The effect of uncertainties in reproducing the ambient solar wind at Earth on forecasting CME arrival times

Nikolai Pogorelov; Syed Raza; Talwinder Singh

REVIEW 4 major objections 4 minor 55 references

The effect of uncertainties in reproducing the ambient solar wind at Earth on forecasting CME arrival times

T0 review · 4 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read Training machine-learning models on the difference between observed and modeled solar wind at the L1 point, just upstream of each CME, cuts mean arrival-time forecast error by about 45 minutes.

desk verdict A plausible, modest ML-correction study whose headline numbers are not yet trustable because the abstract reports no validation protocol or error bars; worth sending to referees, but verify out-of-sample evaluation. read the letter →

arxiv 2508.11217 v1 pith:5C36TX7W submitted 2025-08-15 physics.space-ph

Syed Raza , Talwinder Singh , Nikolai Pogorelov This is my paper · ORCID

classification physics.space-ph

keywords coronalmassejectionCMEtimeofarrivalspaceweatherforecastingmachinelearningcorrectionL1solarwindWSA-ENLIL-ConemeanabsoluteerrorDONKI

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

This paper is trying to establish that the gap between what a physics-based heliospheric model predicts for the solar wind at L1 and what spacecraft actually measure there contains a usable correction signal for CME arrival-time forecasts. Using 122 CME events from 2012 to 2023, the authors train three simple machine-learning models on those measured-versus-modeled discrepancies and find that the best single-feature model reduces mean absolute error by 36.6 minutes and the best multi-feature model by 45 minutes. If the relationship holds beyond the training record, forecast centers could post-process existing physics-based predictions with a cheap ML correction instead of waiting for a more accurate heliospheric model.

What carries the argument

The discrepancy features: per-event differences between observed L1 solar-wind measurements and WEC model outputs, taken just upstream of each CME. These features are fed into three regressors (KNN, SVM, and linear regression) whose target is the model's arrival-time error for that event. The load-bearing mechanism is the learned mapping from ambient-wind discrepancy to transit-time error; the univariate-versus-multivariate comparison tests whether any single solar-wind quantity carries the correction signal alone.

What would settle it

Hold out all CMEs after a fixed date (for example, 2020), train the ML models only on earlier events, and compare corrected versus uncorrected MAE on the held-out set. The central claim would be contradicted if the corrected MAE is not below the uncorrected MAE on that set, and weakened if the 45-minute gain is not reproducible out of sample.

Watch

Extended reading notes

Core claim

The central claim is that ambient solar-wind reproduction errors at L1, sampled immediately upstream of an Earth-directed CME, carry systematic information about that CME's time-of-arrival error, and that simple machine-learning regressors can extract it. The authors demonstrate this with k-nearest neighbors, support vector machines, and linear regression applied to 122 events, comparing univariate and multivariate feature sets. Their headline results are an MAE improvement of 36.6 minutes for the best univariate model and 45 minutes for the best multivariate model, both measured against the uncorrected WEC-based forecast error.

Load-bearing premise

The learned correction only works if the mismatch between modeled and measured solar wind at L1, sampled just before a CME arrives, is a stable proxy for that CME's own arrival-time error; if that mismatch is mostly unrelated noise, the correction collapses.

Editorial extensions

If this is right

Operational WEC-based forecasts could be post-processed with a small ML model, adding no new physics and only minutes of computation.
Combining multiple discrepancy features beats any single feature, so the correction signal is distributed across solar-wind quantities rather than concentrated in one.
An error reduction of about 45 minutes is modest against the >10-hour baseline MAE but comes essentially for free on top of the existing forecast pipeline.
The same archived forecast-error record can be reused to test whether newer heliospheric models share the same discrepancy-to-error relationship.

Reading between the lines

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

Because the discrepancy is measured at L1 rather than at the CME itself, the learned correction may transfer to other cone-model forecasts whose ambient-solar-wind errors resemble WEC's; testing that directly would show how event-specific the correction is.
The reported gains are averages, and if arrival-time errors are heteroscedastic, the same correction could give larger absolute improvements for the high-error events that matter most for warnings.
A natural stress test is to train on 2012–2023 events and evaluate on Solar Cycle 25 events through 2025, checking whether the discrepancy-to-error mapping drifts with solar activity.

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

Quick take: this is a modest, believable operational-forecast study, and the abstract alone is not enough to trust the headline numbers. The claimed result—trained KNN/SVM/linear-regression corrections using WEC-vs-observed solar-wind discrepancies at L1, cutting CME arrival-time MAE by 36.6–45 minutes on 122 events—is concrete and, as far as I can tell, new. That kind of model-output-stats correction is standard in other forecasting fields, so the approach is sensible and the experiment is worth doing. Credit where due: the dataset is public (DONKI), the models are simple and comparable, and the authors are not overclaiming; they put "correct" in scare quotes, which is honest about the statistical nature of the correction.

Soft spots, stated in proportion: the abstract gives no baseline MAE, no error bars, no significance tests, and no description of train/test splitting or cross-validation. With 122 events spread over eleven years and univariate/multivariate feature selection, the 36.6- and 45-minute gains could easily come from in-sample fitting or from a few influential CMEs. The stress-test note is right on this—the central claim depends on out-of-sample evaluation, and the abstract does not establish it. The physical premise, that L1 discrepancies carry learnable information about downstream CME arrival error, is plausible but not argued; it could be forecast noise. I would not call any of these fatal from the abstract, but they make the headline numbers unverifiable as presented.

I could only read the abstract; the full text came through as mojibake in the provided rendering, so my assessment is limited. The structure is coherent and the framing is honest. For whom is this paper? Space-weather forecasters who run WEC operationally and want cheap correction layers—they should care. The broader CME physics community will find it incremental.

Recommendation: send it to peer review, but the referee should insist on seeing the validation protocol and error bars. If the gains survive a temporal or blocked out-of-sample split, this is a useful, citable result. If not, it's an in-sample artifact. The paper deserves a serious referee to make that call.

Referee Report

4 major / 4 minor

Summary. The paper proposes a post-processing correction for CME arrival-time forecasts by training machine-learning models (k-nearest neighbors, support vector machine, linear regression) on discrepancies between observed and WSA-ENLIL-Cone modeled solar wind at L1. Using 122 CME events from the NASA CCMC DONKI database between March 2012 and March 2023, the authors report that their best univariate and multivariate models reduce the mean absolute error of CME time-of-arrival forecasts by 36.6 and 45 minutes, respectively. The abstract does not describe the validation protocol, baseline MAE, or uncertainty estimates, and the full text is unreadable in the provided rendering, so the central claim cannot be independently assessed.

Significance. If the reported improvement is out-of-sample and statistically robust, the result is practically valuable: it would demonstrate that systematic L1 solar-wind model discrepancies carry portable, learnable information for correcting WEC-based CME arrival predictions. The use of public DONKI data and three standard ML models makes the approach easy to reproduce and extend. However, the significance is conditional on a missing evaluation protocol; without explicit out-of-sample testing, the headline numbers could be in-sample artifacts and would not support the paper's conclusion.

major comments (4)

[Abstract] The central claim, 'The best univariate and multivariate models improved the forecast by 36.6 and 45 minutes, respectively,' is reported without any statement of whether these improvements are out-of-sample. The abstract also does not give the baseline MAE, the train/test split (e.g., temporal, random, or leave-one-out), or whether hyperparameter tuning and feature selection were performed on the same 122 events. With 122 events and three models, in-sample evaluation or selection overfitting could easily produce gains of this magnitude. Specify the evaluation protocol and provide error bars or a significance test, or the headline improvement is not interpretable.
[Full text] The submitted manuscript body is garbled/mojibake in the provided version; only the abstract is readable. Consequently, the definitions of the discrepancy features, the sampling window upstream of each CME, the model hyperparameters, and the feature-selection procedure cannot be verified. For the review to proceed, a clean, readable version of the full text is required. This is not a scientific criticism but a necessary condition for assessing the paper's methodology.
[Abstract / Methods (unreadable)] The abstract states that features are discrepancies between observed and modeled solar wind 'at the L1 point, upstream of CMEs.' It is not specified when these L1 observations are available relative to the CME forecast issue time and the CME's actual arrival. If the L1 data used as features become available only after the CME has passed L1, then the proposed correction is not a forecast correction but an ex-post diagnosis. Clarify the timing and ensure that the feature set would be available in an operational setting.
[Dataset / Feature selection] The dataset comprises 122 events over 11 years, and the study compares three models and univariate/multivariate feature combinations. This involves many multiple comparisons. The reported 45-minute multivariate gain over the best univariate model should be accompanied by a test of whether the difference is statistically meaningful (e.g., paired bootstrap, Diebold-Mariano test) and by controlling for feature-selection bias, for example through nested cross-validation. Without such analysis, the multivariate improvement may reflect selection overfitting rather than a portable signal.

minor comments (4)

[Abstract] The abstract should state the baseline MAE (the '>10 hours' value) explicitly so the reader can see the relative improvement. It should also specify the units of '36.6 and 45 minutes' and report uncertainties on these values.
[Abstract] The term 'discrepancies' is vague. Please specify which solar-wind parameters (density, speed, magnetic field?) and what time window relative to the CME are used to construct the features.
[References] Please cite the relevant WSA-ENLIL-Cone model description and the DONKI/CCMC database documentation, and describe the version of WEC used for the archived forecasts.
[Full text] The rendered equations and tables are illegible in the provided copy; please ensure the final submission has clean typesetting for all equations and table entries.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the ML correction uses L1 SW discrepancy features that are distinct from the TOA-error target; missing validation details are a correctness concern, not circularity.

full rationale

The paper's central derivation is an empirical ML correction: the features are discrepancies between observed and WEC-modeled solar wind at L1, and the target is the WEC CME time-of-arrival error. These are not the same quantity by construction; the TOA error comes from DONKI/WEC arrival-time outputs, while the discrepancy features come from comparing L1 SW observations with the model's SW output. No equation in the available text defines the features in terms of the target or vice versa, and no self-citation chain, uniqueness theorem, or ansatz-importing citation is visible. The abstract's scare quotes around 'correct' signal that the correction is approximate, which is a caveat, not circularity. The main limitation is that the abstract does not describe the evaluation protocol; if the 36.6- and 45-minute improvements were computed on the training set, they would be overfitting artifacts, but that would be a statistical validity problem, not a circular reduction of the prediction to its inputs. Since no in-sample protocol is stated and the available text does not exhibit a specific reduction, the appropriate finding is no significant circularity.

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

The paper's contribution is a statistical correction layer on top of public physics-model outputs; the ledger shows what that layer depends on. Four modeling elements are fitted or hand-selected (hyperparameters, regression coefficients, feature combinations); five domain assumptions carry the transferability of the result; no invented physical entities appear. The heaviest unexamined load is the assumption that L1 ambient solar wind discrepancies causally or statistically drive CME arrival-time error across solar cycle conditions.

free parameters (4)

KNN neighborhood size k
Selected to minimize MAE on the training events; value not disclosed in the abstract.
SVM hyperparameters (kernel, C, regularization)
Tuned on the 122-event dataset; values not disclosed in the abstract.
Linear regression coefficients
Fitted to map L1 solar wind discrepancy features to WEC arrival-time error on the historical events.
Feature set and feature combinations
Choice of which L1 solar wind properties enter the univariate and multivariate schemes is a modeling choice; the 'best' combination is selected by comparing outcomes, so it is a hand-fitted element of the method.

assumptions (5)

domain assumption Archived WSA-ENLIL-Cone (WEC) runs in DONKI faithfully represent the operational model state for every event (same model version, cone parameters, ambient input).
The entire study treats DONKI's archived WEC outputs as the model prediction to be corrected; heterogeneous archived runs would make the discrepancy features incomparable across the 122 events. Invoked by the abstract's use of DONKI WEC inputs and outputs.
domain assumption The discrepancy between modeled and observed solar wind at L1, sampled upstream of each CME, captures a learnable share of the CME arrival-time error.
This is the methodological premise of the whole study, stated in the abstract as training on 'discrepancies between observed and modeled solar wind at the L1 point, upstream of CMEs'; if the dominant error sources are internal to the CME (cone geometry, speed, drag), the L1 ambient signal will not carry the needed information.
domain assumption The learned discrepancy-to-error mapping is stationary across the 2012-2023 study window and transferable to future CMEs.
The training window spans solar cycle 24 maximum into cycle 25; the abstract offers no argument that the statistical relationship survives regime changes in solar wind structure.
standard math Textbook ML performance assumptions hold for KNN, SVM, and linear regression on this small tabular sample.
The paper claims no new mathematical machinery; the corrections rely on standard supervised learning assumptions, including exchangeable training samples and stable distributions.
domain assumption The 122-event sample is large enough for stable estimation of the correction mapping.
With many candidate features and three model families, 122 events is a small sample; the abstract gives no confidence intervals to confirm stability.

how reviews work

0 comments

Cite this review

Pith. "Pith review of The effect of uncertainties in reproducing the ambient solar wind at Earth on forecasting CME arrival times." pith.science (2026). https://pith.science/paper/5C36TX7W

@misc{pith2026250811217,
  author       = {Pith},
  title        = {Pith review of: The effect of uncertainties in reproducing the ambient solar wind at Earth on forecasting CME arrival times},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5C36TX7W}},
  note         = {Machine review of arXiv:2508.11217}
}

read the original abstract

Coronal Mass Ejections (CMEs) are the major drivers of Space Weather (SWx), so predicting their arrival at Earth is a major aspect of SWx forecasting. Despite increasingly complex models proposed over the past decades, the mean absolute error (MAE) for predictions of CME arrival still surpasses 10 hours. In this study, we use machine learning (ML) techniques trained on the discrepancies between observed and modeled solar wind (SW) at the L1 point, upstream of CMEs, to quantify and ''correct'' the errors in CME Time of Arrival (TOA) associated with these discrepancies. We use CME data from the Database Of Notifications, Knowledge, Information (DONKI) developed by the NASA Community Coordinated Modeling Center (CCMC) for our investigation. The WSA-ENLIL-Cone (WEC) model inputs and outputs are available on DONKI for each CME, along with the associated forecast errors. The dataset consists of 122 CME events observed between March 2012 and March 2023. SW properties at L1 and publicly available simulation results based on the WEC model are obtained. Three machine learning (ML) models are employed: 1) k-nearest neighbors (KNN), 2) support vector machine (SVM), and 3) linear regression (LR). Univariate and multivariate ML schemes were developed to examine how individual features and their combinations contribute to reducing the MAE in CME TOA forecasts. The best univariate and multivariate models improved the forecast by 36.6 and 45 minutes, respectively.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 33 canonical work pages

[1]

, Abduallah , Y

Alobaid2022 APACrefauthors Alobaid , K A. , Abduallah , Y. , Wang , J T L. , Wang , H. , Jiang , H. , Xu , Y. Jing , J. APACrefauthors \ 2022 10 . Predicting CME arrival time through data integration and ensemble learning Predicting CME arrival time through data integration and ensemble learning . Frontiers in Astronomy and Space Sciences 9 . APACrefDOI d...

work page arXiv 2022
[2]

, Luhmann, J

Arge04 APACrefauthors Arge, C. , Luhmann, J. , Odstrcil, D. , Schrijver, C. \ Li, Y. APACrefauthors \ 2004 . Stream structure and coronal sources of the solar wind during the May 12th, 1997 CME Stream structure and coronal sources of the solar wind during the may 12th, 1997 cme . Journal of Atmospheric and Solar-Terrestrial Physics 66 15 1295 - 1309 . APA...

2004
[3]

\ Pizzo, V J

Arge00 APACrefauthors Arge, C N. \ Pizzo, V J. APACrefauthors \ 2000 . Improvement in the prediction of solar wind conditions using near-real time solar magnetic field updates Improvement in the prediction of solar wind conditions using near-real time solar magnetic field updates . Journal of Geophysical Research: Space Physics 105 A5 10465-10479 . APACre...

doi:10.1029/1999ja000262 2000
[4]

, Delaboudiniere , J P

Brueckner1998 APACrefauthors Brueckner , G E. , Delaboudiniere , J P. , Howard , R A. , Paswaters , S E. , St. Cyr , O C. , Schwenn , R. Wang , D. APACrefauthors \ 1998 Jan . Geomagnetic storms caused by coronal mass ejections (CMEs): March 1996 through June 1997 Geomagnetic storms caused by coronal mass ejections (CMEs): March 1996 through June 1997 . Ge...

work page doi:10.1029/98gl00704 1998
[5]

APACrefauthors \ 2011 Apr 08

Chen11 APACrefauthors Chen, P F. APACrefauthors \ 2011 Apr 08 . Coronal Mass Ejections: Models and Their Observational Basis Coronal mass ejections: Models and their observational basis . Living Reviews in Solar Physics 8 1 1 . APACrefURL https://doi.org/10.12942/lrsp-2011-1 APACrefURL APACrefDOI doi:10.12942/lrsp-2011-1 APACrefDOI

doi:10.12942/lrsp-2011-1 2011
[6]

, Shen , C

Chi2016 APACrefauthors Chi , Y. , Shen , C. , Wang , Y. , Xu , M. , Ye , P. \ Wang , S. APACrefauthors \ 2016 09 . Statistical Study of the Interplanetary Coronal Mass Ejections from 1995 to 2015 Statistical Study of the Interplanetary Coronal Mass Ejections from 1995 to 2015 . Solar Physics 291 8 2419-2439 . APACrefDOI doi:10.1007/s11207-016-0971-5 APACrefDOI

doi:10.1007/s11207-016-0971-5 2016
[7]

, Liu , J

Chierichini2024 APACrefauthors Chierichini , S. , Liu , J. , Korsós , M B. , Del Moro , D. \ Erdélyi , R. APACrefauthors \ 2024 03 . CME Arrival Modeling with Machine Learning CME Arrival Modeling with Machine Learning . The Astrophysical Journal 963 1 121 . APACrefDOI doi:10.3847/1538-4357/ad1cee APACrefDOI

doi:10.3847/1538-4357/ad1cee 2024
[8]

\ Smart, D

Dryer1984 APACrefauthors Dryer, M. \ Smart, D. APACrefauthors \ 1984 . Dynamical models of coronal transients and interplanetary disturbances Dynamical models of coronal transients and interplanetary disturbances . Advances in Space Research 4 7 291 - 301 . APACrefURL http://www.sciencedirect.com/science/article/pii/027311778490200X APACrefURL APACrefDOI ...

work page arXiv 1984

Show all 55 references

[9]

, Acebal , A

Emmons2013 APACrefauthors Emmons , D. , Acebal , A. , Pulkkinen , A. , Taktakishvili , A. , MacNeice , P. \ Odstrcil , D. APACrefauthors \ 2013 . Ensemble forecasting of coronal mass ejections using the WSA‐ENLIL with CONED Model Ensemble forecasting of coronal mass ejections ...

2013 doi
[10]

, Zheng , Y

Fu2021 APACrefauthors Fu , H. , Zheng , Y. , Ye , Y. , Feng , X. , Liu , C. \ Ma , H. APACrefauthors \ 2021 04 . Joint Geoeffectiveness and Arrival Time Prediction of CMEs by a Unified Deep Learning Framework Joint Geoeffectiveness and Arrival Time Prediction of CMEs by a Unif...

2021 doi
[11]

, Lara , A

Gopalswamy05 APACrefauthors Gopalswamy , N. , Lara , A. , Manoharan , P K. \ Howard , R A. APACrefauthors \ 2005 01 . An empirical model to predict the 1-AU arrival of interplanetary shocks An empirical model to predict the 1-AU arrival of interplanetary shocks . Advances in S...

2005 doi
[12]

, Lara , A

Gopalswamy01 APACrefauthors Gopalswamy , N. , Lara , A. , Yashiro , S. , Kaiser , M L. \ Howard , R A. APACrefauthors \ 2001 12 . Predicting the 1-AU arrival times of coronal mass ejections Predicting the 1-AU arrival times of coronal mass ejections . Journal of Geophysical Re...

2001 doi
[13]

, Shen , C

Gui2011 APACrefauthors Gui , B. , Shen , C. , Wang , Y. , Ye , P. , Liu , J. , Wang , S. \ Zhao , X. APACrefauthors \ 2011 06 . Quantitative Analysis of CME Deflections in the Corona Quantitative Analysis of CME Deflections in the Corona . Solar Physics 271 1-2 111-139 . APACr...

2011 doi
[14]

, Barnard , L A

Jones2017 APACrefauthors Jones , S R. , Barnard , L A. , Scott , C J. , Owens , M J. \ Wilkinson , J. APACrefauthors \ 2017 09 . Tracking CMEs using data from the Solar Stormwatch project: observing deflections and other properties Tracking CMEs using data from the Solar Storm...

2017 doi
[15]

, Opher , M

Kay15 APACrefauthors Kay , C. , Opher , M. \ Evans , R M. APACrefauthors \ 2015 06 . Global Trends of CME Deflections Based on CME and Solar Parameters Global Trends of CME Deflections Based on CME and Solar Parameters . The Astrophysical Journal 805 2 168 . APACrefDOI doi:10....

2015 doi
[16]

, Palmerio , E

Kay24 APACrefauthors Kay , C. , Palmerio , E. , Riley , P. , Mays , M L. , Nieves‐Chinchilla , T. , Romano , M. Chulaki , A. APACrefauthors \ 2024 07 . Updating Measures of CME Arrival Time Errors Updating Measures of CME Arrival Time Errors . Space Weather 22 e2024SW003951 . ...

2024 doi
[17]

Liu2018CATPUMA APACrefauthors Liu, J. , Ye, Y. , Shen, C. , Wang, Y. \ Erdélyi, R. APACrefauthors \ 2018 . A New Tool for CME Arrival Time Prediction using Machine Learning Algorithms: CAT-PUMA A new tool for cme arrival time prediction using machine learning algorithms: Cat-p...

2018 doi
[18]

, Kuznetsova , M M

Maddox2014 APACrefauthors Maddox , M M. , Kuznetsova , M M. , Pulkkinen , A A. , Zheng , Y. , Rastaetter , L. , Chulaki , A. MacNeice , P J. APACrefauthors \ 2014 . The NASA Community Coordinated Modeling Center (CCMC) Next Generation Space Weather Data Warehouse The NASA Comm...

2014
[19]

, Gopalswamy , N

Manoharan2004 APACrefauthors Manoharan , P K. , Gopalswamy , N. , Yashiro , S. , Lara , A. , Michalek , G. \ Howard , R A. APACrefauthors \ 2004 Jun . Influence of coronal mass ejection interaction on propagation of interplanetary shocks Influence of coronal mass ejection inte...

2004 doi
[20]

, MacNeice, P J

Mays2020 APACrefauthors Mays, M L. , MacNeice, P J. , Taktakishvili, A. , Wiegand, C P. , Merka, J. , Adamson, E T. Wallace, S. APACrefauthors \ 2020 May . NASA/NOAA MOU ANNEX FINAL REPORT: EVALUATING MODEL ADVANCEMENTS FOR PREDICTING CME ARRIVAL TIME Nasa/noaa mou annex final...

2020
[21]

, Taktakishvili , A

Mays15 APACrefauthors Mays , M L. , Taktakishvili , A. , Pulkkinen , A. , MacNeice , P J. , Rast \"a tter , L. , Odstrcil , D. Kuznetsova , M M. APACrefauthors \ 2015 06 . Ensemble Modeling of CMEs Using the WSA-ENLIL+Cone Model Ensemble Modeling of CMEs Using the WSA-ENLIL+Co...

2015 doi
[22]

, Biesecker, D

Millward13 APACrefauthors Millward, G. , Biesecker, D. , Pizzo, V. \ de Koning, C A. APACrefauthors \ 2013 . An operational software tool for the analysis of coronagraph images: Determining CME parameters for input into the WSA-Enlil heliospheric model An operational software ...

2013 doi
[23]

, Wang , Y

Mishra2016 APACrefauthors Mishra , W. , Wang , Y. \ Srivastava , N. APACrefauthors \ 2016 11 . ON UNDERSTANDING THE NATURE OF COLLISIONS OF CORONAL MASS EJECTIONS OBSERVED BY STEREO ON UNDERSTANDING THE NATURE OF COLLISIONS OF CORONAL MASS EJECTIONS OBSERVED BY STEREO . The As...

2016 doi
[24]

, Dryer , M

Moon2002 APACrefauthors Moon , Y J. , Dryer , M. , Smith , Z. , Park , Y D. \ Cho , K S. APACrefauthors \ 2002 May . A revised shock time of arrival (STOA) model for interplanetary shock propagation: STOA-2 A revised shock time of arrival (STOA) model for interplanetary shock ...

2002 doi
[25]

, Brito, T V

morley2018measures APACrefauthors Morley, S K. , Brito, T V. \ Welling, D T. APACrefauthors \ 2018 . Measures of Model Performance Based On the Log Accuracy Ratio Measures of model performance based on the log accuracy ratio . Space Weather 16 1 69--88 . APACrefURL https://doi...

2018 doi
[26]

, Linker , J A

Odstrcil2003 APACrefauthors Odstrcil , D. , Linker , J A. , Lionello , R. , Mikic , Z. , Riley , P. , Pizzo , V J. \ Luhmann , J G. APACrefauthors \ 2003 . Merging of coronal and heliospheric numerical two-dimensional MHD models Merging of coronal and heliospheric numerical tw...

2003 doi
[27]

\ Pizzo , V J

Odstrcil99b APACrefauthors Odstrcil , D. \ Pizzo , V J. APACrefauthors \ 1999 01 . Three-dimensional propagation of coronal mass ejections (CMEs) in a structured solar wind flow: 2. CME launched adjacent to the streamer belt Three-dimensional propagation of coronal mass ejecti...

1999 doi
[28]

, Riley , P

Odstrcil04 APACrefauthors Odstrcil , D. , Riley , P. \ Zhao , X P. APACrefauthors \ 2004 02 . Numerical simulation of the 12 May 1997 interplanetary CME event Numerical simulation of the 12 May 1997 interplanetary CME event . Journal of Geophysical Research (Space Physics) 109...

2004 doi
[29]

\ Pizzo , V J

Odstrcil99a APACrefauthors Odstr c il , D. \ Pizzo , V J. APACrefauthors \ 1999 01 . Three-dimensional propagation of CMEs in a structured solar wind flow: 1. CME launched within the streamer belt Three-dimensional propagation of CMEs in a structured solar wind flow: 1. CME la...

1999 doi
[30]

\ Cargill, P

OwensandCargill2004 APACrefauthors Owens, M. \ Cargill, P. APACrefauthors \ 2004 . Predictions of the arrival time of Coronal Mass Ejections at 1 AU: an analysis of the causes of errors Predictions of the arrival time of coronal mass ejections at 1 au: an analysis of the cause...

2004 doi
[31]

, Spence , H E

Owens2008 APACrefauthors Owens , M J. , Spence , H E. , McGregor , S. , Hughes , W J. , Quinn , J M. , Arge , C N. Odstrcil , D. APACrefauthors \ 2008 . Metrics for solar wind prediction models: Comparison of empirical, hybrid, and physics‐based schemes with 8 years of L1 obse...

2008 doi
[32]

, Bilitza , D

Papitashvili2014 APACrefauthors Papitashvili , N. , Bilitza , D. \ King , J. APACrefauthors \ 2014 . OMNI: A Description of Near-Earth Solar Wind Environment OMNI: A Description of Near-Earth Solar Wind Environment . 40th COSPAR Scientific Assembly 40th cospar scientific assem...

2014
[33]

\ King, J H

OMNI1hour2020 APACrefauthors Papitashvili, N E. \ King, J H. APACrefauthors \ 2020 . OMNI 1-hour Data Set Omni 1-hour data set \ [Dataset]. NASA Space Physics Data Facility . APACrefURL https://omniweb.gsfc.nasa.gov/ APACrefURL APACrefDOI doi:10.48322/45bb-8792 APACrefDOI

2020 doi
[34]

APACrefauthors \ 2007 05

Pulkkinen07 APACrefauthors Pulkkinen , T. APACrefauthors \ 2007 05 . Space Weather: Terrestrial Perspective Space Weather: Terrestrial Perspective . Living Reviews in Solar Physics 4 1 1 . APACrefDOI doi:10.12942/lrsp-2007-1 APACrefDOI

2007 doi
[35]

APACrefauthors \ 2025 May

Raza2025 APACrefauthors Raza, S. APACrefauthors \ 2025 May . Ambient Solar Wind Effect on CME Propagation Ambient solar wind effect on cme propagation \ [Software]. Zenodo . APACrefURL https://doi.org/10.5281/zenodo.15290983 APACrefURL APACrefDOI doi:10.5281/zenodo.15290983 APACrefDOI

2025 doi
[36]

\ Ben-Nun , M

Riley&BenNun2021 APACrefauthors Riley , P. \ Ben-Nun , M. APACrefauthors \ 2021 . On the Sources and Sizes of Uncertainty in Predicting the Arrival Time of Interplanetary Coronal Mass Ejections Using Global MHD Models On the Sources and Sizes of Uncertainty in Predicting the A...

2021 doi
[37]

, Mays, M L

Riley2018 APACrefauthors Riley, P. , Mays, M L. , Andries, J. , Amerstorfer, T. , Biesecker, D. , Delouille, V. Zhao, X. APACrefauthors \ 2018 . Forecasting the Arrival Time of Coronal Mass Ejections: Analysis of the CCMC CME Scoreboard Forecasting the arrival time of coronal ...

2018 doi
[38]

, Cécere , M

Sahade2020 APACrefauthors Sahade , A. , Cécere , M. \ Krause , G. APACrefauthors \ 2020 . Influence of Coronal Holes on CME Deflections: Numerical Study Influence of Coronal Holes on CME Deflections: Numerical Study . The Astrophysical Journal 896 53 . APACrefURL https://doi.o...

2020 doi
[39]

, Owens , M J

Savani2010 APACrefauthors Savani , N P. , Owens , M J. , Rouillard , A P. , Forsyth , R J. \ Davies , J A. APACrefauthors \ 2010 05 . Observational Evidence of a Coronal Mass Ejection Distortion Directly Attributable to a Structured Solar Wind Observational Evidence of a Coron...

2010 doi
[40]

, Srivastava , N

Sharma2013 APACrefauthors Sharma , R. , Srivastava , N. , Chakrabarty , D. , Möstl , C. \ Hu , Q. APACrefauthors \ 2013 07 . Interplanetary and geomagnetic consequences of 5 January 2005 CMEs associated with eruptive filaments Interplanetary and geomagnetic consequences of 5 J...

2013 doi
[41]

, Wang , Y

Shen2017 APACrefauthors Shen , F. , Wang , Y. , Shen , C. \ Feng , X. APACrefauthors \ 2017 08 . On the Collision Nature of Two Coronal Mass Ejections: A Review On the Collision Nature of Two Coronal Mass Ejections: A Review . Solar Physics 292 104 . APACrefDOI doi:10.1007/s11...

2017 doi
[42]

, Wu , S T

Shen2012 APACrefauthors Shen , F. , Wu , S T. , Feng , X. \ Wu , C C. APACrefauthors \ 2012 11 . Acceleration and deceleration of coronal mass ejections during propagation and interaction Acceleration and deceleration of coronal mass ejections during propagation and interactio...

2012 doi
[43]

, Benson, B

Singh2023Improving APACrefauthors Singh, T. , Benson, B. , Raza, S A Z. , Kim, T K. , Pogorelov, N V. , Smith, W P. \ Arge, C N. APACrefauthors \ 2023 2 . Improving the Arrival Time Estimates of Coronal Mass Ejections by Using Magnetohydrodynamic Ensemble Modeling, Heliospheri...

2023 arXiv
[44]

, Kim , T K

Singh2020 APACrefauthors Singh , T. , Kim , T K. , Pogorelov , N V. \ Arge , C N. APACrefauthors \ 2020 . Application of a Modified Spheromak Model to Simulations of Coronal Mass Ejection in the Inner Heliosphere Application of a Modified Spheromak Model to Simulations of Coro...

2020 doi
[45]

, Yalim, M S

Singh18 APACrefauthors Singh, T. , Yalim, M S. \ Pogorelov, N V. APACrefauthors \ 2018 aug . A Data-constrained Model for Coronal Mass Ejections Using the Graduated Cylindrical Shell Method A data-constrained model for coronal mass ejections using the graduated cylindrical she...

2018 doi
[46]

, Scott, C J

TuckerHood2015 APACrefauthors Tucker-Hood, K. , Scott, C J. , Owens, M J. , Jackson, D. , Barnard, L. , Davies, J A. \ et al. APACrefauthors \ 2015 . Validation of a priori CME arrival predictions made using real-time heliospheric imager observations Validation of a priori cme...

2015 doi
[47]

, Fischer , S

Vandas1996 APACrefauthors Vandas , M. , Fischer , S. , Dryer , M. , Smith , Z. \ Detman , T. APACrefauthors \ 1996 Feb . Simulation of magnetic cloud propagation in the inner heliosphere in two dimensions 2. A loop parallel to the ecliptic plane and the role of helicity Simula...

1996 doi
[48]

APACrefauthors \ 2001 08

Vrsnak01 APACrefauthors Vr s nak , B. APACrefauthors \ 2001 08 . Deceleration of Coronal Mass Ejections Deceleration of Coronal Mass Ejections . Solar Physics 202 1 173-189 . APACrefDOI doi:10.1023/A:1011833114104 APACrefDOI

2001 doi
[49]

\ Gopalswamy , N

Vrsnak02 APACrefauthors Vr s nak , B. \ Gopalswamy , N. APACrefauthors \ 2002 02 . Influence of the aerodynamic drag on the motion of interplanetary ejecta Influence of the aerodynamic drag on the motion of interplanetary ejecta . Journal of Geophysical Research (Space Physics...

2002 doi
[50]

, Liu , J

Wang2019 APACrefauthors Wang , Y. , Liu , J. , Jiang , Y. \ Erdélyi , R. APACrefauthors \ 2019 08 . CME Arrival Time Prediction Using Convolutional Neural Network CME Arrival Time Prediction Using Convolutional Neural Network . The Astrophysical Journal 881 1 15 . APACrefDOI d...

2019 doi
[51]

, Ye, P Z

Wang2002 APACrefauthors Wang, Y M. , Ye, P Z. , Wang, S. , Zhou, G P. \ Wang, J X. APACrefauthors \ 2002 . A statistical study on the geoeffectiveness of Earth-directed coronal mass ejections from March 1997 to December 2000 A statistical study on the geoeffectiveness of earth...

2002 doi
[52]

, Gopalswamy , N

Xie2006 APACrefauthors Xie , H. , Gopalswamy , N. , Ofman , L. , St. Cyr , O C. , Michalek , G. , Lara , A. \ Yashiro , S. APACrefauthors \ 2006 . Improved input to the empirical coronal mass ejection (CME) driven shock arrival model from CME cone models Improved input to the ...

2006 doi
[53]

, Ofman , L

Xie2004 APACrefauthors Xie , H. , Ofman , L. \ Lawrence , G. APACrefauthors \ 2004 03 . Cone model for halo CMEs: Application to space weather forecasting Cone model for halo CMEs: Application to space weather forecasting . Journal of Geophysical Research: Space Physics 109 A0...

2004 doi
[54]

, Richardson , I G

Zhang2007 APACrefauthors Zhang , J. , Richardson , I G. , Webb , D F. , Gopalswamy , N. , Huttunen , E. , Kasper , J C. Zhukov , A N. APACrefauthors \ 2007 10 . Solar and interplanetary sources of major geomagnetic storms (Dst <= -100 nT) during 1996-2005 Solar and interplanet...

2007 doi
[55]

\ Dryer , M

Zhao&Dryer2014 APACrefauthors Zhao , X. \ Dryer , M. APACrefauthors \ 2014 . Current status of CME/shock arrival time prediction Current status of CME/shock arrival time prediction . Space Weather 12 7 448-469 . APACrefDOI doi:10.1002/2014SW001060 APACrefDOI

2014 doi

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [1]

, Abduallah , Y

Alobaid2022 APACrefauthors Alobaid , K A. , Abduallah , Y. , Wang , J T L. , Wang , H. , Jiang , H. , Xu , Y. Jing , J. APACrefauthors \ 2022 10 . Predicting CME arrival time through data integration and ensemble learning Predicting CME arrival time through data integration and ensemble learning . Frontiers in Astronomy and Space Sciences 9 . APACrefDOI d...

work page arXiv 2022

[2] [2]

, Luhmann, J

Arge04 APACrefauthors Arge, C. , Luhmann, J. , Odstrcil, D. , Schrijver, C. \ Li, Y. APACrefauthors \ 2004 . Stream structure and coronal sources of the solar wind during the May 12th, 1997 CME Stream structure and coronal sources of the solar wind during the may 12th, 1997 cme . Journal of Atmospheric and Solar-Terrestrial Physics 66 15 1295 - 1309 . APA...

2004

[3] [3]

\ Pizzo, V J

Arge00 APACrefauthors Arge, C N. \ Pizzo, V J. APACrefauthors \ 2000 . Improvement in the prediction of solar wind conditions using near-real time solar magnetic field updates Improvement in the prediction of solar wind conditions using near-real time solar magnetic field updates . Journal of Geophysical Research: Space Physics 105 A5 10465-10479 . APACre...

doi:10.1029/1999ja000262 2000

[4] [4]

, Delaboudiniere , J P

Brueckner1998 APACrefauthors Brueckner , G E. , Delaboudiniere , J P. , Howard , R A. , Paswaters , S E. , St. Cyr , O C. , Schwenn , R. Wang , D. APACrefauthors \ 1998 Jan . Geomagnetic storms caused by coronal mass ejections (CMEs): March 1996 through June 1997 Geomagnetic storms caused by coronal mass ejections (CMEs): March 1996 through June 1997 . Ge...

work page doi:10.1029/98gl00704 1998

[5] [5]

APACrefauthors \ 2011 Apr 08

Chen11 APACrefauthors Chen, P F. APACrefauthors \ 2011 Apr 08 . Coronal Mass Ejections: Models and Their Observational Basis Coronal mass ejections: Models and their observational basis . Living Reviews in Solar Physics 8 1 1 . APACrefURL https://doi.org/10.12942/lrsp-2011-1 APACrefURL APACrefDOI doi:10.12942/lrsp-2011-1 APACrefDOI

doi:10.12942/lrsp-2011-1 2011

[6] [6]

, Shen , C

Chi2016 APACrefauthors Chi , Y. , Shen , C. , Wang , Y. , Xu , M. , Ye , P. \ Wang , S. APACrefauthors \ 2016 09 . Statistical Study of the Interplanetary Coronal Mass Ejections from 1995 to 2015 Statistical Study of the Interplanetary Coronal Mass Ejections from 1995 to 2015 . Solar Physics 291 8 2419-2439 . APACrefDOI doi:10.1007/s11207-016-0971-5 APACrefDOI

doi:10.1007/s11207-016-0971-5 2016

[7] [7]

, Liu , J

Chierichini2024 APACrefauthors Chierichini , S. , Liu , J. , Korsós , M B. , Del Moro , D. \ Erdélyi , R. APACrefauthors \ 2024 03 . CME Arrival Modeling with Machine Learning CME Arrival Modeling with Machine Learning . The Astrophysical Journal 963 1 121 . APACrefDOI doi:10.3847/1538-4357/ad1cee APACrefDOI

doi:10.3847/1538-4357/ad1cee 2024

[8] [8]

\ Smart, D

Dryer1984 APACrefauthors Dryer, M. \ Smart, D. APACrefauthors \ 1984 . Dynamical models of coronal transients and interplanetary disturbances Dynamical models of coronal transients and interplanetary disturbances . Advances in Space Research 4 7 291 - 301 . APACrefURL http://www.sciencedirect.com/science/article/pii/027311778490200X APACrefURL APACrefDOI ...

work page arXiv 1984