Pith. sign in

REVIEW 4 major objections 5 minor 51 references

A Symplectic Map Approach to Magnetic Field-Line Dynamics in Tokamaks

T0 review · 4 major / 5 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read In the conservative Tokamap, lowering magnetic shear slows chaotic transport as $\tau_c\propto x_q^{-0.213}$.

desk verdict Plausible scaling exponents for the conservative Tokamap, but the headline recurrence scaling rests on a single exponential fit that excludes the sticky tail. read the letter →

arxiv 2608.05414 v1 pith:F4FXWN35 submitted 2026-08-05 nlin.CD physics.plasm-ph

Diego F. M. Oliveira , Edson D. Leonel This is my paper · ORCID
classification nlin.CDphysics.plasm-ph
keywords Tokamapsymplecticmappingmagneticfield-linetransportHamiltonianchaosPoincarérecurrencestatisticsstickinessdynamicscalingshear
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

The paper aims to establish that long-time magnetic field-line transport in a tokamak is governed by the geometry of invariant structures, not just by local chaos. Working with the conservative Tokamap, an exact area-preserving map of field-line motion, the authors show that lowering the magnetic-shear parameter $x_q$ systematically slows transport. The quantitative core is the algebraic law $\tau_c \propto x_q^{-0.213}$: the characteristic time between returns to a fixed phase-space region grows as the shear decreases. Because the dynamics stay chaotic throughout, the slower transport is attributed to stickiness near KAM islands, resonance chains, and cantori. If correct, the result connects microscopic phase-space structure to macroscopic confinement-relevant transport times and implies that Lyapunov instability alone cannot predict transport efficiency.

What carries the argument

The argument is carried by the conservative Tokamap itself, the exact symplectic map defined by $\Psi_{n+1}=\tfrac{1}{2}(P_n+\sqrt{P_n^2+4\Psi_n})$ and $T_{n+1}=T_n+1/q(\Psi_{n+1})-\tfrac{x_L}{4\pi^2}\cos(2\pi T_n)/(1+\Psi_{n+1})^2$ (mod 1), with $P_n=\Psi_n-1-\tfrac{x_L}{2\pi}\sin(2\pi T_n)$ and the safety-factor profile $q(\Psi)=x_q(1+\Psi^2)^2$. This map preserves phase-space area exactly, so the invariant tori, islands, and cantori are genuine features of the dynamics rather than numerical artifacts. The transport analysis rests on two further pieces of machinery: the homogeneous scaling hypothesis $\Psi(n,x_q)=\ell F(\ell^a n,\ell^b x_q)$, which yields the exponent relation $z=\delta/\beta$, and the survival probability $S(\tau)=P(T>\tau)$, whose exponential decay constant $B$ in the interval $S(\tau)>0.08$ defines the characteristic transport time $\tau_c=1/|B|$.

What would settle it

Repeat the recurrence measurement with many independent chaotic orbits at $x_q=10^{-4}$ and $x_q=10^{-8}$, fitting $|B|$ to each survival curve over $S(\tau)>0.08$ after the same $10^3$-iteration transient; if the orbit-to-orbit spread in $|B|$ is comparable to the factor of $10^4$ separating the two parameter values, or if the early tail is visibly non-exponential, then $\tau_c\propto x_q^{-0.213}$ would not survive.

Watch

Extended reading notes

Core claim

The central discovery is that the conservative Tokamap's mixed phase space produces a scale-invariant transport law with a well-defined long-time exponent. For ensembles of $10^4$ trajectories, the mean radial coordinate grows as $\Psi\propto n^{\beta}$, saturates as $\Psi_{\mathrm{sat}}\propto x_q^{\delta}$, and crosses over at $n_x\propto x_q^{z}$, with $\beta=0.847(9)$, $\delta=-0.198(1)$, and $z=-0.234(1)$; the homogeneous scaling assumption predicts $z=\delta/\beta$, and the measured curves collapse onto one universal function under the rescalings $n\to n/x_q^z$ and $\Psi\to\Psi/x_q^{\delta}$. At long times, the survival probabilities $S(\tau)$ of returns to a fixed box shift monotonically toward longer recurrence times as $x_q$ decreases from $10^{-4}$ to $10^{-8}$. Fitting the early portion $S(\tau)>0.08$ to $S(\tau)=P_0 e^{B\tau}$ gives $|B|\propto x_q^{0.213}$, hence $\tau_c=1/|B|\propto x_q^{-0.213}$; the paper interprets this as progressively dominant stickiness generated by KAM islands, resonance chains, and cantori.

Load-bearing premise

The load-bearing premise is that the early tail of the return-time distribution from one long chaotic orbit, after a $10^3$-iteration transient, is a converged single exponential; if the survival probability is not converged or is genuinely non-exponential on $S(\tau)>0.08$, the fitted $B$ and the exponent $0.213$ would not be a robust transport law.

Editorial extensions

If this is right

  • The scaling law $z=\delta/\beta$ means only two of the three transport exponents are independent; the observed collapse of the transport curves onto one master function is the direct test of this prediction.
  • Because $\tau_c\propto x_q^{-0.213}$, reducing magnetic shear over the investigated range slows the exploration of phase space by a power law rather than by an abrupt transition.
  • Positive largest Lyapunov exponents coexist with progressively slower transport, so local exponential instability is not a sufficient measure of how efficiently field lines explore the plasma.
  • The same combination of phase-space portraits, Lyapunov exponents, scaling collapse, and recurrence statistics can be applied to other Tokamap variants and to generic Hamiltonian systems with mixed phase space.

Reading between the lines

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

  • Inference: The paper's exponent $0.213$ is measured with a single long orbit per parameter value; an ensemble of independent orbits would show whether the survival probability and its decay constant are orbit-independent or only typical.
  • Inference: Varying the recurrence box's location and size at fixed $x_q$ would test whether $\tau_c\propto x_q^{-0.213}$ is a global property of the chaotic sea or a local diagnostic tied to the chosen window.
  • Inference: A quantitative link to cantorus flux formulas is left implicit; computing the flux across the relevant partial barriers and comparing it with $1/\tau_c$ would place the exponent on firmer mechanistic ground.
  • Inference: For reversed-shear or dissipative Tokamaps, the same analysis could reveal a different exponent or a non-exponential survival tail, since separatrix reconnection and crisis-induced intermittency change the sticky structures; the paper flags these systems as future work but does not predict the outcome.
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.

Referee Report

4 major / 5 minor

Summary. The paper studies the conservative Tokamap, an exactly symplectic map for magnetic field-line dynamics, and combines phase-space portraits, Lyapunov-exponent calculations, ensemble-averaged radial transport, and Poincaré recurrence statistics. The central claims are: (i) ensemble-averaged transport exhibits dynamic scaling with growth, saturation, and crossover exponents satisfying z = δ/β and producing a universal data collapse; and (ii) the characteristic recurrence time grows algebraically as τ_c ∝ x_q^{-0.213}, which the authors attribute to increasing stickiness near KAM islands and cantori as the magnetic-shear parameter x_q is reduced. The dynamic-scaling part is internally consistent and the collapse is visually convincing. The recurrence part, however, is the weakest link: it rests on a single orbit per parameter value, an exponential fit restricted to S(τ) > 0.08, and no reported uncertainties, so the headline exponent 0.213 is not yet robustly established.

Significance. If the results are confirmed, the paper would provide a useful quantitative bridge between phase-space geometry and long-time transport in a widely studied plasma-physics map. The dynamic-scaling framework is a genuine strength: the exponents β, δ, and z are measured independently from the ensemble transport, the scaling relation z = δ/β is satisfied within the reported errors, and the collapse in Fig. 2(b) is a nontrivial consistency test. The recurrence analysis aims at an important and timely question—how partial barriers and stickiness control asymptotic transport—and a validated power law for τ_c would be a valuable addition. However, the current evidence for the headline recurrence law is incomplete, so the paper's significance depends on additional numerical work rather than on the present manuscript alone.

major comments (4)
  1. [Section 2.2, Eq. (43) and Fig. 4] The central result τ_c ∝ x_q^{-0.213} is not yet supported with adequate statistical rigor. For each x_q, the survival probability S(τ) is computed from a single orbit after discarding only 10^3 iterations, and the decay coefficient B is obtained by fitting S(τ) = P0 e^{Bτ} over the restricted range S(τ) > 0.08. No error bars are reported for |B|, no check of convergence with respect to orbit length or number of recurrences is given, and the sensitivity of the fitted exponent to the fit window and to the choice of recurrence box is not examined. With five points spanning five decades, a claimed exponent of 0.213 should be accompanied by a stability analysis; otherwise the value could be an artifact of the fit window rather than a transport law.
  2. [Section 2.2, Eqs. (40)-(46)] The interpretation of the exponential fit is in tension with the paper's own discussion of stickiness. The text states that the long-time tail of S(τ) carries information about sticky motion near KAM islands and cantori, but the fit deliberately discards precisely that tail (S(τ) ≤ 0.08). If the true recurrence-time distribution develops a power-law tail, the exponential decay constant is not a robust characteristic transport time and may not reflect the asymptotic trapping that the paper invokes to explain the slowdown. The manuscript should either justify the exponential model over the fitted range, quantify the tail separately, or present a complementary analysis of the long-time behavior before attributing τ_c ∝ x_q^{-0.213} to stickiness.
  3. [Section 2, Eq. (17) and Section 2.2, Eq. (40)] There is a direct inconsistency in the definition of the phase-space domain. The text repeatedly states that the radial variable satisfies 0 ≤ Ψ ≤ 1, with Ψ = 1 at the plasma boundary, yet the recurrence box is centered at Ψ = 5.0 and Fig. 1(a) shows Ψ ranging up to about 40, while Fig. 1's caption introduces an undefined parameter δ = 10^{-6}. The recurrence statistics are therefore obtained in a region that appears to lie outside the stated physical domain. This must be resolved: either the normalization of Ψ is different from what is written, or the recurrence box is misplaced, in which case the reported scaling may be specific to a particular (possibly non-physical) phase-space region.
  4. [Section 2.1, Fig. 3] The dynamic-scaling exponents are presented with standard errors from least-squares fits, but the universal collapse in Fig. 2(b) is assessed only visually. Given that the collapse is a central supporting result, the authors should provide a quantitative measure of collapse quality, for example the maximum or rms deviation of the rescaled curves from a master curve. This would also serve as a check on the claim that the scaling function is universal over the full range of x_q studied.
minor comments (5)
  1. [Fig. 1 caption] The caption introduces the parameter δ = 10^{-6} without defining it anywhere in the text; if it is a leftover from a previous version, it should be removed or defined.
  2. [Figs. 2 and 4] Several axis labels and tick labels are garbled or hard to read, for example '10°1' in Fig. 2 and the exponent notation in Fig. 4. The figures should be regenerated with clear mathematical notation, including negative exponents.
  3. [Section 2.2, Eq. (43)] The notation S(τ) = P0 e^{Bτ} with B negative is unconventional; using S(τ) = P0 e^{-τ/τ_c} would make the interpretation of τ_c as a decay time immediate and would avoid the absolute-value notation |B| introduced later.
  4. [Section 2.1, Eq. (23)] The ensemble average in Eq. (23) averages over the first N iterations of each trajectory, but the text does not state whether the initial transient is discarded for the ensemble transport analysis; for consistency with the recurrence analysis, this should be specified.
  5. [References] Reference [9] is cited as 'Rechester and Rosenbluth 2020' in the bibliography, but the original work is much older; the citation should point to the original publication or clarify that it is a reprint.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the scaling exponents are empirical fits and the z=δ/β relation is an internal consistency check of an explicitly stated homogeneity ansatz, not a prediction forced by construction.

full rationale

The paper's central scaling claims are measured numerical characterizations rather than results that reduce to their own inputs. The dynamic-scaling relation z=δ/β (Eq. 35) follows algebraically from the explicitly stated generalized homogeneous ansatz (Eq. 27); testing this relation by comparing a directly fitted crossover exponent with the value obtained from independently fitted β and δ is a legitimate consistency check of the ansatz, not a circular derivation. The recurrence result τ_c ∝ x_q^{−0.213} (Eq. 46) is obtained from a numerical fit of the exponential decay coefficient |B| versus x_q (Eq. 44) together with the definition τ_c = 1/|B| (Eq. 45); this is a reported empirical scaling law, not a disguised input presented as an independent prediction. The self-citations [44,45] motivate the scaling framework but are not load-bearing, because the scaling hypotheses are stated and tested against the present Tokamap data. Several internal inconsistencies exist—Ψ is introduced as lying in [0,1] while figures and the recurrence box use Ψ values around 5–40, and the Fig. 1 caption introduces an undefined δ=10^{−6}—and the restriction of the exponential fit to S(τ)>0.08 raises a legitimate robustness concern about whether the discarded tail, where stickiness would be expected, is adequately captured. These are correctness and statistical-robustness issues, not circularity; the derivation chain does not reduce to its own conclusion by construction.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new entities. Its claims rest on a phenomenological scaling ansatz and on several fitted exponents, plus a specific safety-factor profile and an exponential fit to recurrence statistics. The homogeneous scaling relation is a generic consistency condition, not a first-principles derivation.

free parameters (4)
  • growth exponent β = 0.847(9)
    Fitted to the power-law growth regime of the ensemble-averaged radial coordinate Ψ(n, x_q) in Fig. 2(a).
  • saturation exponent δ = -0.198(1)
    Fitted to the saturation value Ψ_sat versus x_q in Fig. 3(a).
  • crossover exponent z = -0.234(2)
    Fitted to crossover iteration n_x versus x_q in Fig. 3(b); also predicted as δ/β = -0.233(3).
  • recurrence decay exponent = 0.213
    Fitted to |B| versus x_q in Fig. 4(b), giving τ_c ∝ x_q^{-0.213}.
assumptions (4)
  • domain assumption The ensemble-averaged radial coordinate is a generalized homogeneous function, Ψ(n, x_q) = ℓ F(ℓ^a n, ℓ^b x_q) (Eq. 27).
    Self-similarity of transport is assumed, not derived from the map; it is the basis for the scaling relation z = δ/β.
  • domain assumption The survival probability follows a single exponential, S(τ) = P0 e^{Bτ}, over the interval S(τ) > 0.08 (Eq. 43).
    This fit form discards the long-time tail, which is precisely where sticky dynamics would appear; the fitted B defines τ_c.
  • domain assumption The safety-factor profile q(Ψ) = x_q (1 + Ψ^2)^2 is a representative monotonic tokamak equilibrium (Eq. 16).
    Results are for this specific profile; the paper does not test robustness to other profiles.
  • domain assumption Recurrence statistics from a single long orbit approximate the equilibrium ensemble (Section 2.2).
    No error bars or multiple-orbit averaging are provided, so convergence of S(τ) is assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Symplectic Map Approach to Magnetic Field-Line Dynamics in Tokamaks." pith.science (2026). https://pith.science/paper/F4FXWN35

@misc{pith2026260805414,
  author       = {Pith},
  title        = {Pith review of: A Symplectic Map Approach to Magnetic Field-Line Dynamics in Tokamaks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F4FXWN35}},
  note         = {Machine review of arXiv:2608.05414}
}
abstract

Magnetic field-line transport in tokamaks is governed by the interplay between chaotic dynamics and invariant phase-space structures that act as partial transport barriers. We investigate the conservative Tokamap, an exact symplectic mapping for magnetic field-line dynamics, using a unified geometrical, dynamical, and statistical framework. Phase-space portraits and Lyapunov exponents characterize the mixed Hamiltonian dynamics, while ensemble-averaged transport exhibits universal dynamic scaling with growth, crossover, and saturation regimes connected through a homogeneous scaling theory. Poincar\'e recurrence statistics reveal that decreasing the magnetic-shear parameter systematically slows transport, with the characteristic transport time following the algebraic scaling $\tau_c\propto x_q^{-0.213}$. This behavior reflects increasingly dominant stickiness associated with KAM islands, resonance chains, and cantori. Our results establish a direct connection between the geometrical organization of Hamiltonian phase space and macroscopic transport properties, providing a comprehensive framework for understanding long-time magnetic field-line transport in tokamaks and other Hamiltonian systems with mixed phase space.

Figures

Figures reproduced from arXiv: 2608.05414 by the authors.

Figure 1
Figure 1. (a) Phase-space portrait showing invariant spanning curves, KAM islands, and a chaotic sea. The spanning curves act as transport barriers, the islands correspond to regular motion around stable periodic orbits, and the chaotic sea consists of irregular trajectories exhibiting stochastic transport. (b) Distribution of the largest Lyapunov exponent, yielding a mean value of λ¯ = 4.12 ± 0.01, consistent with chaotic be… view at source ↗
Figure 2
Figure 2. (a) Different curves of Ψ for five different values of the control parameter xq. (b) Collapse of the data onto a single universal curve after the appropriate scaling transformations. 2.1 Ensemble-Averaged Transport Dynamics Chaotic transport in Hamiltonian systems is intrinsically statistical. Although trajectories initiated from nearby initial conditions obey the same deterministic equations of motion, they may exp… view at source ↗
Figure 3
Figure 3. (a) Behavior of the saturation value Ψsat as a function of xq. (b) Behavior of the crossover iteration number nx as a function of xq. Power-law fits yield the critical exponents δ = −0.198(1) in (a) and z = −0.234(1) in (b). Using Eqs. (30) and (33), the exponent ratio becomes a b = δ β , (34) and comparison with Eq. (26) yields the scaling law z = δ β . (35) Equation (35) demonstrates that only two of the three cri… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Recurrence statistics for the conservative Tokamap. (a) Survival probability S(τ) for different values of the perturbation parameter xq, computed using a recurrence box centered at (T, Ψ) = (0.5, 5.0) with ∆T = 0.01 and ∆Ψ = 0.10. The systematic shift of the curves tow…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

51 extracted references · 51 canonical work pages

  1. [1]

    Wesson J and Campbell D J 2011Tokamaksvol 149 (Oxford university press)

    work page

  2. [2]

    Ongena J, Koch R, Wolf R and Zohm H 2016Nature Physics12398–410 13 Diego F. M . Oliveira & Edson D. Leonel

    work page

  3. [3]

    Stacey W M 2010Fusion: An introduction to the physics and technology of magnetic confinement fusion(John Wiley & Sons)

    work page

  4. [4]

    Miyamoto K 2005Plasma physics and controlled nuclear fusion(Springer)

    work page

  5. [5]

    White R B 2013Theory Of Toroidally Confined Plasmas, The(World Scientific Publishing Company)

    work page

  6. [6]

    Hazeltine R D and Meiss J D 2003Plasma confinement(Courier Corporation)

    work page

  7. [7]

    Freidberg J P 2014ideal MHD(Cambridge University Press)

    work page

  8. [8]

    Freidberg J P 2008Plasma physics and fusion energy(Cambridge university press)

    work page

Show all 51 references
  1. [9]

    Rechester A and Rosenbluth M 2020 Electron heat transport in a tokamak with destroyed magnetic surfacesHamiltonian Dynamical Systems(CRC Press) pp 684–7

    2020

  2. [10]

    Abdullaev S S 2006Construction of Mappings for Hamiltonian Systems and Their Applications(Berlin: Springer)

  3. [11]

    Evans T E, Moyer R A, Burrell K H, Fenstermacher M E, Joseph I, Leonard A W, Osborne T H, Porter G D, Schaffer M J, Snyder P Bet al.2006nature physics2419–23

  4. [12]

    Evans T, Burrell K, Fenstermacher M, Moyer R, Osborne T, Schaffer M, West W, Yan L, Boedo J, Doyle Eet al.2006Physics of plasmas13

  5. [13]

    Loarte A, Lipschultz B, Kukushkin A, Matthews G, Stangeby P, Asakura N, Counsell G, Federici G, Kallenbach A, Krieger Ket al.2007Nuclear Fusion47S203–63

  6. [14]

    Kolmogorov A N 1954 On conservation of conditionally periodic motions for a small change in hamilton’s functionDokl. akad. nauk Sssrvol 98 pp 527–30

    1954

  7. [15]

    Vladimir I 1963Russian Mathematical Surveys1885–191

  8. [16]

    M¨ oser J 1962Nachr. Akad. Wiss. G¨ ottingen, II1–20

  9. [17]

    Chirikov B V 1979Physics reports52263–379

  10. [18]

    Lichtenberg A J and Lieberman M A 2013Regular and chaotic dynamicsvol 38 (Springer Science & Business Media)

  11. [19]

    Meiss J 1992Reviews of Modern Physics64795

  12. [20]

    Leonel E D, Galia M V C, Barreiro L A and Oliveira D F 2016Physical Review E94062211

  13. [21]

    Greene J M 1979Journal of Mathematical Physics201183–201

  14. [22]

    MacKay R, Meiss J and Percival I 1984Physica D: Nonlinear Phenomena1355–81

  15. [23]

    Zaslavsky G M 2002Physics reports371461–580

  16. [24]

    Zaslavsky G M 2007The physics of chaos in Hamiltonian systems(world scientific)

  17. [25]

    Altmann E G, Motter A E and Kantz H 2005Chaos: An Interdisciplinary Journal of Nonlinear Science15

  18. [26]

    Boozer A H 1979 Guiding center drift equations Tech. rep. Princeton Univ., NJ (USA). Plasma Physics Lab

    1979

  19. [27]

    Cary J R and Littlejohn R G 1983Annals of Physics1511–34

  20. [28]

    Constantinescu D 2005Romanian Journal of Physics50325

  21. [29]

    Hudson S, Dewar R, Dennis G, Hole M, McGann M, Von Nessi G and Lazerson S 2012 Physics of Plasmas19

    2012

  22. [30]

    Abdullaev S S 2006Construction of mappings for Hamiltonian systems and their applications (Springer)

  23. [31]

    Ullmann K and Caldas I L 2000Chaos, Solitons & Fractals112129–40 14 Diego F. M . Oliveira & Edson D. Leonel

  24. [32]

    Morrison P 2000Physics of Plasmas72279–89

  25. [33]

    Abdullaev Set al.2014Magnetic stochasticity in magnetically confined fusion plasmasvol 78 (Springer)

  26. [34]

    Balescu R, Vlad M and Spineanu F 1998Physical Review E58951

  27. [35]

    Eberhard M 2005Physical Review E—Statistical, Nonlinear, and Soft Matter Physics71 026411

  28. [36]

    Bartoloni B, Schelin A and Caldas I L 2016Physics Letters A3802416–21

  29. [38]

    Wingen A, Spatschek K and Abdullaev S 2005Contributions to plasma physics45500–13

  30. [39]

    Misguich J 2001Physics of Plasmas82132–8

  31. [40]

    Balescu R 1998Physical Review E583781

  32. [41]

    Shinohara S and Aizawa Y 1998Progress of theoretical physics100219–33

  33. [42]

    Eckmann J P and Ruelle D 1985Reviews of modern physics57617

  34. [43]

    Benettin G, Galgani L, Giorgilli A and Strelcyn J M 1980Meccanica159–20

  35. [44]

    Oliveira D F, Bizao R A and Leonel E D 2009Mathematical Problems in engineering2009 213857

  36. [45]

    Oliveira D F, Silva M R and Leonel E D 2015Physica A: Statistical Mechanics and its Applications436909–15

  37. [46]

    Meiss J D and Ott E 1986Physica D: Nonlinear Phenomena20387–402

  38. [47]

    Karney C F 1983Physica D: Nonlinear Phenomena8360–80

  39. [48]

    Chirikov B V and Shepelyansky D L 1984Physica D: Nonlinear Phenomena13395–400

  40. [49]

    Altmann E G and Kantz H 2005Physical Review E—Statistical, Nonlinear, and Soft Matter Physics71056106

  41. [50]

    Viana R L, Mugnaine M and Caldas I L 2023Physics of Plasmas30

  42. [51]

    Balescu R 2005Aspects of anomalous transport in plasmas(CRC Press)

  43. [52]

    Wootton A, Carreras B, Matsumoto H, McGuire K, Peebles W, Ritz C P, Terry P and Zweben S 1990Physics of Fluids B: Plasma Physics22879–903 15

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.