{"id":"485f0994-4d8d-4442-a21b-177f2402c0c6","arxiv_id":"2505.04900","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A Gamma mixture model of overshooting plume lengths in 2D simulations of six Cepheid models yields a calibrated diffusion coefficient and a tentative relationship between overshooting ratio and shell depth and width.","lead":"Six 2D simulations of convection in Cepheid interiors show that overshooting plumes from an inner convective shell can reach nearly to the outer convective envelope. The authors fit the plume penetration distances with a mixture of two Gamma distributions and propose a new diffusion coefficient for stellar evolution models.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Upper-side overshooting distances are right-censored by the outer simulation boundary (Sec. 3.2.2), so the Gamma-mixture SO mean and the lower/upper ratio in Figs. 12a,b may be biased; a censoring-aware refit is needed before Eq. 18 is used.","rationale":"The paper's intended contribution is a statistically grounded characterization of overshooting from an interior convective shell, culminating in a ratio law (Figs. 12) and a diffusion coefficient (Eq. 18). The reader's weakest-assumption analysis correctly identifies the upper-boundary truncation as the load-bearing soft spot. Section 3.2.2 explicitly states that for M > 6 M⊙ the extreme overshooting plumes often hit the outer boundary and the EVT length saturates at the full radiative-layer depth; the same truncation must affect the PDF of first-zero positions used for the Gamma mixture, even though the mixture section does not discuss it. Because the non-penetrative radial boundary forces Fk = 0 at Ro, all un-truncated plumes are measured as stopping at the boundary. The SO component is the tail of the distribution, so its mean and the lower/upper ratio are the quantities most vulnerable to right-censoring. Deeper shells in the sample are also the ones with more frequent boundary contact, so the 'depth' and 'width' trends in Fig. 12 may be conflated with a censoring trend. This is not a claim that the result is wrong; the authors' long time series (25–138 τconv), high radial resolution (>100 grid points/Hp), and prior 2D/3D overshoot comparisons give the simulations real evidential weight, and the paper states its own parametric limitations clearly. But the censoring correction is analytically straightforward and should be applied before the ratio law or Eq. 18 is used. A censoring-aware refit (treating boundary hits as right-censored) is the single check that would settle this; if the corrected ratios retain the monotonic decrease, the central claim survives. The reader's conditional verdict is appropriate; I do not propose moving it.","tokens_in":27877,"tokens_out":7343,"duration_ms":74836,"concrete_test":"Refit the upper-side mixture model with right-censored likelihood: for plumes whose first Fk zero is at or outside the outer boundary, contribute the survival function at Δro_max = (Ro − R_U_CB)/Hp,U_CB instead of the boundary density. Compare censoring-corrected ℓU_so and ℓL_so/ℓU_so across all six models to Table 3 and Fig. 12. If the corrected ratio no longer decreases monotonically with depth/width, or if any corrected ℓU_so changes by more than the trend differences (~0.1–0.3 Hp), the central scaling and Eq. 18 need revision. A complementary check: rerun ceph8 with the outer boundary extended outward by ~0.1 R and compare the SO parameters.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim is that the mean of the strong-overshooting (SO) Gamma component defines a characteristic overshooting length on each side of the shell, and that the ratio ℓL_so/ℓU_so decreases with shell depth and width (Figs. 12a,b; Eq. 18). This claim is most vulnerable where the data are truncated: for all models except ceph5, the simulation outer boundary lies just below the convective envelope and overshooting plumes 'often hit the outer boundary' (Sec. 3.2.2). Since the radial boundaries are non-penetrative (Sec. 2), the vertical kinetic energy flux Fk (Eq. 6) vanishes at Ro, so the measured first-zero position ro for any plume that would have penetrated farther is recorded at Ro. The PDF of Δro used for the Gamma-mixture fit (Eqs. 12–16, Fig. 11) is therefore right-censored at Δro_max = (Ro − R_U_CB)/Hp,U_CB. The fitting procedure described in Sec. 3.3.1 fits the CDF by least squares with no censoring term; it cannot distinguish a plume that stops at Ro from one that would have gone beyond. The SO component is precisely the tail that is cut off, so its mean ℓU_so, the 90th percentile, and the mixture weight are all biased, and ℓU_so is systematically underestimated. The bias is expected to be largest for the deeper, more massive shells (ceph8, ceph9), where the EVT analysis already finds δ-like distributions at the boundary. The apparent drop of ℓL_so/ℓU_so with depth/thickness in Fig. 12 could therefore be partly an artifact of the domain size rather than a stellar-structure effect. Because the diffusion coefficient D_shell (Eq. 18) inherits the upper-side SO parameters from these fits (Eq. 19), this censoring undermines the proposed mixing prescription as well. The authors acknowledge the boundary issue for EVT but do not account for it in the mixture model.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents 2D MUSIC simulations of six 5.75–9 M_sun Cepheid models that contain an interior convective shell and an outer convective envelope, with the simulation domain truncated just below the envelope. The authors characterize the shell convection through filling factors and plume widths, then analyze convective boundary mixing using extreme value theory and a newly proposed mixture of two Gamma distributions fitted to the PDF of plume penetration distances. They report that strong overshooting above the shell fills the radiative layer, that the ratio of lower to upper strong-overshooting lengths decreases with shell depth and width (Figs. 12a,b), and they propose a diffusion coefficient for 1D stellar evolution models (Eq. 18), including the speculative idea of a 'super-mixing layer' formed by overlapping overshooting from the shell and the envelope.","tokens_in":28250,"tokens_out":7615,"duration_ms":76913,"significance":"The paper addresses an important gap: multidimensional simulations of the interior convective shells of Cepheids are rare, and 1D overshooting prescriptions for such shells are poorly calibrated. The simulations are long (25-138 convective turnover times) and well resolved (more than 100 grid cells per pressure scale height at the convective boundaries), and the mixture-model decomposition of overshooting PDFs is a genuinely new statistical tool that could provide physically motivated overshooting lengths and a smooth diffusion profile. If the censoring and statistical-significance issues are resolved, the proposed relationship between shell geometry and overshooting asymmetry would be directly useful for stellar evolution modelling, and the super-mixing-layer hypothesis is stimulating even though its quantitative form relies on assumed envelope parameters.","major_comments":[{"comment":"The upper-side overshooting statistics are right-censored by the artificial outer boundary. For all models except ceph5, plumes frequently reach the outer radius Ro, and because the radial boundaries are non-penetrative (Sec. 2) the vertical kinetic energy flux vanishes there; any plume that would have penetrated farther is recorded at Delta_ro_max = (Ro - R_U_CB)/Hp_U_CB. The mixture-model fit in Sec. 3.3.1 minimizes least squares on the CDF and contains no censoring term, so the SO component of the upper PDF, which is precisely the tail affected by the cutoff, has biased values of the mean ell_U_so, the 90th percentile, and the mixture weight. The bias is expected to be largest for ceph8 and ceph9, where the EVT already gives delta-like distributions at the boundary. Consequently the decrease of ell_L_so/ell_U_so with depth and width shown in Fig. 12 could be partly a domain-size artifact rather than a physical trend. Please provide a censoring-aware refit (e.g., maximum likelihood with a survival contribution for Delta_ro = Delta_ro_max) or repeat with a larger outer boundary, and quantify how ell_U_so and the ratios in Fig. 12 change.","section":"Sec. 3.2.2 and Sec. 3.3.1, Table 3, Fig. 12"},{"comment":"The central trend is inferred from only six models in which shell depth (R_L_CB/R) and shell width ((R_U_CB - R_L_CB)/Hp) are correlated with each other. The error bars in Fig. 12 are the standard deviations of the fitted SO Gamma distributions, not the uncertainties of the means or of the fit parameters, and no significance test, correlation coefficient, or confidence interval is reported for the linear regression line. Given the scatter and the overlap of the error bars, the statement that the ratio 'depends directly' on depth and width is not statistically established. The authors' caveat at the end of Sec. 3.3.2 acknowledges this, but the abstract and conclusions present the trend as a result. Please add significance testing (e.g., Spearman or Pearson correlation with bootstrap or a hierarchical model) and report parameter uncertainties rather than distribution widths.","section":"Sec. 3.3.2, Fig. 12"},{"comment":"The proposed diffusion coefficient is dimensionally inconsistent as written. The variable Delta_ro in Eq. (12) is normalized by Hp,CB, and the mixture parameters in Eqs. (19) are quoted in units of Hp,CB, but the arguments of the incomplete Gamma functions in Eq. (18) contain (R_CB - r)/R, i.e., a stellar-radius normalization. As a result, D_shell is not the CDF-based quantity described in the text, and Fig. 13 cannot be reproduced from Eq. (18). Please replace the /R factors with /Hp,CB (or otherwise consistently define the dimensionless radial coordinate in the Gamma CDF) and regenerate the figure. The same issue affects Eq. (21).","section":"Sec. 3.3.3, Eqs. (18) and (21)"},{"comment":"The super-mixing layer and the full-star diffusion coefficient D_MM are constructed by assuming the envelope overshooting parameters are equal to the lower-shell parameters (mu_E = mu_L, alpha_E = alpha_L, lambda_E = lambda_L) and that D_envelope_MLT = 0.8 D_MLT, with no simulation containing the envelope. The text does state these are assumptions, but the conclusion that a super-mixing layer with D/D_MLT > 1 could exist is an extrapolation, not a simulation result. Please mark this explicitly in the figure caption and in Sec. 4 so that readers do not mistake the assumed envelope model for a calibrated prediction.","section":"Sec. 3.3.3, Fig. 13"}],"minor_comments":[{"comment":"The claim that overshooting above the convective shell 'fills the space between these convectively unstable layers' should be qualified: for M > 6 M_sun this is inferred from plumes that hit the outer boundary of a domain truncated just below the envelope, so the filled region is the simulated gap, not an independently converged overshooting length.","section":"Abstract and Sec. 3.2.2"},{"comment":"The caption describes the standard deviation of the SO Gamma distribution as an 'error on the mean'; it is the distribution width, not a standard error of the mean. Please reword to avoid implying a smaller statistical uncertainty.","section":"Table 3 caption"},{"comment":"The linear regression line is drawn without reporting its equation, R-squared value, or uncertainty; either report these quantities or remove the line to avoid implying a quantified fit.","section":"Fig. 12"},{"comment":"The mixture model has seven free parameters fitted by nonlinear least squares, but no goodness-of-fit statistic or parameter covariance is reported; adding at least a residual metric and a bootstrap uncertainty on ell_so would strengthen the comparison between models.","section":"Sec. 3.3.1"}],"recommendation":"major_revision","confidential_remarks":"The manuscript fits the scope of A&A and the work is a useful contribution to a relatively understudied problem. The main risk is the outer-boundary censoring of the upper overshooting statistics, which directly affects the central quantitative claim and the diffusion-coefficient calibration. If the censoring-aware analysis or a larger-domain simulation confirms the trend, the paper would be a solid publication; as it stands, the reported ratio and D_shell rest on a biased fit that the current text does not acknowledge."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: the Gamma-mixture decomposition is a real methodological step, and the paper is honest about its own limits. Second: the stress-test concern is correct. The upper overshooting statistics are right-censored by a non-penetrative outer boundary, and the mixture fit does not account for it. That puts the headline trend (ℓL_so/ℓU_so decreasing with shell depth/thickness) on shaky ground, and it also contaminates the proposed diffusion coefficient.\n\nWhat the paper does well: long 2D MUSIC simulations of six Cepheid-like structures, high radial resolution at the convective boundaries, more than 25 convective turnover times per run. The diagnostics (filling factor, inflow width, vertical kinetic flux) are clearly presented. The finding that overshoot above the shell reaches the outer boundary is a real result, even if it complicates the statistics. The mixture model is a sensible, if standard, way to split a skewed positive PDF into a frequent weak component and a rare strong component; I think the Gamma choice is well motivated by maximum entropy. The authors also explicitly flag that the six-model sample is too small to separate depth and width effects, and that the super-mixing layer needs confirmation. The self-citations to the group's earlier EVT work are appropriate rather than excessive.\n\nThe soft spots are the load-bearing ones. Section 3.3.1 fits the CDF with least squares and no censoring term. Plumes that would have gone past Ro are recorded at Ro. The SO component is exactly the tail being truncated, so its mean, 90th percentile, and weight are biased downward on the upper side. The bias is probably worst for ceph8 and ceph9, where the EVT already returns δ-like distributions at the boundary. The apparent drop of the ratio with depth/width could therefore come from domain geometry rather than stellar structure. The same censoring feeds Eq. 18 through Eq. 19, so the proposed mixing prescription inherits the bias. The authors acknowledge the boundary issue for EVT, but they do not carry it into the mixture fit. A censoring-aware likelihood, or one model with a larger outer radius, would resolve the question.\n\nMinor point: the six points in Fig. 12 have correlated depth and width and no significance test; the authors say this themselves. Also, D_shell is a calibrated fit to the same runs, so it should be presented as calibration, not prediction; the authors mostly do this. The boundary issue is the one that needs fixing before the ratio can be trusted.\n\nWho should read it: anyone working on convective boundary mixing in shell convection, and developers of 1D overshoot prescriptions. I would cite the mixture-model approach, but not the ratio trend as it stands. Verdict: this deserves a serious referee, not a desk reject. The referee should request a censoring-aware refit and a clear statement of how much of the trend survives.","headline":"A genuinely new statistical decomposition of overshoot plumes, but the paper's central ratio trend is undermined by right-censoring at the outer simulation boundary and needs a censoring-aware refit.","tokens_in":28898,"tokens_out":4453,"would_cite":true,"duration_ms":44225,"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":"This paper claims that in Cepheid interiors, overshooting above the inner convective shell reaches all the way to the outer envelope, while the ratio of lower to upper overshoot lengths drops as the shell deepens and thickens.","keywords":["convective overshooting","convective shells","Cepheid variables","hydrodynamic simulations","extreme value theory","Gamma mixture model","convective boundary mixing","super-mixing layer"],"falsifier":"Re-run the same six stellar structures with the simulation domain extended through the outer convective envelope, or with a large radiative buffer above the shell, and rebuild the PDF of upper penetration distances; if the mean of the strong-overshooting Gamma component, the 90th-percentile length, or the ratio $\\ell^L_{\\rm so}/\\ell^U_{\\rm so}$ shifts by more than the quoted 1$\\sigma$ errors, the truncated boundary is distorting the central result.","tokens_in":27659,"feed_emoji":"⭐","tokens_out":13189,"duration_ms":106335,"temperature":0.7,"pith_summary":"Cepheid variable stars serve as cosmic distance markers, yet their evolution models disagree with pulsation-based masses by 10–20%. The paper argues that one missing ingredient is a realistic treatment of the inner convective shell that appears in the Z-bump opacity region of intermediate-mass Cepheids. Using 2D hydrodynamic simulations of six stars from 5.75 to 9 solar masses, it shows that overshooting plumes above this shell travel far enough to fill the entire radiative layer up to the outer convective envelope. To quantify the mixing, the paper fits the distribution of plume penetration distances as a mixture of two Gamma distributions, one for frequent weak overshooting and one for rare strong overshooting; the mean of the strong component defines a characteristic overshooting length on each side of the shell. The ratio of lower to upper strong-overshooting lengths stays below one and decreases as the shell becomes deeper and thicker, leading to a new diffusion coefficient, $D_{\\rm shell}$ (Eq. 18), and to the idea of a 'super-mixing layer' where shell and envelope overshoot overlap.","feed_headline":"Cepheid overshoot below the shell shrinks as shells deepen","feed_subtitle":"Six 2D simulations yield a mixing recipe that lets shell and envelope overshoot overlap.","key_machinery":"The central object is the mixture-of-two-Gamma model fitted to the PDF of the plume penetration distance $\\Delta r_o = |r_o - R_{\\rm CB}|/H_{p,\\rm CB}$ (Eqs. 12–16). The weak Gamma component captures the 85–95% of plumes that stop just past the convective boundary; the strong Gamma component captures the 5–15% of rare, deep plumes, and its mean $\\ell_{\\rm so}$ becomes the characteristic overshooting length. The same fitted CDF feeds the new diffusion coefficient $D_{\\rm shell}$ (Eq. 18), which keeps the whole strong-overshooting distribution instead of only the deepest plume used by the extreme-value-theory approach. The Weibull/GEVD analysis (Eq. 11) remains as a consistency check and provides the maximal overshoot length.","core_discovery":"The paper's central claim is that convective boundary mixing around a Cepheid's inner shell is not one process but two: frequent shallow plumes that barely cross the Schwarzschild boundary, and rare deep plumes that dominate mixing in the radiative zone. By fitting the probability density function of the dimensionless penetration distance $\\Delta r_o$ with a mixture of two Gamma distributions, the analysis isolates the strong-overshooting population; its mean, $\\ell_{\\rm so}$, is proposed as the characteristic overshooting length on each side of the shell, while its 90th percentile matches the maximal length from extreme-value theory. Across the six simulations, the ratio $\\ell^L_{\\rm so}/\\ell^U_{\\rm so}$ is close to unity for shallow, thin shells and drops steadily as the shell lies deeper and becomes wider, and it is always below one. The paper further claims that for stars above about 6 solar masses the upper overshooting layer reaches through the whole radiative zone, and that when shell and envelope overshooting layers overlap the diffusion coefficient can exceed the local mixing-length value, so the two convective layers may behave as one super-mixing layer. From the mixture-model CDF, it proposes the diffusion coefficient $D_{\\rm shell}$ (Eq. 18) as a calibrated input for 1D stellar evolution models.","pith_inferences":["If the trend in Figs. 12a,b continues beyond the simulated range, then for thicker and deeper shells the ratio $\\ell^L_{\\rm so}/\\ell^U_{\\rm so}$ would approach the factor of about 0.2 found in simulations of massive stars, unifying the previously scattered literature values; this extrapolation is not made in the paper.","Because the simulations omit radial pulsations, a testable extension is to impose the Cepheid pulsation on the shell boundaries and remeasure the two Gamma components; the characteristic lengths might then become phase-dependent, which would time-modulate $D_{\\rm shell}$.","The mixture-model split into weak and strong overshooting could be applied to other convective boundaries, such as cores and envelopes, giving a common two-population language for overshooting across stellar regimes; the paper only demonstrates it for shells.","A 3D simulation of the same shells at comparable radial resolution would show whether the convection-roll geometry that dominates these thin shells is a 2D artefact; if the strong-overshooting Gamma mean changes, the absolute values of $\\ell_{\\rm so}$ but not necessarily the ratio trend would need revision."],"forward_implications":["1D Cepheid evolution models can adopt $D_{\\rm shell}$ (Eq. 18) with the parameter ranges of Eq. (19), replacing an uncalibrated overshooting length with a prescription derived from plume statistics on both sides of the shell.","For stars above about 6 $M_\\odot$, the upper overshooting layer occupies the entire radiative zone between the shell and the outer envelope, so evolution models that leave that zone unmixed underestimate the chemical communication between the two convective layers.","The lower-to-upper overshoot ratio is not a single constant: the trend with shell depth and width explains why massive-star shells and thin A-type shells had previously yielded ratios of about 0.2 and 0.5.","Overlapping overshoot layers from the shell and the envelope can produce local diffusion larger than the mixing-length value, so the two adjacent convective zones may effectively merge into a super-mixing layer.","Using the 90th percentile of the strong-overshooting Gamma, at most about 1.5% of all overshooting plumes can reach the outer convective envelope directly, quantifying how much mass exchange crosses the radiative zone."],"supporting_citations":[{"why":"supplies the EVT-based maximal overshoot length and the diffusion-coefficient formula $D_{\\rm EVT}=D_{\\rm MLT}(1-F(r))$ that the new mixture-based coefficient extends.","marker":"Pratt et al. (2017)"},{"why":"provides 3D box-in-star simulations of massive-star convective shells finding lower overshoot about one fifth of upper; the paper's new ratio is compared with this baseline.","marker":"Cristini et al. (2017, 2019)"},{"why":"extends the massive-shell simulations and is another source of the 0.2 lower/upper overshoot ratio used as a comparison.","marker":"Rizzuti et al. (2022)"},{"why":"predicts a lower/upper overshoot ratio near 0.5 for thin shallow shells and overshoot filling the radiative zone in A-type stars; it motivates the claim that upper overshoot fills the whole layer.","marker":"Guo & Li (2021)"},{"why":"supplies the inflow-width and filling-factor diagnostics, and the argument that 2D and 3D overshoot lengths agree near convective boundaries, justifying the 2D setup.","marker":"Dethero et al. (2024)"},{"why":"provides the maximum-entropy rationale for choosing Gamma distributions as the components of the mixture model.","marker":"Park & Bera (2009)"}],"fun_headline_variants":["Cepheid overshoot splits into weak and strong layers","Deep shells shrink below-shell overshoot in Cepheids","Shell and envelope overshoot may merge into one mixing zone","New diffusion recipe from 2D Cepheid simulations","Extreme-value stats reveal dual overshoot in Cepheid interiors"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The results assume that truncating each simulation just below the outer convective envelope does not change how far overshooting plumes would really travel above the shell, even though for stars above 6 solar masses the strongest plumes regularly hit this artificial outer boundary and the measured upper overshooting distribution is therefore cut off at the edge of the domain.","fun_headline_variants_meta":{"raw":{"variants":["Cepheid overshoot splits into weak and strong layers","Deep shells shrink below-shell overshoot in Cepheids","Shell and envelope overshoot may merge into one mixing zone","New diffusion recipe from 2D Cepheid simulations","Extreme-value stats reveal dual overshoot in Cepheid interiors"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00076,"raw_usage":{"total_tokens":3463,"prompt_tokens":1123,"completion_tokens":2340,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":739,"completion_tokens_details":{"reasoning_tokens":2257}},"tokens_in":739,"tokens_out":2340,"duration_ms":18753,"temperature":1.0,"reasoning_tokens":2257,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T23:18:45.462973+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the same six stellar structures with the simulation domain extended through the outer convective envelope, or with a large radiative buffer above the shell, and rebuild the PDF of upper penetration distances; if the mean of the strong-overshooting Gamma component, the 90th-percentile length, or the ratio $\\ell^L_{\\rm so}/\\ell^U_{\\rm so}$ shifts by more than the quoted 1$\\sigma$ errors, the truncated boundary is distorting the central result.","supporting_citations":[{"cited_title":"2017, A&A, 604, A125","cited_arxiv_id":null,"evidence_quote":"supplies the EVT-based maximal overshoot length and the diffusion-coefficient formula $D_{\\rm EVT}=D_{\\rm MLT}(1-F(r))$ that the new mixture-based coefficient extends."},{"cited_title":"2022, MNRAS, 515, 4013","cited_arxiv_id":null,"evidence_quote":"extends the massive-shell simulations and is another source of the 0.2 lower/upper overshoot ratio used as a comparison."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"predicts a lower/upper overshoot ratio near 0.5 for thin shallow shells and overshoot filling the radiative zone in A-type stars; it motivates the claim that upper overshoot fills the whole layer."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the maximum-entropy rationale for choosing Gamma distributions as the components of the mixture model."}],"review_version":1}