Pith. sign in

REVIEW 3 major objections 4 minor 42 references

Magnetic model reveals why a 500 Mm solar filament erupted sideways

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

A neural-network magnetic field reconstruction shows that a 500 Mm filament's pre-eruption flux rope had an extended eastern footprint connected to the flare ribbon and coronal dimming, explaining the asymmetric eruption.

T0 review reviewed 2026-08-05 challenge →

load-bearing objection A careful single-event NLFFF study with a genuinely robust ~500 Mm flux-rope reconstruction; the dimming-expansion interpretation leans on the least stable connectivity in the ensemble, so it is plausible but not confirmed. the 3 major comments →

arxiv 2508.18121 v1 pith:S7IJI4QL submitted 2025-08-25 astro-ph.SR

Magnetic structure and asymmetric eruption of a 500 Mm filament rooted in weak-field regions

Stefan Purkhart , Astrid M. Veronig , Robert Jarolim , Karin Dissauer , Julia K. Thalmann This is my paper
classification astro-ph.SR
keywords solar filament eruptionmagnetic flux ropenonlinear force-free field extrapolationphysics-informed neural networkscoronal dimmingflare ribbonsweak-field regionsAR 13229
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

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 establish that a single pre-eruptive magnetic structure—a flux rope roughly 500 megameters long that threads the active region and reaches into quiet-Sun plasma—can account for the strongly asymmetric eruption of the February 24, 2023 filament. The paper argues that the rope's eastern leg fans out into a broad weak-field footprint, so when it lifts off it carves out the large coronal dimming and draws the inverse J-shaped flare ribbon, while the western leg stays multiply anchored under strong overlying field and therefore confines that side of the eruption. If the reconstruction is right, dimming location and ribbon shape were fixed before the flare began, by the rope's footpoint geometry. It also makes the methodological point that physics-informed neural network extrapolations can recover meaningful connectivity even in weak-field regions where classical force-free extrapolations typically fail.

Core claim

Using a PINN-based nonlinear force-free extrapolation of the pre-eruption photospheric field, the paper finds a channel of high current density whose length and shape match the observed 500 Mm filament. Field lines from the eastern portion of this channel form a magnetic flux rope with an extended, fanned-out footprint in a weak-field negative-polarity region; that footprint coincides with the area enclosed by the inverse J-shaped flare ribbon and with the initial coronal dimming. The same model shows overlying strapping field lines anchored in the region into which the dimming later expands, and a compact, multiply anchored western leg beneath strong sunspot-connected overlying fields. The

What carries the argument

The central object is a 500 Mm magnetic flux rope delivered by the NF2 physics-informed neural network nonlinear force-free extrapolation, in its vector-potential form. The neural network maps coordinates to a vector potential A, the field is B = curl A so divergence-free by construction, and training minimises a weighted sum of force-free, boundary, and potential-boundary losses; the lower boundary is the HMI vector magnetogram and the side and top boundaries are potential field. This mesh-free representation lets the method fill a roughly 730 x 550 x 300 Mm volume while allowing local departures from strict force-freeness in weak-field regions. The interpretation hinges on tracing field li

Load-bearing premise

The reconstruction stands or falls on the corona being close enough to force-free in the 5-G region that field lines traced from weak-field boundary pixels reflect real magnetic connections rather than numerical artifacts.

What would settle it

Take the same HMI boundary and seed regions and compute the field with a different NLFFF code or with chromospheric boundary data: the interpretation fails if the southern dimming region connects to the unconstrained twisted flux bundle in the preferred solution, or if the eastern MFR footprint does not fall inside the inverse J-shaped ribbon.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • The pre-eruptive flux rope's eastern footprint, not the flare reconnection alone, sets where the core dimming appears and how large it can grow.
  • Dimming growth into the strapping-field region is the observable signature of strapping-strapping reconnection, linking ribbon expansion to flux addition to the erupting rope.
  • The western leg's multiple anchor points and overlying sunspot-connected field explain the suppressed dimming and partial confinement, so asymmetric eruptions can be diagnosed from pre-eruptive footpoint structure.
  • PINN-based NLFFF extrapolation can be used on filaments that extend well beyond active-region cores, opening very large or quiet-Sun-rooted structures to quantitative coronal field modeling.
  • The reconstructed free energy of about 3 x 10^32 erg and negative helicity are consistent with the inverse-S filament and its northern-hemisphere sign preference.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If footpoint geometry controls dimming, then for other whipping-like eruptions the pre-eruptive dimming region should be predictable from the fan of MFR field lines; a survey of similar events could test this without waiting for new instrumentation.
  • The southern part of the dimming region is the fragile piece: in some sensitivity runs it connects to an unconstrained twisted flux bundle. A natural next test is to rerun the same event with chromospheric magnetogram constraints and see whether that connection disappears as the boundary information improves.
  • This result suggests the NLFFF limit for filaments may lie beyond active-region cores; applying the same method to a fully quiescent filament, where one leg has no strong-field anchor at all, would test how far the force-free assumption can be pushed.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper analyzes the pre-eruptive magnetic configuration and asymmetric eruption of a roughly 500 Mm inverse S-shaped filament on 2023 February 24, partially rooted in AR 13229 and extending into weak-field regions. Using AIA, HMI, KSO, GONG, and STIX observations together with a physics-informed neural network (PINN) NLFFF extrapolation (NF2), the authors identify a large-scale magnetic flux rope whose high-current-density channel matches the observed filament. They report an extended eastern MFR footprint associated with the inverse J-shaped flare ribbon and the initial coronal dimming, while overlying strapping field lines connect to the region into which the dimming later expands. They interpret the dimming as stationary flux-rope and strapping-flux dimming, with later expansion driven by strapping-strapping and rope-strapping reconnection. The western leg shows multiple anchor points and stronger overlying fields, explaining the partial confinement and lack of dimming there. A 16-run sensitivity study over the force-free weighting factor lambda_ff is included in Appendix A.

Significance. If the interpretation holds, the paper provides a valuable demonstration that PINN-based NLFFF extrapolation can model large-scale filaments extending into weak-field regions, and it offers a physically coherent scenario connecting pre-eruptive MFR geometry to flare-ribbon and dimming asymmetries. The observational analysis is well documented, and the ensemble sensitivity study is a clear strength: the main MFR channel is robust for lambda_ff > 0.2, and the quantitative force-free metrics are reported for the full volume and for strong- and weak-field subregions. The central risk is that the load-bearing connectivity underlying the dimming-expansion scenario is the least robust part of the ensemble, and the model parameter lambda_ff was selected in part by matching the same observations used to validate the scenario. This does not undermine the existence of the MFR, but it weakens the specific claim about the strapping-field connectivity through the dimming region.

major comments (3)
  1. [Appendix A; Fig. A.2; Section 5] The interpretation in Section 5 that 'the area into which the dimming expanded is connected to strapping field lines that overlay the MFR's extended eastern leg' rests on exactly the part of the NLFFF ensemble that Appendix A shows to be unstable. The text states that for certain lambda_ff values the southernmost part of the dimming region is connected to an unconstrained twisted flux bundle, with no correlation to lambda_ff, and that lambda_ff = 0.60 gives a steep connectivity gradient that does not match observations. Since lambda_ff = 0.40 was selected partly because it produces the observed dimming boundary, the agreement of the selected model with the observed dimming is not an independent confirmation. I request that the authors either (i) reformulate the dimming-connectivity claim as one of several possible topologies with explicit uncertainty, or (ii) provide an independent test,
  2. [Sections 2.4, 4; Table A.1] The weak-field region that carries the dimming interpretation is also where the NLFFF assumptions are least secure. The mean unsigned flux in the final dimming mask is only ~5 G (Section 3.2), close to HMI noise levels, and Section 4 acknowledges that the real corona may deviate from a force-free state in such regions. Table A.1 shows that in the weak-field subregion E_div/E is about 4.9e-2 for the adopted lambda_ff = 0.40, i.e., near the 5% threshold used for reliable helicity computations, and the current-weighted angle theta_j does not improve systematically with lambda_ff there. The paper should quantify how the inferred connectivity changes under plausible perturbations of the weak-field boundary data (e.g., masking to the noise level) and should temper statements that the model 'effectively captures the essential large-scale connectivity' in this specific region.
  3. [Section 3.2; Fig. 6; Fig. A.2] The dimming mask is defined by a fixed LBR threshold of -0.5 (Eq. 1), and the field-line connectivity shown in Fig. 6 (right column) and Fig. A.2 (bottom row) uses that mask as seed regions. No sensitivity analysis is provided for this threshold. Because the connectivity conclusions concern a weak-field region where the dimensions and location of the mask are likely threshold-dependent, a modest change in the threshold could alter the seed region and hence the inferred strapping-field connectivity. The authors should show that their conclusions are robust to a reasonable range of LBR thresholds, or explicitly justify the chosen threshold physically.
minor comments (4)
  1. [Abstract and throughout] There are several typographical artifacts, e.g., 'e ffective' in the Abstract and 'whith' in Appendix A. These should be corrected in the final version.
  2. [Section 2.4, Eq. (7)] The total loss in Eq. (7) includes lambda_B0, which is decayed from 1000 to 1, but the text does not explain why this particular schedule is chosen or how sensitive the final solution is to the decay endpoint. A brief justification would improve reproducibility.
  3. [Section 2.3] The description of the CEA submap and its non-alignment with Carrington longitude/latitude is clear but would benefit from a small schematic or a more explicit statement of the projection-induced distortion in the top-right corner, since the authors note the 50-degree limitation there.
  4. [Section 4] The paper relies on the dimming classification of Veronig et al. (2025). It would be helpful to state explicitly which observational signatures, independent of the NLFFF model, support the stationary flux-rope versus moving flux-rope classification, so that the reader can separate the model-dependent and observation-based parts of the argument.

Circularity Check

1 steps flagged

Mild model-selection circularity for the southern dimming boundary, but the central MFR result is robust and independently corroborated.

specific steps
  1. fitted input called prediction [Appendix A (sensitivity analysis); selection used in Sect. 2.4; connectivity used in Sect. 3.3/Fig. 6 and Sect. 4]
    "The value λff = 0.4 we selected for the results presented in this paper represents a balance between satisfying the quality metrics and aligning well with the observations."

    The force-free weighting λff is a free parameter. The paper selects λff = 0.4 partly because it produces a connectivity boundary that aligns with the observed southern dimming boundary, and rejects λff = 0.60 because it 'does not match the observed evolution.' The same selected model is then used to infer the dimming-region connectivity (strapping fields) and to conclude that the area into which the dimming expanded is connected to strapping field lines. For this specific subregion, the model's agreement with the observation is therefore partly a selection criterion rather than an independent confirmation. However, the core MFR and the eastern footprint are robust across the ensemble, so the circularity is localized and not fatal to the main conclusion.

full rationale

The central claim—that a ~500 Mm MFR with an extended eastern footprint existed before the eruption—is an inference from a boundary-value NLFFF calculation and is robust across the 16-member λff ensemble. It is also corroborated by independent Hα/EUV observations of the filament shape and by the spatial correspondence of the MFR footprint with the inverse J-shaped ribbon. The self-citations to Jarolim et al. (2023) for the PINN method and to Veronig et al. (2025) for dimming categories are not load-bearing in a circular sense: the method is externally validated and applied to a new event, and the classification is a descriptive framework rather than a proof mechanism. The main circularity concern is the selection of λff = 0.4, which is chosen in part by visual agreement with the observed dimming boundary. Appendix A transparently shows that the southernmost part of the dimming region is connected to an unconstrained twisted flux bundle for some λff values, with no correlation to λff, indicating genuine difficulty finding a local force-free solution there. Because the paper uses the selected model to interpret exactly that dimming boundary, there is a mild validation circularity. But the broader MFR and strapping-flux interpretation remain supported by ensemble-robust features and by the qualitative agreement with the filament and flare-ribbon observations. Thus the paper is largely self-contained and not significantly circular, aside from this localized model-selection issue.

Axiom & Free-Parameter Ledger

2 free parameters · 5 axioms · 0 invented entities

No new physical entities (particles, forces, dimensions) are introduced. The magnetic flux rope, strapping fields, and flux-rope dimming categories are pre-existing concepts (e.g., Veronig et al. 2025) applied to this event. The main load-bearing inputs are the NLFFF approximations, the boundary data choice, and one model weighting parameter.

free parameters (2)
  • lambda_ff (force-free weighting factor) = 0.4
    The neural network loss combines force-free residual with boundary matching; lambda_ff = 0.4 was selected as the best compromise between force-free quality metrics and visual agreement with observations (Section 2.4, Appendix A). The MFR is present for all lambda_ff > 0.2, so the main conclusion does not hinge on this value, but the final field-line connectivity shown in the paper is tied to it.
  • Dimming threshold (LBR < -0.5) = -0.5 log10 relative counts
    A pixel is classified as dimming if LBR < -0.5; this threshold was intentionally set low to isolate the strong eastern dimming from false detections caused by moving filament plasma (Section 2.2). The derived dimming area (9.25e9 km2) and mean unsigned flux (5.2 G) depend on this cut.
axioms (5)
  • domain assumption The coronal magnetic field is force-free (J x B = 0) over most of the volume, including weak-field regions.
    Core assumption of NLFFF modeling; justified in Section 1 via the Gary (2001) beta model for heights 2-40 Mm. The authors explicitly note deviations may occur where plasma-beta is not small, particularly in weak-field regions (Section 4).
  • domain assumption The side and top boundaries of the computational volume can be approximated by a potential field solution.
    Section 2.4 states the side and top boundaries are approximated by a potential field; this is standard for NLFFF but can bias connectivity near volume edges.
  • domain assumption The pre-eruption magnetic configuration at 19:34 UT is representative of the structure that erupted at about 20:03 UT.
    The extrapolation uses an HMI magnetogram from 19:34 UT, about 30 min before flare onset. The authors support this by noting similar filament structure between KSO H-alpha at 14:59 UT and AIA 304 at 19:34 UT (Section 3.1), and the eruption starts around 20:00 UT.
  • domain assumption HMI photospheric vector magnetogram, after error-masked clipping (Bdiff,clipped), is a reliable lower boundary in weak-field regions.
    Weak-field transverse components have low signal-to-noise; the method removes pixels below the error threshold (Section 2.4). The paper acknowledges the challenge for NLFFF modeling of weak-field regions in Section 1.
  • domain assumption Cartesian geometry is adequate for this field of view because the filament lies within 50 degrees of disk center.
    Section 2.3 states that projection effects and spherical geometry become important beyond about 50 degrees from disk center, and that the Cartesian approach introduces inaccuracies in peripheral areas but is suitable for the filament under study.

reviewed 2026-08-05 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Magnetic structure and asymmetric eruption of a 500 Mm filament rooted in weak-field regions." pith.science (2026). https://pith.science/paper/S7IJI4QL

@misc{pith2026250818121,
  author       = {Pith},
  title        = {Pith review of: Magnetic structure and asymmetric eruption of a 500 Mm filament rooted in weak-field regions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S7IJI4QL}},
  note         = {Machine review of arXiv:2508.18121}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

We performed a detailed analysis of the magnetic structure and asymmetric eruption of a large (about 500 Mm) inverse S-shaped filament partially located in AR 13229 on February 24, 2023. We linked the filament's pre-eruptive magnetic configuration to its highly asymmetric eruption dynamics and the formation of a large-scale coronal dimming in a weak-field region (mean unsigned flux of about 5 G). To reconstruct the coronal magnetic field, we applied a physics-informed neural network (PINN)-based nonlinear force-free field (NLFFF) extrapolation method to a pre-eruption HMI vector magnetogram. The NLFFF extrapolation reveals a large-scale magnetic flux rope (MFR) of about 500 Mm in length, consistent with the filament. We identified an extended MFR footprint to the east that connects to the J-shaped flare ribbon, outlining where the coronal dimming began. Overlying strapping fields connect to the area into which the dimming and flare ribbon later expand. This configuration explains the formation of the dimming as a stationary flux rope and strapping flux dimming, with subsequent expansion driven by the growth of the MFR footprint through strapping-strapping reconnection. Conversely, the western filament leg shows multiple anchor points and strong overlying magnetic fields, which suppressed the dimming and partially confined the eruption on that side. The reconstructed pre-eruptive NLFFF configuration offers a clear physical explanation for the asymmetries seen in the eruption, flare geometry, and coronal dimming. This demonstrates that PINN-based NLFFF extrapolation can effectively model large-scale filaments extending into weak-field regions, enhancing our understanding of complex solar eruptions.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

42 extracted references · 29 canonical work pages

  1. [1]

    N., Fan, Y ., Kazachenko, M

    Afanasyev, A. N., Fan, Y ., Kazachenko, M. D., & Cheung, M. C. M. 2023, ApJ, 952, 136

    work page 2023

  2. [2]

    A., Subramanian, S

    Ahamed, A. A., Subramanian, S. P., Rahman, A. M., et al. 2025, New A, 114, 102312

    work page 2025

  3. [3]

    G., Sun, X., Hoeksema, J

    Bobra, M. G., Sun, X., Hoeksema, J. T., et al. 2014, Sol. Phys., 289, 3549 Démoulin, P., Priest, E. R., & Lonie, D. P. 1996, J. Geophys. Res., 101, 7631

    work page 2014

  4. [4]

    L., Wheatland, M

    DeRosa, M. L., Wheatland, M. S., Leka, K. D., et al. 2015, ApJ, 811, 107

    2015

  5. [5]

    Gary, G. A. 2001, Sol. Phys., 203, 71

    work page 2001

  6. [6]

    Gary, G. A. & Hagyard, M. J. 1990, Sol. Phys., 126, 21

    1990

  7. [7]

    Gibson, S. E. 2018, Living Reviews in Solar Physics, 15, 7

    work page 2018

  8. [8]

    M., Török, T., Vršnak, B., Manchester, W., & Veronig, A

    Green, L. M., Török, T., Vršnak, B., Manchester, W., & Veronig, A. 2018, Space Sci. Rev., 214, 46

    work page 2018

  9. [9]

    W., Hill, F., Hubbard, R

    Harvey, J. W., Hill, F., Hubbard, R. P., et al. 1996, Science, 272, 1284

    1996

  10. [10]

    & van Ballegooijen, A

    Hillier, A. & van Ballegooijen, A. 2013, ApJ, 766, 126

    work page 2013

  11. [11]

    T., Liu, Y ., Hayashi, K., et al

    Hoeksema, J. T., Liu, Y ., Hayashi, K., et al. 2014, Sol. Phys., 289, 3483

    2014

  12. [12]

    S., Acton, L

    Hudson, H. S., Acton, L. W., & Freeland, S. L. 1996, ApJ, 470, 629

    1996

  13. [13]

    2024, Sol

    Idrees, S., Su, J., Chen, J., & Deng, Y . 2024, Sol. Phys., 299, 91

    work page 2024

  14. [14]

    2014, ApJ, 788, 60

    Janvier, M., Aulanier, G., Bommier, V ., et al. 2014, ApJ, 788, 60

    work page 2014

  15. [15]

    K., Veronig, A

    Jarolim, R., Thalmann, J. K., Veronig, A. M., & Podladchikova, T. 2023, Nature Astronomy, 7, 1171

    work page 2023

  16. [16]

    2022, The Innovation, 3, 100236

    Jiang, C., Feng, X., Guo, Y ., & Hu, Q. 2022, The Innovation, 3, 100236

    work page 2022

  17. [17]

    T., Feng, X., & Hu, Q

    Jiang, C., Wu, S. T., Feng, X., & Hu, Q. 2014, ApJ, 786, L16

    work page 2014

  18. [18]

    J., Grimm, O., et al

    Krucker, S., Hurford, G. J., Grimm, O., et al. 2020, A&A, 642, A15

    2020

  19. [19]

    R., Title, A

    Lemen, J. R., Title, A. M., Akin, D. J., et al. 2012, Sol. Phys., 275, 17

    2012

  20. [20]

    Liu, R., Alexander, D., & Gilbert, H. R. 2009, ApJ, 691, 1079

    work page 2009

  21. [21]

    Martens, P. C. & Zwaan, C. 2001, ApJ, 558, 872

    work page 2001

  22. [22]

    E., & Wheatland, M

    Mastrano, A., Yang, K. E., & Wheatland, M. S. 2020, Sol. Phys., 295, 97 Müller, D., St. Cyr, O. C., Zouganelis, I., et al. 2020, A&A, 642, A1

    work page 2020

  23. [23]

    J., Freij, N., Stansby, D., et al

    Mumford, S. J., Freij, N., Stansby, D., et al. 2023, SunPy

    work page 2023

  24. [24]

    2014, Living Reviews in Solar Physics, 11, 1

    Parenti, S. 2014, Living Reviews in Solar Physics, 11, 1

    2014

  25. [25]

    D., Thompson, B

    Pesnell, W. D., Thompson, B. J., & Chamberlin, P. C. 2012, Sol. Phys., 275, 3

    2012

  26. [26]

    P., & Cameron, R

    Peter, H., Warnecke, J., Chitta, L. P., & Cameron, R. H. 2015, A&A, 584, A68 Pötzi, W., Veronig, A., Jarolim, R., et al. 2021, Sol. Phys., 296, 164

    work page 2015

  27. [27]

    A., et al

    Purkhart, S., Collier, H., Hayes, L. A., et al. 2025, A&A, 698, A318

    2025

  28. [28]

    M., Dickson, E

    Purkhart, S., Veronig, A. M., Dickson, E. C. M., et al. 2023, A&A, 679, A99

    work page 2023

  29. [29]

    M., Kliem, B., et al

    Purkhart, S., Veronig, A. M., Kliem, B., et al. 2024, A&A, 689, A259

    work page 2024

  30. [30]

    Raissi, M., Perdikaris, P., & Karniadakis, G. E. 2019, Journal of Computational Physics, 378, 686 Rodríguez-Gómez, J. M., Kuckein, C., González Manrique, S. J., et al. 2024, ApJ, 964, 27

    work page 2019

  31. [31]

    H., Bush, R

    Schou, J., Scherrer, P. H., Bush, R. I., et al. 2012, Sol. Phys., 275, 229

    2012

  32. [32]

    J., De Rosa, M

    Schrijver, C. J., De Rosa, M. L., Metcalf, T. R., et al. 2006, Sol. Phys., 235, 161

    work page 2006

  33. [33]

    2024, Sol

    Song, Y ., Su, J., Zhang, Q., et al. 2024, Sol. Phys., 299, 85

    work page 2024

  34. [34]

    Sterling, A. C. & Hudson, H. S. 1997, ApJ, 491, L55

    1997

  35. [35]

    K., Dumbovi´c, M., Dissauer, K., et al

    Thalmann, J. K., Dumbovi´c, M., Dissauer, K., et al. 2023, A&A, 669, A72

    work page 2023

  36. [36]

    K., Linan, L., Pariat, E., & Valori, G

    Thalmann, J. K., Linan, L., Pariat, E., & Valori, G. 2019, ApJ, 880, L6 The SunPy Community, Barnes, W. T., Bobra, M. G., et al. 2020, The Astrophys- ical Journal, 890, 68 Trujillo Bueno, J. & del Pino Alemán, T. 2022, ARA&A, 60, 415

    work page 2019

  37. [37]

    2013, A&A, 553, A38

    Valori, G., Démoulin, P., Pariat, E., & Masson, S. 2013, A&A, 553, A38

    work page 2013

  38. [38]

    M., Dissauer, K., Kliem, B., et al

    Veronig, A. M., Dissauer, K., Kliem, B., et al. 2025, Living Reviews in Solar Physics, 22, 2

    work page 2025

  39. [39]

    Wheatland, M. S. & Régnier, S. 2009, ApJ, 700, L88

    work page 2009

  40. [40]

    2006, Sol

    Wiegelmann, T., Inhester, B., & Sakurai, T. 2006, Sol. Phys., 233, 215

    work page 2006

  41. [41]

    Wiegelmann, T., Petrie, G. J. D., & Riley, P. 2017, Space Sci. Rev., 210, 249

    2017

  42. [42]

    K., Inhester, B., et al

    Wiegelmann, T., Thalmann, J. K., Inhester, B., et al. 2012, Sol. Phys., 281, 37 Article number, page 13 of 15 A&A proofs: manuscript no. aa55860-25corr Appendix A: Sensitivity analysis for PINN-based NLFFF modeling To assess the robustness of the NLFFF extrapolation results shown in Sect. 3.3, especially their sensitivity to the force-free weighting facto...

    work page 2012

This paper was first reviewed by deepseek-v4-flash on August 5, 2026.