{"id":"1f24ea7d-c0c1-4a12-8e85-4a9f40d0801e","arxiv_id":"2411.19160","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The authors construct enriched Galerkin discretizations with monolithic convex limiting and entropy fixes that preserve local bounds and entropy stability for nonlinear scalar conservation laws.","lead":"This paper develops bound-preserving and entropy-stable enriched Galerkin schemes for nonlinear scalar hyperbolic equations by limiting both cell averages and nodal values. It matters because it extends robust algebraic flux correction tools from continuous and discontinuous Galerkin methods to the less expensive enriched Galerkin framework, with 2D tests including the rotating KPP wave.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The optimal-order claim rests on the unproved assertion that the modified high-order target (24) is an O(h^2) perturbation of the original EG weak form (20); if this fails, the limiters cannot restore second-order accuracy.","rationale":"The DMP and entropy-stability parts of the construction are largely algebraic and appear internally consistent: equations (46)-(51) show the limited scheme is a convex combination of bound-preserving intermediate states, and conditions (44),(45) are imposed by construction, with the final scaling step only reducing the magnitude of already-constrained contributions. The genuine soft spot is the accuracy link between the modified high-order target and the original EG method. The paper provides no a priori error estimate for (24); it merely asserts the perturbation is second-order. The convergence tables in Section 7 are consistent with this assertion, which is real evidence, but they do not close the gap for the 'proving' language in the abstract. A direct numerical comparison against the unmodified weak form is the cheapest decisive check. This does not change the reader's CONDITIONAL verdict: the concern is addressable and not an observed failure, so REJECT is too strong, and ACCEPT would require either the missing estimate or the proposed comparison.","tokens_in":20389,"tokens_out":18550,"duration_ms":179650,"concrete_test":"Implement the original EG weak form (12)/(20) without the Taylor, GFE, lumped-quadrature, and dot-u_i modifications, and compare its solutions with the modified HO scheme (24) on the smooth Burgers and KPP test cases of Section 7 on meshes h = 2^-4 to 2^-7 with the same dt/h. Compute the L1 norm of the difference of the two RHS vectors or of the final solutions at T = 0.1; if the difference does not decay as O(h^2) relative to the solution error, the target scheme is not a second-order perturbation and the optimal-order claim for BP-ES is unproven. If the difference is O(h^2), the gap is closed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central accuracy claim is that BP-ES preserves optimal convergence order. That claim presupposes that the high-order target (24), obtained from the EG weak form (20) by replacing f(u_h+delta u_h) with f(u_h)+f'(u_h)delta u_h, by the group finite element interpolation (22), by lumped boundary quadrature, and by the reconstructed derivative dot u_i in (23), is a second-order perturbation of (20). The text asserts this immediately after (24) but gives no estimate and no regularity conditions. The Taylor error is controlled only when delta u_e = U_e - mean(u_h) is small, i.e., in smooth regions, and the lumped boundary and dot u_i modifications are not analyzed. If any of these changes is only first-order, no limiter can repair the loss of accuracy, so the measured O(h^2) convergence of HO/BP/BP-ES in Section 7 is not supported by proof. The numerical experiments are consistent with the assumption, but the abstract's 'proving' is not backed by a theorem.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops bound-preserving and entropy stable enriched Galerkin (EG) methods for nonlinear scalar hyperbolic conservation laws. The spatial discretization is split into a property-preserving low-order local Lax--Friedrichs part and an antidiffusive high-order correction. A flux limiter constrains the cell averages of the piecewise-constant enrichment, while a clip-and-scale monolithic convex limiter with an algebraic entropy fix constrains the nodal values of the continuous component. The fully discrete scheme uses SSP-RK2 time stepping. Numerical experiments for linear advection, inviscid Burgers, and the two-dimensional KPP problem are reported, showing optimal convergence rates for smooth solutions and correct entropy solutions for the rotating KPP benchmark.","tokens_in":20605,"tokens_out":5133,"duration_ms":51997,"significance":"If fully established, the paper would provide the first property-preserving AFC/MCL-type EG method for nonlinear scalar hyperbolic conservation laws, extending the authors' earlier linear-advection EG work. The limiter construction is explicit and parameter-free: the constraints are enforced through algebraic inequalities (44), (45), (47), (49), and (50), and the zero-sum conditions (39) guarantee discrete conservation. The numerical KPP test convincingly demonstrates that the entropy fix is needed, not just cosmetic. However, the central proof claims in the abstract and conclusions are not matched by formal theorem statements in the body, and the optimal-order claim rests on an unproved consistency assertion about the modified high-order target.","major_comments":[{"comment":"The statement immediately after Eq. (24) that the modified high-order scheme 'represents a second-order perturbation of (20)' is load-bearing but unproved. The perturbation involves four distinct substitutions: the Taylor approximation (21), the group finite element interpolant (22), lumped boundary-face quadrature, and the reconstructed time derivative (23). No estimate in h, no regularity assumptions, and no reference are supplied. Since all subsequent limiting and entropy analysis applies to this modified target, the claimed optimal convergence of the BP and BP-ES schemes in Section 7 and the Conclusions depends on this assertion. Please provide a lemma with a proof (or a precise citation) showing that the difference between the right-hand sides of (20) and (24) is O(h^2) in the relevant norm for solutions of the regularity used in the convergence tests.","section":"§3.2.1, Eq. (24)"},{"comment":"The paper contains no formal theorem statements, despite the abstract's claim 'In addition to proving the claimed properties of the proposed approach'. The arguments in Section 5 are conditional: inequalities (44), (45), (47), (49), and (50) are stated as sufficient conditions, and the statements 'It is easy to verify' carry the entropy-flux correction in Section 5.2 and the clipping step in Section 5.3. In particular, the final scaling stage (51) modifies the clipped contributions, and its preservation of both (49) and (50) is asserted but not demonstrated. Please state explicit theorems for the semi-discrete scheme (38): local DMP/LED, invariance of the admissible interval, and the discrete entropy inequality, with proofs that include the scaling stage. The same applies to the SSP-RK2 statement in Section 6, where the convex-combination argument should be formalized under the CFL conditions (42)–(43).","section":"§5.1–§5.3"},{"comment":"Remark 1 concedes that the cited linear stability analysis of the EG method is valid only for triangular meshes, while all numerical experiments in Section 7 use uniform rectangular (Q1) meshes. The paper then asserts that the piecewise-constant enrichment of CG-Q1 'has a stabilizing effect' on quadrilaterals, without proof or reference. This is load-bearing because the HO scheme (24) is the target that the limiters are designed not to degrade; if the HO baseline is unstable or only first-order accurate on quadrilaterals, the optimal-order numerical results cannot be explained by the limiting analysis. Please either supply a stability or consistency analysis for the quadrilateral case, or clearly restrict the theoretical claims to the setting in which they are proven and rephrase the accuracy claim as a numerical observation.","section":"Remark 1 (Section 3.1)"}],"minor_comments":[{"comment":"There is a typo in 'In the remainder if this section', which should read 'In the remainder of this section'.","section":"Section 4, introductory paragraph"},{"comment":"The caption of Figure 3 says 'BP schemes' but the figure includes BP-ES results; please update the caption to 'BP and BP-ES schemes'.","section":"Figure 3 caption"},{"comment":"Reference [44] appears to be a duplicate of reference [20]: both cite Kuzmin, Hajduk, and Rupp, 'Limiter-based entropy stabilization of semi-discrete and fully discrete schemes for nonlinear hyperbolic problems', CMAME 389 (2022) 114428. Please remove the duplicate.","section":"References"},{"comment":"The set E'_e is used in Eq. (24) before its definition later in the same subsection; move the definition immediately before the equation or add a forward reference.","section":"Section 3.2.1, Eq. (24)"}],"recommendation":"major_revision","confidential_remarks":"The paper fits the journal's scope and the numerical study is convincing, but the gap between the abstract's 'proving' claim and the absence of any theorem statements is substantial. The unproved O(h^2) consistency of the modified high-order target is the most serious issue, because it directly underpins the optimal-order claim. I believe the results are likely correct and fixable, but the manuscript needs a genuine consistency proof and formal statements of the DMP and entropy-stability results before it can be accepted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Best quick read: this paper does what it claims at the numerical level. It combines the flux-limited cell-average constraint from the authors' DG work with an entropic clip-and-scale limiter for the CG nodal component, giving the first bound-preserving and entropy-stable EG scheme for nonlinear scalar conservation laws. The KPP rotating wave test is the strongest piece of evidence: the BP-only version fails to reproduce the wave structure while the entropy-stable version gets it right.\n\nThe construction is explicit and the conservation properties are handled correctly: the zero-sum conditions (39) are preserved by the scaling stage, and the limiting inequalities (44)–(50) do imply the stated DMP and entropy constraints if the low-order baselines are as cited. The paper builds honestly on the authors' own MCL/AFC toolbox, and the self-citations are appropriate.\n\nWhere it gets softer: the abstract says 'proving the claimed properties,' but the paper contains no formal theorem statements. Several key steps are dismissed with 'it is easy to verify.' That is normal for a JCP methods paper if the claims are worded as algorithmic properties supported by cited lemmas, but the word 'proving' overstates what is actually demonstrated.\n\nThe more substantive gap is the asserted second-order perturbation property of the modified high-order target (24) relative to the original EG weak form (20). The Taylor linearization, GFE interpolation, lumped boundary quadrature, and reconstructed time derivative are each plausible, but the paper gives no estimate, no regularity conditions, and no quantification of the errors. This is load-bearing for the optimal-convergence claim: if any of these modifications is only first-order, no limiter can recover second-order accuracy. The numerical experiments are consistent with second-order convergence, so I suspect the estimate is true under smoothness assumptions, but the text needs to state and prove it or explicitly assume it.\n\nRemark 1 is a smaller caveat: the stability analysis cited for EG only covers triangles, while all experiments use quadrilaterals. The authors acknowledge this, but it should be discussed as a limitation rather than a side remark.\n\nNo code is shipped, but the algorithm is detailed enough to reimplement.\n\nBottom line: this is a competent methods paper for the AFC/MCL community. I would send it to a serious referee, not desk-reject it. I would ask the authors to prove or clearly state the perturbation estimate, add explicit theorem statements (or soften the claim), and address the quadrilateral gap. With those revisions it is a solid JCP contribution.","headline":"A solid, careful extension of MCL to nonlinear EG that deserves refereeing but overstates its proof content; the main gap is an unproved O(h^2) perturbation claim.","tokens_in":21104,"tokens_out":3804,"would_cite":true,"duration_ms":35382,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["65M60","35L65","65M12"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper develops a limiting strategy that makes enriched Galerkin discretizations of nonlinear scalar conservation laws both bound-preserving and entropy stable, while preserving near-optimal convergence in smooth tests.","keywords":["enriched Galerkin method","nonlinear hyperbolic conservation laws","discrete maximum principle","entropy stability","monolithic convex limiting","flux-corrected transport","algebraic flux correction","local Lax-Friedrichs scheme"],"falsifier":"On a smooth manufactured solution with rectangular meshes, compute the difference between the original EG weak form (20) and the modified target (24) under the same initial data and boundary conditions; if that difference decays more slowly than $O(h^2)$ as the mesh is refined, the claimed second-order perturbation, and with it the optimal accuracy of the limited scheme, fails.","tokens_in":20147,"feed_emoji":"🧮","tokens_out":8299,"duration_ms":72683,"temperature":0.7,"pith_summary":"This paper aims to establish that an enriched Galerkin (EG) space discretization of a nonlinear scalar conservation law can be made both bound-preserving and entropy stable without abandoning its advantages over discontinuous Galerkin methods. The authors split the high-order EG semi-discretization into a low-order local Lax-Friedrichs scheme that is provably stable and a remainder of antidiffusive fluxes and element contributions, then limit that remainder so that local discrete maximum principles and entropy inequalities hold simultaneously. They prove that the constrained semi-discrete problems satisfy these properties, and in two-dimensional tests the limited schemes keep the physically correct weak solution while converging at nearly second order on smooth problems. This matters because EG methods are locally conservative and cheaper than DG but had not previously been equipped with nonlinear stability constraints.","feed_headline":"Enriched Galerkin schemes get bounds and entropy with new limiters","feed_subtitle":"Flux and clip-and-scale limiters enforce both nonlinear stability constraints at near-optimal convergence rates.","key_machinery":"The load-bearing mechanism is an algebraic splitting of the high-order EG discretization into a property-preserving low-order part and limited antidiffusive corrections. For cell averages, the low-order part is the local Lax-Friedrichs finite volume scheme and the limiter is a monolithic convex limiting formula constrained by an entropy production bound on each interface. For nodal values, the low-order part is an algebraic CG-LLF scheme and the new element is an entropic clip-and-scale limiter that clips each antidiffusive element contribution to local intermediate-state bounds, scales it to satisfy the entropy production bound, and then rescales the element vector to preserve the zero-sum condition that guarantees conservation. The low-order schemes supply the baseline bound and entropy guarantees, and the limiters are designed not to destroy them.","core_discovery":"For the initial value problem $\\partial_t u + \\nabla\\cdot f(u)=0$, the paper claims that one can evolve the EG cell averages $U_e$ and the continuous nodal values $u_i$ through a corrected system whose right-hand sides are weighted sums toward locally bound-preserving intermediate states. The cell-average flux limiter enforces local bounds together with an entropy production bound on each interface flux, while a new clip-and-scale limiter on element contributions enforces nodal bounds and a discrete analogue of the entropy inequality for the continuous component. The zero-sum conservation property is maintained by a final scaling step, and a strong stability preserving two-stage time integrator keeps the bounds at every stage under CFL conditions. The numerical experiments claim optimal second-order convergence for smooth solutions and show that, in a rotating two-dimensional benchmark with nonconvex flux, the entropy fix is what prevents convergence to the wrong weak solution.","pith_inferences":["If the unproved assertion that the modified high-order target is a second-order perturbation of the original EG weak form is correct, the same splitting should extend to systems with a known entropy pair, such as the shallow water equations, with the entropy fix applied in entropy variables; the paper only sketches this direction.","A direct comparison of the original EG weak form with the linearized target on smooth manufactured solutions would settle whether the limiting stage is masking an accuracy loss introduced by the group finite element and linearization step.","Since the underlying EG stability analysis is proven only on triangles while all experiments use quadrilateral meshes, a quadrilateral-specific stability proof would close the gap between the theory and the reported computations."],"forward_implications":["Cell averages and nodal values of an EG solution can satisfy local bounds and semi-discrete entropy stability at the same time, in a single monolithic solve, rather than through operator splitting or postprocessing.","The entropy fix is load-bearing for correct weak solutions: in the rotational benchmark, bound-preserving limiting alone converges to the wrong two-shock solution, while the entropy-limited version captures the spiral wave.","Constrained EG schemes recover near-optimal accuracy in smooth tests, with reported $\\ell^\\infty(L^1)$ rates about 1.91 for linear advection and $L^1$ rates around 1.82--1.93 for the two-dimensional KPP-type problem.","The same splitting framework carries modern FCT/MCL limiting tools from continuous and discontinuous Galerkin settings into enriched Galerkin discretizations."],"supporting_citations":[{"why":"Supplies the EG-specific MCL/FCT limiters for cell averages and nodal values that this paper extends from linear advection to nonlinear problems.","marker":"[33]"},{"why":"Supplies the flux limiting framework for DG cell averages and the LLF low-order scheme used as the baseline for cell-average constraints.","marker":"[34]"},{"why":"Supplies the monolithic convex limiting formula and the bar-state bounds used in the cell-average limiter.","marker":"[19]"},{"why":"Supplies the algebraic entropy fix and the entropy-production-bound limiting used to make cell-average fluxes entropy stable.","marker":"[20]"},{"why":"Provides the CG entropy stability condition and the zero-sum argument for element contributions that the nodal limiter adapts.","marker":"[38]"},{"why":"Supplies the algebraic flux correction framework, notation, and proofs that the analysis builds on and references.","marker":"[23]"},{"why":"Defines the original reduced-P1/EG method and the linear stability analysis that motivates the enrichment.","marker":"[24]"},{"why":"Supplies the theorem converting local bound preservation plus CFL conditions into invariant-domain preservation for fully discrete schemes.","marker":"[41]"}],"fun_headline_variants":["Clip-and-scale limiters bound enriched Galerkin","Flux and clip limiters give EG stability and accuracy","Entropy-stable limiters for nonlinear enriched Galerkin","Monotone limiters enforce bounds and entropy in EG"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The accuracy claim rests on the unproved assertion that the modified high-order target (equation (24), built from a linearized flux and a group finite element interpolant) differs from the original EG weak form only by second-order terms; if that perturbation is larger, neither the optimal convergence rate nor the fidelity to the original EG method is established.","fun_headline_variants_meta":{"raw":{"variants":["Clip-and-scale limiters bound enriched Galerkin","Flux and clip limiters give EG stability and accuracy","Entropy-stable limiters for nonlinear enriched Galerkin","Monotone limiters enforce bounds and entropy in EG"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000221,"raw_usage":{"total_tokens":1448,"prompt_tokens":939,"completion_tokens":509,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":555,"completion_tokens_details":{"reasoning_tokens":442}},"tokens_in":555,"tokens_out":509,"duration_ms":5782,"temperature":1.0,"reasoning_tokens":442,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T10:28:12.776884+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On a smooth manufactured solution with rectangular meshes, compute the difference between the original EG weak form (20) and the modified target (24) under the same initial data and boundary conditions; if that difference decays more slowly than $O(h^2)$ as the mesh is refined, the claimed second-order perturbation, and with it the optimal accuracy of the limited scheme, fails.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the flux limiting framework for DG cell averages and the LLF low-order scheme used as the baseline for cell-average constraints."},{"cited_title":"Kuzmin, M","cited_arxiv_id":null,"evidence_quote":"Provides the CG entropy stability condition and the zero-sum argument for element contributions that the nodal limiter adapts."},{"cited_title":"Becker, E","cited_arxiv_id":null,"evidence_quote":"Defines the original reduced-P1/EG method and the linear stability analysis that motivates the enrichment."},{"cited_title":"Kuzmin, M","cited_arxiv_id":null,"evidence_quote":"Supplies the theorem converting local bound preservation plus CFL conditions into invariant-domain preservation for fully discrete schemes."}],"review_version":1}