Modeling and Stabilization of Transport-Dominated Flows

Abhijnan Dikshit; Anshi Gupta; Bhavana Morankar; Christopher Agesen; Jack Murphy; Karnav Raval; Lane H. Rogers; Nanda N. Raghunathan; Sam Bradford; Valeria Barra

REVIEW 3 major objections 5 minor 17 references

Modeling and Stabilization of Transport-Dominated Flows

T0 review · 3 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read On the slotted-cylinder advection benchmark, streamline-upwind Petrov-Galerkin stabilization does not improve final-time errors over the unstabilized finite-element baseline, while quasi-monotone limiters nearly eliminate overshoots and und

desk verdict A competent camp report whose headline negative result about SUPG is provisional because the implementation uses an acknowledged lagged time-derivative proxy and no code is provided. read the letter →

arxiv 2607.22455 v1 pith:UGR53QY5 submitted 2026-07-24 physics.flu-dyn cs.NAmath-phmath.MPmath.NA

Christopher Agesen , Sam Bradford , Abhijnan Dikshit , Anshi Gupta , Bhavana Morankar , Jack Murphy , Nanda N. Raghunathan , Karnav Raval

show 2 more authors

Lane H. Rogers Valeria Barra

This is my paper · ORCID

classification physics.flu-dyncs.NAmath-phmath.MPmath.NA MSC 65M6065M1276M1086A10

keywords SUPGstabilizationstreamline-upwindtransport-dominatedflowsslotted-cylinderadvectionquasi-monotonelimitershyperdiffusionfiniteelementmethodasymptoticexpansion

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to settle how different stabilization strategies behave on a standard deformational-flow benchmark: advection of slotted cylinders on a sphere. Its empirical claim is that SUPG, despite being the theoretically favored streamline-diffusion method, shows no error-metric improvement over no stabilization at final time in this test, while quasi-monotone limiters reduce spurious extrema to near machine precision at the price of smearing the tracer field. It also proves that the conservative semidiscrete SUPG formulation conserves total mass but does not preserve pointwise bounds or TVD, so the observed ringing is structural. Finally it proposes an asymptotic-expansion SUPG algorithm that replaces the explicit previous-step proxy for the time derivative with a series of explicit Galerkin solves.

What carries the argument

The key mechanism is the Petrov-Galerkin test-function perturbation v + τ u·∇v, where τ is a stabilization parameter assembled from advective, temporal, and diffusive timescales in an inverse-norm formula. The SUPG correction adds residual-based diffusion along streamlines only; the tracer equation is stabilized in mass form with the advective residual weighted by density. The paper's asymptotic proposal replaces the previous-time-step proxy for ∂Q/∂t with an expansion Q = Q0 + εQ1, so each time step solves two explicit Galerkin problems instead of one implicit SUPG problem.

What would settle it

Run the same slotted-cylinder benchmark with a fully implicit SUPG (or sufficiently small Δt so the time-derivative proxy is accurate) and recompute the error table. If SUPG then improves on the unstabilized baseline in ℓ1 or ℓ2, the paper's negative verdict is implementation-driven; if it still does not, the verdict is robust.

Watch

Extended reading notes

Core claim

The central discovery is comparative and partly negative: for the slotted-cylinder test with the supplied resolution and error metrics, the SUPG stabilization term (added mass-weighted to the tracer residual) does not beat the unstabilized Galerkin baseline in ℓ1, ℓ2, ℓ∞, overshoot, or undershoot; it does, however, give the smallest global errors among stabilized runs and preserves sharp features. Quasi-monotone limiters alone or in combination drive overshoot and undershoot down to about 1e-7 but roughly double the ℓ1 and ℓ2 errors. The paper's theoretical section explains why: SUPG is consistent residual-based dissipation that conserves total mass but is not a monotone or bound-preserving

Load-bearing premise

The empirical comparison rests on using the previous time step's ∂Q/∂t as a proxy in the SUPG residual; if that lag injects error, the measured non-improvement of SUPG may be an artifact rather than a property of the method.

Editorial extensions

If this is right

On this benchmark, the quasi-monotone limiter's extrema control comes with large global shape error; no tested method wins on every metric.
Standard linear SUPG conserves total density and tracer mass on the closed sphere, but does not guarantee nonnegativity, a discrete maximum principle, or TVD behavior; the observed overshoots are consistent with the method's structure.
Combining the limiter with hyperdiffusion, SU, or SUPG leaves extrema metrics at the limiter-only level, so the limiter dominates extrema behavior.
An asymptotic-expansion SUPG scheme can in principle include a non-delayed time derivative in the residual at the cost of solving one extra explicit Galerkin problem per expansion term.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the explicit previous-step proxy for ∂Q/∂t introduces O(Δt) error, a fairer test of SUPG would use small time steps or an implicit treatment; the no-improvement verdict may be implementation-limited rather than intrinsic.
Tracking error at every time step (not just final) would likely show transient differences, since the benchmark returns to the initial state at final time, and may reveal when each method's bias enters.
The nondimensional stabilization size ε ≈ 4e-3 suggests SUPG's correction is genuinely weak at atmospheric scales; this predicts its effect stays small on this test regardless of implementation, and higher-order polynomial degree or sharper initial data would be needed to expose its benefit.
A testable extension: implement the asymptotic Q0 + εQ1 algorithm and compare against implicit SUPG; if the series converges quickly, it offers near-implicit accuracy at explicit cost.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this is a well-scoped computational study from a graduate modeling camp, with a clean theoretical section and a useful cross-framework consistency check. But the central negative result—SUPG not improving on unstabilized advection—rests on an implementation detail the paper itself flags as inaccurate, and the conclusions don't carry that caveat. Take the empirical claims conditionally until the code is available or the residual is fixed.

What's actually new: the specific comparison of SU/SUPG against limiters and hyperdiffusion in ClimaCore on the slotted-cylinder test, the Firedrake SU cross-check, and a proposed asymptotic SUPG algorithm (not implemented). The theoretical analysis in §5.1 is standard but correct: the conservation proof via the constant test function is clean, and the distinction between conservation and pointwise boundedness is worth stating. The authors also honestly acknowledge that the metrics are final-time only and that higher-order tests are missing.

Where the paper gets shaky: the SUPG result. Section 5.2 explicitly says the time derivative in the residual is 'computed from the previous time step as a proxy' and that this 'suffers from inaccuracy for large time steps.' That is exactly the setup used for Table 4. If the proxy error is the same order as the stabilization benefit, the measured non-improvement could be an artifact. The conclusions in §4 and §6 do not mention this caveat. In addition, the paper never states precisely which residual form the ClimaCore code uses—the factored ρ R_adv(q) with ∂_t q from §3.3, or the conservative ∇·(Qu) + ∂_t Q used in the §5.2 asymptotic derivation. At the discrete level these differ by a density-residual term, and the lagged variable appears to be Q, not q. No code or commit hash is provided, so a reader cannot check which one was implemented. That makes the headlining empirical claim load-bearing but unverified.

The asymptotic algorithm is a reasonable idea but unvalidated, and the expansion parameter ε ≈ 4×10⁻³ is only loosely justified. Fine as future-work speculation.

For whom: researchers working on FEM-based atmospheric dycores who want a quick map of trade-offs on one benchmark. It is not a definitive study. A serious referee could engage with this, but only with the request to supply code, correct the SUPG residual, and rerun the benchmark. As is, the empirical conclusion should not be cited as established.

Referee Report

3 major / 5 minor

Summary. The manuscript compares stabilization strategies for transport-dominated spherical advection using the slotted-cylinder benchmark. SU and SUPG are implemented in ClimaCore.jl, and SU also in Firedrake; these are compared against hyperdiffusion, quasi-monotone limiters, and combinations. The central empirical claim is that quasi-monotone limiters best control unphysical extrema (q_over/q_under ~ 1e-7) at the cost of larger global errors, while SUPG gives the lowest global errors among stabilized runs but does not improve on the unstabilized baseline in the computed metrics. A theoretical section proves semidiscrete conservation for the conservative SUPG formulation, explains why standard linear SUPG is not bound-preserving, and proposes an asymptotic-expansion-based time-dependent SUPG algorithm as future work.

Significance. If the negative SUPG result is robust, it is a useful caution for climate-dycore practitioners: streamline stabilization may not pay off on this benchmark under final-time global-error metrics. The paper has clear strengths: the semidiscrete conservation proof in §5.1 is clean; the stabilization parameter uses the standard Tezduyar–Osawa formula with no fitted parameters; the comparison includes two independent implementations; and the authors explicitly acknowledge limitations regarding final-time-only metrics and the lack of higher-order tests. However, the central empirical claim rests on an implementation detail that the paper itself flags as inaccurate, and the conclusions do not carry that caveat. The theoretical analysis is largely independent of this issue and is sound in its main claims, though the asymptotic algorithm is only proposed, not implemented or tested.

major comments (3)

[§5.2, Table 4, Abstract] The SUPG negative result depends on the lagged residual time derivative. The text states that ∂Q/∂t in the SUPG correction is 'computed from the previous time step as a proxy' and that this 'suffers from inaccuracy for large time steps.' At the reported Δt = 345.6 s, there is no convergence study in Δt and no comparison with a non-lagged or implicit SUPG implementation. Without evidence that the lagged term is small compared with the stabilization effect, the conclusion 'SUPG was not found to improve upon the no stabilization case' cannot be distinguished from an artifact of the proxy. This caveat is absent from the Abstract and §4/§6 conclusions. The authors should either verify the result with a non-lagged residual or explicitly state that the result is conditional on this implementation.
[§3.3, Eq. (6), Table 1, §5.2] There is an inconsistency between the residual forms used in the discrete implementation and in the theoretical derivation. Equation (6) defines the SUPG correction using ρ_h R_adv(q_h), with the 'consistent' option in Table 1 being ∂_t q + u·∇q. In contrast, §5.2 describes the implemented stabilizing term as involving ∇·(Qu) + ∂Q/∂t, with ∂Q/∂t lagged. At the discrete level, where ρ_h does not satisfy the continuity equation exactly (as acknowledged in §3.3), these two forms differ by a term proportional to q_h times the density residual. The code's actual choice is not stated. If the lagged variable is Q rather than q, the correction is not consistent with the factored form in Eq. (6). The manuscript should clarify which residual is actually used and, if necessary, rerun the SUPG case with the consistent advective residual.
[§4, §6] The central conclusions overstate the reliability of the SUPG comparison. The statements that SUPG 'does not improve on the no stabilization baseline for any of the metrics' and 'was not found to improve upon the no stabilization case' are presented without the caveat that the result is conditional on the lagged-residual proxy. Because no code or commit hash is provided, the implementation cannot be independently checked. The manuscript should either provide the implementation details needed for reproducibility or soften the claims so that they are explicitly scoped to the implemented lagged form.

minor comments (5)

[Table 4, §4] The Firedrake/ClimaCore SU comparison reports ℓ∞ = 4.50 versus 6.17×10⁻¹, a factor of about 7.3. The text says these differ by 'a single order of magnitude'; this is borderline. More importantly, the cause of this discrepancy is not investigated, which weakens the claim that the two implementations provide a strong consistency check.
[§5.2, Eq. (32)] The finite-difference formula (32) writes ∂Q_i^n/∂t ≈ (Q_i^n − Q_i^{n−1})/Δt, but Step 2 of the algorithm uses ∂Q_0^{n+1}/∂t computed from Q_0^n and Q_0^{n+1}, which corresponds to (Q_0^{n+1} − Q_0^n)/Δt. The indexing should be made consistent.
[§5.2] The sentence 'Substituting into (19)' appears to refer to the nondimensionalized SUPG equation (29), not to the exact mixing-ratio equation (19). This is a typographical error that should be corrected.
[§2] The initial-condition formula uses logical notation such as '∧' without definition and defines D(s) without specifying its domain for negative arguments. This makes the initialization harder to parse than necessary.
[References] Reference [9] is cited as 'Stabilization Parameters in SUPG and PSPG Formulations' but the year is given as 2003; the actual Tezduyar–Sathe reference should be checked and the full bibliographic details completed. Minor formatting issues appear in several URLs (e.g., 'doi:https://doi.org/...' with line breaks).

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; empirical benchmark and theoretical analysis are self-contained.

full rationale

The paper's central empirical comparison (Table 4) is an implementation benchmark, not a derivation from fitted parameters. The stabilization parameter uses the standard Tezduyar–Osawa formula (8); no parameter is fitted to the error metrics that are then reported. The SUPG/SU residual forms follow Brooks–Hughes [4], and the theoretical conservation/boundedness analysis in §5.1 is derived directly from the PDE (17)–(19). The asymptotic SUPG algorithm in §5.2 is explicitly proposed as future work and is not used to produce Table 4, so no circularity arises from it. The genuine caveat is the lagged ∂Q/∂t proxy in §5.2, which could weaken the SUPG comparison, but this is an implementation-validity threat, not circularity: the negative SUPG result is conditional on that proxy rather than entailed by the inputs. No self-citation is load-bearing; [5] and [7] are references to the computational environment and dycore paper, not to a result that the present claims reduce to. Hence there is no circular step to report.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

The central comparison uses standard numerical schemes; no constants are fitted to the target data. The only hand-chosen coefficient is the hyperdiffusion value, which affects the hyperdiffusion-only and combined runs. The theory relies on standard FEM assumptions and the closed-sphere domain.

free parameters (1)

Hyperdiffusion coefficient = 6.6×10^14
Chosen from the ClimaCore example (not fitted to target result) for the H runs; it affects the hyperdiffusion and combined runs but not the SUPG conclusion.

assumptions (3)

domain assumption The computational domain is a closed sphere and the finite element space contains the constant test function wh=1.
Used in §5.1 to prove semidiscrete conservation; if the domain had boundary or constants were not in the space, the boundary term would not vanish and conservation would not follow.
domain assumption The discrete density satisfies the continuity equation closely enough to factor R(ρq) = ρ R_adv(q); otherwise the conservative and advective forms decouple.
§3.3: 'if ρh does not satisfy (9) exactly ... conservative and advective forms decouple', so the implementation and theory assume ρh is near-exact.
standard math Standard finite element spaces, L2 inner products, and the divergence theorem on the sphere are valid.
Used throughout the weak formulation in §2.1 and §3.1.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Modeling and Stabilization of Transport-Dominated Flows." pith.science (2026). https://pith.science/paper/UGR53QY5

@misc{pith2026260722455,
  author       = {Pith},
  title        = {Pith review of: Modeling and Stabilization of Transport-Dominated Flows},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UGR53QY5}},
  note         = {Machine review of arXiv:2607.22455}
}

read the original abstract

This report explores and compares numerical stabilization methods for transport-dominated flows arising in atmospheric modeling. The Streamline-Upwind (SU) and Streamline-Upwind Petrov-Galerkin (SUPG) stabilization formulations are implemented in Julia's ClimaCore.jl package and Python's Firedrake package for a test problem with slotted-cylinder initial conditions. These methods are compared for their ability to mitigate spurious oscillations while preserving sharp features against existing hyperdiffusion and quasi-monotone limiter methods, alongside various combinations. For this benchmark, quasi-monotone limiters were most effective at minimizing un-physical extrema, at the expense of diffusing the overall structure. The SUPG method was not found to improve upon the no stabilization case, for this test case in the computed error metrics. However, it performs best at preserving sharp feature and overall structure, among tested stabilized runs. Theoretical properties of SUPG are analyzed, and an asymptotics-based SUPG algorithm is proposed as future work.

Figures

Figures reproduced from arXiv: 2607.22455 by the authors.

**Figure 2.** Comparison of the streamline-upwind (SU) solutions computed with Clima [PITH_FULL_IMAGE:figures/full_fig_p013_2.png] view at source ↗

**Figure 3.** Tracer fields at final time point t = T for four stabilization methods implemented in ClimaCore.jl. Now turning our attention to [PITH_FULL_IMAGE:figures/full_fig_p014_3.png] view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: Diagrammatic representation of the solution algorithm for the asymptotic [PITH_FULL_IMAGE:figures/full_fig_p020_4.png]

**Figure 5.** Figure 5: Tracer fields at the intermediate time t = T/2 for all stabilization choices implemented in ClimaCore.jl. 25 [PITH_FULL_IMAGE:figures/full_fig_p025_5.png]

**Figure 6.** Figure 6: Tracer fields at the final time t = T for all stabilization choices implemented in ClimaCore.jl. 26 [PITH_FULL_IMAGE:figures/full_fig_p026_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

17 extracted references · 4 canonical work pages

[1]

Impact and importance of hyperdiffusion on the spectral element method: A linear dispersion analysis

Paul A. Ullrich et al. “Impact and importance of hyperdiffusion on the spectral element method: A linear dispersion analysis”. In:Journal of Computational Physics375 (Dec. 2018), pp. 427–446.issn: 10902716.doi:10.1016/j.jcp. 2018.06.035. 22

doi:10.1016/j.jcp 2018
[2]

Optimization-based limiters for the spectral element method

Oksana Guba, Mark Taylor, and Amik St-Cyr. “Optimization-based limiters for the spectral element method”. In:Journal of Computational Physics(2010).doi: https://doi.org/10.1016/j.jcp.2014.02.029

work page doi:10.1016/j.jcp.2014.02.029 2010
[3]

Fully multidimensional flux-corrected transport algorithms for fluids

Steven T Zalesak. “Fully multidimensional flux-corrected transport algorithms for fluids”. In:Journal of Computational Physics31.3 (1979), pp. 335–362.issn: 0021-9991.doi:https://doi.org/10.1016/0021- 9991(79)90051- 2.url: https://www.sciencedirect.com/science/article/pii/0021999179900512

arXiv 1979
[4]

Streamline upwind/Petrov- Galerkin formulations for convection dominated flows with particular emphasis on the incompressible Navier-Stokes equations

Alexander N. Brooks and Thomas J.R. Hughes. “Streamline upwind/Petrov- Galerkin formulations for convection dominated flows with particular emphasis on the incompressible Navier-Stokes equations”. In:Computer Methods in Ap- plied Mechanics and Engineering(1982).doi:https : / / doi . org / 10 . 1016 / 0045-7825(82)90071-8

1982
[5]

Version v0.14.51

CliMA.ClimaCore.jl. Version v0.14.51. Apr. 2025.url:https://github.com/ CliMA/ClimaCore.jl.git

2025
[6]

Ham et al.Firedrake User Manual

David A. Ham et al.Firedrake User Manual. First edition. Imperial College London et al. May 2023.doi:10.25561/104839

doi:10.25561/104839 2023
[7]

The Climate Modeling Alliance Atmosphere Dynamical Core: Concepts, Numerics, and Scaling

Dennis Yatunin et al. “The Climate Modeling Alliance Atmosphere Dynamical Core: Concepts, Numerics, and Scaling”. In:Journal of Advances in Modeling Earth Systems18.3 (2026), e2025MS005014.doi:https://doi.org/10.1029/ 2025MS005014.url:https://agupubs.onlinelibrary.wiley.com/doi/abs/ 10.1029/2025MS005014

work page doi:10.1029/2025ms005014 2026
[8]

A class of deformational flow test cases for linear transport problems on the sphere

Ramachandran D. Nair and Peter H. Lauritzen. “A class of deformational flow test cases for linear transport problems on the sphere”. In:Journal of Computa- tional Physics(2010).doi:https://doi.org/10.1016/j.jcp.2010.08.014

work page doi:10.1016/j.jcp.2010.08.014 2010

Show all 17 references

[9]

Stabilization Parameters in SUPG and PSPG Formulations

Tayfun Tezduyar and Sunil Sathe. “Stabilization Parameters in SUPG and PSPG Formulations”. In:Journal of Computational and Applied Mathematics(2003)

2003
[10]

The Gusto Development Team.Gusto: A library for dynamical cores using com- patible finite element discretisations.https://github.com/firedrakeproject/ gusto. 2026

2026
[11]

Efficient implementation of essentially non- oscillatory shock-capturing schemes

Chi-Wang Shu and Stanley Osher. “Efficient implementation of essentially non- oscillatory shock-capturing schemes”. In:Journal of computational physics(1988). doi:https://doi.org/10.1016/0021-9991(88)90177-5

1988 doi
[12]

The Effect of the Consistent Mass Matrix on the Maximum-Principle for Scalar Conservation Equations

Jean-Luc Guermond, Bojan Popov, and Yong Yang. “The Effect of the Consistent Mass Matrix on the Maximum-Principle for Scalar Conservation Equations”. In: Journal of Scientific Computing70.3 (2017), pp. 1358–1366.doi:10 . 1007 / s10915-016-0285-7

2017
[13]

A Difference Method for Numerical Calculation of Dis- continuous Solutions of the Equations of Hydrodynamics

Sergei K. Godunov. “A Difference Method for Numerical Calculation of Dis- continuous Solutions of the Equations of Hydrodynamics”. In:Matematicheskii Sbornik47.3 (1959). In Russian; English translation published as U.S. Joint Pub- lications Research Service report JPRS 7226 (1...

1959
[14]

High Resolution Schemes for Hyperbolic Conservation Laws

Ami Harten. “High Resolution Schemes for Hyperbolic Conservation Laws”. In: Journal of Computational Physics49.3 (1983), pp. 357–393.doi:10.1016/0021- 9991(83)90136-5

1983 doi
[15]

A new finite element formulation for computational fluid dynamics: II. Beyond SUPG

Michel Mallet Thomas J.R. Hughes and Akira Mizukami. “A new finite element formulation for computational fluid dynamics: II. Beyond SUPG”. In:Computer Methods in Applied Mechanics and Engineering(1986).doi:https://doi.org/ 10.1016/0045-7825(86)90110-6

1986 doi
[16]

Flux Correction Tools for Finite Elements

Dmitri Kuzmin and Stefan Turek. “Flux Correction Tools for Finite Elements”. In:Journal of Computational Physics175.2 (2002), pp. 525–558.doi:10.1006/ jcph.2001.6955

2002
[17]

Invariant Domains and Second-Order Continuous Finite Element Approximation for Scalar Conservation Equations

Jean-Luc Guermond and Bojan Popov. “Invariant Domains and Second-Order Continuous Finite Element Approximation for Scalar Conservation Equations”. In:SIAM Journal on Numerical Analysis55.6 (2017), pp. 3120–3146.doi:10. 1137/16M1106560. 24 A Tracer Fields for All Stabilization ...

2017

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.

[1] [1]

Impact and importance of hyperdiffusion on the spectral element method: A linear dispersion analysis

Paul A. Ullrich et al. “Impact and importance of hyperdiffusion on the spectral element method: A linear dispersion analysis”. In:Journal of Computational Physics375 (Dec. 2018), pp. 427–446.issn: 10902716.doi:10.1016/j.jcp. 2018.06.035. 22

doi:10.1016/j.jcp 2018

[2] [2]

Optimization-based limiters for the spectral element method

Oksana Guba, Mark Taylor, and Amik St-Cyr. “Optimization-based limiters for the spectral element method”. In:Journal of Computational Physics(2010).doi: https://doi.org/10.1016/j.jcp.2014.02.029

work page doi:10.1016/j.jcp.2014.02.029 2010

[3] [3]

Fully multidimensional flux-corrected transport algorithms for fluids

Steven T Zalesak. “Fully multidimensional flux-corrected transport algorithms for fluids”. In:Journal of Computational Physics31.3 (1979), pp. 335–362.issn: 0021-9991.doi:https://doi.org/10.1016/0021- 9991(79)90051- 2.url: https://www.sciencedirect.com/science/article/pii/0021999179900512

arXiv 1979

[4] [4]

Streamline upwind/Petrov- Galerkin formulations for convection dominated flows with particular emphasis on the incompressible Navier-Stokes equations

Alexander N. Brooks and Thomas J.R. Hughes. “Streamline upwind/Petrov- Galerkin formulations for convection dominated flows with particular emphasis on the incompressible Navier-Stokes equations”. In:Computer Methods in Ap- plied Mechanics and Engineering(1982).doi:https : / / doi . org / 10 . 1016 / 0045-7825(82)90071-8

1982

[5] [5]

Version v0.14.51

CliMA.ClimaCore.jl. Version v0.14.51. Apr. 2025.url:https://github.com/ CliMA/ClimaCore.jl.git

2025

[6] [6]

Ham et al.Firedrake User Manual

David A. Ham et al.Firedrake User Manual. First edition. Imperial College London et al. May 2023.doi:10.25561/104839

doi:10.25561/104839 2023

[7] [7]

The Climate Modeling Alliance Atmosphere Dynamical Core: Concepts, Numerics, and Scaling

Dennis Yatunin et al. “The Climate Modeling Alliance Atmosphere Dynamical Core: Concepts, Numerics, and Scaling”. In:Journal of Advances in Modeling Earth Systems18.3 (2026), e2025MS005014.doi:https://doi.org/10.1029/ 2025MS005014.url:https://agupubs.onlinelibrary.wiley.com/doi/abs/ 10.1029/2025MS005014

work page doi:10.1029/2025ms005014 2026

[8] [8]

A class of deformational flow test cases for linear transport problems on the sphere

Ramachandran D. Nair and Peter H. Lauritzen. “A class of deformational flow test cases for linear transport problems on the sphere”. In:Journal of Computa- tional Physics(2010).doi:https://doi.org/10.1016/j.jcp.2010.08.014

work page doi:10.1016/j.jcp.2010.08.014 2010