Pith. sign in

REVIEW 3 major objections 5 minor 33 references

Performance of radio-based detection to operational monitoring M5+ class solar flares

T0 review · 3 major / 5 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read Ground radio can detect major solar flares minutes before X-rays

desk verdict A solid feasibility study with a real predictive-information finding at 8.8 GHz, but the operational monitoring claim is built on a sample that never sees quiet-Sun periods. read the letter →

arxiv 2608.02199 v1 pith:RQXC56XF submitted 2026-08-03 astro-ph.SR astro-ph.IMphysics.plasm-phphysics.space-ph

Julia van Ravenswaaij , Ines Wilms , Ivan Ricardo , Michaela Brchnelova This is my paper
classification astro-ph.SRastro-ph.IMphysics.plasm-phphysics.space-ph
keywords solarflaresM5+radiobursts8.8GHzmonitoringelasticnetlogisticregressionspaceweatherearlywarninggyrosynchrotronemissionground-basedflaredetection
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 sets out to show that ground-based radio observations can replace or back up space-based X-ray measurements for monitoring major solar flares (class M5 and above). Using two years of radio flux at eight frequencies and X-ray flare classifications, it trains a regularized logistic regression that must decide, minute by minute, whether an M5+ flare is underway. It finds that the 8.8 GHz channel carries the signal: elevated radio flux appears on average 3–4 minutes before the X-ray threshold is crossed, and minute-level detection lands at 53% precision and 65% recall. If the approach holds, defense and communication users would gain a ground-based early-warning channel that does not depend on satellite availability.

What carries the argument

The machinery is an elastic net regularized logistic regression—a binary classifier that estimates the minute-level probability of an M5+ flare from radio flux while penalizing coefficients to select a sparse set of predictors. The model includes each of eight radio channels at lags 0 through 10 minutes, uses class weights to counterbalance that only ~6% of minutes are flare minutes, and tunes its penalty λ via five-fold time-series cross-validation on the F1 score. The elastic net's selection is what pins the argument to the 8800 MHz channel: all eleven coefficients for that channel (instantaneous plus ten lags) survive, and the lag-3 and lag-4 terms carry the largest weights, which is what

What would settle it

A concrete test: apply the fitted model to a continuous stretch of radio and X-ray data that includes at least a few months with no M5+ flares, and measure the false-alarm rate per day. If the daily rate is too high for operators, the operational claim fails regardless of within-window recall. Alternatively, check whether flares without a high-frequency radio signature (e.g., limb flares) are systematically missed, which would cap recall.

Watch

Extended reading notes

Core claim

The central claim is that super-high-frequency radio measurements, especially at 8800 MHz, constitute a workable alternative for monitoring M5-class and stronger solar flares when soft X-ray data are unavailable. In the authors' model, the 8.8 GHz channel is the only predictor retained across all lags; its strongest coefficients sit 3 and 4 minutes before the M5 X-ray threshold is crossed, matching the expected timing of gyrosynchrotron emission from the accelerated electrons that later produce the gradual soft X-ray rise. On a held-out 30% of the data, the model detects 877 of 1,346 flare minutes (65% recall) with 53% precision, and at the flare level it forecasts 17 of 30 events, half of t

Load-bearing premise

The load-bearing premise is that the model's behavior inside flare-adjacent windows represents how it will behave in real operations, but the model has never been tested on the long quiet stretches that an operational alarm would actually face.

Editorial extensions

If this is right

  • If correct, flare monitoring no longer depends on space-borne X-ray sensors; a ground receiver can issue alerts during satellite outages.
  • The 3–4 minute lead time gives automated systems a small but real window to warn HF, VHF/UHF/SHF users before the M5 threshold is crossed.
  • A single 8.8 GHz channel carries most of the predictive value; lower-frequency channels (below about 1 GHz) are poor predictors because they pick up many non-flare radio bursts.
  • For high-risk operations, lowering the decision threshold below 0.5 would trade false alarms for fewer missed flares, since the default threshold misses about one third of flare minutes.
  • A multiclass extension beyond the binary M5+ threshold would be needed for global applicability at lower latitudes, where weaker flares can matter.

Reading between the lines

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

  • Because the training windows all contain at least one known M5+ flare, the reported 93% accuracy and 53% precision are conditional on an enriched sample; a real operational deployment would also encounter long quiet periods, so the false-alarm rate in ordinary conditions remains untested.
  • The authors note that some false negatives were observed by a second station, suggesting that data fusion across stations—rather than only adding frequency channels—could lift recall further.
  • For flares without gyrosynchrotron emission (for example limb flares or radio-quiet events), no ground radio signature will appear; a robust operational system would need to combine radio with other ground-based proxies or accept these misses.
  • A testable extension is to apply the same model to continuous data spanning solar minimum to measure how the false-alarm count changes when the flare base rate is much lower.
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

3 major / 5 minor

Summary. The paper investigates whether ground-based radio observations can monitor and predict M5+ solar flares as a possible alternative to space-based soft X-ray monitoring. Using two years of GOES-16 X-ray data and RSTN radio data (8 channels, 1-minute resolution), the authors fit an elastic-net regularized logistic regression with up to 10-minute lags, using time-series cross-validation and class weighting. They report that the 8.8 GHz channel is the dominant predictor, with minute-level precision/recall of 53%/65%, an overall accuracy of 93%, and claim that radio signals appear 3–4 minutes before the X-ray M5 threshold is crossed. The paper concludes that radio measurements at super high frequencies can serve as an alternative to monitor major flaring activity.

Significance. If the operational claims were supported, this would be a valuable contribution, providing a ground-based backup for space-based flare monitoring, with clear relevance to the DISTURB system and HF users. The study is carefully designed in several respects: a temporal hold-out test set, time-series cross-validation for tuning, post-double-selection inference, and a channel-by-channel comparison. The finding that 8.8 GHz carries the strongest predictive signal is physically plausible and is a concrete, falsifiable result. However, a key limitation in the sampling frame means the central operational conclusion is not currently supported by the evidence.

major comments (3)
  1. [Section 2.2, Appendix C.1] The dataset is constructed by stacking time windows around known M5+ flares; no quiet-Sun intervals are included. The test set therefore has a 7.52% flare-minute base rate, roughly 20 times the ~0.35% operational base rate over the same two years. Consequently the reported 93% accuracy and 53% precision are conditional on an enriched sample. If the false-positive rate observed in flare windows (775/16,564 ≈ 4.7%) applied to quiet periods, precision under the true base rate would be approximately (0.65×0.0035)/(0.65×0.0035 + 0.047×0.9965) ≈ 4.6%, about an order of magnitude lower. The abstract's claim that radio 'can serve as an alternative method to monitor major solar flaring activity' is not supported by the measured metrics. The model must be evaluated on continuous data that includes quiet minutes, or the conclusions must be restricted to 'periods of known flare activity.'
  2. [Section 3.4, Eq. (3), Section 3.2] The claim that radio signals appear '3 to 4 minutes before' the M5 threshold is based on the magnitudes of the lagged coefficients at lags 3 and 4 in Eq. (3), not on a direct measurement of lead times. These coefficients describe the conditional association in a standardized logistic regression and do not themselves constitute an observed temporal lead. The paper's own event-level analysis (Section 3.2) shows that only 7 of 30 test flares were predicted before onset, with a median lead time of 4 minutes for those 7. The abstract's 'on average' phrasing is not supported by the data. Please report the distribution of actually observed lead times between radio-based positive predictions and X-ray threshold crossings, or substantially qualify the early-warning claim.
  3. [Section 3.2, Table 4, Section 5] The paper switches between minute-level and flare-level performance without clarifying which is the operational metric. The abstract and conclusion cite minute-level precision/recall (53%/65%, Table 4) and infer that 'roughly one third of flares were not detected.' At the flare level, however, only 17 of 30 test flares were correctly forecast (recall 56.7%) and 4 false alarms were recorded. A monitoring system would be evaluated on event-level detection and false-alarm rates, not on minute-level counts. The authors should explicitly specify the intended operational use case and report both levels consistently, explaining which is the relevant performance metric.
minor comments (5)
  1. [Section 3.3] The text says 'A total of 468 observations were flagged as false negatives,' but Table 3 reports 469 FN. Please correct the inconsistency.
  2. [Table D.17] In the Channel 8 row, the Recall column appears as '007667' — likely a typo for 0.7667. Please verify.
  3. [Section 3.2] The event-level results for the combined model (17 TP, 13 FN, 4 FA, 2 NM) are reported only in prose. A table similar to Appendix D would improve reproducibility and comparison across channels.
  4. [General] The term 'flare events' is used for both one-minute observations and individual flares. Please define these terms explicitly to avoid confusion between minute-level and flare-level analyses.
  5. [Section 2.3.2] The choice α=0.5 is stated to be fixed for computational convenience; this is acknowledged later in Section 4.2, but could be mentioned earlier as a limitation of the hyperparameter search.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central feasibility claim is supported by a temporal hold-out evaluation on external RSTN/GOES data.

full rationale

The paper's central claim—that RSTN radio flux, especially at 8.8 GHz, carries predictive information about M5+ X-ray flares—is supported by a supervised logistic regression with a temporal hold-out: the model is trained on the first 41,251 minutes and evaluated on the last 17,910 test minutes (Appendix C.1). The radio predictors and X-ray labels are independent measurements, and the reported precision/recall are out-of-sample, so the core feasibility claim is not circular. The '3 to 4 minutes before' lead-time statement is partly a reading of the largest fitted lag coefficients (ch7_{t-3}, ch7_{t-4}), but the paper also reports test-set flare-level lead times (7 of 17 detected flares were identified before onset, with a median lead time of 4 minutes), so it is not purely a re-description of the fit. The substantive concern is external validity, not circularity: all training/test windows are stacked around M5+ flares ('we stacked all time windows with M5+solar flare activity'), so quiet-Sun false-alarm behavior is unmeasured and the operational monitoring claim may be overstated. That is a sampling/generalization limitation, not a derivation that reduces to its own inputs. No self-citation is load-bearing; the external references are context, not premises of the statistical result.

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

The central claim rests on five modeling choices and domain assumptions: the GOES ground truth, RSTN cross-station comparability, linearity of the logistic link, stationarity over the two-year window, and the validity of post-selection inference. No new physical entities are introduced. The free parameters are mostly standard regularization choices; the sample-frame assumption (only windows around known flares) is the most consequential and is not captured in this ledger because it is a data-construction choice rather than a fitted number.

free parameters (5)
  • lambda (elastic net penalty strength) = 2836
    Selected by time-series 5-fold CV maximizing F1 on tuning folds (Appendix C.2); controls sparsity and directly shapes the reported model.
  • alpha (elastic net mixing) = 0.5
    Fixed by hand to reduce computational burden (Section 2.3.2); not tuned, though the authors acknowledge other values may improve performance.
  • class weights = 1/p0 and 1/p1 from the training sample
    Inverse-frequency class weighting used in the loss to counter imbalance (Appendix C.1); depends on the sample's flare proportion.
  • decision threshold = 0.5
    Standard logistic cutoff used for all confusion matrices; the authors note that operational thresholds might be lower (Section 4.2).
  • maximum lag = 10 minutes
    Chosen to capture the pre-flare radio lead; the authors note larger windows could change performance and complexity.
assumptions (5)
  • domain assumption GOES-16 soft X-ray long-channel flux is the accepted ground truth for whether an M5+ flare occurred.
    Section 2.1; the paper uses GOES XRS as the classification standard, even though its space-based availability is the motivation for the study.
  • domain assumption RSTN stations are sufficiently cross-calibrated that fusing their data into one time series is valid.
    Section 2.2; data from San Vito, Sagamore Hill, and Palehua are combined after per-minute max aggregation.
  • domain assumption A linear logistic relationship between radio flux and the log-odds of M5+ flare occurrence is adequate.
    Equation (1); nonlinear effects are not modeled, and the authors list this as a limitation (Section 4.2).
  • domain assumption The March 2023–March 2025 period is sufficiently stationary for time-series cross-validation to be valid.
    Appendix C.1; the data cover only the SC25 maximum, and the model does not test for distribution shift across solar cycle phases.
  • standard math Post-double-selection p-values are valid for inference after elastic-net variable selection.
    Section 3.1; relies on the Belloni et al. (2016, 2017) post-selection inference framework.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Performance of radio-based detection to operational monitoring M5+ class solar flares." pith.science (2026). https://pith.science/paper/RQXC56XF

@misc{pith2026260802199,
  author       = {Pith},
  title        = {Pith review of: Performance of radio-based detection to operational monitoring M5+ class solar flares},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RQXC56XF}},
  note         = {Machine review of arXiv:2608.02199}
}
read the original abstract

Early detection of major solar flares is critical for defense operations due to their potential to disturb radar and radio systems. Typically, soft X-ray flux is used to monitor and classify solar flares, but since this flux has to be measured in space, it means that its availability itself is dependent on space weather conditions. For this reason, in this paper, we investigated the feasibility of using ground radio observations to monitor major (M5+ class) solar flares. We made use of datasets from the GOES-16 satellite and the Radio Solar Telescope Network in the time range between March 2023 and March 2025. An elastic net regularized logistic regression model was trained on this data, optimized through a grid search and with incorporated class weighting for class imbalance. It was found that especially higher frequencies (8800 MHz) had a reasonable ability in monitoring and predicting major flares (precision and recall for flare events are 53% and 65%, respectively - implying that roughly one third of flares were not detected - with signals appearing, on average, 3 to 4 minutes before the M5 threshold is exceeded). Radio measurements at super high frequencies can thus serve as an alternative method to monitor major solar flaring activity.

Figures

Figures reproduced from arXiv: 2608.02199 by the authors.

Figure 1
Figure 1. X-ray and corresponding radio data during a solar flare at 28 March 2024 regression model with elastic net regularization. For details on the statistical methodology, we refer the interested readers to Hastie et al. (2009), Chapter 4. 2.3.1. Logistic regression Let yt denote the response variable that takes on the value one if an M5+ solar flare occurs (i.e. M5 threshold exceeded) during minute t, and zero otherwise… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

33 extracted references · 12 canonical work pages

  1. [1]

    10.1029/2024JA032925

    Design and Pre-Flight Calibration.Journal of Geophysical Research: Space Physics,129(11). 10.1029/2024JA032925. 1, 2.1 Wu, Z., A. Kuznetsov, S. Anfinogentov, V . Melnikov, R. Sych, et al.,

    doi:10.1029/2024ja032925

  2. [2]

    1 Benz, A

    10.1007/s41116-016- 0004-3. 1 Benz, A. O., R. Brajˇsa, and J. Magdaleni´c,

    doi:10.1007/s41116-016-

  3. [5]

    Solar Physics,226(1), 143–151

    Callisto A New Concept for Solar Radio Spectrometers. Solar Physics,226(1), 143–151. 10.1007/s11207-005-5688-9,astro-ph/0410437. 4.2 Brown, J. C.,

    arXiv

  4. [8]

    Precision Recall F1-score Support 0 (No Flare) 0.9744 0.7784 0.8655 16 564 1 (Flare) 0.2153 0.7481 0.3344 1 346 Accuracy 0.7762 17 910 Table D.17.Classification performance at the flare level: M5+flares forecasted out of a total of 30 flares in the test set (True Positives; TP) thereby indicating how many of these were predicted to occur before, at or aft...

    2000

  5. [11]

    10.1086/306101

    Reconstructing the Thermal and Spatial Form of a Solar Flare from Scaling Laws and Soft X-Ray Measurements.The Astrophysical Journal,504(2), 1051–1066. 10.1086/306101. 1 Giersch, O., and J. Kennewell,

    work page doi:10.1086/306101

  6. [16]

    10.1051/0004- 6361:20077643,0707.2322

    The correlation between soft and hard X-rays component in flares: from the Sun to the stars.Astronomy&Astrophysics,472(1), 261–268. 10.1051/0004- 6361:20077643,0707.2322. 1, 4.1 Kane, S. R.,

    arXiv

  7. [22]

    10.1051/0004- 6361/201423731,1410.2410

    Large-scale simulations of solar type III radio bursts: flux density, drift rate, duration, and bandwidth.Astronomy and Astrophysics,572, A111. 10.1051/0004- 6361/201423731,1410.2410. 1 Reid, H. A. S., and N. Vilmer,

    arXiv

  8. [23]

    10.1051/0004-6361/201527758,1609.04743

    Coronal type III radio bursts and their X-ray flare and interplanetary type III counterparts.Astronomy&Astrophysics,597, A77. 10.1051/0004-6361/201527758,1609.04743. 1 Shamsutdinova, J. N., L. K. Kashapova, J. Zhang, H. Reid, and D. A. Zhdanov,

    arXiv

Show all 33 references
  1. [24]

    10.1093/mnras/stae1899,https://academic.oup.com/ mnras/article-pdf/533/2/1453/58883301/stae1899.pdf

    Relationship be- tween microwave and metre ranges during an impulsive solar flare.Monthly Notices of the Royal Astronomical Society,533(2), 1453–1462. 10.1093/mnras/stae1899,https://academic.oup.com/ mnras/article-pdf/533/2/1453/58883301/stae1899.pdf. 1, 4.1 Trottet, G.,

    doi

  2. [26]

    10.1007/s11214-010-9708- 1,1109.6629

    The Relationship Between Solar Radio and Hard X-ray Emission.Space Science Reviews,159(1-4), 225–261. 10.1007/s11214-010-9708- 1,1109.6629. 1 Woods, T. N., T. Eden, F. G. Eparvier, A. R. Jones, D. L. Woodraska, P. C. Chamberlin, and J. L. Machol,

    arXiv

  3. [28]

    1 Xu, X., Z

    10.3847/1538-4357/ad46ff,2405.03116. 1 Xu, X., Z. Zhao, Q. Liu, Q. Li, Z. Wu, G. Lu, Y . Su, Y . Chen, and F. Yan,

    arXiv

  4. [29]

    1 Zou, H., and T

    10.3847/1538-4365/ade0b5. 1 Zou, H., and T. Hastie,

    doi

  5. [30]

    10.1111/j.1467- 9868.2005.00503.x

    Regularization and variable selection via the elastic net.Journal of the Royal Statistical Society: Series B (Statistical Methodology),67(2), 301–320. 10.1111/j.1467- 9868.2005.00503.x. 2.3.2 18 van Ravenswaaij et al.: Radio-based solar flare detection Appendix A: Pre-analysis...

    2005

  6. [31]

    A.2.Number of solar flares per day (M5-class and above) from March 2023 to March

    19 van Ravenswaaij et al.: Radio-based solar flare detection Fig. A.2.Number of solar flares per day (M5-class and above) from March 2023 to March

    2023

  7. [32]

    A.3.Distribution of solar flare durations (M5-class and above) from March 2023 to March

    Fig. A.3.Distribution of solar flare durations (M5-class and above) from March 2023 to March

    2023

  8. [112]

    1 17 van Ravenswaaij et al.: Radio-based solar flare detection Ratcliffe, H., E

    10.3847/1538-4357/ac98b4, 2210.03364. 1 17 van Ravenswaaij et al.: Radio-based solar flare detection Ratcliffe, H., E. P. Kontar, and H. A. S. Reid,

    arXiv

  9. [483]

    1 Matsumoto, K., S

    10.1071/PH570483. 1 Matsumoto, K., S. Masuda, M. Shimojo, and H. Hayakawa,

    doi

  10. [1113]

    1 Li, B., and I

    10.1086/159121. 1 Li, B., and I. H. Cairns,

    doi

  11. [1971]

    10.1007/BF00149070

    The Deduction of Energy Spectra of Non-Thermal Electrons in Flares from the Observed Dynamic Spectra of Hard X-Ray Bursts.Solar Physics,18(3), 489–502. 10.1007/BF00149070. 4.1 Defensie,

    doi

  12. [1972]

    10.1007/BF00153384

    Thick-Target Processes and White-Light Flares.Solar Physics,24(2), 414–428. 10.1007/BF00153384. 4.1 Isola, C., F. Favata, G. Micela, and H. S. Hudson,

    doi

  13. [1986]

    10.1007/BF00159956

    Relative Timing of Hard X-Rays and Radio Emissions during the Different Phases of Solar Flares - Consequences for the Electron Acceleration.Solar Physics,104(1), 145–163. 10.1007/BF00159956. 1 White, S. M., A. O. Benz, S. Christe, F. F´arn´ık, M. R. Kundu, et al.,

    doi

  14. [1988]

    10.1007/BF00148588

    Solar Flare Hard X-Ray Observations.Solar Physics,118(1-2), 49–94. 10.1007/BF00148588. 1, 4.1 16 van Ravenswaaij et al.: Radio-based solar flare detection Ferrente, F., C. Quintero Noda, F. Zuccarello, and S. L. Guglielmino,

    doi

  15. [1998]

    10.1146/annurev.astro.36.1.131

    Radio Emission from Solar Flares.Annual Review of Astronomy and Astrophysics,36, 131–188. 10.1146/annurev.astro.36.1.131. 1 Belloni, A., V . Chernozhukov, I. Fern´andez-Val, and C. Hansen,

    doi

  16. [2005]

    10.1007/s11207-005-5254-5

    Survey on Solar X-ray Flares and Associated Coherent Radio Emissions.Solar Physics,226(1), 121–142. 10.1007/s11207-005-5254-5. 1 Benz, A. O., C. Monstein, and H. Meyer,

    doi

  17. [2007]

    10.1007/s11207-007-0365-9,astro-ph/0701570

    Are There Radio-quiet Solar Flares?Solar Physics,240(2), 263–270. 10.1007/s11207-007-0365-9,astro-ph/0701570. 3.3, 4.1, 5 Benz, A. O., P. C. Grigis, A. Csillaghy, and P. Saint-Hilaire,

    arXiv

  18. [2009]

    Springer, New York, NY , USA, 2 edn

    The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Springer, New York, NY , USA, 2 edn. ISBN Hardcover: 978-0-387-84857-0, ISBN eBook: 978-0-387-84858-7. 10.1007/978-0-387-84858-7. 2.3 Hudson, H. S.,

    doi

  19. [2011]

    10.1007/s11214-010-9701-8,1109.5932

    An Observational Overview of Solar Flares.Space Science Reviews,159(1-4), 19–106. 10.1007/s11214-010-9701-8,1109.5932. 4.1 Garcia, H. A.,

    arXiv

  20. [2014]

    10.1007/s11207-013-0375-8

    Fundamental Emission of Type III Bursts Produced in Non-Maxwellian Coronal Plasmas with Kappa-Distributed Background Particles.Solar Physics,289(3), 951–976. 10.1007/s11207-013-0375-8. 1 Loughhead, R. E., J. A. Roberts, and M. K. McCabe,

    doi

  21. [2017]

    10.1002/2017SW001658, URLhttps://doi.org/ 10.1002/2017SW001658

    Solar Radio Burst Statistics and Implications for Space Weather Effects.Space Weather,15, 1511–1522. 10.1002/2017SW001658, URLhttps://doi.org/ 10.1002/2017SW001658. 1, 4.1, 5 Hastie, T., R. Tibshirani, and J. H. Friedman,

    doi

  22. [2022]

    E2022RS007456 2022RS007456, https://doi.org/10.1029/2022RS007456,https:// agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2022RS007456

    Analysis of the Radio Solar Telescope Network’s Noon Flux Observations Over Three Solar Cycles (1988–2020).Radio Science,57(8), e2022RS007,456. E2022RS007456 2022RS007456, https://doi.org/10.1029/2022RS007456,https:// agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2022RS00745...

    1988 doi

  23. [2023]

    10.1093/pasj/psad058,https://academic.oup

    Relationship of peak fluxes of solar radio bursts and X-ray class of solar flares: Application to early great solar flares.Publications of the Astronomical Society of Japan,75(6), 1095–1104. 10.1093/pasj/psad058,https://academic.oup. com/pasj/article-pdf/75/6/1095/54151707/psa...

    doi

  24. [2024]

    10.1051/0004-6361/202449512,2404.06231

    Understanding the thermal and magnetic properties of an X-class flare in the low solar atmosphere.Astronomy&Astrophysics,686, A244. 10.1051/0004-6361/202449512,2404.06231. 4.1 Fletcher, L., B. R. Dennis, H. S. Hudson, S. Krucker, K. Phillips, et al.,

    arXiv

  25. [2025]

    URLhttps://www.defensie.nl/actueel/ nieuws/2024/05/30/instrument-disturb-bepaalt-zon-of-vijand

    Instrument Disturb bepaalt: zon of vijand. URLhttps://www.defensie.nl/actueel/ nieuws/2024/05/30/instrument-disturb-bepaalt-zon-of-vijand. 1 Dennis, B. R.,

    2024

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.