Pith. sign in

REVIEW 2 major objections 5 minor 37 references

Modeling of Ionization and Recombination Processes in Plasma with Arbitrary Non-Maxwellian Electron Distributions

T0 review · 2 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Two Maxwellian decomposition methods compute ionization and recombination rates for arbitrary electron distributions, showing that flare and solar-wind non-thermal tails can bias temperature diagnostics.

desk verdict A solid Maxwellian-decomposition tool for non-Maxwellian rates, but the flare temperature-bias claim rests on an assumed low-energy core. read the letter →

arxiv 2506.14668 v1 pith:JGR5UKGQ submitted 2025-06-17 astro-ph.SR physics.plasm-ph

Chengcai Shen , Xiaocan Li , Yuan-Kuen Ko , John C. Raymond , Fan Guo , Vanessa Polito , Viviane Pierrard This is my paper · ORCID
classification astro-ph.SRphysics.plasm-ph
keywords Maxwelliandecompositionnon-Maxwellianelectrondistributionionizationraterecombinationkappasolarflarereconnectionwindequilibrium
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 argues that computing ionization and recombination rates under the usual Maxwellian assumption can mislead temperature diagnostics whenever the electron population has a non-thermal tail, and it provides two practical ways to compute the correct rates for any electron distribution. The trick is to decompose the arbitrary distribution into a sum of Maxwellian components, compute rates at each component's temperature, and add them up. Applied to electrons accelerated in a simulated solar-flare reconnection site, the method finds ionization rates significantly higher and recombination rates lower than a Maxwellian at the same core temperature, which would make Fe emission lines imply an inflated plasma temperature. Applied to a truncated kappa distribution from an exospheric solar-wind model, it finds ionization rates below the standard kappa case, again biasing temperature estimates. The authors also validate their decomposition against existing kappa-distribution tables and against direct cross-section integration.

What carries the argument

The machinery is the Maxwellian decomposition identity: any normalized electron energy distribution f(E) is approximated by f(E) ≈ Σ_i c_i f_M(E, a_i T), where f_M is the Maxwellian distribution at temperature T_i = a_i T and the c_i are weights summing to one. Because ionization and recombination rate coefficients are linear in the electron distribution, the rate for the full distribution is just Σ_i c_i α(T_i), using tabulated Maxwellian rates at the component temperatures. The paper implements two ways to find the coefficients: a classic iterative least-squares solver (LSQR) over a dense temperature grid, and a single-layer linear regression trained like a neural network, which allows constraints such as non-negative coefficients. The decomposition carries the argument because it converts arbitrary non-Maxwellian rate calculations into sums of standard tabulated Maxwellian rates.

What would settle it

Take a standard kappa distribution with κ=2 at a low temperature, e.g., T=$10^{5}$ K, and compute the carbon ionization rate two ways: by directly integrating the ionization cross section against the full kappa distribution, and by the Maxwellian decomposition using extrapolated CHIANTI rates. If the two rates disagree by more than the few percent level claimed, the extrapolation that underpins the method fails. A simpler observational falsifier: in a solar flare region where the electron distribution has been independently measured (e.g., by hard X-ray spectroscopy), compare the Fe ion fractions predicted by this method with those inferred from observed emission lines; systematic disagreement would indicate the rate calculation is missing something.

Watch

Extended reading notes

Core claim

The central discovery is that the Maxwellian decomposition method, already known for kappa distributions, can be turned into a general tool for arbitrary electron energy distributions, and that doing so reveals qualitatively different ionization balance in realistic flare and solar-wind plasmas than Maxwellian or standard-kappa assumptions predict. For the accelerated reconnection electrons, the high-energy tail boosts ionization of Fe ions while depletion of low-energy electrons lowers recombination, so equilibrium ion fractions correspond to apparent temperatures roughly twice the actual core temperature. For the truncated kappa distribution, the missing high-energy electrons suppress ionization relative to the standard kappa case, shifting equilibrium populations toward higher apparent temperatures.

Load-bearing premise

The method requires Maxwellian rate coefficients at component temperatures that can lie far outside the tabulated atomic data, and the paper bridges that gap by extrapolating using the known asymptotic behavior of ionization rates while ignoring density effects; if that extrapolation is inaccurate, the non-Maxwellian rates and the resulting temperature biases will shift.

Editorial extensions

If this is right

  • If the reconnection-site result is correct, Maxwellian-based analyses of flare Fe lines will systematically overestimate the emitting plasma temperature, by roughly a factor of two for the simulated case, so EUV and X-ray diagnostics of flares may need revision.
  • If the truncated-kappa result is correct, assuming a standard kappa distribution in the solar wind overestimates electron temperatures; O VI and Fe charge-state measurements should show the signature of truncation.
  • Because the decomposition is linear in the distribution, the same coefficients can be reused to compute emissivities and line ratios, extending the method beyond ionization balance to full spectral synthesis.
  • The method's speed and generality mean it can be coupled directly to time-dependent non-equilibrium ionization simulations, where the electron distribution evolves on the same timescale as the ionization state.

Reading between the lines

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

  • The apparent temperature bias identified here suggests that some super-hot plasma components reported in flare observations could be artifacts of assuming Maxwellian distributions, effectively transferring 'temperature' into the non-thermal electron population; the paper does not test observed spectra, so this is an extension.
  • A natural testable extension is to apply the decomposition to the actual output of a kinetic particle-in-cell reconnection simulation at many time frames, then forward-model synthetic line spectra and compare with observations from current EUV spectrometers.
  • If the extrapolation dependence is quantified and published, the same framework could be ported to other atomic databases, making non-Maxwellian rates a standard option in plasma models.
  • The finding that standard kappa fits miss up to about 20% in some Fe ion fractions suggests that for high-precision work, fitting a kappa to a simulated or observed spectrum is not equivalent to using the true distribution; direct decomposition should be preferred.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 5 minor

Summary. The paper presents two numerical methods for fitting non-Maxwellian electron distributions as superpositions of Maxwellian components, then computes ionization and recombination rate coefficients by summing Maxwellian rates at the corresponding temperatures. For standard kappa distributions the results are benchmarked against the Hahn & Savin (2015) tables and the KAPPA package for carbon, oxygen, and iron in ionization equilibrium. The methods are then applied to two scenarios: an electron distribution from a combined MHD-particle simulation of magnetic reconnection, and a high-energy truncated kappa distribution from an exospheric solar-wind model. The paper reports that in the flare case ionization rates increase and recombination rates decrease relative to a Maxwellian, shifting Fe ion fractions and potentially biasing temperature diagnostics, while in the solar-wind case truncation lowers ionization rates relative to standard kappa and leads to a similar temperature-overestimation effect.

Significance. If the method is correct, it is a practically useful and flexible tool: it offers Maxwellian decomposition for distributions beyond the standard kappa family, is extensible to updated atomic databases, and is accompanied by public code. The standard-kappa validation at kappa=2 against Hahn & Savin (2015) and the KAPPA package for dominant C, O, and Fe ions is a genuine strength, and the fitting errors are quantified in Figures 1, 2, and 3. The two applications address scientifically important questions about non-Maxwellian effects on Fe-line temperature diagnostics and solar-wind charge states. However, the most striking conclusion—that flare-accelerated electrons reduce recombination and thereby bias inferred temperatures—depends on an assumed low-energy Maxwellian core rather than on the simulation, and the extrapolation of Maxwellian rates outside the tabulated atomic-data range is not validated. These gaps are load-bearing for the abstract's astrophysical claims, so the paper is not yet ready for acceptance.

major comments (2)
  1. [§3.2, Fig. 7] The conclusion that recombination rates decrease for accelerated flare electrons is imposed by construction rather than measured. In Fig. 7(a), the particle-simulation distribution is replaced by a Maxwellian for E ≤ 0.1 keV, with the text stating this is done 'to reduce the impact of numerical noise.' Since radiative recombination for the Fe ions considered is controlled mainly by the low-energy population, the lower recombination rates in Fig. 7(e) and the lower equilibrium Fe ion fractions in Fig. 7(f) follow directly from the assumed Maxwellian core, not from the simulation. The authors acknowledge in §3.2 that 'the detailed particle distribution behaviors in lower energy core are not well addressed yet,' making the abstract's claim of a temperature overestimation conditional on an unconstrained assumption. Please add a sensitivity study varying the low-energy cutoff and core temperature, or compute recombination rates from the full simulated distribution with a validated noise treatment, and qualify the abstract and Section 4 accordingly.
  2. [§3.1, Eq. (6)] The extrapolation of Maxwellian rate coefficients for temperatures outside the tabulated CHIANTI range is not validated. The text states that extrapolation is performed because the asymptotic behavior is 'well known when ignoring the density effects,' but no error estimate is supplied for the extrapolated region. This matters because the decomposition coefficients ai in Fig. 1(c) span several orders of magnitude, forcing Ti far outside the tables for low-temperature applications, and the same issue affects the Fe rates used in the flare application. Unlike §3.3, where direct integration against ionization cross sections via Eq. (8) is used to validate the decomposition (Fig. 9d), no such direct check is provided for the standard kappa cases. Please quantify the extrapolation error, for example by comparing the decomposition-based rates with direct cross-section integration for kappa=2 and for the Fe ions used in the flare case.
minor comments (5)
  1. [Section 4] The Summary states that the largest relative differences in ion populations reach up to ~50%, whereas Figure 6's right panels have a y-axis maximum of 0.35 and the text describing Fig. 6 emphasizes that manifest differences appear only for low populations; these numbers should be harmonized.
  2. [§2.2] The text says the LSQR solver is 'packaged in Scikit-learn.LinearRegression,' but scikit-learn's LinearRegression does not expose an LSQR solver under that name; please clarify the actual solver used, such as scipy.sparse.linalg.lsqr or Ridge(solver='lsqr').
  3. [Fig. 7(a)] The legend and text refer to 'T=2×10^7 k' with a lowercase 'k' in one place; the units should be written consistently as K.
  4. [Eqs. (3)–(4)] The normalization constant Aκ is defined with a Gamma-function expression that is somewhat ambiguous because of the parentheses; please add an explicit statement of how the temperature scale in the kappa distribution is defined, since the decomposition uses Ti = aiT with k_B absorbed into the energy units.
  5. [Title and Abstract] The phrase 'arbitrary non-Maxwellian electron distributions' is broader than the demonstrated cases (standard kappa, one particle-simulation distribution with an imposed Maxwellian core, and truncated kappa); consider tempering the wording or adding a short limitations discussion for distributions with sharp cutoffs or negative decomposition coefficients.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the Maxwellian-decomposition rate calculation is self-contained and externally benchmarked.

full rationale

The paper's central derivation is a numerical method rather than a fitted prediction: an arbitrary electron distribution is approximated as a weighted sum of Maxwellians (Eq. 1), the coefficients are obtained by least-squares fitting (Eq. 5), and the ionization and recombination rates are then formed as the same weighted sum of tabulated Maxwellian rates (Eq. 6). This is an approximation scheme, not a circular claim: the resulting rate coefficients are consequences of the chosen input distribution, not quantities used to define that distribution. The kappa-distribution results are validated against the independently published Hahn & Savin (2015) tables and the KAPPA package, which are external benchmarks. The truncated solar-wind case includes a direct consistency check: Maxwellian-decomposition rates are compared with direct integration of the same CHIANTI ionization cross sections (Eq. 8 and Figure 9d), confirming that the decomposition approximation does not silently inject the rate outcome. The flare application's statement that recombination rates may decrease is explicitly treated by the authors as an unresolved input assumption rather than as an established prediction: they state that 'the detailed particle distribution behaviors in lower energy core are not well addressed yet' and that 'the decrease in recombination rates and the resulting shift of ionization equilibrium states require further exploration in future work.' Replacing the low-energy simulation spectrum with a Maxwellian is an acknowledged modeling assumption; it is an input to the calculation, not a conclusion derived from the output, so it is a limitation or conditional caveat rather than a circular step. The self-citations to previous MHD and particle-transport work (Shen et al. 2023; Li et al. 2018, 2022) supply numerical algorithms and simulation setups, but they are not invoked as uniqueness theorems or used to forbid alternative methods, and the load-bearing reasoning is carried by the equations and external comparisons. Accordingly, the paper does not reduce any claimed prediction to its own inputs by construction.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The core method rests on the completeness and accuracy of a Maxwellian basis and on the atomic data and rate extrapolations. The free parameters are computational thresholds and modeling choices rather than hidden physical constants. No new physical entities are introduced.

free parameters (4)
  • Accuracy threshold for Maxwellian components: Min(c_i) ~ 1e-5 = 1e-5
    Used in Section 3.1 to decide which components are retained for reliable ionization rates, especially for kappa=2 where small components in the high-energy tail affect rates.
  • Low-energy cutoff for particle-simulation distribution = E <= 0.1 keV
    Section 3.2 replaces simulation electrons below 0.1 keV with a Maxwellian approximation to reduce numerical noise; this choice affects recombination rates.
  • Truncation energy for exospheric distribution = E ~ 3 kBT
    Section 3.3 represents the exospheric model's abrupt truncation as a quick decrease near 3 kBT; the exact shape is a modeling choice that affects ionization rates.
  • Best-matched kappa parameters for particle distribution comparison = kappa=8.18, T=4.41e7 K
    Fitted via Eq. (7) to the particle simulation distribution in Section 3.2; used only for the approximation comparison in Figure 8, not for the central conclusion.
assumptions (5)
  • domain assumption The target electron distribution can be represented as a finite sum of Maxwellians with non-negative or small negative coefficients (Eq. 1).
    Established in prior work (Ko et al. 1996; Hahn & Savin 2015). The paper checks fitting error empirically but does not prove convergence for arbitrary distributions.
  • domain assumption CHIANTI atomic data for ionization and recombination cross sections and rates are accurate.
    All rate calculations use CHIANTI data; any systematic atomic-data errors propagate into the computed rates and conclusions.
  • domain assumption Maxwellian rate coefficients can be extrapolated beyond the tabulated CHIANTI temperature range using known asymptotic behavior.
    Section 3.1 invokes this because decomposition temperature nodes can reach very large values; the paper does not quantify the extrapolation error.
  • domain assumption The exospheric truncated kappa distribution from Pierrard et al. (2023) is a valid representation of solar wind electrons at 5 R_sun.
    Taken as input for the solar wind application in Section 3.3 without independent validation in this paper.
  • domain assumption The Parker transport equation particle simulation (Li et al. 2018, 2022) with turbulence parameters (correlation length 1e3 km, normalized amplitude 0.2) produces a realistic accelerated electron distribution.
    The electron distribution in Section 3.2 is taken from this simulation; the turbulence parameters are inputs from the prior model and are not tested here.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Modeling of Ionization and Recombination Processes in Plasma with Arbitrary Non-Maxwellian Electron Distributions." pith.science (2026). https://pith.science/paper/JGR5UKGQ

@misc{pith2026250614668,
  author       = {Pith},
  title        = {Pith review of: Modeling of Ionization and Recombination Processes in Plasma with Arbitrary Non-Maxwellian Electron Distributions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JGR5UKGQ}},
  note         = {Machine review of arXiv:2506.14668}
}
read the original abstract

In astronomical environments, the high-temperature emission of plasma mainly depends on ion charge states, which requires accurate analysis of the ionization and recombination processes. For various phenomena involving energetic particles, the non-Maxwellian distributions of electrons exhibiting high-energy tails can significantly enhance the ionization process. Therefore, accurately computing ionization and recombination rates with non-Maxwellian electron distributions is essential for emission diagnostic analysis. In this work, we report two methods for fitting various non-Maxwellian distributions by using the Maxwellian decomposition strategy. For standard \{kappa} distributions, the calculated ionization and recombination rate coefficients show comparable accuracy to other public packages. We apply the above methods to two specific non-Maxwellian distribution scenarios: (I) accelerated electron distributions due to magnetic reconnection revealed in a combined MHD-particle simulation; (II) the high-energy truncated \{kappa} distribution predicted by the exospheric model of the solar wind. During the electron acceleration process, ionization rates of high-temperature iron ions increase significantly compared to their initial Maxwellian distribution, while the recombination rates may decrease due to the electron distribution changes in low-energy ranges. This can potentially lead to an overestimation of the plasma temperature when analyzing the Fe emission lines under the Maxwellian distribution assumption. For the truncated \{kappa} distribution in the solar wind, the ionization rates are lower than those for the standard \{kappa} distribution, while the recombination rates remain similar. This leads to an overestimation of plasma temperature when assuming a \{kappa} distribution.

Figures

Figures reproduced from arXiv: 2506.14668 by the authors.

Figure 1
Figure 1. The Maxwellian decomposition fitting results for the chosen κ-distribution profiles. In panel (a), the solid lines are kappa distribution f(E) and the dots show the results by summing multiple Maxwellian components. The E is in units of kBT (eV) in this plot; (b) The fitting errors defined by |ff itting −fκ|/fκ for each kappa case; (c) The Maxwellian components coefficients: ci and ai . The summing distribution prof… view at source ↗
Figure 2
Figure 2. The maximum fitting errors covering different electron distribution profiles f(E) versus κ-values. The blue dots are results including f(E) with at least 99.999% number of electrons at the lower energy side, and the orange dots count all (E) points with f(E) > 10−14 [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. The Maxwellian decomposition fitting by employing the neural network model for a typical κ￾distribution. In panel (a), the solid line is κ = 6 distribution f(E) and the dots show the results by summing multiple Maxwellian components from the training results. (b) The relative error between the training results and input defined by |f(E)f itting −f(E)|/f(E); (c) The trainable parameters (or Maxwellian decomposition c… view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: The ionization and recombination rates calculated by using different Maxwellian decomposition coefficients for the κ = 2 case. (a) The chosen Maxwellian decomposition coefficients with different accuracy levels: low to 10−10 , 10−5 , and 10−4 as shown by blue, green, a…
Figure 5
Figure 5. Figure 5: Same as in [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Ion population of carbon, oxygen, and iron in ionization equilibrium states for the κ = 2 case. The left panels show the ion fraction and the right panels are relative differences between our results (solid lines) and others using the table in Hahn & Savin (2015) (dash…
Figure 7
Figure 7. Figure 7: The Maxwellian decomposition fitting results for an electron distribution from particle simula￾tions, and the corresponding ionization and recombination rates. (a) The normalized electron distribution profiles; (b) The relative differences between the fitting results a…
Figure 8
Figure 8. Figure 8: Comparison between particle simulation fitting results and the Kappa-distribution approximation. (a) Electron distribution profiles obtained from particle simulations (orange solid line) and the best matched Kappa distribution with κ = 8.18 and T = 4.41 × 107 K (blue d…
Figure 9
Figure 9. Figure 9: The Maxwellian decomposition fitting results for a truncated Kappa distribution (κ = 6) pre￾sented in the exospheric model of the solar wind. (a) Electron distribution profiles of the Kappa distribution (cyan dotted line) and truncated case (black solid line). The fitt…
Figure 10
Figure 10. Figure 10: Ionization and recombination rates in truncated κ = 6 distribution in the slow solar wind as shown by solid lines. For comparison, the rates for Maxwellian and standard Kappa distribution cases are shown by dashed and dotted lines, respectively. The different colors i…
Figure 11
Figure 11. Figure 11: The time cost for performing linear fitting model using single CPU core of Intel(R) Xeon(R) Silver 4214 running at 2.2GHz. The lower horizontal axis shows the interval of temperature grids in a logarithmic scale, and the upper horizontal axis is the corresponding samp…
Figure 12
Figure 12. Figure 12: The single-layer neural network employed in this work. The input layer is represented by multiple Maxwellian distributions (f(Ti)) corresponding to different temperatures Ti . The output layer shows the sum of these Maxwellian distributions with different weights (ci)…

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 20 canonical work pages

  1. [1]

    1996, ApJ, 457, 939, doi: 10.1086/176787 Astropy Collaboration, Robitaille, T

    Ballegooijen, A. 1996, ApJ, 457, 939, doi: 10.1086/176787 Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., et al. 2013, A&A, 558, A33, doi: 10.1051/0004-6361/201322068 Astropy Collaboration, Price-Whelan, A. M., Sip˝ ocz, B. M., et al. 2018, AJ, 156, 123, doi: 10.3847/1538-3881/aabc4f Del Zanna, G., Dere, K. P., Young, P. R., &

    work page doi:10.1086/176787 1996

  2. [2]

    2021, ApJ, 909, 38, doi: 10.3847/1538-4357/abd8ce

    Landi, E. 2021, ApJ, 909, 38, doi: 10.3847/1538-4357/abd8ce

    doi:10.3847/1538-4357/abd8ce 2021

  3. [3]

    P., Del Zanna, G., Young, P

    Dere, K. P., Del Zanna, G., Young, P. R., Landi, E., & Sutherland, R. S. 2019, ApJS, 241, 22, doi: 10.3847/1538-4365/ab05cf

    doi:10.3847/1538-4365/ab05cf 2019

  4. [4]

    C., & Young, P

    Fossi, B. C., & Young, P. R. 1997, A&AS, 125, 149, doi: 10.1051/aas:1997368 Dzifˇ c´ akov´ a, E., Dud´ ık, J., Kotrˇ c, P., F´ arn´ ık, F., & Zemanov´ a, A. 2015, ApJS, 217, 14, doi: 10.1088/0067-0049/217/1/14 Dzifˇ c´ akov´ a, E., Dud´ ık, J., Zemanov´ a, A., L¨ orinˇ c´ ık, J., & Karlick´ y, M. 2021, ApJS, 257, 62, doi: 10.3847/1538-4365/ac2aa7 Dzifˇ c´...

    doi:10.1051/aas:1997368 1997

  5. [5]

    454, Hinode-3: The 3rd Hinode Science Meeting, ed

    Series, Vol. 454, Hinode-3: The 3rd Hinode Science Meeting, ed. T. Sekii, T. Watanabe, & T. Sakurai, 167 Dzifˇ c´ akov´ a, E., Zemanov´ a, A., Dud´ ık, J., & Mackovjak, ˇS. 2018, ApJ, 853, 158, doi: 10.3847/1538-4357/aaa426

    work page doi:10.3847/1538-4357/aaa426 2018

  6. [6]

    2017, ApJ, 835, 124, doi: 10.3847/1538-4357/835/2/124

    Effenberger, F., Rubio da Costa, F., Oka, M., et al. 2017, ApJ, 835, 124, doi: 10.3847/1538-4357/835/2/124

    work page doi:10.3847/1538-4357/835/2/124 2017

  7. [7]

    2021, Introduction and Motivation, ed

    Fichtner, H., & Lazar, M. 2021, Introduction and Motivation, ed. M. Lazar & H. Fichtner (Cham: Springer International Publishing), 3–12, doi: 10.1007/978-3-030-82623-9 1

    doi:10.1007/978-3-030-82623-9 2021

  8. [8]

    S., et al

    Foster, A., Smith, R., Brickhouse, N. S., et al. 2018, American Astronomical Society Meeting Abstracts #231, 231, 253.03. https://ui.adsabs.harvard.edu/abs/2018AAS... 23125303F/abstract

    work page 2018

Show all 37 references
  1. [9]

    2024, SSRv, 220, 43, doi: 10.1007/s11214-024-01073-2

    Guo, F., Liu, Y.-H., Zenitani, S., & Hoshino, M. 2024, SSRv, 220, 43, doi: 10.1007/s11214-024-01073-2

    2024 doi

  2. [10]

    Hahn, M., & Savin, D. W. 2015, ApJ, 809, 178, doi: 10.1088/0004-637X/809/2/178

    2015 doi

  3. [11]

    Jeffrey, N. L. S., Fletcher, L., & Labrosse, N. 2016, A&A, 590, A99, doi: 10.1051/0004-6361/201527986 26 —. 2017, ApJ, 836, 35, doi: 10.3847/1538-4357/836/1/35

    2016 doi

  4. [12]

    S., Bykov, A

    Kaastra, J. S., Bykov, A. M., & Werner, N. 2009, A&A, 503, 373, doi: 10.1051/0004-6361/200912492

    2009 doi

  5. [13]

    K., Fisk, L

    Ko, Y. K., Fisk, L. A., Gloeckler, G., & Geiss, J. 1996, Geophys. Res. Lett., 23, 2785, doi: 10.1029/96GL02449

    1996 doi

  6. [14]

    2023, in AGU Fall Meeting Abstracts, Vol

    Ko, Y.-K., Pierrard, V., & Shen, C. 2023, in AGU Fall Meeting Abstracts, Vol. 2023, SH51E–2653

    2023

  7. [15]

    J., et al

    Krucker, S., Battaglia, M., Cargill, P. J., et al. 2008, A&A Rv, 16, 155, doi: 10.1007/s00159-008-0014-9

    2008 doi

  8. [16]

    2022, ApJ, 932, 92, doi: 10.3847/1538-4357/ac6efe

    Li, X., Guo, F., Chen, B., Shen, C., & Glesener, L. 2022, ApJ, 932, 92, doi: 10.3847/1538-4357/ac6efe

    2022 doi

  9. [17]

    2018, ApJ, 866, 4, doi: 10.3847/1538-4357/aae07b

    Li, X., Guo, F., Li, H., & Li, S. 2018, ApJ, 866, 4, doi: 10.3847/1538-4357/aae07b

    2018 doi

  10. [18]

    A., et al

    Lionello, R., Downs, C., Linker, J. A., et al. 2019, SoPh, 294, 13, doi: 10.1007/s11207-019-1401-2

    2019 doi

  11. [19]

    Livadiotis, G., & McComas, D. J. 2013, SSRv, 175, 183, doi: 10.1007/s11214-013-9982-9

    2013 doi

  12. [20]

    P., Lavarra, M

    Lomazzi, P., Rouillard, A. P., Lavarra, M. A., et al. 2025, arXiv e-prints, arXiv:2504.05177, doi: 10.48550/arXiv.2504.05177

    2025 doi arXiv

  13. [21]

    Y., et al

    Maksimovic, M., Zouganelis, I., Chaufray, J. Y., et al. 2005, Journal of Geophysical Research (Space Physics), 110, A09104, doi: 10.1029/2005JA011119

    2005 doi

  14. [22]

    2017, Physics of Plasmas, 24, 062906, doi: 10.1063/1.4985302

    Montag, P., Egedal, J., Lichko, E., & Wetherton, B. 2017, Physics of Plasmas, 24, 062906, doi: 10.1063/1.4985302

    2017 doi

  15. [23]

    Oka, M., Ishikawa, S., Saint-Hilaire, P., Krucker, S., & Lin, R. P. 2013, ApJ, 764, 6, doi: 10.1088/0004-637X/764/1/6

    2013 doi

  16. [24]

    1968, in Astrophysics and Space Science

    Olbert, S. 1968, in Astrophysics and Space Science

    1968

  17. [25]

    10, Physics of the Magnetosphere, ed

    Library, Vol. 10, Physics of the Magnetosphere, ed. R. D. L. Carovillano & J. F. McClay, 641, doi: 10.1007/978-94-010-3467-8 23

    doi

  18. [26]

    P., & Scudder, J

    Owocki, S. P., & Scudder, J. D. 1983, ApJ, 270, 758, doi: 10.1086/161167

    1983 doi

  19. [27]

    C., & Saunders, M

    Paige, C. C., & Saunders, M. A. 1982, ACM Trans. Math. Software, 8, 43

    1982

  20. [28]

    2011, Journal of Machine Learning Research, 12, 2825

    Pedregosa, F., Varoquaux, G., Gramfort, A., et al. 2011, Journal of Machine Learning Research, 12, 2825

    2011

  21. [29]

    Pierrard, V., & Lemaire, J. 1996, J. Geophys. Res., 101, 7923, doi: 10.1029/95JA03802

    1996 doi

  22. [30]

    2023, Plasma, 6, 518, doi: 10.3390/plasma6030036

    Pierrard, V., P´ eters de Bonhome, M., Halekas, J., et al. 2023, Plasma, 6, 518, doi: 10.3390/plasma6030036

    2023 doi

  23. [31]

    2018, ApJ, 864, 63, doi: 10.3847/1538-4357/aad62d

    Polito, V., Dud´ ık, J., Kaˇ sparov´ a, J., et al. 2018, ApJ, 864, 63, doi: 10.3847/1538-4357/aad62d

    2018 doi

  24. [32]

    J., Lepri, S

    Rivera, Y. J., Lepri, S. T., Raymond, J. C., et al. 2021, ApJ, 921, 93, doi: 10.3847/1538-4357/ac1676

    2021 doi

  25. [33]

    C., Miki´ c, Z., et al

    Shen, C., Raymond, J. C., Miki´ c, Z., et al. 2017, ApJ, 850, 26, doi: 10.3847/1538-4357/aa93f3

    2017 doi

  26. [34]

    C., & Murphy, N

    Shen, C., Raymond, J. C., & Murphy, N. A. 2023, ApJ, 943, 111, doi: 10.3847/1538-4357/aca6e7

    2023 doi

  27. [35]

    K., Raymond, J

    Shen, C., Reeves, K. K., Raymond, J. C., et al. 2013, ApJ, 773, 110, doi: 10.1088/0004-637X/773/2/110

    2013 doi

  28. [36]

    2022, ApJ, 926, 35, doi: 10.3847/1538-4357/ac391810

    Szente, J., Landi, E., & van der Holst, B. 2022, ApJ, 926, 35, doi: 10.3847/1538-4357/ac391810. 1002/essoar.10508664.1

    2022 doi

  29. [37]

    P., Hunana, P., Mostafavi, P., et al

    Zank, G. P., Hunana, P., Mostafavi, P., et al. 2015, ApJ, 814, 137, doi: 10.1088/0004-637X/814/2/137

    2015 doi

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.