signs, which suggests the computations used the standard form, but the printed equation must be corrected and its sign consistency with Eqs. (4)-(7) checked explicitly.","section":"Equation (3)"}],"minor_comments":[{"comment":"Please state how the standard errors were computed (block averaging, bootstrap, or repeat runs) and how the resampling of the LDA and cold-wire signals affects the reported covariances.","section":"Section 2.2"},{"comment":"The SBL approaching-flow reference is a single profile at x_T/H = -35, while the CBL reference is the average of two profiles at x_T/H = 1.4 and 22.4. Since the two cases use different reference locations, please explain why this does not affect the comparability of the SBL and CBL results.","section":"Figures 3 and 8"},{"comment":"Please clarify whether the neutral reference cases in Tables 1 and 2 were measured with the same spire sets as the stratified cases to which they are compared.","section":"Section 2.1"},{"comment":"The sentence 'the sum of the two contributions considered singularly is larger than the increment resulting by their combined effect' is unclear and should be rephrased for precision.","section":"Section 5"},{"comment":"The data availability statement should be moved to a dedicated section and cite the figshare DOI in the standard format.","section":"Acknowledgments"}],"recommendation":"major_revision","confidential_remarks":"The experimental dataset is likely valuable and the paper has clear strengths, but the central causal attribution in the CBL case (and plausibly in the SBL case) is not supported by the present comparison because the thermal boundary conditions differ between the reference and array configurations. This can likely be addressed with additional analysis of heat fluxes and surface temperatures rather than new experiments, but the current wording overstates the uniqueness of the roughness mechanism. The abstract inconsistency and the sign error in Eq. (3) also need correction. I recommend major revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know about arXiv:1908.05463. It gives us the first stratified boundary-layer wind-tunnel data over an H x 2H x H rectangular-block array at a 45-degree wind direction, with data publicly archived. The stable-stratification results—reduced Reynolds stresses, Monin-Obukhov length up to 80% larger, roughness length down 27%, displacement height up 5%—are plausible, consistent with cube-array studies, and worth having as a benchmark.\n\nThe weak spot is the convective case. The headline claim is that increased roughness reduces surface stratification, doubling the Monin-Obukhov length relative to the approaching flow. But the comparison is not apples to apples. The CBL floor is heated by mats; the wooden blocks are not heated. So the array removes heated floor area and replaces it with cold roughness. That alone reduces the surface heat flux and raises the Monin-Obukhov length. The paper even admits the heat flux is reduced 'also as consequence of the wooden buildings not being heated.' Since L goes as u*^3 over (w'θ')0, the doubling is mechanically expected from the changed thermal surface condition; it does not uniquely implicate roughness-modified stability. The same confound likely affects the SBL runs, where the blocks are not temperature-controlled and may be warmer than the cooled floor.\n\nSecondary issues: the abstract still quotes 35%/12% while the full text says 27%/5%, and the CBL percentages differ between abstract (50%) and text (55%). Also, the fitted surface parameters come without error bars, even though the covariances feeding them have 10–25% standard error. The practice of fitting the log-law to get L, z0, and d and then interpreting stability through those same fitted values is internal to the framework; the external comparisons to field data and prior experiments help, but the CBL interpretation needs an independent thermal check.\n\nThis paper deserves a referee, not a desk reject. The SBL dataset is solid and useful for validating LES and urban dispersion models. The CBL part can be salvaged if the authors quantify the building surface temperature contribution and frame the result as the combined effect of roughness and partial heating rather than a pure roughness effect. I'd ask for those changes, plus uncertainty estimates and a consistent abstract. Bring it to the next journal-club slot; the confound discussion is instructive.","headline":"Useful stable-stratification dataset over a novel urban array, but the headline convective result is confounded by unheated building blocks.","tokens_in":16776,"tokens_out":3949,"would_cite":true,"duration_ms":36288,"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":"The paper claims that adding urban-scale roughness to a stratified boundary layer changes the surface stability itself: the Monin–Obukhov length grows up to 80% in stable conditions and doubles in convective conditions, with roughness…","keywords":["stable boundary layer","convective boundary layer","urban canopy","Monin-Obukhov length","roughness length","displacement height","wind tunnel","Reynolds shear stress"],"falsifier":"Measure the same array with dense three-dimensional sampling, or simulate it numerically with resolved surface heat flux, and compare the true area-averaged surface shear stress and heat flux with the linear extrapolation from $1.5H$ to $4H$; a systematic mismatch would mean the reported changes in $L$, $z_0$, and $d$ are artifacts of the extrapolation rather than real stratification-roughness interactions.","tokens_in":15724,"feed_emoji":"🌆","tokens_out":11744,"duration_ms":107026,"temperature":0.7,"pith_summary":"The paper reports wind-tunnel experiments in which stable, neutral, and convective boundary layers approach a regular array of rectangular building blocks, and it asks how the urban surface itself alters stratification. In stable conditions the array reduces Reynolds stresses and canopy turbulence, cuts the aerodynamic roughness length by up to 27%, raises the displacement height by up to 5%, and increases the Monin–Obukhov length $L$ by up to 80% relative to the approaching flow. In convective conditions friction velocity rises, roughness length increases by up to 55%, displacement height falls by about half, and the Monin–Obukhov length doubles over the array. The authors read these changes as the roughness damping the effective surface stratification, and conclude that urban canopy parameterizations should use stability-dependent roughness length and displacement height rather than neutral values.","feed_headline":"City roughness doubles Monin-Obukhov length in convective air","feed_subtitle":"Wind-tunnel flows show stable air sees up to 80% longer Monin-Obukhov length over the urban array.","key_machinery":"The Monin–Obukhov length $L$, the height at which mechanical and buoyant turbulence production become comparable, is the central diagnostic. The paper estimates $L$, friction velocity $u_*$, friction temperature $\\theta_*$, roughness lengths $z_0$ and $z_{0h}$, and displacement heights $d$ and $d_h$ by fitting stability-corrected logarithmic profiles for wind and temperature to point measurements, and by linearly extrapolating measured Reynolds shear stress and vertical heat flux profiles to the surface. Comparisons between lower-roughness approaching flow and higher-roughness array flow isolate the effect of the urban geometry on these surface-layer parameters.","core_discovery":"The paper shows experimentally that urban-like roughness modifies how Monin–Obukhov similarity applies to an already stratified flow. Over a regular array of $H \\times 2H \\times H$ blocks at 45° wind direction, stable stratification lowers the friction velocity $u_*$, cuts Reynolds stresses and in-canopy turbulence, reduces $z_0$ by 16–27%, raises $d$ by up to 5%, and increases the Monin–Obukhov length $L$ by up to 80% compared with the same flow without the array. Convective stratification raises $u_*$ through the combined action of roughness and instability, increases $z_0$ by up to 55%, reduces $d$ by about half, and doubles $L$. The authors conclude that urban surfaces reduce the effective stability felt by the flow, and that stability-dependent $z_0$ and $d$ are needed in urban parameterizations.","pith_inferences":["Implicit in the results, a single-layer urban scheme using neutral $z_0$ and $d$ would need stability-dependent corrections; the measured changes in $L$ give a direct target for such a parameterization.","A testable extension is to repeat the experiment with heated building surfaces: the unheated wooden blocks reduced upward heat flux, so real cities with heated walls and roofs would likely show a smaller doubling of $L$.","A further consequence is that dense roughness damps stratification in both directions, so urban arrays may act as a self-limiting buffer on extreme surface stability; varying building density and aspect ratio would test this."],"forward_implications":["In stable conditions the urban array reduces the friction velocity and Reynolds stresses, cuts $z_0$ by up to 27%, raises $d$ by up to 5%, and increases $L$ by up to 80% compared with the approaching flow.","In convective conditions the array increases $u_*$, raises $z_0$ by up to 55%, lowers $d$ by about half, and doubles $L$, meaning the surface feels weaker convection than the incoming flow does.","Urban canopy parameterizations that keep $z_0$ and $d$ fixed at neutral values will misrepresent surface fluxes whenever the approaching boundary layer is stratified.","The internal boundary layer developing over the array reaches about $2.5H$ in stable conditions and $3$\\textendash $4H$ in convective conditions, setting a height limit for single-layer canopy models.","Stable stratification suppresses in-canopy turbulence and slows canopy flow without changing its direction, while convective stratification increases in-canopy velocity variances."],"supporting_citations":[{"why":"Supplies the wind-tunnel method for generating stable and convective boundary layers and the linear-extrapolation technique used to estimate surface fluxes.","marker":"Marucci et al. (2018)"},{"why":"Provides the neutral reference flow and displacement-height estimate for the same H×2H×H array that the stratified cases are compared against.","marker":"Castro et al. (2017)"},{"why":"Earlier stratified wind-tunnel experiment over a cube array that supplies the 2.5H internal-boundary-layer height and canopy-flow trends used for comparison.","marker":"Uehara et al. (2000)"},{"why":"Stratified approaching-flow experiment over staggered cubes used as a comparison for constant-flux-layer height, length scales, and CBL turbulence profiles.","marker":"Kanda and Yamao (2016)"},{"why":"Foundational similarity theory from which the logarithmic wind and temperature profiles and the Monin–Obukhov length are taken.","marker":"Monin and Obukhov (1954)"},{"why":"Supplies the simplified stability-correction functions used to fit the diabatic profiles in stable and convective cases.","marker":"Högström (1988)"},{"why":"Documents the LDA/cold-wire measurement and resampling setup used to acquire simultaneous velocity–temperature statistics.","marker":"Marucci and Carpentieri (2019)"},{"why":"Field data for the stable boundary layer used as an external comparison for the normalized turbulence profiles.","marker":"Caughey et al. (1979)"}],"fun_headline_variants":["Urban arrays stretch stable-air Monin-Obukhov length by 80%","Convective flow over city blocks doubles Monin-Obukhov length","Urban roughness cuts effective stability, boosting Monin-Obukhov up to 80%","City arrays increase Monin-Obukhov length in both stable and convective flows","Urban surfaces reduce effective stratification in wind-tunnel flows"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole analysis rests on the assumption that, averaged across space, the downward momentum flux and upward heat flux decrease nearly linearly with height just above the building tops, so extending those measured trends to the surface gives the true surface fluxes; the paper itself notes the data were too coarse for a direct spatial average.","fun_headline_variants_meta":{"raw":{"variants":["Urban arrays stretch stable-air Monin-Obukhov length by 80%","Convective flow over city blocks doubles Monin-Obukhov length","Urban roughness cuts effective stability, boosting Monin-Obukhov up to 80%","City arrays increase Monin-Obukhov length in both stable and convective flows","Urban surfaces reduce effective stratification in wind-tunnel flows"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000774,"raw_usage":{"total_tokens":3411,"prompt_tokens":920,"completion_tokens":2491,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":536,"completion_tokens_details":{"reasoning_tokens":2390}},"tokens_in":536,"tokens_out":2491,"duration_ms":17317,"temperature":1.0,"reasoning_tokens":2390,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:12:37.549036+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the same array with dense three-dimensional sampling, or simulate it numerically with resolved surface heat flux, and compare the true area-averaged surface shear stress and heat flux with the linear extrapolation from $1.5H$ to $4H$; a systematic mismatch would mean the reported changes in $L$, $z_0$, and $d$ are artifacts of the extrapolation rather than real stratification-roughness interactions.","supporting_citations":[{"cited_title":", author Obukhov, A.M","cited_arxiv_id":null,"evidence_quote":"Foundational similarity theory from which the logarithmic wind and temperature profiles and the Monin–Obukhov length are taken."}],"review_version":1}