{"id":"3d4858b1-7153-4887-b400-29dbc1cc48a1","arxiv_id":"1908.11784","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Edge trajectories in a spatially developing Blasius boundary layer can be tracked three times longer than before with a moving box, revealing streak switching and coexistence of bypass and Tollmien-Schlichting transition routes near the edge.","lead":"This paper uses a moving computational box to track the edge state, the boundary between laminar and turbulent dynamics, in a Blasius boundary layer over thousands of time units. It reports the first observation of streak switching in this flow and finds that at long times both bypass and Tollmien-Schlichting routes to turbulence coexist near the edge.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Moving-box equivalence to a fixed long domain is asserted without metrics; all long-time edge observations depend on it.","rationale":"The reader's weakest assumption is exactly the moving-box equivalence, and I agree with that assessment. I considered whether the abstract's overstatement about TS waves coexisting on the edge trajectory itself is more damaging; the TS waves are in fact observed on the below-edge bracketing trajectory after the streak decays (Fig 6 right, Fig 7), while the edge trajectory is only tracked to t≈4700. This is a real interpretive overreach, but it is secondary: if the moving-box equivalence were demonstrated, the core numerical observations on the bracketing trajectories would stand, and the reinterpretation of the edge as separating two routes is a conceptual proposal. Conversely, if the moving box is not equivalent, even the moderate-time edge trajectory and the streak-switching claim are unsupported. The validation sentence in Section 2.2 is the only direct evidence offered for equivalence, and no code or data are provided for independent checks. The absence of metrics is particularly important because the moving box changes the fringe target profile in time, which is not a symmetry of the Blasius base flow. A targeted reproduction of the stated validation would settle the concern. I therefore keep the reader's CONDITIONAL verdict (no change).","tokens_in":13642,"tokens_out":5889,"duration_ms":57031,"concrete_test":"Recompute the edge trajectory for at least one full switching cycle (e.g., t=0-3300) with a fixed domain Lx=12000 at the stated resolution, and compare pointwise with the moving-box run restricted to the common spatial overlap. Quantify the maximum relative difference in a(t) and xG(t), the timing of the second streak-switching event, and the amplitude/wavelength of any downstream wake. Also run the below-edge bracketing trajectory to t≈12000 in both frames and compare the TS wave emergence time and phase speed (Fig 8). If the normalized field difference is larger than the bisection precision (δλ/λ* ≈ 1e-4) or if the switching/wave timings differ by more than a few percent, the moving box is not equivalent and the long-time claims fail.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claims—long-time edge tracking, streak switching, and coexistence of bypass and TS routes—all rely on the moving box reproducing the dynamics of a much longer fixed domain. Section 2.2 states 'The present computation of the edge trajectory using the moving box technique was validated against the case of a non-moving domain of size (Lx,Ly,Lz)=(12000,60,100)' but gives no quantitative comparison, no convergence data, and no description of how the validation was performed. The issue is not routine Galilean invariance: in a spatially developing Blasius layer the base flow is not translation-invariant, and the moving box updates the fringe target profile in time (Appendix A, x0←x0+cbox tm). The equivalence of the time-dependent fringe forcing in the moving frame to the physical inflow/outflow of a long fixed domain is a nontrivial assumption. If the moving box artificially modifies the streak core, the wake where TS waves grow, or the fringe-induced inflow noise, the streak-switching timing and the later TS wave emergence could be numerical artifacts rather than physical edge dynamics. The one-sentence validation is insufficient support for results that are not otherwise reproduced.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents large-scale edge-tracking simulations in a spatially developing Blasius boundary layer, using a novel moving-box technique that advects the computational domain downstream at piecewise-constant speed. The authors report that edge trajectories can be bracketed for about 4700 time units, observe streak-switching dynamics (claimed as a first for the Blasius boundary layer), and find that slightly off-edge trajectories evolve either into streak breakdown (bypass route) or into Tollmien–Schlichting wave growth (classical route) at later times. The paper also compares the measured TS wave phase speed with linear stability theory and uses the results to argue for a reinterpretation of the edge as a separator of the two transition routes, with practical relevance to bypass-transition experiments.","tokens_in":13829,"tokens_out":8225,"duration_ms":69461,"significance":"If the moving-box method is truly equivalent to a much longer fixed-domain computation, this is a significant methodological contribution that enables long-time studies of localized coherent structures in spatially developing flows. The streak-switching observation extends earlier results from the asymptotic suction boundary layer and from symmetric Blasius computations to an unconstrained Blasius boundary layer, and the quantitative TS wave speed comparison is a convincing check. The conceptual interpretation of the edge in a flow whose base state is linearly unstable is thought-provoking and likely to stimulate further work. However, the strength of these claims is currently limited by the very brief, non-quantitative validation of the moving-box technique and by the fact that the coexistence of the two routes is demonstrated only on off-edge trajectories beyond the tracking limit.","major_comments":[{"comment":"The sentence 'The present computation of the edge trajectory using the moving box technique was validated against the case of a non-moving domain of size (Lx,Ly,Lz)=(12000,60,100)' is the only evidence for the central enabling assumption, yet no quantitative comparison, convergence data, or description of the validation procedure is given. Because the Blasius base flow is not Galilean-invariant and the fringe forcing becomes time-dependent through x0(t) (Appendix A), the equivalence between the moving box and a fixed long domain is a nontrivial assumption, not a routine change of reference frame. All long-time results -- the streak-switching cycles and the apparent coexistence of TS waves -- are produced with the moving box, so the paper's main claims depend on this equivalence. Please provide a quantitative validation: e.g., compare time series of the bisection observable a(t), the center-of-mass position x_G(t), and the two bracketing trajectories over t in [0,4700] between the moving-box run and the fixed-domain run, and report the level of agreement. The statement that 'different histories of cbox yield exactly the same results' should also be substantiated with actual trajectories.","section":"Section 2.2"},{"comment":"The abstract states that 'At long enough times, TS waves co-exist with the coherent structure characteristic of edge trajectories,' but the TS wave growth is actually observed on an off-edge trajectory (the δλ<0 case, Fig. 6 right) after the streak structure begins to decay, and the edge trajectory itself is only tracked up to t≈4700. The paper's own discussion in Section 4 acknowledges this limitation ('we have no further information on the nature (bypass or classical) of the trajectories within these 2%'). As it stands, the evidence supports coexistence of the two routes for trajectories infinitesimally below the edge, not on the edge trajectory itself. Please either (i) provide evidence that the edge trajectory itself, tracked further with a different observable or method, exhibits simultaneous streaks and TS waves, or (ii) rephrase the conclusion (and abstract) to state that the two routes coexist in the neighbourhood of the edge, which is what the data show.","section":"Section 3.2 and abstract"}],"minor_comments":[{"comment":"The text 'In the present study we used und by bisection, the two dotted lines are the observable bounds...' contains an incomplete/garble ('und by bisection'); please restore the intended sentence.","section":"Section 2.3"},{"comment":"The phrase 'The structure of the paper is at follows' should read 'is as follows.'","section":"Section 1"},{"comment":"Please define how the '2% or less' relative difference in the main observable is computed; this is the operational criterion for the tracking limit and should be specified precisely.","section":"Section 4"},{"comment":"The word 'substracting' should be 'subtracting.'","section":"Appendix A"},{"comment":"The reference 'Jordison 1970' is likely a misspelling of 'Jordinson' (J. Fluid Mech. 43, 801-811); please verify the correct spelling.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The main risk to the paper's central contribution is the one-sentence validation of the moving box. If the authors can provide a rigorous quantitative comparison against the fixed-domain computation, I would be willing to accept the paper after revision. The coexistence claim should be aligned with the evidence. The paper is within the scope of the journal and, if the validation issues are resolved, would be a useful contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Hi [Colleague],\n\nThe paper you should know about is a computational edge-tracking study in a spatially developing Blasius boundary layer. The headline result is that a moving box technique, new to spatially developing flows, extends edge tracking threefold and reveals streak switching in the Blasius edge for the first time. At longer times, the paper argues that TS waves and streaks coexist near the edge, and it reinterprets the edge as the boundary between bypass and classical transition routes.\n\nWhere it earns its keep: the moving box is a sensible and well-motivated extension of Galilean transforms, and the tracking time gains are concrete. The streak switching observation is genuinely new for this configuration. The TS wave identification is cross-checked against linear stability results, and the measured phase speed (0.32 ± 0.005) matches within the error bars. The self-similar scaling of the structure with δ* is convincing. The discussion of state space is nuanced and doesn’t overclaim the existence of a true edge state in the asymptotic sense.\n\nThe soft spots are two, and they are addressable rather than fatal.\n\nFirst, the moving box validation is asserted in one sentence: a comparison to a non-moving domain of size 12000, with no metrics, no convergence test, and no description of what was compared. This is the load-bearing assumption for everything past t≈4700, since the base flow is not Galilean invariant in a spatially developing layer. I don’t think the technique is suspect, but the evidence as presented is insufficient for the claim that it exactly reproduces a longer fixed domain.\n\nSecond, the coexistence claim is slightly overstated in the abstract. The TS waves actually appear on a trajectory bracketing the edge from below (δλ<0) after the edge-tracking limit, not on the edge trajectory itself. The abstract’s phrase “TS waves co-exist with the coherent structure characteristic of edge trajectories” is loose, though the body text is more careful. This should be clarified.\n\nNo code or data is provided, which is common for 2019 but would help address the validation concern.\n\nBottom line: the paper deserves a serious referee. The central edge-tracking results for t<4700 look internally consistent, the new observations are real, and the moving box idea is broadly useful for localised structures in spatially developing flows. I’d recommend sending it to review with a request for a real description of the moving-box validation and a tightening of the coexistence wording.","headline":"A genuinely useful moving-box technique and first streak switching in the Blasius edge, but the validation of the technique is too thin and the coexistence claim is slightly ahead of the data.","tokens_in":14370,"tokens_out":2693,"would_cite":true,"duration_ms":22938,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["47.27.Cn","47.20.Ft"],"model":"deepseek-v4-flash","headline":"A moving box lets edge trajectories in a Blasius boundary layer reveal streak switching, Tollmien–Schlichting waves, and the edge as a divider of two transition routes.","keywords":["edge state","edge manifold","Blasius boundary layer","bypass transition","Tollmien–Schlichting waves","streak switching","moving box technique","bisection method"],"falsifier":"Perform the same edge-bracketing bisection in a fixed domain of size $(L_x,L_y,L_z)=(12000,60,100)$ with the same resolution and compare, up to $t\\approx 4700$, the observable $a(t)$, the streak-switch times, the structure position $x_G(t)$, and the TS wave crest trajectories against the moving-box run; if the trajectories differ by more than the bisection uncertainty ($\\delta\\lambda/\\lambda^*\\sim 10^{-4}$) in any of these diagnostics, the equivalence between moving and fixed boxes fails.","tokens_in":13429,"feed_emoji":"🌊","tokens_out":16883,"duration_ms":139380,"temperature":0.7,"pith_summary":"This paper adapts the “edge” concept from parallel shear flows to a spatially developing Blasius boundary layer, i.e. the flat-plate boundary layer whose laminar profile itself becomes linearly unstable downstream. In such a flow the edge—the manifold of states that separate laminar and turbulent behaviour—does not separate two stable attractors, because the laminar state is unstable. Using a computational box that moves with the localized flow structure, the authors track an edge trajectory to $Re_x\\approx 9\\times 10^5$, roughly ten times beyond the theoretical onset of Tollmien–Schlichting (TS) waves, the two-dimensional waves of the classical route. Along this trajectory the central streak switches its sign repeatedly—streak switching, observed here for the first time in a spatially developing Blasius layer—and at late times TS waves grow spontaneously ahead of the decaying streaks. The paper therefore proposes that in this regime the edge should be read as the boundary between the bypass (streak-breakdown) and classical (TS-wave) routes to turbulence, not as a basin boundary between two attracting states.","feed_headline":"Edge divides the two transition routes in a boundary layer","feed_subtitle":"Bypass and Tollmien–Schlichting waves meet on a single finite-time edge trajectory.","key_machinery":"The machinery is the moving box technique combined with bisection on a specially chosen observable. The box translates streamwise at piecewise-constant speed $c_{\\mathrm{box}}$ (here alternating between 0 and $0.8U_\\infty$), implemented by Galilean transformations of the velocity and by updating the Blasius base flow used in the fringe (damping) region at the outflow, so the localized structure remains inside a domain of length $L_x=6000$ for times that would otherwise need a box twice as long; this extends the tracking horizon roughly threefold compared with earlier Blasius edge computations. The scalar observable used in the bisection is the volume-averaged root-mean-square of the streamwise vorticity, which vanishes for the laminar profile and is exactly zero for spanwise-invariant TS waves in their linear stage, so the bisection continues to bracket the edge even after the base-flow instability has set in. These two elements produce a finite-time edge trajectory that stays spatially localized, scales with the local displacement thickness $\\delta^*(x_G(t))$, and lives long enough for streak switching and the later coexistence of TS waves with the streak core.","core_discovery":"The central discovery is that both classical and bypass transition mechanisms can be reached from a single finite-time trajectory on the edge manifold, selected by bisection in the amplitude of an initial pair of counter-rotating vortices. At moderate times the trajectory is a localized self-sustained streaky structure that regenerates and switches the sign of its central streak, with no discrete symmetry imposed. At long times, varying the bisection parameter by a relative amount of order $10^{-4}$ decides the outcome: slightly above the edge the streaks break down locally, reproducing bypass transition; slightly below, the streak core decays while TS waves emerge upstream of it, travel at $c\\approx 0.32U_\\infty$, match the wavelength and phase speed predicted by linear stability analysis, and eventually undergo a secondary Klebanoff-type instability that forms a turbulent spot. The bisection stays meaningful until the two routes become indistinguishable to the chosen observable, which occurs at $t\\approx 4700\\,\\delta_0^*/U_\\infty$. The paper’s conclusion is a reinterpretation of the edge as the manifold separating the basins of the two transition routes, both leading to one common turbulent attractor.","pith_inferences":["If the moving-box equivalence holds quantitatively, the same frame-tracking idea could be applied to other spatially developing shear flows with a finite-Reynolds-number linear instability—for instance adverse-pressure-gradient boundary layers or channel entrance regions—to locate the route-dividing manifold without computing enormous fixed domains.","The choice of observable is what postpones the bisection failure; using an observable that is sensitive to TS waves from the start would presumably make the two routes distinguishable earlier, possibly sharpening the definition of the route boundary at the cost of a shorter tracking horizon.","The paper’s reinterpretation suggests testable predictions for experiments: in the weak-bypass regime, turbulent spots should appear either from streak breakdown or from TS wavepackets, and the boundary between those two origins should sit on a state-space manifold whose finite-time shadow is the computed edge trajectory.","Because the edge trajectory is computed only for finite times, its status as a genuine invariant set in the $t\\to\\infty$ limit is open; tracking the rescaled switching cycle over more periods in self-similar variables would indicate whether the streak-switching recurrence is asymptotic or a transient feature."],"forward_implications":["For free-stream turbulence levels $Tu\\gtrsim 2\\%$, the achieved horizon $Re_x^{\\max}\\approx 9\\times 10^5$ covers the experimentally observed intermittency range, so the computed edge trajectories can serve as an alternative, unsteady, localized base flow for stability analysis in that regime.","Because only one attracting turbulent state is evident, the edge here does not separate two basins of attraction; the edge-tracking observable loses discrimination once the two routes converge, so the bisection time limit is not a numerical artifact but a property of the state space.","Streak switching is not an artifact of parallel-flow approximations: it occurs in the spatially developing Blasius layer as well, with a switching period that grows with local boundary-layer thickness.","TS waves can be generated spontaneously from the wake of a decaying streak core without any external forcing, and they subsequently undergo a secondary instability and spot formation—so classical and bypass mechanisms can coexist in one unforced simulation.","The mixed $Tu<2\\%$ regime is precisely where the one-dimensional observable cannot distinguish the routes, motivating observables that explicitly weight TS-wave growth or local methods that avoid global state-space information."],"supporting_citations":[{"why":"Introduced the bisection procedure used to bracket the edge manifold by iterating between laminar and turbulent initial conditions.","marker":"Itano & Toh (2001)"},{"why":"Established the edge-of-chaos bisection framework in shear flow, which the present algorithm carries over to a spatially developing base flow.","marker":"Skufca et al. (2006)"},{"why":"Provided the earlier long-domain Blasius edge computation whose localized streaky structure and self-similar scaling are extended here.","marker":"Duguet et al. (2012)"},{"why":"Established edge states as mediators of bypass transition and reported streak switching in the asymptotic suction boundary layer, the dynamics reproduced here.","marker":"Khapko et al. (2016)"},{"why":"Studied the coexistence of Tollmien–Schlichting and bypass routes in plane Poiseuille flow, the precedent for the paper’s reinterpretation of the edge.","marker":"Zammert & Eckhardt (2017b)"},{"why":"Supplies the experimental intermittency range in Reynolds number versus turbulence level used to show that the achieved horizon covers strong bypass transition.","marker":"Shahinfar & Fransson (2011)"},{"why":"Provides the linear-stability onset of TS waves with non-parallel corrections, used to identify the waves and their phase speed.","marker":"Berlin (1998)"},{"why":"Shows that finite-amplitude streaks damp TS-wave growth, supporting the observed spatial separation between the wavepacket and the streak core.","marker":"Cossu & Brandt (2004)"}],"fun_headline_variants":["Edge trajectory reveals switch between boundary layer transition routes","Single edge path bifurcates into bypass or TS-wave transition","Edge manifold splits laminar, bypass, and classical routes","Boundary layer edge: one trajectory, two transition fates","Streak switching on edge precedes Tollmien–Schlichting takeover"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The moving box is assumed to reproduce exactly the dynamics of a much longer fixed computational domain over the whole tracking horizon; the paper states a validation against the fixed domain of length 12000 but gives no metrics, so if the frame changes or the time-updated base flow in the fringe distorts the streak core or the wake where TS waves grow, the long-time observations would change.","fun_headline_variants_meta":{"raw":{"variants":["Edge trajectory reveals switch between boundary layer transition routes","Single edge path bifurcates into bypass or TS-wave transition","Edge manifold splits laminar, bypass, and classical routes","Boundary layer edge: one trajectory, two transition fates","Streak switching on edge precedes Tollmien–Schlichting takeover"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000888,"raw_usage":{"total_tokens":3842,"prompt_tokens":963,"completion_tokens":2879,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":579,"completion_tokens_details":{"reasoning_tokens":2795}},"tokens_in":579,"tokens_out":2879,"duration_ms":21542,"temperature":1.0,"reasoning_tokens":2795,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:06:44.047745+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Perform the same edge-bracketing bisection in a fixed domain of size $(L_x,L_y,L_z)=(12000,60,100)$ with the same resolution and compare, up to $t\\approx 4700$, the observable $a(t)$, the streak-switch times, the structure position $x_G(t)$, and the TS wave crest trajectories against the moving-box run; if the trajectories differ by more than the bisection uncertainty ($\\delta\\lambda/\\lambda^*\\sim 10^{-4}$) in any of these diagnostics, the equivalence between moving and fixed boxes fails.","supporting_citations":[{"cited_title":"& Toh, S","cited_arxiv_id":null,"evidence_quote":"Introduced the bisection procedure used to bracket the edge manifold by iterating between laminar and turbulent initial conditions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Established the edge-of-chaos bisection framework in shear flow, which the present algorithm carries over to a spatially developing base flow."},{"cited_title":", Schlatter, P","cited_arxiv_id":null,"evidence_quote":"Provided the earlier long-domain Blasius edge computation whose localized streaky structure and self-similar scaling are extended here."},{"cited_title":", Kreilos, T","cited_arxiv_id":null,"evidence_quote":"Established edge states as mediators of bypass transition and reported streak switching in the asymptotic suction boundary layer, the dynamics reproduced here."},{"cited_title":"& Fransson, J","cited_arxiv_id":null,"evidence_quote":"Supplies the experimental intermittency range in Reynolds number versus turbulence level used to show that the achieved horizon covers strong bypass transition."},{"cited_title":"1998 Oblique waves in boundary layer transition","cited_arxiv_id":null,"evidence_quote":"Provides the linear-stability onset of TS waves with non-parallel corrections, used to identify the waves and their phase speed."},{"cited_title":"& Brandt, L","cited_arxiv_id":null,"evidence_quote":"Shows that finite-amplitude streaks damp TS-wave growth, supporting the observed spatial separation between the wavepacket and the streak core."}],"review_version":1}