{"id":"2a771dd2-21de-4991-9bb8-ad3116d51347","arxiv_id":"2506.16296","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A compressible RANS mixing model with entropy-gradient and counter-gradient heat-flux corrections reproduces density-stratified Rayleigh-Taylor mixing widths for stratification parameters 0.5 to 3.","lead":"This paper extends a Reynolds-averaged mixing transition model to compressible flows by adding entropy-gradient and heat-flux corrections to the turbulent mass flux term. For density-stratified Rayleigh-Taylor mixing, the revised model matches high-fidelity simulation mixing widths better than the incompressible baseline.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The validation is in-sample: C4=C3 is fixed by calibrating on the same HiFi cases later used as ground truth, so the predictive claim and the 'no new coefficient' statement are not yet established; an out-of-sample test is the decisive missing check.","rationale":"The reader's weakest assumption matches the main risk I find. I verified the derivation path: equations (2.22)–(2.35) are dimensionally and algebraically consistent, and (2.37) follows from requiring (2.35) to collapse to the baseline local Atwood number when c→∞. Nothing in the derivation fixes C4; it is not derived from the Gibbs relation or from the counter-gradient argument. The phrase 'based on calibrations with HiFi data' states explicitly that C4=C3 was chosen to fit the same simulations that are later used as ground truth. This makes the headline agreement in Figs. 8–10 a demonstration of calibration consistency rather than predictive skill. Because the model has only one calibration relation, this is not by itself a reason to reject: most RANS closures contain fitted coefficients. The concern is that the paper's framing — 'without introducing additional model coefficient' and 'systematic validations confirm efficacy' — converts a fitted relation into a general claim without an out-of-sample test. The proposed leave-one-out or a priori extraction of C4 from HiFi budgets would settle whether C4≈C3 is physically robust or a convenience fit. I therefore recommend the same conditional posture as the reader, with the condition being an out-of-sample validation and release of code and data.","tokens_in":18768,"tokens_out":10069,"duration_ms":119076,"concrete_test":"Leave-one-out calibration: re-estimate r = C4/C3 by minimizing the mixing-width error on only three of the four Sr cases (e.g., Sr=0.5, 1, 2), then predict the held-out case (Sr=3) with r fixed; cycle through all four hold-outs. If the optimal r differs from 1 by more than about 20%, or if held-out errors substantially exceed in-sample errors, then C4=C3 is dataset-specific. The stronger variant is to repeat without recalibration on an independent compressible RT or RM case with a different Atwood number or adiabatic index.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The algebraic chain leading to (2.38) is internally consistent: the incompressible limit of (2.35) indeed forces C3 = C_A(Γ-1)/Γ, and the entropy-gradient and counter-gradient heat-flux terms reduce to the stated forms. The load-bearing weak point is not the algebra but the status of C4. Equation (2.33) introduces a second closure coefficient, and the paper fixes it by the statement that 'optimal predictions' are obtained when C4=C3, 'based on calibrations with HiFi data' (§2.3). All validation in §3 (Figs. 8–10) then compares against those same HiFi cases. The agreement therefore cannot be read as an independent test of the closure; the relation C4=C3 is, in effect, one extra fitted condition, even if the final model contains no separately tunable constant. The §4 claim that the model adds no model coefficient is correspondingly overstated. The central predictive assertion — that the closure transfers to other compressible mixing flows — rests entirely on the unverified assumption that the optimal ratio C4/C3 is close to 1 outside the Luo–Wang dataset.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper extends the incompressible K-L-gamma mixing transition model of Xie et al. (2025) to density-stratified compressible Rayleigh-Taylor flows. Using the perfect-gas equation of state and the Gibbs relation, the authors derive a closure for the turbulent mass flux that replaces the density-gradient term in the local Atwood number with an entropy-gradient term and adds a counter-gradient turbulent heat-flux term. The final closure in Eq. (2.38) preserves the original coefficient C_A by setting C3 = C_A(Gamma-1)/Gamma and C4 = C3. The model is validated against the Luo-Wang HiFi simulations for Sr = 0.5, 1, 2, and 3, showing good agreement for mixing widths and mean species/density profiles.","tokens_in":19037,"tokens_out":6525,"duration_ms":60019,"significance":"If the predictive claim survives, this is a useful contribution to RANS modeling of compressible interfacial mixing: the derivation is systematic, the incompressible limit of Eq. (2.35) correctly constrains C3, and the final form (2.38) retains the baseline coefficient structure on paper. The emphasis on counter-gradient turbulent heat flux in stratified compressible RT flows is a physically interesting feature that standard gradient-diffusion closures miss. However, the central evidence is weakened by the in-sample calibration of C4 and by the absence of any out-of-sample test, so the general predictive claim for compressible mixing flows is not yet established.","major_comments":[{"comment":"The closure coefficient C4 is not fixed by the incompressible-limit constraint; the paper states in Section 2.3 that 'optimal predictions' are obtained when C4=C3, 'based on calibrations with HiFi data'. This makes C4 effectively a fitted parameter, despite the claim in Sections 2.3 and 4 that the model 'maintains the original model coefficient C_A without introducing additional parameters'. Because the same HiFi data are then used as ground truth in all four validation cases in Section 3, the agreement shown in Figs. 8-10 is an in-sample measure of the calibration, not an independent test of the closure. A decisive out-of-sample test, such as a different Atwood number, a non-isothermal initial stratification, or an RM or KH case, is needed before the predictive claim can be sustained.","section":"Sections 2.3 and 3, Eqs. (2.33)-(2.38) and Figs. 8-10"},{"comment":"All four validated cases share the same flow configuration: isothermal density-stratified RT with Atwood number 0.5, Gamma=5/3, and Sr = 0.5, 1, 2, and 3. Since C4=C3 is calibrated against these very cases, the close agreement in mixing widths, mass-fraction profiles, and density profiles is expected and does not demonstrate generalization. The abstract and title claim a model for 'compressible mixing flows' generally, but no RM, KH, or different-Atwood case is shown. The final paragraph of Section 4 acknowledges this limitation, but the absence of out-of-sample validation is a load-bearing gap for the central claim of the paper.","section":"Section 3, Figs. 8-10"}],"minor_comments":[{"comment":"The sentence 'In the incompressible limit c to 0' should read 'c to infinity'.","section":"Section 2.3, after Eq. (2.38)"},{"comment":"The phrase 'It it attributed' is a typo and should read 'It is attributed'.","section":"Section 2.2"},{"comment":"The term 'contour-gradient diffusion' appears to be a typo for 'counter-gradient diffusion'.","section":"Section 2.3"},{"comment":"The speed of sound c-bar is used before being defined; the definition c-bar^2 = Gamma-bar p-bar / rho-bar should be stated near its first use.","section":"Eq. (2.31)"}],"recommendation":"major_revision","confidential_remarks":"The paper's novelty relative to Cloutman (1987, 2003) and Moran-Lopez & Schilling (2013) is acknowledged by the authors and is not in itself a problem. The load-bearing weakness is the calibration-validation overlap: C4=C3 is chosen from the HiFi data and then the same data are used for validation. If the authors can add an out-of-sample case, the paper would be considerably stronger; without one, the central predictive claim is not yet supported."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague, read this if you care about RANS closures for compressible RT mixing. The paper does something genuinely useful: it extends the K-L-gamma mixing transition model to density-stratified compressible RT flows by replacing the baroclinic density-gradient local Atwood number with an entropy-gradient term plus a counter-gradient turbulent heat-flux term. The assembled closure, eq. (2.38), is new in this combination, and the derivation from the perfect-gas EOS, the Gibbs relation, and the compressible instability criterion is mostly coherent. The incompressible limit correctly forces C3 = C_A(Γ−1)/Γ, and the final model keeps the original coefficient C_A with no separately tunable constant. The HiFi budget analysis that motivates the turbulent mass flux as the dominant term is a solid piece of evidence. Credit is also due for citing the older pieces (Cloutman, Morán-López and Schilling, Zhou) that contain partial versions of these ideas, and for stating in Section 4 that RM and KH cases remain to be tested.\n\nThe load-bearing soft spot is the status of C4. Section 2.3 explicitly says that optimal predictions are obtained when C4 = C3, based on calibrations with HiFi data. All four validation cases come from the same Luo–Wang dataset. So the agreement in Figs. 8–10 measures the quality of that fit, not an independent prediction. The claim that the model adds no new coefficient is therefore overstated; the relation C4 = C3 is, in effect, one extra fitted condition. There is no out-of-sample test: no different Atwood number, no RM or KH case, no stratification outside Sr = 0.5–3. The general claim of a 'framework for compressible mixing flows' is broader than the evidence presented. Minor issue: the text says the incompressible limit is c → 0, which should be c → ∞. These caveats do not undermine the derivation, but they do limit what the paper can currently claim.\n\nWho is this for? People working on RANS modeling of RT/RM mixing or engineering predictions for ICF and supernova applications. They will find a usable starting point and a good literature thread. My recommendation: send it to peer review rather than desk-reject. A serious referee should make acceptance conditional on an out-of-sample test—different Atwood number, RM, or new stratification—and on clearly labeling C4 = C3 as a calibrated relation. Releasing code and data would strengthen the contribution considerably. As it stands, it is a solid conditional paper, not an independent validation.","headline":"A coherent compressible RANS closure with real derivation value, but the validation is in-sample because C4=C3 is tuned on the same HiFi data used as ground truth.","tokens_in":19549,"tokens_out":2617,"would_cite":false,"duration_ms":32744,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["76F25","76F55","76N15","76E17"],"pacs":["47.20.Ma","47.27.-i","47.40.-x"],"model":"deepseek-v4-flash","headline":"Extending an incompressible K-L-gamma mixing-transition RANS model to stratified compressible Rayleigh-Taylor flows, this paper replaces the baroclinic density-gradient source with entropy-gradient and counter-gradient heat-flux terms and…","keywords":["compressible Rayleigh-Taylor instability","RANS turbulence modeling","K-L-gamma mixing transition model","turbulent mass flux closure","turbulent heat flux","counter-gradient diffusion","density stratification","interfacial mixing"],"falsifier":"Measure the turbulent heat flux $T'u'_i$ in the high-fidelity $S_r=3$ case at $t/\\tau=4$ and test whether the counter-gradient closure (2.33) with $C_4=C_A(\\bar{\\Gamma}-1)/\\bar{\\Gamma}$ reproduces its sign and magnitude in the mixing zone; if the counter-gradient term is mis-signed, the corrected model's agreement with mixing widths would have to come from error cancellation rather than the claimed mechanism. A complementary test is to run the same closure on a compressible Richtmyer-Meshkov case with unchanged coefficients, an extension the paper leaves for future work.","tokens_in":18593,"feed_emoji":"🌊","tokens_out":12564,"duration_ms":117102,"temperature":0.7,"pith_summary":"This paper extends the incompressible $K$-$L$-$\\gamma$ mixing-transition RANS model to density-stratified compressible Rayleigh-Taylor flows, which matter for supernova and inertial-confinement-fusion applications. It identifies the failure of the baseline model in the closure of the turbulent mass flux: the baroclinic density-gradient product that feeds turbulent kinetic energy production switches off when stratification reverses the density gradient, even though the flow remains unstable. Using the perfect-gas equation of state and the Gibbs relation, the authors rewrite the turbulent mass flux as an entropy-gradient term plus a counter-gradient turbulent heat-flux term and fold both into the local Atwood number closure (2.38). The corrected closure keeps all original model coefficients, reduces to the incompressible form as the sound speed becomes large, and matches high-fidelity mixing widths and density and mass-fraction profiles for stratification parameters $S_r = 0.5$ to $3$. If the claim holds, compressible mixing predictions in engineering-scale settings gain a practical RANS route without a new empirical-coefficient calibration campaign.","feed_headline":"RANS mixing model now tracks compressible Rayleigh-Taylor flows","feed_subtitle":"Swap the baroclinic density gradient for entropy and counter-gradient heat flux; no new coefficients needed.","key_machinery":"The central object is the modified local Atwood number closure (2.38): $A_{ssi} = \\frac{\\tilde{L}}{\\bar{\\rho}+\\tilde{L}} |\\partial\\bar{\\rho}/\\partial x_i| [ C_A\\,\\partial\\bar{\\rho}/\\partial x_i + \\frac{C_A(\\bar{\\Gamma}-1)}{\\bar{\\Gamma}} \\frac{1}{\\bar{c}^2} (\\bar{\\rho}\\bar{c}_p\\,\\partial\\bar{T}/\\partial x_i - \\frac{2\\bar{\\Gamma}-1}{\\bar{\\Gamma}-1}\\,\\partial\\bar{p}/\\partial x_i) ]$, with $C_3=C_4=C_A(\\bar{\\Gamma}-1)/\\bar{\\Gamma}$. In one algebraic change, the raw density-gradient factor is replaced by the compressible instability combination plus a counter-gradient heat-flux term. This routes the compressibility corrections into the buoyancy production term $S_{Kf}$ without adding model coefficients, and it reduces exactly to the original closure as $\\bar{c}\\to\\infty$.","core_discovery":"On its own terms, the paper claims that the failure of incompressible RANS models on density-stratified compressible Rayleigh-Taylor flows has a specific cause: the buoyancy source in the TKE equation is closed through the baroclinic product $\\partial\\bar{p}/\\partial x_i\\,\\partial\\bar{\\rho}/\\partial x_i$, and in stratified flows the density gradient reverses sign too early, so the production term switches off while the flow is still unstable. Using the perfect-gas equation of state and the Gibbs relation, the authors derive the turbulent mass flux velocity $u''_i$ as a sum of an entropy-gradient contribution, which mirrors the compressible instability criterion, and a turbulent heat-flux contribution that exhibits counter-gradient transport. Folding these into the local Atwood number produces closure (2.38), with $C_3=C_4=C_A(\\bar{\\Gamma}-1)/\\bar{\\Gamma}$ so that no new model coefficients enter. With this closure, the $K$-$L$-$\\gamma$ mixing-transition model matches high-fidelity bubble and spike mixing widths and mean density and mass-fraction profiles for stratification parameters $S_r = 0.5, 1, 2, 3$, including the strongly stratified cases where the original baroclinic source vanishes.","pith_inferences":["Because $C_4=C_3$ was inferred by optimal matching to high-fidelity data rather than measured independently, a sharper test would extract $C_4$ directly from turbulent heat-flux statistics in the same or different flows.","The entropy-gradient/heat-flux mechanism is likely not specific to Rayleigh-Taylor flow; applying closure (2.38) unchanged to compressible Richtmyer-Meshkov and Kelvin-Helmholtz mixing would test that generalization, which the paper explicitly leaves open.","The derivation neglects fluctuations of the adiabatic index, so flows with strongly dissimilar molecular structures or extreme temperatures, such as some inertial-confinement-fusion mixtures, may need a $\\Gamma'$ term; this is an untested edge of the parameter-free claim.","The counter-gradient heat flux correlates with organized large-scale structures at high $S_r$, which suggests an observable link between the closure's success and flow organization that conditional statistics of high-fidelity data could check."],"forward_implications":["For stratification parameters $S_r=0.5$, $1$, $2$, and $3$, the corrected model reproduces high-fidelity bubble and spike mixing widths and mean density and mass-fraction profiles, while the baseline model stagnates at late times.","The original incompressible coefficients are retained unchanged, so existing calibrated parameter sets transfer directly to the compressible regime.","The closure supplies turbulent kinetic energy production in regions where the baroclinic product vanishes, which is the specific failure mode identified in the baseline model.","In the incompressible limit $\\bar{c}\\to\\infty$, the modified local Atwood number reduces exactly to the original form, preserving the baseline model's incompressible behavior.","The derivation identifies turbulent entropy flux and turbulent heat flux as the two channels through which compressibility enters the dominant production term of the TKE equation."],"supporting_citations":[{"why":"Supplies the baseline K-L-gamma mixing-transition model that this paper extends.","marker":"Xie et al. 2025"},{"why":"Provides the high-fidelity density-stratified compressible RT simulations used both to diagnose the baseline failure and to validate the corrected model.","marker":"Luo & Wang 2022"},{"why":"Introduces the buoyancy-drag style modeling of the turbulent mass flux term that the new closure modifies.","marker":"Dimonte & Tipton 2006"},{"why":"Gives the local Atwood number and weighting construction into which the compressibility correction is inserted.","marker":"Kokkinakis et al. 2015"},{"why":"Provides the counter-gradient turbulent heat-flux closure adopted for the temperature-fluctuation term.","marker":"Cloutman 2003"},{"why":"Supplies the exact TKE equation and Gibbs-relation framework used in the derivation.","marker":"Chassaing et al. 2002"},{"why":"Documents the non-quadratic growth and compressibility effects in stratified RT layers that motivate the correction.","marker":"Gauthier 2017"},{"why":"States the compressible local instability criterion that replaces the incompressible baroclinic criterion.","marker":"Gamalii et al. 1980"}],"fun_headline_variants":["Entropy and heat flux corrections fix compressible RANS mixing","Compressible RANS mixing model without new coefficients","Counter-gradient heat flux enables compressible mixing model","Turbulent entropy improves compressible RANS mixing"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the fitted equality $C_4=C_3$: the heat-flux coefficient was calibrated to the same high-fidelity simulations that later serve as validation, so the model's predictive accuracy outside that calibration set remains an assumption.","fun_headline_variants_meta":{"raw":{"variants":["Entropy and heat flux corrections fix compressible RANS mixing","Compressible RANS mixing model without new coefficients","Counter-gradient heat flux enables compressible mixing model","Turbulent entropy improves compressible RANS mixing"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000563,"raw_usage":{"total_tokens":2749,"prompt_tokens":1101,"completion_tokens":1648,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":717,"completion_tokens_details":{"reasoning_tokens":1584}},"tokens_in":717,"tokens_out":1648,"duration_ms":14616,"temperature":1.0,"reasoning_tokens":1584,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T23:44:14.208852+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the turbulent heat flux $T'u'_i$ in the high-fidelity $S_r=3$ case at $t/\\tau=4$ and test whether the counter-gradient closure (2.33) with $C_4=C_A(\\bar{\\Gamma}-1)/\\bar{\\Gamma}$ reproduces its sign and magnitude in the mixing zone; if the counter-gradient term is mis-signed, the corrected model's agreement with mixing widths would have to come from error cancellation rather than the claimed mechanism. A complementary test is to run the same closure on a compressible Richtmyer-Meshkov case with unchanged coefficients, an extension the paper leaves for future work.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the baseline K-L-gamma mixing-transition model that this paper extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the high-fidelity density-stratified compressible RT simulations used both to diagnose the baseline failure and to validate the corrected model."},{"cited_title":"& Tipton, R","cited_arxiv_id":null,"evidence_quote":"Introduces the buoyancy-drag style modeling of the turbulent mass flux term that the new closure modifies."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the local Atwood number and weighting construction into which the compressibility correction is inserted."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the counter-gradient turbulent heat-flux closure adopted for the temperature-fluctuation term."},{"cited_title":", Antonia, R","cited_arxiv_id":null,"evidence_quote":"Supplies the exact TKE equation and Gibbs-relation framework used in the derivation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents the non-quadratic growth and compressibility effects in stratified RT layers that motivate the correction."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"States the compressible local instability criterion that replaces the incompressible baroclinic criterion."}],"review_version":1}