A Comparison of Flare Forecasting Methods. III. Systematic Behaviors of Operational Solar Flare Forecasting Systems

Leka, K. D.; Park, Sung-Hong; Kusano, K.; Andries, Jesse; Barnes, G.; Bingham, Suzy; Bloomfield, D. Shaun; McCloskey, Aoife E.; Delouille, Véronique; Falconer, D. A.; Gallagher, P. T.; Georgoulis, Manolis K.; Kubo, Yûki; Lee, Kangjin; Lee, Sang‐Woo; Lobzin, V. V.; Mun, JunChul; Murray, Sophie A.; Nageem, Tarek A.M. Hamad; Qahwaji, Rami; Sharpe, Michael; Steenburgh, Rob; Steward, Graham; Terkildsen, Michael

doi:10.3847/1538-4357/ab2e11

Cited by 55 publications

(29 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We emphasize that many, but by no means all, of the existing flare-prediction studies did not consider these magnetic field parameters as time series. Instead, forecasting relied on cross-sectional, or point-in-time (snapshot) parameter values 42 – 44 . There are a few exceptions: Gallagher et al .…”

Section: Methodsmentioning

confidence: 99%

Multivariate time series dataset for space weather data analytics

et al. 2020

Self Cite

View full text Add to dashboard Cite

We introduce and make openly accessible a comprehensive, multivariate time series (MVTS) dataset extracted from solar photospheric vector magnetograms in Spaceweather HMI Active Region Patch (SHARP) series. Our dataset also includes a cross-checked NOAA solar flare catalog that immediately facilitates solar flare prediction efforts. We discuss methods used for data collection, cleaning and pre-processing of the solar active region and flare data, and we further describe a novel data integration and sampling methodology. Our dataset covers 4,098 MVTS data collections from active regions occurring between May 2010 and December 2018, includes 51 flare-predictive parameters, and integrates over 10,000 flare reports. Potential directions toward expansion of the time series, either “horizontally” – by adding more prediction-specific parameters, or “vertically” – by generalizing flare into integrated solar eruption prediction, are also explained. The immediate tasks enabled by the disseminated dataset include: optimization of solar flare prediction and detailed investigation for elusive flare predictors or precursors, with both operational (research-to-operations), and basic research (operations-to-research) benefits potentially following in the future.

show abstract

Section: Methodsmentioning

confidence: 99%

Multivariate time series dataset for space weather data analytics

et al. 2020

Self Cite

View full text Add to dashboard Cite

show abstract

“…For SWx, the off-SEL viewpoints are more important the SEL ones, since the former can help decipher the CME structure. (Leka et al, 2019).…”

Section: Paradigm Shiftsmentioning

confidence: 99%

Improving the Medium-Term Forecasting of Space Weather: A Big Picture Review From a Solar Observer's Perspective

Vourlidas

2021

Front. Astron. Space Sci.

View full text Add to dashboard Cite

We have improved considerably our scientific understanding of the key solar drivers of Space Weather, i.e., Coronal Mass Ejections, flares, in the last 20+ years thanks to a plethora of space missions and modeling advances. Yet, a major breakthrough in assessing the geo-effectiveness of a given CME and associated phenomena still escapes us, holding back actionable medium-term (up to 7 days) forecasting of Space Weather. Why is that? I adopt a two-pronged approach to search for answers. First, I assess the last 20+ years of research on solar drivers by identifying lessons-learned and paradigm shifts in our view of solar activity, always in relation to Space Weather concerns. Then, I review the state of key observation-based quantities used in forecasting to isolate the choke points and research gaps that limit medium-term forecasting performance. Finally, I outline a path forward along three vectors—breakthrough capabilities, geo-effective potential, and actionable forecast—with the strongest potential to improve space weather forecasting horizon and robustness.

show abstract

“…Optimizing the threshold is accomplished using validation data, as described in Section 3.4. Some methods choose an arbitrary threshold of 0.5 to generate the categorical prediction (Leka et al 2019a) while others use an automatically determined threshold based on optimizing the Receiver Operating Characteristic (ROC, Jolliffe & Stephenson 2012). Interestingly, this coincides with the threshold that maximizes the TSS score.…”

Section: Stage I: Convolutional Neural Networkmentioning

confidence: 99%

Decreasing False Alarm Rates in ML-based Solar Flare Prediction using SDO/HMI Data

Deshmukh,

Flyer,

Van Der Sande

et al. 2021

Preprint

View full text Add to dashboard Cite

A hybrid two-stage machine learning architecture that addresses the problem of excessive false positives (false alarms) in solar flare prediction systems is investigated. The first stage is a convolutional neural network (CNN) model based on the VGG-16 architecture that extracts features from a temporal stack of consecutive Solar Dynamics Observatory (SDO) Helioseismic and Magnetic Imager (HMI) magnetogram images to produce a flaring probability. The probability of flaring is added to a feature vector derived from the magnetograms to train an extremely randomized trees (ERT) model in the second stage to produce a binary deterministic prediction (flare/no flare) in a 12-hour forecast window. To tune the hyperparameters of the architecture a new evaluation metric is introduced, the "scaled True Skill Statistic". It specifically addresses the large discrepancy between the true positive rate and the false positive rate in the highly unbalanced solar flare event training datasets. Through hyperparameter tuning to maximize this new metric, our two-stage architecture drastically reduces false positives by ≈ 48% without significantly affecting the true positives (reduction by ≈ 12%), when compared with predictions from the first stage CNN alone. This, in turn, improves various traditional binary classification metrics sensitive to false positives such as the precision, F1 and the Heidke Skill Score. The end result is a more robust 12-hour flare prediction system that could be combined with current operational flare forecasting methods. Additionally, using the ERT-based feature ranking mechanism, we show that the CNN output probability is highly ranked in terms of flare prediction relevance.

show abstract

A Comparison of Flare Forecasting Methods. III. Systematic Behaviors of Operational Solar Flare Forecasting Systems

Cited by 55 publications

References 33 publications

Multivariate time series dataset for space weather data analytics

Multivariate time series dataset for space weather data analytics

Improving the Medium-Term Forecasting of Space Weather: A Big Picture Review From a Solar Observer's Perspective

Decreasing False Alarm Rates in ML-based Solar Flare Prediction using SDO/HMI Data

Contact Info

Product

Resources

About