Analysis of 42 years of Cosmic Ray Measurements by the Neutron Monitor at Lomnick\'y st\'it Observatory

Filip \v{S}tempel; Igor Strh\'arsk\'y; Imre Kisv\'ardai; J\'an Kuban\v{c}\'ak; Luk\'a\v{s} Randu\v{s}ka; Ronald Langer; \v{S}imon Mackovjak

REVIEW 4 major objections 5 minor 12 references

Analysis of 42 years of Cosmic Ray Measurements by the Neutron Monitor at Lomnick\'y st\'it Observatory

T0 review · 4 major / 5 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read A 42-year neutron-monitor record links cosmic rays to solar activity and geomagnetic storms, with a 7-21 hour lead that may enable storm forecasting.

desk verdict A genuinely useful 42-year neutron monitor dataset release wrapped in an over-sold prediction story; the correlation work is fine as description, but the 7–21 h lead-time claim and the 0.22 predictive power score do not support real-time storm forecasting as presented. read the letter →

arxiv 2502.09627 v1 pith:LT5KASXB submitted 2025-01-30 physics.space-ph astro-ph.EP

Imre Kisvárdai , Filip Štempel , Lukáš Randuška , Šimon Mackovjak , Ronald Langer , Igor Strhárský , Ján Kubančák This is my paper · ORCID

classification physics.space-phastro-ph.EP

keywords cosmicraysneutronmonitorspaceweathergeomagneticstormpredictionForbushdecreasethunderstormgroundenhancementDstindexsolarradioflux

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 claims that the Lomnický štít neutron monitor, a ground-based cosmic-ray detector in Slovakia, has now produced the first continuous, publicly accessible 42-year record of cosmic-ray flux, from December 1981 to July 2023 in hourly resolution and from 2001 in minute resolution. Using this record, the authors show that neutron count rates are anticorrelated with the solar radio flux (Pearson -0.743) and positively correlated with geomagnetic activity (Dst), with the best match occurring when Dst is shifted 7-21 hours later than the neutron data. During Forbush decreases, the correlation with Dst rises to 0.589 with a 7-hour delay. A predictive power score of 0.22 for neutron flux against Dst is presented as evidence that cosmic-ray measurements could feed real-time geomagnetic storm prediction models. The paper also reports the discovery of three thunderstorm ground enhancements from June 2022.

What carries the argument

The load-bearing object is the curated, barometrically corrected neutron-monitor count-rate series itself, produced by averaging the four best tubes of the LMKS 8NM64 monitor, interpolating gaps (less than 0.016% of samples), and applying the exponential pressure-correction $N_{\rm cor}=N_{\rm raw}\exp(\beta(p-p_m))$ with the station-specific barometric coefficient. The statistical machinery draws on Pearson correlation with time-shifted Dst and F10.7 indices, and on a decision-tree predictive power score that can capture non-linear relations; the 7-21 hour correlation peak is the specific mechanism claimed to enable storm warnings, physically attributed to the velocity difference between relativistic cosmic rays and the solar wind.

What would settle it

Re-run the prediction test with a time-series split: train the decision tree on the first 30 years of neutron flux and Dst data, then predict the last 12 years; if the normalized RMSE is no better than the naive median baseline (predictive power score near 0), the claim of real-time storm forecasting from neutron measurements is not supported.

Watch

Extended reading notes

Core claim

The central discovery is that a cleaned, pressure-corrected time series from the 8NM64 neutron monitor at Lomnický štít is stable and complete enough to serve as a long-term space-weather dataset, and that this dataset carries an early-warning signature: neutron flux is best correlated with the Dst index when the Dst series is delayed by 7-21 hours, meaning cosmic-ray variations tend to precede the ring-current response by roughly a day. The authors interpret this as a consequence of relativistic cosmic rays outrunning the solar wind that drives the magnetic disturbance. On the event side, the data contain Forbush decreases with amplitudes of 10-20%, ground-level enhancements including the 220% spike of 1989-09-29, and thunderstorm ground enhancements, including the 2023-06-14 event and three previously unreported 2022 events. The strong -0.743 correlation with F10.7 confirms solar modulation of galactic cosmic rays at this station.

Load-bearing premise

The forecasting argument assumes that an in-sample predictive power score, computed without holding out any data, measures the real ability of neutron counts to predict future Dst rather than merely describing the historical dataset.

Editorial extensions

If this is right

Because neutron flux appears to lead Dst by 7-21 hours, neutron monitor time series could be used as an input for early warnings of geomagnetic storms, giving forecasters roughly a day of lead time.
The released 42-year dataset enables other researchers to test space-weather correlations, track long-term solar modulation, and study rare events without reprocessing raw archives.
The correlation between neutron flux and Dst strengthens during Forbush decreases, so storm-time intervals are the most informative periods for training prediction models.
The three newly reported 2022 thunderstorm ground enhancements expand the catalog of TGE events at Lomnický štít and, if confirmed, provide additional samples for studying thunderstorm-driven neutron bursts.
The predictive power score of 0.22 indicates that non-linear models using neutron data may extract more storm information than linear correlations alone, motivating machine-learning-based Dst forecasting.

Reading between the lines

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

Because the predictive power score was computed by fitting and evaluating the decision tree on the full dataset, as the paper states, the 0.22 figure likely overstates genuine out-of-sample forecasting skill; a proper train/test split would be needed to know what a real forecast could achieve.
The 7-21 hour lead window is broad; conditioning the analysis on solar-wind speed or on the phase of the Forbush decrease could sharpen the delay estimate and improve any operational prediction module.
If the electric field can distort the analog electronics of the neutron monitor during thunderstorms, as the paper suspects for the 2022 TGEs, then some reported TGE-like signals might be instrumental rather than genuine neutron bursts; cross-checks with similar detectors at other mountain stations could adjudicate this.
The persistent -0.743 correlation with F10.7 over 42 years suggests this neutron monitor could serve as a long-term solar-activity proxy, potentially extending records of the solar cycle beyond the era of direct radio observations.

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.

Desk Editor's Note

read the letter

The dataset is the real contribution here. The authors have done a service by consolidating 42 years of Lomnický štít neutron monitor data into a continuous, pressure-corrected, publicly archived product with verified event signatures. That alone justifies attention from the neutron monitor and space weather communities. The descriptive correlations—especially the -0.74 with F10.7—are consistent with the established literature and are likely robust. The three 2022 TGE candidates are honestly labeled as uncertain, with the authors openly discussing the tube discrepancy and the SEVAN amplitude issue. That is the right way to report marginal events.

The soft spots are all in the statistical inference attached to the prediction narrative. The paper explicitly says it did not split train and test data for the decision tree, so the reported PPS of 0.22 is an in-sample description, not a measure of forecasting skill. The abstract and discussion nevertheless call it exciting for real-time geomagnetic storm prediction. That overreach should be fixed.

The 7–21 h delay is the bigger problem. It comes from scanning 201 one-hour shifts and picking the maximum Pearson r, which moves from 0.306 to 0.314. That difference is tiny, there is no multiple-comparison correction, and the p-values are computed as if ~365,000 hourly samples were independent. Both the Dst and neutron monitor series are strongly autocorrelated, so the effective sample size is far smaller and the 0.008 increase in r is plausibly within the null variability of lagged correlations. The same issue applies to the 0.555→0.589 improvement during Forbush decreases. The physical story about relativistic cosmic rays arriving ahead of solar plasma may be true, but this analysis does not establish it. I would not let the lead-time claim stand without a proper significance test that accounts for autocorrelation and the scan.

Minor issues: the paper says the decision tree was implemented with SciPy, but DecisionTreeRegressor is from scikit-learn; the delay is sometimes called 7 hours and sometimes 7–21 hours; and the p-values in Table 1 should be reported with effective degrees of freedom rather than a blanket p<0.01.

Who this is for: anyone working with neutron monitor data or building empirical space weather models. The dataset deserves citation and use. The paper deserves a serious referee, but it needs substantial revision before the prediction language can be taken at face value. If the authors reframe the PPS as descriptive, add a proper lag-significance test, and soften the forecasting claims, this becomes a solid data-and-catalog paper.

Referee Report

4 major / 5 minor

Summary. The manuscript introduces a pressure-corrected, gap-interpolated, continuous 42-year dataset of neutron monitor count rates from the Lomnický štít Observatory (hourly from December 1981 to July 2023; minute resolution from January 2001), made publicly available on Zenodo. The authors validate the dataset through case studies of Forbush decreases, ground-level enhancements, and newly reported thunderstorm ground enhancement candidates, and then perform Pearson correlation analyses with F10.7 and Dst, a lag-scan analysis of the neutron-Dst correlation, and a predictive power score (PPS) analysis using decision trees. The paper reports a strong anti-correlation with F10.7 (-0.743), a positive correlation with Dst (0.306 at zero lag, rising to 0.314 at a 7-21 hour delay), and a PPS of 0.22 for neutron flux as a predictor of Dst, which the abstract and discussion interpret as support for real-time geomagnetic storm prediction.

Significance. The primary contribution is the curated, openly available 42-year continuous dataset, which should be of genuine value to the neutron monitor and space weather communities. The descriptive correlations with F10.7 and Dst are broadly consistent with known physics and likely reproducible. However, the paper's central predictive narrative is not supported by the statistical evidence as presented: the PPS is explicitly in-sample, and the 7-21 hour delay is a post-hoc maximum of a lag scan on autocorrelated data without multiple-comparison correction. If the authors were to downgrade the predictive claims and properly quantify uncertainty, the dataset and descriptive results would form a useful contribution; as it stands, the overreach is a substantial issue.

major comments (4)

[Section 3.2, Eq. (3)] The predictive power score is computed by fitting a decision tree on the entire dataset and evaluating it on the same data, as the text states: 'we were using the whole data, and not separating it into train and test datasets, because we did not want to actually predict.' Consequently, the PPS of 0.22 in Table 2 measures in-sample description, not predictive skill. This does not support the abstract's mention of 'exciting possibilities for developing real-time geomagnetic storm prediction' or the Discussion's statement that cosmic ray measurements 'can be used to build forecasting models.' The authors should replace this with a proper out-of-sample evaluation (e.g., temporal cross-validation) or remove the predictive framing from the claims.
[Section 4.2, Figure 6] The 7-21 hour delay is the maximum of a lag scan over 201 one-hour shifts, selected without multiple-comparison correction. The improvement from r=0.306 to r=0.314 is small, and the hourly neutron monitor and Dst series are strongly autocorrelated on timescales of hours to days, so the effective sample size is far below the nominal N and the reported p<0.01 is not meaningful. To support the lead-time claim, the authors need to provide a null distribution (e.g., block-bootstrap or surrogate series), report confidence intervals, and correct for the multiple comparisons; otherwise the delay should be presented as exploratory rather than as a robust finding.
[Section 5, second paragraph] The physical interpretation that neutron monitor measurements provide information about an onsetting magnetic storm 'well before it could reach Earth' is based entirely on the unvalidated 7-21 hour lag-scan maximum. Even if that correlation were robust, a peak in a long-term lagged correlation does not establish that individual storm onsets are preceded by cosmic-ray changes; this would require event-by-event superposed epoch analysis. The text also conflates the full-series lag result with the Forbush-decrease subset by stating the best correlation occurs 'at times of Forbush decreases,' although the 7-21 hour interval comes from the entire dataset, not the FD subset.
[Table 1, p-value statement] The blanket statement that 'the p-values for every coefficient in Table 1 are below 0.01' is not reliable for correlations computed on autocorrelated hourly data, and it is effectively meaningless for the 2-hour TGE windows where F10.7 is constant (as the authors themselves acknowledge for the -0.533 value). The significance of each reported coefficient should be assessed with effective sample sizes or permutation methods that account for the data's temporal dependence.

minor comments (5)

[Section 3.2] The DecisionTreeRegressor() method is attributed to the SciPy package, but the ppscore library uses scikit-learn's implementation; the reference should be corrected.
[Abstract and Section 4.1] The phrase 'TGE-s' should be written as 'TGEs' for consistency.
[Table 1] The columns 'Delayed F10.7' and 'Delayed Dst' are ambiguous because the applied delays (7 h for F10.7, 7-21 h for Dst) are only explained in the text; the table footer should state the exact delays.
[Section 2] The selection of 'the 4 best response characteristic tubes' should be justified with the criterion used, since this choice affects the final count-rate time series.
[Figure 4] In the description of the 2023-06-14 event, the phrase 'the agreement of each neutron monitor counter tubes' should be reworded for grammatical correctness and clarity.

Circularity Check

2 steps flagged · score 6.0 of 10

The in-sample predictive power score and the post-hoc lag-scan maximum are presented as predictive evidence, so the forecasting claims reduce to fits on the same data.

fitted input called prediction [Section 3.2; Table 2; Abstract]
"In this case we were using the whole data, and not separating it into train and test datasets, because we did not want to actually predict, we just wanted to see how accurately it can describe the criterion (even if the tree has already seen all the values)."

The predictive power score is computed by fitting a decision tree to the full neutron monitor and Dst series and then evaluating it on the exact same data. With no train/test split, the tree has already seen every value it is asked to reproduce, so the reported nRMSE and the resulting PPS of 0.22 are in-sample goodness-of-fit measures, not out-of-sample forecast skill. The abstract nevertheless presents this score as evidence for 'developing real-time geomagnetic storm prediction models based on cosmic ray measurements.' The prediction claim is therefore a renamed description of the fitted model's training error.
fitted input called prediction [Section 4.2; Figure 6; Table 1; Section 5]
"Within this analysis, the F10.7 and Dst indices were shifted in one-hour increments across a range from -100 to +100 hours, to examine how potential time delays between the datasets, likely due to physical mechanisms in near-Earth space, might affect their correlation. This resulted in the maximum value of the Pearson coefficients for the neutron flux - Dst correlations of 0.314. These values are present at the interval of 7 to 21 hours of delay in the Dst data"

The 7-21 h lead is not an independently estimated physical quantity; it is the argmax of the correlation-versus-lag curve over 201 scanned shifts. By construction, the maximum of a scanned set is at least as large as the zero-lag value, so the reported improvement from 0.306 to 0.314 is a property of taking the maximum rather than evidence of a real lead time. The Discussion then uses this selected delay to claim that neutron monitor measurements 'provide information about an onsetting magnetic storm well before it could reach Earth,' making the forecasting narrative depend on a post-hoc fitted delay rather than on an out-of-sample or pre-registered test.

full rationale

The descriptive core of the paper—the 42-year dataset curation, the event case studies, and the raw Pearson correlations (F10.7: -0.743; Dst: 0.306)—is not circular: these numbers are computed directly from externally supplied time series, and the paper does not define the indices in terms of the correlations. The self-citations (Kudela and Langer 2009; Kudela et al. 2017; Chum et al. 2020) are calibration and instrumentation references, not load-bearing uniqueness arguments, and the 2022 TGE classification is explicitly hedged by the authors. However, two prediction-oriented quantities are circular. The predictive power score is computed by fitting a decision tree on the full dataset and evaluating it on the same data, as the paper itself admits; 0.22 is therefore an in-sample goodness-of-fit score renamed as 'prediction power.' Similarly, the 7-21 h delay is the maximum of a 201-lag scan over the same series, so the reported improvement is a property of the maximization rather than an independent estimate of a physical lead time. Because the abstract and Discussion build the real-time storm prediction narrative on these two quantities, the predictive claims reduce by construction to fits on the same data. I therefore score 6 (partial circularity), while noting that the descriptive statistics and dataset work remain self-contained.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

No invented entities are introduced. The analysis rests on public data (LMKS neutron monitor, OMNIWeb F10.7 and Dst), a barometric coefficient taken from prior literature, and three methodological choices: the data-fitted delay, the in-sample PPS evaluation, and the unverified TGE explanation.

free parameters (1)

Dst time delay (7-21 h) = 7-21 hours; 7 h used in Table 1
The delay was selected by scanning shifts from -100 to +100 hours and taking the range where the Pearson correlation with neutron flux is maximal; this is a data-fitted parameter with no independent validation.

assumptions (3)

domain assumption Sample independence for Pearson significance tests
The paper reports p<0.01 for all coefficients in Table 1 (Section 4.2) but uses hourly samples from autocorrelated time series, so effective degrees of freedom are far lower than N.
ad hoc to paper In-sample decision-tree fit measures predictive power
Section 3.2 states the tree was fitted and evaluated on the whole dataset with no train/test split, explicitly 'we did not want to actually predict'; PPS therefore measures description, not forecasting.
ad hoc to paper Thunderstorm electric fields can affect the NM analog electronics
The identification of the June 2022 events as TGEs relies on this unverified hypothesis, acknowledged in Section 5.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Analysis of 42 years of Cosmic Ray Measurements by the Neutron Monitor at Lomnick\'y st\'it Observatory." pith.science (2026). https://pith.science/paper/LT5KASXB

@misc{pith2026250209627,
  author       = {Pith},
  title        = {Pith review of: Analysis of 42 years of Cosmic Ray Measurements by the Neutron Monitor at Lomnick\'y st\'it Observatory},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LT5KASXB}},
  note         = {Machine review of arXiv:2502.09627}
}

read the original abstract

The correlation and physical interconnection between space weather indices and cosmic ray flux has been well-established with extensive literature on the topic. Our investigation is centered on the relationships among the solar radio flux, geomagnetic field activity, and cosmic ray flux, as observed by the Neutron Monitor at the Lomnick\'y \v{s}t\'it Observatory in Slovakia. We processed the raw neutron monitor data, generating the first publicly accessible dataset spanning 42 years. The curated continuous data are available in .csv format in hourly resolution from December 1981 to July 2023 and in minute resolution from January 2001 to July 2023 (Institute of Experimental Physics SAS, 2024). Validation of this processed data was accomplished by identifying distinctive events within the dataset. As part of the selection of events for case studies, we report the discovery of TGE-s visible in the data. Applying the Pearson method for statistical analysis, we quantified the linear correlation of the datasets. Additionally, a prediction power score was computed to reveal potential non-linear relationships. Our findings demonstrate a significant anti-correlation between cosmic ray and solar radio flux with a correlation coefficient of -0.74, coupled with a positive correlation concerning geomagnetic field strength. We also found that the neutron monitor measurements correlate better with a delay of 7-21 hours applied to the geomagnetic field strength data. The correlation between these datasets is further improved when inspecting periods of extreme solar events only. Lastly, the computed prediction power score of 0.22 for neutron flux in the context of geomagnetic field strength presents exciting possibilities for developing real-time geomagnetic storm prediction models based on cosmic ray measurements.

Figures

Figures reproduced from arXiv: 2502.09627 by the authors.

**Figure 1.** Top panel: The continuous time series data measured by the neutron monitor at Lomnick´y ˇst´ıt Observatory. Middle panel: The F10.7 solar index. Bottom panel: The Dst index. The least squares polynomial fit created using the polyfit() method from the NumPy library (Harris et al., 2020) is plotted for each presented index to highlight longer periodic variability. Histograms of all measured values can be seen on the r… view at source ↗

**Figure 2.** The selected Forbush Decreases (FDs) with amplitudes ranging from 10 to 20%. The timescale is exactly 7 days on each panel and the scale of the y-axes is also uniform. –11– [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗

**Figure 3.** The selected Ground Level Enhancements (GLEs). The time scale for each panel is 22 hours with a resolution of 1 hour. Note the different scaling on the y-axes [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: The selected Thunderstorm Ground Enhancement (TGE) events. They are displayed on a 2-hour long time interval. The graphs are produced by minute-resolution data. The scaling of the vertical axis in the plots differs, as the first event exhibited value changes of consid…

**Figure 5.** Figure 5: Presumably, 3 recent TGEs reported for the first time. The data shown here is displayed on a 2-hour interval. Note the unusually small time difference between the events. –12– [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]

**Figure 6.** Figure 6: The graph demonstrates the change of the Pearson coefficient resulting from the correlation testing of the neutron monitor and the Dst datasets by the amount of time-shift applied to the Dst data. The application of a timeshift of positive value will result in the Dst…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

12 extracted references · 9 canonical work pages

[1]

P., Bochkov, E

Babich, L. P., Bochkov, E. I., Kutsyk, I. M., & Rassoul, H. K. (2014). Analysis of fundamental interactions capable of producing neutrons in thunderstorms. Physical Review D, 89 (9), 093010. Battarbee, M., Guo, J., Dalla, S., Wimmer-Schweingruber, R., Swalwell, B., & Lawrence, D. (2018). Multi-spacecraft observations and transport simula- tions of solar e...

work page 2014
[12]

Shah, G., Razdan, H., Bhat, C., & Ali, Q

Space Weather, 8 (5). Shah, G., Razdan, H., Bhat, C., & Ali, Q. (1985). Neutron generation in lightning bolts. Nature, 313 (6005), 773–775. ˇSlegl, J., Langer, R., Bruncl ´ ık, T., Maˇ sek, P., Strh´ arsk` y, I., Ambroˇ zov´ a, I., . . . Ploc, O. (2022). Spectrometry of high-energy photons on high mountain observa- tory lomnick´ y ˇ st ´ ıt during thunder...

1985
[24]

doi: 10.1007/s11207-022-01951-4 Lockwood, J. A. (1971). Forbush decreases in the cosmic radiation. Space Science Reviews, 12 (5), 658–715. Mandrikova, O., & Mandrikova, B. (2021). Method of wavelet-decomposition to re- search cosmic ray variations: Application in space weather. Symmetry, 13 (12),

work page doi:10.1007/s11207-022-01951-4 1971
[36]

Chilingarian, A., Babayan, V., Karapetyan, T., Mailyan, B., Sargsyan, B., & Za- zyan, M. (2018). The sevan worldwide network of particle detectors: 10 years of operation. Advances in Space Research, 61 (10), 2680–2696. Chilingarian, A., Daryan, A., Arakelyan, K., Hovhannisyan, A., Mailyan, B., Melkumyan, L., . . . Vanyan, L. (2010). Ground-based observati...

work page 2018
[70]

Okike, O. (2020). Forbush decreases: Algorithm generated dataset. Data in Brief , 33 , 106463. Retrieved from https://www.sciencedirect.com/ science/article/pii/S2352340920313457 doi: https://doi.org/10.1016/ j.dib.2020.106463 Panasyuk, M. I. (2001). Cosmic ray and radiation belt hazards for space missions: An introductory overview. Space Storms and Space...

work page arXiv 2020
[135]

–23– manuscript submitted to Earth and Space Science Mewaldt, R. A. (1996). Cosmic rays. Macmillian Encyclopedia Of Physics, 1 . Mueller, D., Marsden, R. G., St. Cyr, O., Gilbert, H. R., & Team, S. O. (2013). So- lar orbiter: exploring the sun–heliosphere connection. Solar Physics, 285 , 25–

work page 1996
[284]

Pandas, d. t. (2020, February). pandas-dev/pandas: Pandas. Zenodo. Retrieved from https://doi.org/10.5281/zenodo.3509134 doi: 10.5281/zenodo.3509134 Papitashvili, N. E., & King, J. H. (2020). Omni hourly data set[Dataset]. NASA Space Physics Data Facility. Retrieved from https://hpde.io/NASA/ NumericalData/OMNI/PT1H doi: 10.48322/1SHR-HT18 Parker, E. (195...

doi:10.5281/zenodo.3509134 2020
[438]

doi: 10.1016/j.newast.2011.03.003 Chum, J., Langer, R., Baˇ se, J., Koll´ arik, M., Strh´ arsk` y, I., Diendorfer, G., & Rusz, J. (2020). Significant enhancements of secondary cosmic rays and electric field at the high mountain peak of lomnick` y ˇ st ´ ıt in high tatras during thunder- storms. Earth, Planets and Space, 72 , 1–20. Clem, J. M., & Dorman, L...

work page doi:10.1016/j.newast.2011.03.003 2020

Show all 12 references

[1108]

Gupta, V., & Badruddin. (2010). High-speed solar wind streams during 1996–2007: sources, statistical distribution, and plasma/field properties. Solar Physics, –21– manuscript submitted to Earth and Space Science 264 , 165–188. Gurevich, A., Milikh, G., & Roussel-Dupre, R. (199...

2010
[1328]

Schober, P., Boer, C., & Schwarte, L. A. (2018). Correlation coefficients: appropriate use and interpretation. Anesthesia & analgesia, 126 (5), 1763–1768. Schwadron, N. A., Boyd, A., Kozarev, K., Golightly, M., Spence, H., Townsend, L., & Owens, M. (2010). Galactic cosmic ray ...

2018
[2005]

Mavromichalaki, H., Gerontidou, M., Paschalis, P., Papaioannou, A., Paouris, E., Papailiou, M., & Souvatzoglou, G

IEEE Transactions on nuclear science, 52 (6), 2307–2312. Mavromichalaki, H., Gerontidou, M., Paschalis, P., Papaioannou, A., Paouris, E., Papailiou, M., & Souvatzoglou, G. (2015). Recent research applications at the athens neutron monitor station. In Journal of physics: Confer...

2015
[2313]

others (2005)

Mavromichalaki, H., Gerontidou, M., Mariatos, G., Plainaki, C., Papaioannou, A., Sarlanis, C., . . . others (2005). Space weather forecasting at the new athens center: The recent extreme events of january

2005

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.

[1] [1]

P., Bochkov, E

Babich, L. P., Bochkov, E. I., Kutsyk, I. M., & Rassoul, H. K. (2014). Analysis of fundamental interactions capable of producing neutrons in thunderstorms. Physical Review D, 89 (9), 093010. Battarbee, M., Guo, J., Dalla, S., Wimmer-Schweingruber, R., Swalwell, B., & Lawrence, D. (2018). Multi-spacecraft observations and transport simula- tions of solar e...

work page 2014

[2] [12]

Shah, G., Razdan, H., Bhat, C., & Ali, Q

Space Weather, 8 (5). Shah, G., Razdan, H., Bhat, C., & Ali, Q. (1985). Neutron generation in lightning bolts. Nature, 313 (6005), 773–775. ˇSlegl, J., Langer, R., Bruncl ´ ık, T., Maˇ sek, P., Strh´ arsk` y, I., Ambroˇ zov´ a, I., . . . Ploc, O. (2022). Spectrometry of high-energy photons on high mountain observa- tory lomnick´ y ˇ st ´ ıt during thunder...

1985

[3] [24]

doi: 10.1007/s11207-022-01951-4 Lockwood, J. A. (1971). Forbush decreases in the cosmic radiation. Space Science Reviews, 12 (5), 658–715. Mandrikova, O., & Mandrikova, B. (2021). Method of wavelet-decomposition to re- search cosmic ray variations: Application in space weather. Symmetry, 13 (12),

work page doi:10.1007/s11207-022-01951-4 1971

[4] [36]

Chilingarian, A., Babayan, V., Karapetyan, T., Mailyan, B., Sargsyan, B., & Za- zyan, M. (2018). The sevan worldwide network of particle detectors: 10 years of operation. Advances in Space Research, 61 (10), 2680–2696. Chilingarian, A., Daryan, A., Arakelyan, K., Hovhannisyan, A., Mailyan, B., Melkumyan, L., . . . Vanyan, L. (2010). Ground-based observati...

work page 2018

[5] [70]

Okike, O. (2020). Forbush decreases: Algorithm generated dataset. Data in Brief , 33 , 106463. Retrieved from https://www.sciencedirect.com/ science/article/pii/S2352340920313457 doi: https://doi.org/10.1016/ j.dib.2020.106463 Panasyuk, M. I. (2001). Cosmic ray and radiation belt hazards for space missions: An introductory overview. Space Storms and Space...

work page arXiv 2020

[6] [135]

–23– manuscript submitted to Earth and Space Science Mewaldt, R. A. (1996). Cosmic rays. Macmillian Encyclopedia Of Physics, 1 . Mueller, D., Marsden, R. G., St. Cyr, O., Gilbert, H. R., & Team, S. O. (2013). So- lar orbiter: exploring the sun–heliosphere connection. Solar Physics, 285 , 25–

work page 1996

[7] [284]

Pandas, d. t. (2020, February). pandas-dev/pandas: Pandas. Zenodo. Retrieved from https://doi.org/10.5281/zenodo.3509134 doi: 10.5281/zenodo.3509134 Papitashvili, N. E., & King, J. H. (2020). Omni hourly data set[Dataset]. NASA Space Physics Data Facility. Retrieved from https://hpde.io/NASA/ NumericalData/OMNI/PT1H doi: 10.48322/1SHR-HT18 Parker, E. (195...

doi:10.5281/zenodo.3509134 2020

[8] [438]

doi: 10.1016/j.newast.2011.03.003 Chum, J., Langer, R., Baˇ se, J., Koll´ arik, M., Strh´ arsk` y, I., Diendorfer, G., & Rusz, J. (2020). Significant enhancements of secondary cosmic rays and electric field at the high mountain peak of lomnick` y ˇ st ´ ıt in high tatras during thunder- storms. Earth, Planets and Space, 72 , 1–20. Clem, J. M., & Dorman, L...

work page doi:10.1016/j.newast.2011.03.003 2020