{"id":"0ff6c927-5ceb-44ad-bef9-c8f41abb971c","arxiv_id":"2512.15536","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Faraday surface waves produce a non-monotonic mixing rate across a weak density interface: mixing accelerates as the interfacial barrier drops, then decelerates as wave energy decays with depth.","lead":"Surface waves driven by vertical shaking mix a fresh-water layer lying on salt water, and the mixing rate first speeds up as the interface weakens, then slows as wave energy fades with depth. Tank experiments, simulations, and a simple model trace this non-monotonic behavior to a race between barrier reduction and energy injection.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Barrier-vs-energy mechanism is not independently tested; the only support is a 1D closure with free parameters, so the non-monotonicity is not yet causally demonstrated.","rationale":"I read the paper as making two distinct claims: (i) the experimental observation of a non-monotonic entrainment response with initial stratification depth and forcing amplitude, and (ii) the causal explanation in terms of barrier reduction versus depth-decaying energy input. Claim (i) is reasonably supported by the experimental Table 2 and Figure 24e, and the DNS reproduces the trend. Claim (ii), however, rests almost entirely on the 1D model in Section 6.1.2. The closure for the turbulent diffusivity uses a Richardson-number suppression function (6.7) whose constants are explicitly chosen to reproduce the target behavior, and the critical Richardson number Ri_crit is never assigned. A model with a free threshold cannot independently confirm the mechanism. In addition, the experimental and DNS sweeps vary h_init, which simultaneously changes the depth-dependent energy input and the barrier strength; nothing in the paper separates these two factors. The passive-tracer DNS is the right idea, but it is a single case, and the non-monotonic panel (Figure 28f) comes from the tuned model, not from a multi-depth passive-tracer DNS. This does not undermine the observational result, but it does mean the paper has not isolated the proposed mechanism. The conditional verdict is therefore appropriate: the model should be treated as a descriptive illustration, or the barrier-vs-energy mechanism should be tested by direct numerical separation of the two effects.","tokens_in":36679,"tokens_out":12803,"duration_ms":135100,"concrete_test":"Run a single DNS with the same free-surface Faraday forcing and initialize several passive tracers at the depths of Series A (h_init = 2.5, 5, 7.5, 10, 12.5, 15 cm), all advected by the same velocity field. If the non-monotonic h_mixed(h_init) curve persists without an active density barrier, the barrier mechanism is not needed; if the curve is monotonic, it supports the barrier hypothesis.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that non-monotonic mixing results from a balance between barrier lowering and depth-decaying energy input. The evidence for this mechanism is underdetermined. The only multi-case comparison that removes the barrier is a single passive-tracer DNS (Fig. 27), not a sweep over h_init; the non-monotonic curve in Fig. 28f is produced by the 1D model with C_D given by (6.6)-(6.7), where C_barrier=0.1 and n=4 are explicitly 'chosen to mimic the trends observed in the experiments' and Ri_crit is never assigned a value. With a free threshold, the model can be tuned to produce or suppress the non-monotonicity, so it cannot validate the causal story. Moreover, no experiment or DNS varies the interfacial barrier while holding the wave-energy input at the interface fixed; varying h_init changes both factors simultaneously. Thus, the abstract's causal explanation is a plausible interpretation, not an established mechanism.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies mixing across a miscible density interface driven by Faraday surface waves, using laboratory experiments, 3D DNS, linear stability analysis, and a 1D turbulent-diffusion model. The central empirical claim is that the long-term entrainment of the interface is non-monotonic in the initial interface depth: for shallow (or very deep) interfaces the mixed layer deepens more than for intermediate depths. The authors interpret this as a competition between two effects: progressive lowering of the interfacial energy barrier as the upper layer homogenizes, and exponential decay of the wave-induced energy flux with depth. A 1D model with a Richardson-number-dependent turbulent diffusivity (Eqs. 6.1–6.7) is shown to reproduce the qualitative non-monotonic behavior.","tokens_in":37009,"tokens_out":6448,"duration_ms":66009,"significance":"If the mechanism is established, the work would provide a new physical picture for wave-driven mixing of stratified systems: a barrier-reduction/energy-input balance. The experimental dataset is substantial, the DNS code is openly available, the measured Faraday growth rates agree with linear theory (Fig. 11b), and the energy diagnostics are carefully constructed. However, the causal explanation is not independently tested. The 1D model's non-monotonic curve is obtained with parameters (C_barrier=0.1, n=4, and an unassigned Ri_crit) that are explicitly chosen to mimic the experiments, so the model agreement does not validate the mechanism. The experimental evidence itself, while suggestive, lacks error bars and does not isolate the two proposed factors. The paper's main contribution is therefore an important experimental observation with a plausible but under-supported interpretation.","major_comments":[{"comment":"The central mechanism—non-monotonic mixing due to competition between barrier lowering and depth-decaying energy input—is supported by a 1D model whose closure is calibrated to reproduce the target behavior. C_barrier=0.1 and n=4 in f(Ri_inside) are explicitly 'chosen to mimic the trends observed in the experiments', and Ri_crit is never assigned a numerical value. Thus Fig. 28(f) cannot serve as independent evidence for the causal explanation. Please provide a sensitivity analysis over the full parameter set and/or calibrate the constants from DNS diagnostics (e.g., a flux–Richardson relation) without directly targeting the non-monotonic h_mixed(h_init) curve. Alternatively, the abstract and conclusions should soften the causal claim to a plausible interpretation.","section":"§6.1.2, Eqs. (6.6)–(6.7) and Fig. 28(f)"},{"comment":"The experimental demonstration of non-monotonic entrainment rests on h∞ versus h_init for the large-forcing series (F≈0.49). No repeats or error bars are reported, and the h_init sweep simultaneously changes the energy reaching the interface and the initial barrier. The passive-tracer DNS in §6.1.1 (Fig. 27) is a single case (h_init=10 cm, F=0.49), not a sweep, so it cannot show that removing the barrier eliminates the non-monotonicity. A controlled test—e.g., varying the Atwood number or initial interface thickness at fixed h_init, or a DNS sweep with a passive scalar—is needed to isolate the proposed effects.","section":"§5.2, Fig. 24(e), Table 2"},{"comment":"The DNS quantitatively overpredicts entrainment relative to experiments (Fig. 26c), and the paper attributes this to Sc=1. Since the DNS is used to support the qualitative mechanisms and to motivate the 1D model, this discrepancy should be explicitly discussed in terms of its impact on the central claim. If the non-monotonicity is not reproduced in DNS at Sc=1 (or only weakly), the experimental result is the sole evidence; the manuscript should state this and consider Sc-sensitivity tests.","section":"§5.3, Fig. 26(c) and §7"}],"minor_comments":[{"comment":"The caption contains blank placeholders: 'indicates L_sat from (4.11), while corresponds to the wave model...' The marker symbols are missing from the rendered text.","section":"Fig. 10 caption"},{"comment":"Several entries in Table 2 are missing or misaligned (e.g., runs 29 and 27 lack values for some columns). Please reformat the table so all quantities are clearly associated with the correct runs.","section":"Table 2"},{"comment":"The phrase 'turbulence intensity to 10%' should specify whether this is relative to the surface wave orbital velocity or to the turbulent velocity scale u_turb, to avoid ambiguity.","section":"§6.1.1"},{"comment":"The abstract states the mechanism as a finding ('revealing a balance'), while the conclusion more cautiously says the model provides 'a simple way to contextualize the competing dynamics'. Please align the wording with the level of support.","section":"Abstract/Conclusions"}],"recommendation":"major_revision","confidential_remarks":"The experimental dataset and DNS are valuable and likely to be cited. The main issue is that the mechanism claim is overinterpreted relative to the evidence: the 1D model is tuned to reproduce the non-monotonicity, and no experiment/DNS isolates the two competing factors. I recommend requiring either independent validation (e.g., a parameter sweep with passive scalar or a flux–Richardson calibration) or a clear re-framing of the model as hypothesis-generating rather than confirmatory. The paper is otherwise well-written and the code availability is commendable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague, read this if you care about wave-driven mixing of weak density interfaces. The key new observation is the non-monotonic entrainment with initial interface depth: shallow interfaces get pushed down more than intermediate ones, in both tank experiments and Basilisk DNS. That observation looks solid — it is visible in the raw h_mixed(t) traces and the h_init sweep in Table 2, and it is not in the prior Faraday-mixing work from this group or from Liu & Lin. The paper also does a good job characterizing the surface dynamics: measured growth rates match linear theory, the POD decomposition of the free surface and interface is clean, and the energy analysis (potential energy, dissipation, mixing efficiency) is thoughtful. Credit also for making the source code and simulation setups available on the Basilisk site.\n\nThe soft spot is exactly where the stress-test note lands. The 'barrier vs energy' explanation is not independently tested. The only case that removes the stratification barrier is a single passive-tracer DNS (Fig. 27), not a sweep. Varying h_init changes both the barrier and the energy reaching the interface at the same time, so the non-monotonicity itself does not isolate the mechanism. The 1D model reproduces the non-monotonic curve, but the closure constants (C_barrier=0.1, n=4) are explicitly chosen to mimic the experiments, and Ri_crit is never assigned. That makes the model a descriptive illustration, not a validation of the causal story. I would not call this fatal — the paper is honest about the model's role — but the abstract's causal claim goes beyond the evidence.\n\nTwo more proportional quibbles: DNS at Sc=1 entrains faster than experiments, so the quantitative validation is soft; and the experiments have no repeats or error bars, which matters when the headline result is a non-monotonic curve. Neither undermines the basic observation.\n\nWho is this for? People working on stratified mixing, ocean mixed-layer parameterizations, and sloshing applications. They will get a useful mechanism catalog and a clear description of the phenomenology. The paper deserves a serious referee: the observation is new and the modeling, while tuned, is a reasonable starting point. A good referee should ask for either a parameter-free scaling prediction or an experiment that varies the barrier while holding the energy input at the interface fixed.","headline":"The non-monotonic entrainment looks real, but the causal mechanism is supported more by a tuned 1D closure than by direct experiments.","tokens_in":37437,"tokens_out":1855,"would_cite":true,"duration_ms":21837,"reading_group":"maybe","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 reports a non-monotonic mixing rate in Faraday-wave-driven stratified fluids: entrainment first accelerates as homogenization lowers the interfacial density jump, then decelerates as the deepening interface receives exponentially","keywords":["Faraday waves","stratified fluid mixing","entrainment","turbulent diffusivity","Richardson number","direct numerical simulation","parametric instability","mixed layer"],"falsifier":"Measure the entrainment velocity d hmixed/dt and the interfacial density jump continuously in one experimental series with varying initial interface depth. If the mixing rate never accelerates before it decays—or if the measured turbulent kinetic energy flux at the interface does not decay exponentially with interface depth—the barrier-reduction/energy-decay explanation is falsified. A cleaner variant: the passive-tracer case without stratification shows no rise phase; any equivalent configuration in which the active-scalar entrainment curve matches the passive-scalar one would rule out the ba","tokens_in":36589,"feed_emoji":"🌊","tokens_out":5827,"duration_ms":61588,"temperature":0.7,"pith_summary":"This paper establishes that Faraday surface waves—standing waves excited by vertical vibration—mix a weak density interface at a rate that first rises and then falls. The rise comes from the erosion of the interfacial energy barrier: as the upper layer homogenizes, the density jump shrinks and less energy is needed to mix across it. The fall comes from geometry: the interface deepens as it is erased, and the surface-driven turbulence that reaches it decays exponentially with depth. The authors show this in matched laboratory experiments and direct numerical simulations, and they encode the two competing effects in a one-dimensional turbulent-diffusion model whose diffusivity drops sharply once a local Richardson-number threshold is crossed. A sympathetic reader would care because this gives a simple, testable explanation for why wave-driven entrainment stalls—an outcome relevant to sloshing LNG tanks, stratified lakes, and the ocean mixed layer.","feed_headline":"Wave-driven mixing speeds up, then stalls","feed_subtitle":"A weaker density jump lowers the mix cost, but the retreating interface drains its own energy supply.","key_machinery":"The key machinery is the one-dimensional diffusion equation for the horizontally averaged concentration, closed by a turbulent eddy diffusivity D_turb = l_turb u_turb C_D, with u_turb decaying exponentially with depth. The dimensionless coefficient C_D takes a constant value C_tracer in the mixed layer and drops inside the interface via the attenuation function f(Ri_inside) = 1 / (1 + C_barrier (Ri_inside / Ri_crit)^n), where Ri_inside is a modified Richardson number built on the local gradient thickness L' of the interface rather than the mixed-layer scale. This modified Richardson number is what allows the model to feel the difference between a thin, sharp interface and a thick, diffuse on","core_discovery":"The central finding is that Faraday-wave-driven entrainment is controlled by a competition between barrier reduction and energy decay. As long as the density interface is close to the free surface, breaking crests and collapsing cavities inject turbulent energy that homogenizes the upper layer; this lowers the Atwood number and the potential energy cost of mixing, so the entrainment rate accelerates. Once the interface has been pushed deeper, the exponentially decaying wave-induced velocity field (scaling as e^{-kz}) delivers less energy than the weakened but still present barrier demands, and the interface decouples from the surface forcing and stalls. In the authors' words, the mixing rate","pith_inferences":["If the barrier-versus-energy balance is generic, the same non-monotonicity should appear in other surface-forced stratified systems, for instance wind-wave-driven mixing layers or sloshing tanks under changing fill levels; the 1D model could be adapted by rescaling the energy-decay law.","The model's threshold Ri_crit, left numerically unspecified, could be pinned down experimentally by measuring the interface depth at which the turbulent flux vanishes; identifying Ri_crit with a standard critical Richardson number would make the model closer to parameter-free.","A testable extension is to follow the scalar dissipation rate: the model implies that the peak in diapycnal mixing should lag the peak in surface kinetic energy by an amount that grows with initial interface depth, which the authors observe in DNS and which could be verified in experiments.","The exponential decay of u_turb with depth is the load-bearing ingredient for the stall; if instead the energy decay were algebraic or if a secondary instability kept injecting energy at depth, the non-monotonicity would be attenuated or lost."],"forward_implications":["The mixed-layer depth in such systems saturates at a finite value rather than growing indefinitely; the saturating value is set by the balance between barrier cost and depth-decaying energy input.","Entrainment rate is not a monotonic function of initial stratification depth: intermediate initial depths can produce the deepest final mixed layers for a fixed forcing.","Weak forcing still produces mixing bursts once the Faraday wave reaches finite amplitude, because the secondary parametric instability and interfacial breaking persist even at small forcing intensities.","The qualitative dynamics—acceleration, peak, stall—can be captured without resolving three-dimensional turbulence, using a 1D model with two fitted constants and one threshold.","The model's asymmetry (constant diffusivity above the interface, attenuating diffusivity inside) predicts that mixing is mostly one-sided: the upper layer homogenizes while the lower layer remains nearly unmixed, consistent with the measured density profiles."],"fun_headline_variants":["Faraday waves mix fast near surface, then stall","Mixing accelerates, then stalls as interface sinks","Standing waves stir then stall: barrier vs energy","Breaking crests boost mixing until energy fades","Wave stirring peaks then fades as interface deepens"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The explanation rests on a turbulent-flux closure whose constants (C_tracer=4, C_barrier=0.1, n=4) and the threshold Ri_crit are chosen to make the 1D model match the experiments and DNS; without a parameter-free derivation, the agreement is consistent with, but does not independently prove, the barrier-versus-energy mechanism.","fun_headline_variants_meta":{"raw":{"variants":["Faraday waves mix fast near surface, then stall","Mixing accelerates, then stalls as interface sinks","Standing waves stir then stall: barrier vs energy","Breaking crests boost mixing until energy fades","Wave stirring peaks then fades as interface deepens"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000531,"raw_usage":{"total_tokens":2359,"prompt_tokens":676,"completion_tokens":1683,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":420,"completion_tokens_details":{"reasoning_tokens":1609}},"tokens_in":420,"tokens_out":1683,"duration_ms":12704,"temperature":1.0,"reasoning_tokens":1609,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T15:46:03.772496+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the entrainment velocity d hmixed/dt and the interfacial density jump continuously in one experimental series with varying initial interface depth. If the mixing rate never accelerates before it decays—or if the measured turbulent kinetic energy flux at the interface does not decay exponentially with interface depth—the barrier-reduction/energy-decay explanation is falsified. A cleaner variant: the passive-tracer case without stratification shows no rise phase; any equivalent configuration in which the active-scalar entrainment curve matches the passive-scalar one would rule out the ba","supporting_citations":[],"review_version":1}