{"id":"bd04f560-d7a3-4922-80c1-39f08ddfb388","arxiv_id":"1908.05462","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"A turbulent front in a rotating tank switches from advection to inertial-wave propagation scale by scale when the local Rossby number kU/(2Ω) crosses one.","lead":"This experiment tracks the turbulent front created by jets entering a rotating tank and finds that each eddy size switches from being carried by the flow to being carried by inertial waves when the local, scale-dependent Rossby number crosses unity. The result gives a concrete, testable rule for when rotation, rather than advection, moves momentum in rotating turbulence.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (4.3)'s max(U,v_g) model is not the correct superposition if inertial waves are advected by the mean flow; the lab-frame group velocity is U+v_g, and the authors' own Sec. 5 caveat admits axial advection of waves. The visual contour match does not yet exclude this alternative.","rationale":"The central claim of the paper is that the advective-to-propagative transition is local and occurs at Ro_k = 1, equivalently U(z) = 2Ω/k, and that the front advances at the maximum of the advection and inertial-wave group velocities. This claim is operationalized by Eq. (4.3), so the validity of that model is load-bearing. The reader's weakest-assumption analysis identified exactly this point: the model assumes no interaction between advection and propagation, and the authors' Sec. 5 caveat admits that axial advection of inertial waves could occur. A direct consequence of linear wave theory is that a wave packet in a moving fluid has absolute group velocity U + v_g, not max(U, v_g); this becomes important precisely where U and v_g are comparable, which is the transition region used to infer the Ro_k = 1 criterion. Because the paper's support for Eq. (4.3) rests on visual contour agreement and no quantitative residual analysis, the empirical evidence does not yet exclude the additive alternative. The z_T scaling law and the momentum-deficit interpretation of advection suppression are secondary and addressable, but the max-versus-additive question is the central correctness risk. The concern therefore keeps the manuscript in CONDITIONAL status; it does not, on its own, force rejection, because a straightforward re-analysis of the existing data can settle which transport law is actually observed.","tokens_in":16355,"tokens_out":8022,"duration_ms":83054,"concrete_test":"Re-analyze the existing PIV data by computing, for each wavenumber k and height z, two predicted arrival contours: (i) the paper's model, z(k,t) = ∫ max(U(z), 2Ω/k) dt from Eq. (4.3), and (ii) the additive advection-plus-propagation model, z(k,t) = ∫ [U(z) + 2Ω/k] dt, using the same measured U(z) from Eq. (3.2) or from the time- and space-averaged rotating PIV field. Quantify the misfit to the measured E(k,z,t) contours as a function of k and z, focusing on the band z/L ≈ 1.5–4 where U and v_g are comparable. If model (ii) fits as well as or better than model (i), the max/Ro_k = 1 criterion is not uniquely supported; if model (i) is significantly better, the additive advection hypothesis is ruled out.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The decisive weakness is in the transition model, Eq. (4.3). The paper claims that the front for each wavenumber k advances at max(U, v_g(k)) with v_g = 2Ω/k, and takes the agreement of this model with the measured spectral contours in Fig. 9 as evidence for the Ro_k = kU/2Ω = 1 criterion (Eq. 4.2). But linear wave theory in a moving medium says a wave packet's absolute group velocity is U + v_g to leading order, not max(U, v_g). If inertial waves are axially advected while they propagate, the front speed should be U + v_g whenever wave transport is active, which is always at least as large as max(U, v_g). The discrepancy between the max and additive predictions is largest precisely in the transition region U ≈ v_g, the region used to identify the Ro_k = 1 boundary. The comparison in Fig. 9 is visual, with contour overlays and no quantitative misfit reported, so the data do not currently distinguish the max model from the additive advection-plus-propagation model. The authors themselves concede in Sec. 5 that 'Axial advection of inertial waves could take place in our setup, but would be shadowed if advection was the fastest mechanism.' This caveat is not incorporated into Eq. (4.3), and if axial advection of waves is not negligible, the statement that the transition occurs when U ≈ 2Ω/k is not established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports an experimental study of the transport of turbulent fluctuations in a rotating fluid, using a four-jet forcing mechanism at the bottom of a tank mounted on a turntable and 2D-PIV measurements of the turbulent front that invades the quiescent fluid. In the non-rotating case the front is scale-independent and follows z − z0 ~ (τU0/L)^0.48, recovering the oscillating-grid front law of Dickinson & Long with an offset z0 of a few centimeters. Under background rotation the front becomes scale-dependent: low-wavenumber modes advance at the inertial-wave group velocity v_g = 2Ω/k while high-wavenumber modes are advected by the jet, with a narrow transition between the two regions. The authors propose that the transition occurs where the scale-dependent Rossby number Ro_k = kU(z)/2Ω equals unity (Eq. 4.2), equivalently where the local advection velocity U(z) equals v_g(k), and they encode this criterion in a front model in which each mode advances at the maximum of the advection and group velocities (Eq. 4.3). Support is provided by single-mode trajectory analysis (Fig. 10), the transition-height scaling zT/L ≈ 8.96 Ro_Q^1/2 (Eq. 4.1, Fig. 11), and a comparison of model contours with measured spectral energy contours (Fig. 9).","tokens_in":16782,"tokens_out":15589,"duration_ms":152825,"significance":"If the criterion is confirmed, the paper delivers a genuinely local and spectral statement about the advection-versus-wave transition in rotating turbulence: the regime boundary is Ro_k = 1, with faster (larger-scale) modes propagating and slower (smaller-scale) modes advected. This goes beyond the spatially averaged critical Rossby numbers (~0.4–0.5) reported in earlier grid studies and is falsifiable in other geometries, as the authors note in connection with the experiments of Burmann & Noir and the numerics of McDermott & Davidson. The experimental effort is a clear strength: five repetitions per condition with ~5% run-to-run standard deviation, six flow rates and four rotation rates spanning Ro_Q from 0.026 to 2.04, and a scale-resolved front-tracking method adapted from Kolvin et al. The transition threshold itself is not fitted; it follows from equating the measured jet-velocity profile with the theoretical group velocity 2Ω/k, and the fitted quantities elsewhere (front-law prefactors and exponents, transition-height prefactor) are reported with uncertainties.","major_comments":[{"comment":"The model-data comparison offered as the whole-spectrum validation of the criterion depends on an assumed transport rule, max(U, v_g), which is not derived in the paper and which is not the superposition predicted by linear wave theory if the mean flow advects the wave packets; in a moving medium the absolute group velocity is U + v_g to leading order. The two prescriptions differ by up to a factor of two precisely where U ≈ v_g, which is the transition zone that defines the Ro_k = 1 boundary, so the visual overlay of the dashed model contours on the measured spectral contours in Fig. 9 cannot discriminate between them; no quantitative misfit is reported anywhere in §4. I request (i) an arrival-time misfit metric as a function of k and z for the model, (ii) the same comparison with the additive U + v_g model, and (iii) a quantitative justification, based on the Sec. 3.2 momentum-deficit argument or otherwise, for neglecting axial advection of the wave packets in the wave-dominated region. The Sec. 5 admission that 'Axial advection of inertial waves could take place in our setup' marks exactly this assumption as the point that needs support; the mode trajectories in Fig. 10(b), which appear to converge to the pure v_g lines, are helpful but cover only the largest mode at high z. If axial advection of waves is not negligible, the statement that the transition occurs at U ≈ 2Ω/k is not established from the contour comparison.","section":"§4.3, Eq. (4.3), Fig. 9"},{"comment":"The paper should specify which velocity field supplies U(z) in the criterion and in the model. Eq. (4.3) integrates with max_x u(x,z,t')·e_z taken from the rotating PIV fields, which include wave-induced and turbulent fluctuations, whereas the physical argument behind Eq. (4.1) and the phrasing of the transition condition in Eq. (4.2) are given in terms of the mean jet advection profile (the non-rotating law U/U0 ≈ 6.41×10^-2 (z/L)^-1.07 and the steady-jet scaling U ~ d/z). These are not the same quantity: rotation and wave emission modify the jet, and taking the maximum over x of the instantaneous vertical velocity in the wave region could partially track wave-induced velocities, so the criterion risks being confirmed by construction. Please clarify whether U in Eq. (4.2) is the measured instantaneous maximum, the ensemble-averaged jet profile, or the non-rotating profile, and quantify how sensitive the Ro_k = 1 boundary is to this choice.","section":"§4.2–4.3, Eqs. (4.1)–(4.3)"},{"comment":"The statement that the Fig. 9 comparison 'tests this criterion on the entire spectrum' is partly circular, since the switching condition of Eq. (4.3) is the criterion Ro_k = 1 itself; agreement of the model with the contours validates the piecewise max-rule synthesis, but it does not independently certify the threshold value. The genuinely independent evidence for the threshold comes from the mode-trajectory convergence in Fig. 10(b) and from the zT ~ Ro_Q^1/2 scaling in Fig. 11, and the interpretation of Fig. 9 should be framed accordingly. To make the contour comparison informative about the threshold, the authors could vary the crossover condition (e.g., replacing v_g in the max rule by c·v_g) and report how the predicted contours, and a misfit metric, respond; this would also quantify how sharply the data constrain the Ro_k = 1 location.","section":"§4.3, test of the criterion"}],"minor_comments":[{"comment":"The sentence 'A clear separation exists between scales advected by inertial waves and by the local mean flow' reads as though inertial waves advect scales; it should read 'between scales propagated by inertial waves and scales advected by the local mean flow'.","section":"§5, item (i)"},{"comment":"The phrase 'not only does advection itself is suppressed' is ungrammatical; the abstract and conclusion contain similar wording, and the manuscript would benefit from a careful proofread.","section":"§3.2, text near Fig. 7"},{"comment":"The points departing from the scaling for Ro_Q > 3×10^-1 are the ones that would anchor the high-Rossby end of the law; please report the range of β over which zT was verified to be stable and show that the conclusion is unchanged when those points are included.","section":"Fig. 11 and Eq. (4.1)"},{"comment":"The quantities zIW, Δz and zT are used in the caption without definition; please define them in the caption for self-containment.","section":"Fig. 10 caption"},{"comment":"The scale-independence claim is demonstrated for the first six Fourier modes (about 0.3–2 in Lk/2π); the conclusion in §5 states the law holds 'regardless of their transversal wavenumber', which is stronger than the displayed range; please add the resolved k-range to the claim.","section":"§3.1 and §5"}],"recommendation":"major_revision","confidential_remarks":"This is a solid experimental paper well within the scope of JFM. The crucial gate is whether the authors can turn the visual contour match of Fig. 9 into a quantitative discrimination between the max rule of Eq. (4.3) and the advected-wave prediction U + v_g; the Sec. 5 caveat shows they are aware of the ambiguity, but the central claim currently rests on the unquantified comparison. If the revision delivers a misfit metric and a clear statement of which velocity field enters the criterion, I expect the paper to be acceptable. I would also encourage the editor to make sure the revision is examined by someone familiar with inertial-wave diagnostics in experiments, since the wave-identification claims (chevron patterns, reflected-wave timing) would benefit from scrutiny that a generalist referee may not provide."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Let me cut to it: this paper's actual contribution is the claim that the advection-to-inertial-wave transition in rotating turbulence is local in wavenumber and space, set by Ro_k = kU(z)/2Ω = 1. That is new and worth taking seriously. The experiments are careful: they recover the non-rotating t^{1/2} front law, the mode-by-mode tracking is convincing, and the spectral splitting—low k rides the inertial-wave group velocity while high k is advected—is clearly visible in the contours. The authors also deserve credit for stating the main caveat at the end: axial advection of inertial waves 'could take place' but would be shadowed by faster advection.\n\nThe soft spot is exactly that caveat. The model in Eq. (4.3) advances the front at max(U, v_g). In a moving fluid, a wave packet's absolute group velocity is U + v_g to leading order, not max(U, v_g). Those two rules differ most when U ≈ v_g—precisely the region used to fix Ro_k = 1. The comparison in Fig. 9 is visual, with no reported misfit, so the data do not yet distinguish the max rule from an advected-wave, additive rule. The authors are aware of the issue but don't fold it into the model or test it. That's the one load-bearing uncertainty.\n\nOther soft spots are minor: the zT ∝ Ro_Q^{1/2} law is fit through a handful of points with a couple of outliers excluded, and the advection-suppression mechanism is inferred from momentum-conservation arguments rather than measured directly. Neither shakes the central criterion.\n\nIf I were refereeing, I would ask for a quantitative comparison between the model and the measured arrival times, including a test of the additive alternative, and a paragraph or two on when wave advection is safely negligible. Those are standard fixes, not showstoppers.\n\nWorth a serious referee? Yes. The result is important enough for the rotating-turbulence subfield, and the authors are honest about their simplifications. I'd read the revision.","headline":"Scale-dependent transition criterion at Ro_k≈1 is novel and plausible, but the max-velocity model needs a quantitative test against an additive wave-advection alternative.","tokens_in":17265,"tokens_out":4502,"would_cite":true,"duration_ms":44892,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["47.27.-i","47.35.-i"],"model":"deepseek-v4-flash","headline":"Advection yields to inertial waves when the local Rossby number hits one.","keywords":["inertial waves","rotating turbulence","turbulent front","scale-dependent Rossby number","advection","group velocity","jet-driven turbulence","spectral front tracking"],"falsifier":"Track a single Fourier mode $k$ in a rotating jet while independently varying the mean jet speed and rotation: if the height where the mode departs from the pure-advection curve does not follow $U(z) = 2\\Omega/k$, the criterion fails. A complementary check is to measure the axial phase speed of individual inertial-wave packets; any Doppler shift showing that waves are advected by the mean flow would indicate the maximum-of-two-speeds rule is only a kinematic approximation.","tokens_in":16180,"feed_emoji":"🌀","tokens_out":8735,"duration_ms":74879,"temperature":0.7,"pith_summary":"This paper tries to establish exactly where and when rotation stops turbulence from spreading by advection and starts carrying it by inertial waves. Tracking the turbulent front made by four jets entering a rotating tank, the authors find the changeover is local in space and in scale: each transverse wavenumber $k$ switches when the local advection speed $U(z)$ drops to the group velocity of inertial waves, $2\\Omega/k$. The criterion matters because rotating turbulence appears in oceans, atmospheres, and industrial flows, where knowing which mechanism carries energy at which scale controls momentum spreading and the formation of columnar structures. In the same experiment the authors show that rotation suppresses advection itself, and that a front model built on the faster of the two speeds reproduces the measured spectral contours for $z/L > 1.5$.","feed_headline":"Advection yields to inertial waves when local Rossby number hits one","feed_subtitle":"The switch happens where the jet's local speed equals the inertial-wave group velocity, scale by scale.","key_machinery":"The key object is the speed comparison at each scale, expressed as $U(z) = V_g(k) = 2\\Omega/k$, equivalently $\\mathrm{Ro}_k = kU(z)/(2\\Omega) = 1$. The authors embody it in a Lagrangian front model, $z(k,t) = \\int_0^t \\max\\{\\max_x u(x,z(t'),t')\\cdot e_z,\\, v_g(k)\\}\\;\\mathrm{d}t'$, which says a Fourier mode advances at whichever is faster, local advection or inertial-wave group velocity, and they test its predicted arrival-time contours against the measured spectral energy front. The measurement machinery tracks arrival times of each horizontal wavenumber $k$ by thresholding the spectral energy $E(k,z,t)$, the same scale-by-scale approach used in earlier rotating-front experiments.","core_discovery":"The central claim is that the transition between advective and propagative transport of turbulent fluctuations is set by a local, scale-dependent Rossby number: fluctuations of wavenumber $k$ travel advectively where $\\mathrm{Ro}_k = k U(z)/(2\\Omega) > 1$ and switch to inertial-wave propagation where $\\mathrm{Ro}_k < 1$, so the boundary sits at $U(z) = 2\\Omega/k$, the equality between the local mean-flow speed and the axial group velocity of inertial waves. In the jet experiment the largest scale $k_1$ switches at height $z_T/L \\simeq 8.96\\,\\mathrm{Ro}_Q^{1/2}$, consistent with a jet whose centreline velocity decays as $U(z)\\sim U_0 d/z$. The paper further reports that rotation suppresses advection itself, because inertial waves carry momentum ahead of the advected front, and that after waves reflect from the top wall and interfere with upward waves, wave transport weakens and advection resumes at a slower pace.","pith_inferences":["The same $U(z) = c_g(k)$ balance may apply to other wave-bearing turbulent systems, such as stratified turbulence with internal gravity waves, where the group velocity has a different wavenumber dependence; the transition would then be a spectral analogue of a critical layer.","A direct numerical simulation with controlled mean shear and rotation could measure the spectral energy flux across the $\\mathrm{Ro}_k = 1$ boundary, testing whether the front model's success reflects genuine wave transport or only the kinematics of taking the faster of two speeds.","The authors' own caveat that axial advection of inertial waves could be shadowed by faster advection suggests a Doppler-shift test: phase measurements along the rotation axis would reveal whether waves are carried by the mean flow, which the front-tracking alone cannot resolve.","The reflected-wave interference phase implies that in taller or open domains the third-phase slowdown should disappear, a prediction that could be checked by varying the vessel height."],"forward_implications":["In any localized turbulent region embedded in a rotating flow, large scales (small $k$) become wave-dominated first, while small scales stay advective until the local mean flow slows enough.","For a jet, the transition height scales as $z_T \\sim (U_0 d/\\Omega)^{1/2}$, giving a testable prediction for other forcing geometries once the advective velocity profile is known.","Rotation does not merely add a transport channel: by carrying momentum ahead of the advected front, inertial waves suppress the usual turbulent spreading and slow the advected front.","Reflected inertial waves can cancel upward momentum transport, so in confined rotating vessels wave transport loses efficiency once reflections arrive, a point relevant to spin-up and quasi-two-dimensional flows.","The spectral criterion offers a way to separate advective from wave contributions in statistically steady rotating turbulence, where the two mechanisms coexist."],"supporting_citations":[{"why":"Supplies the rotating oscillating-grid front law and the critical-Rossby transition that this paper extends to scale-by-scale motion.","marker":"Dickinson & Long (1983)"},{"why":"Supplies the scale-by-scale front-tracking method and the strong-rotation limit where fluctuations propagate at inertial-wave group velocity.","marker":"Kolvin et al. (2009)"},{"why":"Provides the theoretical $z\\sim t^{1/2}$ turbulent-front law that the non-rotating jet measurements recover.","marker":"Long (1978)"},{"why":"Establishes a critical Rossby number around 0.4 for grid-turbulence transition, the context this paper's local criterion builds on.","marker":"Staplehurst et al. (2008)"},{"why":"Shows eddies elongate along the rotation axis at a speed set by rotation, the basis for inertial-wave axial transport.","marker":"Davidson et al. (2006)"},{"why":"Gives the jet centreline velocity profile $U(z)\\sim U_0 d/z$ used to derive the $z_T \\sim \\mathrm{Ro}_Q^{1/2}$ scaling.","marker":"Pope (2000)"},{"why":"Proposes a similar balance for the breakdown of inertial-wave propagation in a spin-up cylinder.","marker":"Burmann & Noir (2018)"},{"why":"Provides numerical evidence that the transition can be spatially localised with a critical Rossby number near 0.5.","marker":"McDermott & Davidson (2019)"}],"fun_headline_variants":["Transition to inertial waves set by local Rossby number of one","Advection yields to waves when scale-wise Rossby number hits unity","Local Rossby number = 1 marks advection-to-wave switch","Inertial waves take over where advection speed matches wave group velocity","Rotation suppresses advection, inertial waves dominate below Ro=1"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that each scale advances at the faster of the local advection speed and the inertial-wave group velocity, with the two mechanisms acting independently; if inertial waves are themselves carried by the mean flow, or if nonlinear interactions change arrival times, the match between this model and the measured fronts would not certify the $\\mathrm{Ro}_k = 1$ criterion.","fun_headline_variants_meta":{"raw":{"variants":["Transition to inertial waves set by local Rossby number of one","Advection yields to waves when scale-wise Rossby number hits unity","Local Rossby number = 1 marks advection-to-wave switch","Inertial waves take over where advection speed matches wave group velocity","Rotation suppresses advection, inertial waves dominate below Ro=1"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000322,"raw_usage":{"total_tokens":1826,"prompt_tokens":976,"completion_tokens":850,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":592,"completion_tokens_details":{"reasoning_tokens":760}},"tokens_in":592,"tokens_out":850,"duration_ms":7939,"temperature":1.0,"reasoning_tokens":760,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:13:16.186522+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Track a single Fourier mode $k$ in a rotating jet while independently varying the mean jet speed and rotation: if the height where the mode departs from the pure-advection curve does not follow $U(z) = 2\\Omega/k$, the criterion fails. A complementary check is to measure the axial phase speed of individual inertial-wave packets; any Doppler shift showing that waves are advected by the mean flow would indicate the maximum-of-two-speeds rule is only a kinematic approximation.","supporting_citations":[{"cited_title":"& Long, R.R","cited_arxiv_id":null,"evidence_quote":"Supplies the rotating oscillating-grid front law and the critical-Rossby transition that this paper extends to scale-by-scale motion."},{"cited_title":", Cohen, K","cited_arxiv_id":null,"evidence_quote":"Supplies the scale-by-scale front-tracking method and the strong-rotation limit where fluctuations propagate at inertial-wave group velocity."},{"cited_title":"1978 Theory of turbulence in a homogeneous fluid induced by an oscillating grid","cited_arxiv_id":null,"evidence_quote":"Provides the theoretical $z\\sim t^{1/2}$ turbulent-front law that the non-rotating jet measurements recover."},{"cited_title":"& Dalziel, B","cited_arxiv_id":null,"evidence_quote":"Establishes a critical Rossby number around 0.4 for grid-turbulence transition, the context this paper's local criterion builds on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows eddies elongate along the rotation axis at a speed set by rotation, the basis for inertial-wave axial transport."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the jet centreline velocity profile $U(z)\\sim U_0 d/z$ used to derive the $z_T \\sim \\mathrm{Ro}_Q^{1/2}$ scaling."},{"cited_title":"& Noir, J","cited_arxiv_id":null,"evidence_quote":"Proposes a similar balance for the breakdown of inertial-wave propagation in a spin-up cylinder."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides numerical evidence that the transition can be spatially localised with a critical Rossby number near 0.5."}],"review_version":1}