Pith. sign in

REVIEW 4 major objections 6 minor 42 references

Waiting Time Distribution of Supra-Arcade Downflows

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper finds that waiting times between consecutive supra-arcade downflows are heavy-tailed power laws in most of seven solar flares, ruling out linear random or quasi-periodic generation processes.

desk verdict A useful first WTD study of SADs with a plausible heavy-tail result, but the abstract's ruling-out language outruns the evidence because detection incompleteness is unquantified and the F/G comparison shows it can flip the winning model. read the letter →

arxiv 2507.16118 v1 pith:WTU6XXV2 submitted 2025-07-22 astro-ph.SR

Zichao Yang , Rui Liu This is my paper · ORCID
classification astro-ph.SR
keywords supra-arcadedownflowswaitingtimedistributionpower-lawsolarflaresmagneticreconnectionself-organizedcriticalityMHDturbulenceSDO/AIA
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

Supra-arcade downflows are dark, sunward-moving voids seen above flare loop arcades, thought to trace outflows from magnetic reconnection. The paper asks whether these downflows appear on a quasi-periodic clock, as a steady random process, or through some intermittent nonlinear mechanism. Using SDO/AIA 131 Å observations of seven limb flares, with SADs tracked by automated, manual, and the authors' own sliding-window method, it finds that most waiting time distributions are best fitted by a power law with slope between 1.7 and 2.4, and often almost as well by a log-normal. If correct, the timing of SADs carries a fingerprint of scale-invariant, bursty energy release rather than a simple random or periodic trigger.

What carries the argument

The central objects are the SAD waiting times themselves: the intervals between the start (or end) times of consecutive SADs identified in AIA 131 Å running-difference images. The analysis fits three candidate probability distributions—power-law $p(x) \propto x^{-\alpha}$, log-normal, and exponential—by maximum likelihood, then discriminates among them with pairwise likelihood ratio tests and the AIC, AICc, and BIC information criteria. A survival-distribution-function fit, built by reordering the raw waiting times without binning, supplies an independent power-law slope estimate. The theoretical machinery used for interpretation is the fractal-diffusive SOC prediction $\alpha_{\Delta t}=1+(d-1)\beta/2=2$ for $d=3$, its pile-up modification for overlapping events, and the non-stationary Poisson formula $P(\Delta t) = \lambda_0^{-1}\int_0^\infty f(\lambda)\lambda^2 e^{-\lambda\Delta t}d\lambda$, which yields power-law tails with slopes 2–3.

What would settle it

Higher-cadence EUV observations (for example, 3–5 s) of a SAD-rich flare, or a completeness-corrected reanalysis of the existing 12-second catalogues, would settle the claim: if the short-waiting-time deficit disappears and the distribution becomes exponential once missed small SADs are included, the exclusion of linear random processes would fail; if the distribution remains power-law, the conclusion survives.

Watch

Extended reading notes

Core claim

The paper's central discovery is that the waiting time distribution (WTD) of supra-arcade downflows is heavy-tailed and non-exponential. In six of seven flare cases the power-law model wins or ties with the log-normal under likelihood ratio tests and AIC/AICc/BIC criteria, with maximum-likelihood power-law slopes $\alpha_{\mathrm{MLE}}$ around 1.7–2.0 and survival-function slopes $\alpha_{\mathrm{SDF}}$ mostly above 2 (up to 2.4). The exponential model is favored only in Case D, the smallest and most projection-affected sample. The paper therefore states that the results rule out linear random or quasi-periodic processes for SAD generation, and instead point to nonlinear mechanisms with intermittency, clustering, and memory—self-organized criticality, non-stationary Poisson driving, or MHD turbulence—possibly coupled at different scales in the reconnection outflow region.

Load-bearing premise

The SAD catalogues are complete enough that missed detections do not change the model ranking; if small SADs are missed at the 12-second cadence, short waiting times are under-sampled and the fitted distribution flattens, which could make power-law or log-normal win over exponential even if SADs came from a steady random process.

Editorial extensions

If this is right

  • Waiting times between SADs are not exponentially distributed, so SAD births cannot be described by a stationary Poisson process with a constant occurrence rate.
  • Power-law slopes around 1.7–2.4 align with predictions from self-organized criticality and non-stationary Poisson processes, supporting scale-invariant, avalanche-like energy release in the reconnection outflow region.
  • Log-normal fits are nearly as good in several cases, so multiplicative stochastic mechanisms, such as turbulent instabilities, remain in play alongside power-law-producing mechanisms.
  • The power-law/log-normal preference holds across automated tracking, manual tracking, and the new sliding-window manual method, suggesting it is not an artifact of a single detection technique.
  • Case D, the only exponential-favored case, has a small SAD sample and strong projection effects, indicating that limited samples can distort waiting-time model selection.

Reading between the lines

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

  • If short waiting times are under-sampled because small SADs are missed between 12-second frames, the intrinsic slope could be steeper than the raw 1.7–2.0, possibly moving even closer to the 2–3 range predicted by SOC and non-stationary Poisson models.
  • A controlled completeness test could quantify this bias: remove the shortest waiting times from a catalogue and measure how often the exponential model is artificially rejected; this would show how much of the power-law preference is observational rather than physical.
  • The heavy-tailed WTD implies that SAD waiting times should anticorrelate with the local flare energy-release rate, a prediction testable by correlating SAD start times with hard X-ray or EUV light curves of the same flares.
  • The sliding-window difference-imaging plus human-tracking method could be applied to other intermittently appearing coronal features, such as coronal jets or EUV brightenings, to see whether they show the same heavy-tailed timing signature.
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 / 6 minor

Summary. The paper studies the waiting time distributions (WTDs) of supra-arcade downflows (SADs) in seven limb flares, using previously published automated and manual catalogs (Cases A–F from Xie et al. 2022; Case G from Tan et al. 2022) plus a new manually detected event (Case H). Waiting times are defined from successive SAD start or end times. The authors fit each WTD with power-law, log-normal, and exponential distributions using maximum likelihood, also estimate power-law slopes via a survival-function method, and evaluate model preference with likelihood ratio tests and AIC/AICc/BIC. They report that most WTDs are best fitted by a power-law with slope 1.7–2.4, with log-normal often comparable, and conclude that linear random or quasi-periodic processes are ruled out in favor of nonlinear mechanisms such as self-organized criticality, non-stationary Poisson processes, or turbulence.

Significance. If the heavy-tailed WTD conclusion is robust, this would be the first systematic waiting-time analysis of SADs and would connect SAD generation to the broader statistics of flare-related reconnection. The paper has the notable strength of assembling data from multiple independent detection methods, including a same-event comparison (Cases F and G) and a new case (H), and it is candid about several observational limitations. However, the statistical evidence presented is not yet strong enough to support the strong exclusionary claim in the Abstract: the analysis lacks an absolute goodness-of-fit test, the detection incompleteness that could flatten the WTD is unquantified, several key likelihood-ratio tests are not significant, and the preferred model changes with start/end time convention and with detection method. The central claim is therefore plausible but not established at the level asserted.

major comments (4)
  1. [Section 3.1, Eqs. (3)–(4), Table 2] The power-law fit fixes xmin to the observed minimum waiting time in each case and no absolute goodness-of-fit statistic (such as a Kolmogorov-Smirnov p-value for the fitted power law, or a comparison of the MLE power-law fit against the empirical CDF) is reported. AIC/AICc/BIC only rank the three parametric families relative to one another, so a 'best' model can still be an absolutely poor description of the data. Since the paper's central claim is that power-law is the best WTD model, the absence of an absolute fit test is load-bearing and should be addressed.
  2. [Section 3.1, Section 3.2 (Cases F and G)] The detection incompleteness invoked in Section 3.1 is never quantified. The manuscript states that missed small-scale SADs would under-sample short waiting times and flatten the distribution, but it does not provide an injection/recovery test, a completeness function, or any bracketing of the effect. The F/G comparison is a direct demonstration of the sensitivity: on the same 2013 May 22 flare, the automated catalog (Case F, 230 SADs) yields a power-law preference for start-time WTDs, whereas the manual catalog (Case G, 81 SADs) yields a log-normal preference. This shows that the detection method can change the selected model, so the Abstract's claim to rule out linear random processes is stronger than the evidence supports.
  3. [Table 2, Section 3.2] The pairwise likelihood-ratio tests are mostly non-significant. For example, the log-normal vs. power-law LRT gives p = 0.057 for Case A, p = 0.108 for Case E, p > 0.5 for Cases B, C, and D, and p = 0.587 for Case G; only Cases F and H show strong preference for power-law over log-normal. The exponential vs. power-law LRT is also non-significant for Cases B, C, D, and G. Thus the data do not strongly reject log-normal or exponential alternatives in the majority of cases, and the 'rule out' language in the Abstract and Section 5 exceeds what the LRTs establish.
  4. [Section 3.2 (end-time statistics, Case D)] The robustness of the conclusion is further weakened by the model-switching between start- and end-time analyses, explicitly noted for Cases E and G, and by the exclusion of Case D as an outlier. Case D, the only case with a small sample (36 waiting times), favors the exponential distribution by AIC/AICc/BIC. Even if the projection-effect justification for excluding Case D is accepted, the paper should present the statistics with and without Case D and should not base the central claim on a classification that depends on post-hoc exclusion.
minor comments (6)
  1. [Figure 2] The axis labels in Figure 2 contain garbled characters such as 'Sta)t Time', 'P)obability Density', 'P owe) Law F it', and 'E−(onential F it'; the figure should be regenerated with proper fonts.
  2. [Abstract and Table 2] The Abstract quotes a slope range of 1.7–2.4, but Table 2 reports two different estimators, with alpha_MLE between 1.48 and 1.97 and alpha_SDF between 1.77 and 2.40. The text should state which estimator(s) the quoted range refers to, since the two methods give systematically different values.
  3. [Introduction] The sentence 'Section 5 offers the conclusions' is inconsistent with the actual section title 'SUMMARY & CONCLUSION'; please align the cross-reference.
  4. [Section 3.1] There is a typo in 'we present the the waiting time distributions' (duplicate 'the').
  5. [Section 3.1, SDF method] The survival-function fitting is described as a linear regression on a log-log scale, but the survival values are not independent and no uncertainties are reported for alpha_SDF. Please clarify the fitting procedure (e.g., whether a weighted least-squares was used) and provide uncertainties.
  6. [Section 2] The definition of waiting time is given only verbally; please state explicitly that start-to-start and end-to-end intervals are both used and that intervals for SADs appearing/disappearing in the same frame are excluded, as this is important for reproducibility.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity; the WTD result is an empirical fit compared with independent theoretical predictions.

full rationale

The paper's central derivation is an empirical model comparison: waiting times measured from previously published SAD catalogs (Xie et al. 2022; Tan et al. 2022) plus one new manual catalog are fit to power-law, log-normal, and exponential distributions via MLE, and the fitted slopes are compared with external theoretical predictions (Aschwanden 2014; Aschwanden & McTiernan 2010; Aschwanden et al. 2021). The theoretical slopes alpha=2 or 2-2.5 are taken from independent publications and are not derived from the SAD data, so there is no self-definitional or fitted-input-called-prediction loop. The choice xmin = minimum observed waiting time is a standard MLE convention, not a parameter fit to the theory; the observed alpha_MLE values are reported as measured, and the discrepancy with theory is attributed to detection incompleteness, which is a stated limitation rather than a circular rescaling. The few self-citations (Liu 2013; Liu & Wang 2021; Awasthi et al. 2022) support background claims about SAD nature and are not load-bearing for the WTD conclusion. The unquantified detection-incompleteness effect noted in Section 3.1 is a validity threat to the 'rule out exponential' claim, but it is not circular because the claim is not equivalent to the input by construction.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The analysis rests on the reliability of the SAD catalogs, the iid assumption behind the fits, a data-chosen lower cutoff, and published theoretical slope ranges. No new physical entities are postulated.

free parameters (1)
  • xmin (power-law lower cutoff) = 12 s for Cases A-F, H; 36 s for Case G
    Set equal to the minimum observed waiting time in each dataset rather than estimated by a goodness-of-fit procedure. Eq. 4 shows alpha_MLE depends on xmin, so this choice directly affects the reported slopes and model comparisons. No sensitivity analysis is given.
assumptions (4)
  • domain assumption SAD catalogs from Xie et al. (2022), Tan et al. (2022), and the authors' own manual tracking are accurate enough for timing statistics.
    All waiting times are derived from these catalogs. False positives inflate short counts; missed SADs deflate them. The paper acknowledges under-sampling but does not quantify its effect on model choice.
  • standard math Waiting times are exchangeable draws from a single candidate distribution during the MLE/AIC/LRT analysis.
    The fitted models assume iid samples. If SAD occurrence rate varies strongly within a flare (non-stationary Poisson), the aggregate fit is a mixture, and the comparison statistics are only approximate.
  • ad hoc to paper xmin equal to the minimum observed waiting time is a valid lower bound for power-law fitting.
    The paper states xmin is 'naturally specified' as the minimum waiting time. This is a data-dependent choice, not the Clauset et al. (2009) xmin estimation, and it affects alpha via Eq. 4.
  • domain assumption The published SOC and non-stationary Poisson slope predictions (alpha about 2 and 2-2.5) apply to SAD generation.
    Used in Section 4 to interpret the measured slopes. The SOC prediction assumes 3D classical diffusion (d=3, beta=1) and independent avalanches (T2 less than mean waiting time), conditions not verified for SADs.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Waiting Time Distribution of Supra-Arcade Downflows." pith.science (2026). https://pith.science/paper/WTU6XXV2

@misc{pith2026250716118,
  author       = {Pith},
  title        = {Pith review of: Waiting Time Distribution of Supra-Arcade Downflows},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WTU6XXV2}},
  note         = {Machine review of arXiv:2507.16118}
}
read the original abstract

Supra-arcade downflows (SADs) are dark, sunward-moving structures above the arcade of flare loops. Naturally they are considered to be associated with outflows resulting from magnetic reconnection at the vertical current sheet beneath the eruptive structure. Despite the extensive investigations on the SAD properties, the timing information, particularly the waiting time between consecutive SADs, has not been systematically examined. Here, using the 131~{\AA} passband of the Atmospheric Imaging Assembly on-board the Solar Dynamics Observatory, we studied the waiting time distribution (WTD) of SADs in 7 eruptive flares. In six of the 7 flares, the SADs are identified and tracked in previous studies, by two different methods; and in the 7th flare, by our optimized manual method. Based on statistical evaluations, we found that most of the WTDs are best fitted by a power-law function with the slope ranging from 1.7 to 2.4, in accordance with the prediction of non-stationary Poisson processes or self-organized criticality; but often they can also be fitted almost equally well by a log-normal function. These results rule out linear random or quasi-periodic processes to be responsible for the generation of SADs, but suggest that several nonlinear mechanisms be coupled together in the reconnection outflow region to shape the heavy-tailed WTD of SADs.

Figures

Figures reproduced from arXiv: 2507.16118 by the authors.

Figure 1
Figure 1. SADs observed in Case H. An AIA 131 ˚A image is shown in the upper panel, and the corresponding 1-min running difference image in the lower panel, where SADs are marked by red ‘+’ signs. Note that the images have been rotated 90 degree clockwise. An animation AIA 131 ˚A and corresponding running difference images is available online. in the flare. The waiting time is defined by the time interval between successive S… view at source ↗
Figure 2
Figure 2. WTDs derived from the start time of SADs. The histograms are normalized to show the probability density, and fitted with the MLE method for power-law (red), log-normal (green), and exponential (blue) functions. The fitting parameters are shown in [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 13 canonical work pages

  1. [1]

    1974, IEEE Transactions on Automatic Control, 19, 716

    Akaike, H. 1974, IEEE Transactions on Automatic Control, 19, 716

    1974

  2. [2]

    Aschwanden, M. J. 2011, Self-Organized Criticality in Astrophysics (Springer Berlin, Heidelberg), doi: 10.1007/978-3-642-15001-2

    work page doi:10.1007/978-3-642-15001-2 2011

  3. [3]

    Aschwanden, M. J. 2014, ApJ, 782, 54, doi: 10.1088/0004-637X/782/1/54

    work page doi:10.1088/0004-637x/782/1/54 2014

  4. [4]

    J., Johnson, J

    Aschwanden, M. J., Johnson, J. R., & Nurhan, Y. I. 2021, ApJ, 921, 166, doi: 10.3847/1538-4357/ac19a9

    work page doi:10.3847/1538-4357/ac19a9 2021

  5. [5]

    J., & McTiernan, J

    Aschwanden, M. J., & McTiernan, J. M. 2010, ApJ, 717, 683, doi: 10.1088/0004-637X/717/2/683

    doi:10.1088/0004-637x/717/2/683 2010

  6. [6]

    K., Liu, R., & Gou, T

    Awasthi, A. K., Liu, R., & Gou, T. 2022, ApJ, 941, 158, doi: 10.3847/1538-4357/aca3a8 10 Boffetta, G., Carbone, V., Giuliani, P., Veltri, P., &

    doi:10.3847/1538-4357/aca3a8 2022

  7. [7]

    1999, PhRvL, 83, 4662, doi: 10.1103/PhysRevLett.83.4662

    Vulpiani, A. 1999, PhRvL, 83, 4662, doi: 10.1103/PhysRevLett.83.4662

    work page doi:10.1103/physrevlett.83.4662 1999

  8. [8]

    C., et al

    Cai, Q., Shen, C., Raymond, J. C., et al. 2019, MNRAS, 489, 3183, doi: 10.1093/mnras/stz2167

    doi:10.1093/mnras/stz2167 2019

Show all 42 references
  1. [9]

    Ciaravella, A., & Raymond, J. C. 2008, ApJ, 686, 1372, doi: 10.1086/590655

    2008 doi

  2. [10]

    R., & Newman, M

    Clauset, A., Shalizi, C. R., & Newman, M. E. J. 2009, SIAM Review, 51, 661, doi: 10.1137/070710111

    2009 doi

  3. [11]

    G., Linker, J

    Forbes, T. G., Linker, J. A., Chen, J., et al. 2006, SSRv, 123, 251, doi: 10.1007/s11214-006-9019-8

    2006 doi

  4. [12]

    Reeves, K. K. 2025, ApJL, 986, L16, doi: 10.3847/2041-8213/adde59

    2025 doi

  5. [13]

    2002, PhRvE, 65, 046203, doi: 10.1103/PhysRevE.65.046203

    Grigolini, P., Leddon, D., & Scafetta, N. 2002, PhRvE, 65, 046203, doi: 10.1103/PhysRevE.65.046203

    2002 doi

  6. [14]

    J., Huang, Y

    Guo, L. J., Huang, Y. M., Bhattacharjee, A., & Innes, D. E. 2014, ApJL, 796, L29, doi: 10.1088/2041-8205/796/2/L29

    2014 doi

  7. [16]

    J., & Reeves, K

    Hanneman, W. J., & Reeves, K. K. 2014b, ApJ, 786, 95, doi: 10.1088/0004-637X/786/2/95

    doi

  8. [17]

    E., McKenzie, D

    Innes, D. E., McKenzie, D. E., & Wang, T. 2003, SoPh, 217, 247, doi: 10.1023/B:SOLA.0000006899.12788.22

    2003

  9. [18]

    R., Title, A

    Lemen, J. R., Title, A. M., Akin, D. J., et al. 2012, Solar Physics, 275, 17, doi: 10.1007/s11207-011-9776-8

    2012 doi

  10. [19]

    2001, ApJL, 555, L133, doi: 10.1086/323178

    Lepreti, F., Carbone, V., & Veltri, P. 2001, ApJL, 555, L133, doi: 10.1086/323178

    2001 doi

  11. [20]

    F., Cheng, X., Ding, M

    Li, Z. F., Cheng, X., Ding, M. D., et al. 2021, ApJ, 915, 124, doi: 10.3847/1538-4357/ac043e

    2021 doi

  12. [21]

    2004, SoPh, 222, 115, doi: 10.1023/B:SOLA.0000036875.14102.39

    Lin, J. 2004, SoPh, 222, 115, doi: 10.1023/B:SOLA.0000036875.14102.39

    2004

  13. [22]

    G., et al

    Lin, J., Li, J., Forbes, T. G., et al. 2007, ApJL, 658, L123, doi: 10.1086/515568

    2007 doi

  14. [23]

    2013, MNRAS, 434, 1309, doi: 10.1093/mnras/stt1090

    Liu, R. 2013, MNRAS, 434, 1309, doi: 10.1093/mnras/stt1090

    2013 doi

  15. [24]

    2021, A&A, 653, A51, doi: 10.1051/0004-6361/202140847

    Liu, R., & Wang, Y. 2021, A&A, 653, A51, doi: 10.1051/0004-6361/202140847

    2021 doi

  16. [25]

    McKenzie, D. E. 2013, ApJ, 766, 39, doi: 10.1088/0004-637X/766/1/39

    2013 doi

  17. [26]

    E., & Hudson, H

    McKenzie, D. E., & Hudson, H. S. 1999, ApJL, 519, L93, doi: 10.1086/312110

    1999 doi

  18. [27]

    D., Thompson, B

    Pesnell, W. D., Thompson, B. J., & Chamberlin, P. C. 2012, SoPh, 275, 3, doi: 10.1007/s11207-011-9841-3

    2012 doi

  19. [28]

    K., Freed, M

    Reeves, K. K., Freed, M. S., McKenzie, D. E., & Savage, S. L. 2017, ApJ, 836, 55, doi: 10.3847/1538-4357/836/1/55

    2017 doi

  20. [29]

    K., T¨ or¨ ok, T., Miki´ c, Z., Linker, J., & Murphy, N

    Reeves, K. K., T¨ or¨ ok, T., Miki´ c, Z., Linker, J., & Murphy, N. A. 2019, ApJ, 887, 103, doi: 10.3847/1538-4357/ab4ce8

    2019 doi

  21. [30]

    2021, The Innovation, 2, 100083, doi: 10.1016/j.xinn.2021.100083 S´ anchez, R., & Newman, D

    Samanta, T., Tian, H., Chen, B., et al. 2021, The Innovation, 2, 100083, doi: 10.1016/j.xinn.2021.100083 S´ anchez, R., & Newman, D. 2018, A Primer on Complex

    2021

  22. [31]

    943, doi: 10.1007/978-94-024-1229-1

    Systems, Vol. 943, doi: 10.1007/978-94-024-1229-1

    doi

  23. [32]

    L., & McKenzie, D

    Savage, S. L., & McKenzie, D. E. 2011, ApJ, 730, 98, doi: 10.1088/0004-637X/730/2/98

    2011 doi

  24. [33]

    L., McKenzie, D

    Savage, S. L., McKenzie, D. E., & Reeves, K. K. 2012, ApJL, 747, L40, doi: 10.1088/2041-8205/747/2/L40

    2012 doi

  25. [34]

    1978, Annals of Statistics, 6, 461

    Schwarz, G. 1978, Annals of Statistics, 6, 461

    1978

  26. [35]

    K., et al

    Shen, C., Chen, B., Reeves, K. K., et al. 2022, Nature Astronomy, 6, 317, doi: 10.1038/s41550-021-01570-2

    2022 doi

  27. [36]

    2010, Physics of plasmas, 17

    Skender, M., & Lapenta, G. 2010, Physics of plasmas, 17

    2010

  28. [37]

    1978, Communications in Statistics-theory and Methods, 7, 13

    Sugiura, N. 1978, Communications in Statistics-theory and Methods, 7, 13

    1978

  29. [38]

    2022, MNRAS, 516, 3120, doi: 10.1093/mnras/stac2470 ˇSvestka, Z., F´ arn ´ ık, F., Hudson, H

    Tan, G., Hou, Y., & Tian, H. 2022, MNRAS, 516, 3120, doi: 10.1093/mnras/stac2470 ˇSvestka, Z., F´ arn ´ ık, F., Hudson, H. S., & Hick, P. 1998, SoPh, 182, 179, doi: 10.1023/A:1005033717284

    2022 doi

  30. [39]

    Wheatland, M. S. 2000, ApJL, 536, L109, doi: 10.1086/312739

    2000 doi

  31. [40]

    2004, in Proceedings of the 2004 Winter Simulation Conference, 2004., Vol

    Willinger, W., Alderson, D., Doyle, J., & Li, L. 2004, in Proceedings of the 2004 Winter Simulation Conference, 2004., Vol. 1, 141, doi: 10.1109/WSC.2004.1371310

    2004 arXiv

  32. [41]

    K., Shen, C., & Ingram, J

    Xie, X., Reeves, K. K., Shen, C., & Ingram, J. D. 2022, ApJ, 933, 15, doi: 10.3847/1538-4357/ac695d

    2022 doi

  33. [42]

    K., et al

    Xie, X., Shen, C., Reeves, K. K., et al. 2025, ApJL, 984, L27, doi: 10.3847/2041-8213/adc91b

    2025 doi

  34. [43]

    2020, ApJ, 898, 88, doi: 10.3847/1538-4357/ab9a3d

    Xue, J., Su, Y., Li, H., & Zhao, X. 2020, ApJ, 898, 88, doi: 10.3847/1538-4357/ab9a3d

    2020 doi

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.