{"id":"f6f8ab10-fda6-45f7-89e2-d502377312b0","arxiv_id":"2501.03135","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Geodesic vortex detection is extended to 2D Riemannian manifolds and used to date the birth, split, and death of the 2002 austral stratospheric polar vortex, with ozone depletion confined by the vortex edge.","lead":"This paper extends geodesic vortex detection, a method for finding flow-invariant vortex boundaries, from flat planes to curved surfaces, and applies it to the 2002 Antarctic polar vortex split. It reconstructs the vortex's birth, split, and death from reanalysis winds, and connects the detected vortex edge to ozone depletion patterns.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Manifold p-loop equation (15) is asserted, not derived: Appendix A verifies only L=p, not stationarity of the action (14); the 'formal equivalence' to HBV13 skips the curved-surface Euler–Lagrange proof.","rationale":"The reader's verdict focuses on the tau=30-day truncated-wedge assumption, which is a legitimate risk for the specific dates but is an application-level assumption: even if the dates shift, the underlying detection method could be sound. The more load-bearing issue is the mathematical foundation of the manifold extension, because the paper's central novelty is Eq. (15) on curved surfaces. Section 2.2 states that p-loops are solutions of the variational principle (14), but the derivation in Appendix A only confirms that curves tangent to (15) have L=p; it never shows δSL=0. In the planar theory (HBV13), the stationarity condition is what makes a uniformly stretching loop a coherent vortex boundary; the null-geodesic interpretation of Remark 3 requires the geodesic equation, not just the null direction. No such geodesic equation is derived for the manifold case. The note that the equivalence with the planar p-loop equation is 'only formal' (Appendix A) is an explicit admission that the argument is not supplied. This matters because all detected 'CLVs' in Sections 4-5 are limit cycles of (15); if (15) does not extremize (14), the objects are not the objective material-coherence boundaries claimed. A synthetic test on a known spherical vortex, or a direct Euler-Lagrange calculation, would settle this. The manuscript gives credit for sharing code and for a novel application, and the planar method is solid, so the paper is not fatally flawed, but the mathematical claim needs proof or a rigorous citation before acceptance. I therefore keep the reader's CONDITIONAL verdict, but with the condition attached to the derivation rather than to the tau assumption.","tokens_in":22007,"tokens_out":18112,"duration_ms":171741,"concrete_test":"Derive the Euler–Lagrange equations for the functional (14) in the chart (29) with metric (30), for a generic velocity field, and check whether every solution of the first-order system (15) satisfies them. If the Christoffel terms do not cancel identically, the manifold p-loop equation is incomplete. Complementary numerical check: take a synthetic divergence-free flow on S² with a known rotating coherent patch (e.g., a solid-body vortex), compute its exact material boundary over T=30 days, and test whether the outermost p-loop of (15) coincides with that boundary; if it does not, the stationarity condition (14) is violated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central methodological claim is that Eq. (15) gives the stationary curves of the variational principle (14) on an arbitrary 2-D Riemannian manifold. What the paper actually proves in Appendix A is weaker: a curve tangent to the line field (15) satisfies the pointwise uniform-stretching condition L(r,r')=p, i.e., it is a null curve of C−pG. It does not verify the Euler–Lagrange equation δSL=0. In the Euclidean case (HBV13), p-loops are null-geodesics of C−p²G; the stationarity condition imposes not just the null direction but the geodesic equation, which includes connection terms. On a curved surface with metric G (e.g., the nonorthogonal metric (30)), those connection terms involve derivatives of G and C. The algebraic eigenvector construction (15) contains no such terms, so it is not self-evident that integral curves of (15) extremize (14). The paper's own note that the equivalence is 'only formal' (Appendix A) concedes this gap. Because the entire application (Section 4 and the ozone analysis in Section 5) identifies 'coherent Lagrangian vortices' as limit cycles of (15), an unproven equivalence would undermine the definition of what is detected. The birth/death dates and ozone barriers could still be useful transport diagnostics, but they would not be the objective CLVs the paper claims. This is distinct from the tau/truncated-wedge concern raised by the reader, which concerns the interpretation of the dates rather than the underlying detection principle.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper extends geodesic vortex detection, previously formulated for Euclidean flows with Cartesian coordinates, to two-dimensional Riemannian manifolds with arbitrary coordinates, and adapts the birth-and-death CLV framing algorithm to the finite-time validity of the isentropic two-dimensional approximation in the stratosphere. The method is applied to ERA5 600-K isentropic winds to characterize the 2002 austral stratospheric polar vortex, reporting a birth around 31 March 2002, a pre-split death around 21 September 2002, a split around 23 September 2002, and an ozone depletion analysis identifying an outer vortex edge with p≈1 and an inner p-loop with p≈1.6 that routes ozone-poor air poleward. The paper also makes a computational implementation available.","tokens_in":22341,"tokens_out":10387,"duration_ms":98231,"significance":"If the methodological claims hold, the paper provides a general framework for objective vortex detection on curved surfaces and delivers the first geodesic-detection-based life cycle of the austral polar vortex, which is a widely studied but difficult case due to the splitting event. The objectivity argument under isometric observer changes is clean and correctly distinguishes coordinate from metric representations. The public code release and the use of a standard reanalysis are strengths. The main significance depends on the unresolved stationarity question for the manifold p-loop equation and on the robustness of the birth/death dates to the assumed truncated-wedge coherence profile.","major_comments":[{"comment":"The central claim that limit cycles of the line field (15) are stationary curves of the variational principle (14) is not established. Appendix A verifies only that curves tangent to (15) satisfy the pointwise uniform-stretching condition L(r,r')=p, i.e., that they are null curves of C−pG, but it does not verify the Euler–Lagrange equations. The authors' own statement that the equivalence to the HBV13 p-loop equation is 'only formal' concedes this gap. On a curved surface the stationarity condition involves connection terms that depend on derivatives of G and C, and the algebraic construction (15) contains no such terms. Because the vortex identifications in Section 4 and the ozone analysis in Section 5 all rely on (15), the paper should either supply a rigorous derivation for arbitrary 2-D Riemannian manifolds or explicitly revise the claim to describe the detected loops as uniformly stretching material loops rather than extremizers of the averaged stretching.","section":"Section 2.2, Eq. (15), Appendix A"},{"comment":"The inferred birth and death dates, including tbirth=31 March 2002 and tdeath=21 September 2002, rest entirely on the assumption that Texp(t0) has a truncated wedge shape at height τ=30 days. This assumption is asserted without sensitivity analysis: no tests for other values of τ, no error bars on the plateau edges t_late^0 and t_early^0, and no discussion of how deviations from a flat plateau would shift the dates. The paper acknowledges that the estimated vortex lifespan exceeds τ, so the objects are only quasi-CLVs. Given that the life-cycle dates are a headline result, the authors should provide evidence that the plateau structure is robust to the choice of τ or reframe the dates as conditional on this modeling assumption.","section":"Section 3, steps 2.a and 2.b"},{"comment":"The validation of the vortex edge as a transport barrier is weakened by two acknowledged issues that should be addressed more directly. First, the advected image of the 21 August 2002 p≈0.9 loop on 19 September exhibits a measured relative stretching of about 2.15, which contradicts the uniform-stretching property defining a p-loop; the explanation in terms of tangential stretching and noise is plausible but unquantified. Second, the inner p≈1.6 loop of 13 August is selected a posteriori to coincide with the poleward boundary of the low-ozone ring, using the same ERA5 wind and ozone fields, so the conclusion that ozone-depleted air mixes poleward is partly constructed from the target data. The authors should add independent support, such as a different tracer, a different reanalysis, or an a priori selection of the inner loop from the line field alone, or explicitly label this part as a consistency check rather than a prediction.","section":"Section 5, Figures 7 and 8"}],"minor_comments":[{"comment":"The heading 'Kinematics of SVPs' and nearby text use 'SVP' inconsistently with the stated acronym 'SPV'; please harmonize the terminology throughout.","section":"Section 1.1.1"},{"comment":"The first sentence contains the typo 'deliniate'; it should read 'delineate'.","section":"Introduction"},{"comment":"The word 'dyapicnic' appears in the paragraph on the temporal validity of the 2-D assumption; it should be 'diapycnic'.","section":"Section 3"},{"comment":"Reference [LR10] contains the typo 'Lyapunoc'; the correct term is 'Lyapunov'.","section":"References"},{"comment":"The eigenvector notation is inconsistent: equations (15) and (17) use a superscript parenthetical index while equation (16) does not; please standardize.","section":"Equations (15)-(17)"}],"recommendation":"major_revision","confidential_remarks":"The missing stationarity proof for the manifold p-loop equation is the key technical obstacle and is load-bearing for the paper's central claim. It appears fixable through an appendix or a precise citation to a result covering the Riemannian case, so I recommend major revision rather than rejection. The application itself is interesting and the code release is a positive feature, but the life-cycle dates and ozone conclusions should be presented with more explicit caveats about the truncated-wedge assumption and the selection of the inner p-loop."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here is the short version. The paper extends geodesic vortex detection to 2-D Riemannian manifolds by replacing C with G^-1 C, and the objectivity argument under isometric observer changes is clean. The 2002 SPV case study is a legitimate new application, the code is public, and the authors are transparent about their assumptions. It is a solid applied paper, but the central methodological equivalence is asserted rather than proven, and the birth-death dates should be treated as 30-day-wedge diagnostics, not hard measurements.\n\nWhat is actually new: the G^-1 C eigenvalue formulation for CLVs on manifolds, the manifold objectivity discussion, the adaptation of birth-death framing to the ~30-day validity of 2-D isentropic flow, and the first geodesic-detection life cycle for the 2002 austral vortex. The authors honestly cite the prior work (HBV12 Appendix C, Kar15) that already pointed at G^-1 C, so they do not oversell.\n\nSoft spots, in order. First, the paper claims that solutions of the variational principle (14) are p-loops, the limit cycles of the line field (15). Appendix A only shows that a curve tangent to (15) has L(r,r')=p, i.e., it is a null curve of C-p^2 G. The equivalence with the HBV13 geodesic derivation is explicitly called 'only formal'. A proper derivation on a curved surface would introduce connection terms that do not appear in the algebraic eigenvector construction. My reading is that the result is probably true, since in 2-D Lorentzian geometry every null curve is a null geodesic up to reparameterization, so the null condition alone may be sufficient. But the paper never says this, and a referee should ask for the argument.\n\nSecond, the birth and death dates rest on the truncated-wedge assumption for Texp(t0), capped at tau=30 days. The figures show noisy plateaus and no error bars; tbirth and tdeath are computed as plateau edge minus/plus tau. This is a diagnostic with a built-in prior, not a measurement.\n\nThird, the ozone story is partly circular: ERA5 ozone advected by ERA5 winds, and the inner p≈1.6 loop is selected after looking at the low-ozone ring. The authors also admit the 21 August p-loop stretches by a factor of about 2.15 by 19 September, which sits uneasily with the CLV label; their tangential-stretching explanation is plausible but untested. And there is no synthetic benchmark on a known curved flow.\n\nWho is this for: anyone doing LCS or transport-barrier analysis on the sphere or other curved surfaces. It deserves a serious referee. The referee should request (1) a derivation or at least an explicit null-geodesic argument on manifolds, (2) a synthetic test, (3) uncertainty on the plateau dates. If those are added, this becomes a useful reference. Send it to review.","headline":"Useful, honest manifold extension of geodesic vortex detection with a solid 2002 SPV case study, but the curved-surface stationarity proof is skipped and the birth/death dates rest on a 30-day wedge assumption - still worth refereeing.","tokens_in":22884,"tokens_out":10686,"would_cite":true,"duration_ms":102188,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37N10"],"pacs":[],"model":"deepseek-v4-flash","headline":"Geodesic vortex detection, extended to curved surfaces and arbitrary coordinates, reconstructs the 2002 austral polar vortex life cycle from isentropic winds.","keywords":["geodesic vortex detection","coherent Lagrangian vortices","p-loops","sudden stratospheric warming","austral polar vortex","Cauchy–Green strain tensor","ozone depletion","isentropic winds"],"falsifier":"Run the birth-and-death framing on the same isentropic winds with the coherence cap set to 20, 30, and 45 days; if the inferred birth and death dates move with the cap, the plateau-edge dates are artifacts of the cap and not intrinsic vortex properties.","tokens_in":21795,"feed_emoji":"🌀","tokens_out":10926,"duration_ms":135976,"temperature":0.7,"pith_summary":"Geodesic vortex detection is a way to identify vortices whose boundaries are material curves that resist stretching, without choosing a preferred observer. This paper extends it from flat Cartesian flows to general two-dimensional curved surfaces, in arbitrary coordinates, by building the detection line field from the metric-corrected Cauchy–Green tensor. Applied to isentropic reanalysis winds on the southern hemisphere, the method frames the full 2002 austral polar vortex life cycle: birth around 31 March 2002, death of the pre-split vortex around 21 September 2002, and split around 23 September 2002. It also shows that the vortex edge confines ozone-depleted air, while a weaker inner loop with $p$ near 1.6 lets that air mix poleward rather than equatorward.","feed_headline":"Curved-surface vortex detection frames the 2002 polar vortex split","feed_subtitle":"New manifold version gives birth, death, and split dates and shows ozone-depleted air mixing poleward.","key_machinery":"The load-bearing object is the p-loop equation (15), $r'=\\ell^\\pm_p(r)$, built from the two eigenvalue–eigenvector pairs of $G^{-1}C$ with $0<\\lambda_1<p^2<\\lambda_2$; the eigenvectors are orthonormal with respect to the metric, and the line-field coefficients make every solution stretch by exactly $p$. The paper's second mechanism is the modified birth-and-death framing: because stratospheric air leaves an isentropic surface after roughly $\\tau=30$ days, the life-expectancy curve $T_{\\mathrm{exp}}(t_0)$ is expected to be a truncated wedge of height $\\tau$, so the pre-split vortex death is read as $t_{\\mathrm{death}}=t_0^{\\mathrm{late}}+\\tau$ and its birth as $t_{\\mathrm{birth}}=t_0^{\\mathrm{early}}-\\tau$ from the plateau edges in forward and backward time.","core_discovery":"The paper's central claim is that geodesic vortex detection remains valid and observer-independent on a 2D Riemannian manifold when the p-loop line field is formed from the eigenvalue–eigenvector pairs of $G^{-1}C$, where $G$ is the coordinate representation of the manifold metric and $C$ is the coordinate representation of the right Cauchy–Green strain tensor. The closed limit cycles of the resulting line field, the p-loops, are material curves stretching uniformly by a factor $p$ over $[t_0,t_0+T]$, and the outermost loop in a nested family is the vortex boundary. With this metric-aware construction and a truncated-wedge birth-and-death algorithm capped at 30 days of valid two-dimensional motion, the authors report the first geodesic-detection-based life cycle of the austral stratospheric polar vortex: birth on 31 March 2002, pre-split death on 21 September 2002, split on 23 September 2002, and ozone-poor air held by the vortex edge ($p\\approx1$) but leaking poleward across an inner loop with $p\\approx1.6$.","pith_inferences":["The authors do not test whether the plateau-edge dates shift under a different coherence cap; re-running the framing with $\\tau=20$ and 45 days would separate a genuine vortex property from a cap artifact.","The same metric-aware construction should extend to a global atlas of charts, making geodesic vortex detection usable on full spherical data; the southern-cap parameterization used here is a natural template.","The kinematic poleward-mixing claim is testable in a chemistry-transport model: seed an inert tracer inside the 13 August 2002 low-ozone ring and see whether it crosses the $p\\approx1.6$ loop poleward within 30 days."],"forward_implications":["The metric-aware p-loop construction can be applied to any two-dimensional flow on a curved surface without projecting to a plane, so ocean and planetary vortices on a sphere become directly detectable.","Vortex boundaries identified this way are observer-independent even in non-orthogonal coordinates, so edge-based diagnostics such as ozone or temperature contrasts can be compared across studies without frame corrections.","The method gives a concrete timetable for the 2002 sudden stratospheric warming—birth 31 March, pre-split death 21 September, split 23 September—that can anchor kinematic comparisons with dynamical explanations of the event.","The ozone analysis indicates that the vortex edge is a stronger transport barrier than the inner $p\\approx1.6$ loop, so ozone-poor air is mixed poleward into the vortex interior; a similar two-barrier geometry may structure other polar vortices."],"supporting_citations":[{"why":"Supplies the variational definition of coherent Lagrangian vortices and the p-loop construction that the manifold version builds on.","marker":"[HBV13]"},{"why":"Provides the geodesic theory of transport barriers and the p-loop equation for plane flows that is generalized here.","marker":"[HBV12]"},{"why":"Introduces the birth-and-death framing algorithm that gives life-expectancy curves and plateau-edge dates.","marker":"[ACKBV20]"},{"why":"Prior geodesic detection of the polar vortex edge, showing ozone and temperature contrast across it, which this paper extends to curved surfaces and to the austral event.","marker":"[SSBVH17]"},{"why":"Establishes the roughly one-month diapycnic mixing timescale that motivates the 30-day coherence cap.","marker":"[Hay05]"},{"why":"Supplies the spherical-cap coordinate chart used to parameterize the southern hemisphere without a pole singularity.","marker":"[LR10]"},{"why":"Documents the reanalysis system whose isentropic winds and ozone fields are analyzed.","marker":"[HBB+20]"}],"fun_headline_variants":["Manifold-adapted vortex detection dates 2002 polar split","Geodesic vortex tool works on spheres, frames 2002 split","Birth, split, death: curved-surface vortex tracking in 2002","2002 polar vortex split timed via geodesic detection on sphere"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The birth and death dates rest on the assumption that the vortex's measured coherence lifetime forms a flat plateau of height 30 days, so the dates can be read off the plateau edges; if the plateau is not flat or the 30-day cap does not match how long the two-dimensional isentropic approximation really holds, the dates are artifacts of the cap rather than properties of the vortex.","fun_headline_variants_meta":{"raw":{"variants":["Manifold-adapted vortex detection dates 2002 polar split","Geodesic vortex tool works on spheres, frames 2002 split","Birth, split, death: curved-surface vortex tracking in 2002","2002 polar vortex split timed via geodesic detection on sphere"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000237,"raw_usage":{"total_tokens":1527,"prompt_tokens":985,"completion_tokens":542,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":601,"completion_tokens_details":{"reasoning_tokens":466}},"tokens_in":601,"tokens_out":542,"duration_ms":5539,"temperature":1.0,"reasoning_tokens":466,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:53:15.229374+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the birth-and-death framing on the same isentropic winds with the coherence cap set to 20, 30, and 45 days; if the inferred birth and death dates move with the cap, the plateau-edge dates are artifacts of the cap and not intrinsic vortex properties.","supporting_citations":[],"review_version":1}