Predicting Solar Flares with Machine Learning: Investigating Solar Cycle Dependence

Wang, Xiantong; Chen, Yang; Tóth, G.; Manchester, W.; Gombosi, T. I.; Hero, Alfred O.; Jiao, Zhenbang; Sun, Hu; Jin, Meng; Liu, Y.

doi:10.3847/1538-4357/ab89ac

Cited by 60 publications

(55 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently, progress has been made in flare forecasting (e.g., Chen & Wang, 2020, Kusano et al., 2020. X. Wang et al., 2020), and it is expected that with machine learning patterns leading up to eruptions (flares and CMEs) can be recognized even more reliably.…”

Section: An L4 Mission That Is Critical For Human Explorationmentioning

confidence: 99%

A Multi‐Purpose Heliophysics L4 Mission

Posner¹,

Arge

Staub

et al. 2021

Space Weather

View full text Add to dashboard Cite

The Earth-Sun Lagrangian point 4 is a meta-stable location at 1 au from the Sun, 60° ahead of Earth's orbit. It has an uninterrupted view of the solar photosphere centered on W60, the Earth's nominal magnetic field connection to the Sun. Such a mission on its own would serve as a solar remote sensing observatory that would oversee the entire solar radiation hemisphere with significant relevance for protecting Moon and Mars explorers from radiation exposure. In combination with appropriately planned observatories at L1 and L5, the three spacecraft would provide 300° longitude coverage of photospheric magnetic field structure, and allow continuous viewing of both solar poles, with >3.6° elevation. Ideally, the L4 and L5 missions would orbit the Sun with a 7.2° inclination out of the heliographic equator, 14.5° out of the ecliptic plane. We discuss the impact of extending solar magnetic field observations in both longitude and latitude to improve global solar wind modeling and, with the development of local helioseismology, the potential for long-term solar activity forecasting. Such a mission would provide a unique opportunity for interplanetary and interstellar dust science. It would significantly add to reliability of operational observations on fast coronal mass ejections directed at Earth and for human Mars explorers on their round-trip journey. The L4 mission concept is technically feasible, and is scientifically compelling. Section 1: IntroductionSpace weather (SWx) forecasting is a complex problem, spanning vastly different regions of space and involving several scientific disciplines within and beyond Heliophysics. The main pillar that all SWx forecasts rely on is rapid availability of sets of observations of the Sun and inner heliosphere measured on the ground and on space platforms. Budgetary realism dictates the locations of observatories and their instrumentation to be chosen carefully so they address short-term needs and lead to long-term improvements.It is important to note that much of the underlying physics of space weather is not yet understood. This includes, to name just a subset of phenomena, the initiation of flares and coronal mass ejections, the acceleration of the solar wind and that of particles to high energies. At the time of writing, the Parker Accepted ArticleThis article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as

show abstract

Section: An L4 Mission That Is Critical For Human Explorationmentioning

confidence: 99%

A Multi‐Purpose Heliophysics L4 Mission

Posner¹,

Arge

Staub

et al. 2021

Space Weather

View full text Add to dashboard Cite

show abstract

“…Liu et al, 2000) and solar flares (cf. Chen et al, 2019b;Jiao et al, 2020;Wang et al, 2020). However, as most machine learning models are not interpretable, they typically do not help us to understand the underlying physics.…”

Section: Black-box Modelsmentioning

confidence: 99%

What sustained multi-disciplinary research can achieve: The space weather modeling framework

Gombosi

Chen

Glocer

et al. 2021

J. Space Weather Space Clim.

Self Cite

View full text Add to dashboard Cite

MHD-based global space weather models have mostly been developed and maintained at academic institutions. While the ``free spirit'' approach of academia enables the rapid emergence and testing of new ideas and methods, the lack of long-term stability and support makes this arrangement very challenging. This paper describes a successful example of a university-based group, the Center of Space Environment Modeling (CSEM) at the University of Michigan that developed and maintained the Space Weather Modeling Framework (SWMF) and its core element, the BATS-R-US extended MHD code. It took a quarter of a century to develop this capability and reach its present level of maturity that makes it suitable for research use by the space physics community through the Community Coordinated Modeling Center (CCMC) as well as operational use by the NOAA Space Weather Prediction Center (SWPC).

show abstract

“…Currently, new physical and geometrical (topological) features are applied to flare prediction using machine learning (e.g., Wang et al 2020a;Deshmukh et al 2020), and it has been noted that training sets may be sensitive to which period in the solar cycle they are drawn from. (Wang et al 2020b).…”

Section: Introductionmentioning

confidence: 99%

Operational solar flare prediction model using Deep Flare Net

Nishizuka

Kubo

Sugiura

et al. 2021

Earth Planets Space

View full text Add to dashboard Cite

We developed an operational solar flare prediction model using deep neural networks, named Deep Flare Net (DeFN). DeFN can issue probabilistic forecasts of solar flares in two categories, such as ≥ M-class and < M-class events or ≥ C-class and < C-class events, occurring in the next 24 h after observations and the maximum class of flares occurring in the next 24 h. DeFN is set to run every 6 h and has been operated since January 2019. The input database of solar observation images taken by the Solar Dynamic Observatory (SDO) is downloaded from the data archive operated by the Joint Science Operations Center (JSOC) of Stanford University. Active regions are automatically detected from magnetograms, and 79 features are extracted from each region nearly in real time using multiwavelength observation data. Flare labels are attached to the feature database, and then, the database is standardized and input into DeFN for prediction. DeFN was pretrained using the datasets obtained from 2010 to 2015. The model was evaluated with the skill score of the true skill statistics (TSS) and achieved predictions with TSS = 0.80 for ≥ M-class flares and TSS = 0.63 for ≥ C-class flares. For comparison, we evaluated the operationally forecast results from January 2019 to June 2020. We found that operational DeFN forecasts achieved TSS = 0.70 (0.84) for ≥ C-class flares with the probability threshold of 50 (40)%, although there were very few M-class flares during this period and we should continue monitoring the results for a longer time. Here, we adopted a chronological split to divide the database into two for training and testing. The chronological split appears suitable for evaluating operational models. Furthermore, we proposed the use of time-series cross-validation. The procedure achieved TSS = 0.70 for ≥ M-class flares and 0.59 for ≥ C-class flares using the datasets obtained from 2010 to 2017. Finally, we discuss the standard evaluation methods for operational forecasting models, such as the preparation of observation, training, and testing datasets, and the selection of verification metrics.

show abstract

Predicting Solar Flares with Machine Learning: Investigating Solar Cycle Dependence

Cited by 60 publications

References 38 publications

A Multi‐Purpose Heliophysics L4 Mission

A Multi‐Purpose Heliophysics L4 Mission

What sustained multi-disciplinary research can achieve: The space weather modeling framework

Operational solar flare prediction model using Deep Flare Net

Contact Info

Product

Resources

About