Context. Machine-learning methods for predicting solar flares typically employ physics-based features that have been carefully cho- sen by experts in order to capture the salient features of the photospheric magnetic fields of the Sun.
Aims. Though the sophistication and complexity of these models have grown over time, there has been little evolution in the choice of feature sets, or any systematic study of whether the additional model complexity leads to higher predictive skill.
Methods. This study compares the relative prediction performance of four different machine-learning based flare prediction models with increasing degrees of complexity. It evaluates three different feature sets as input to each model: a “traditional” physics-based feature set, a novel “shape-based” feature set derived from topological data analysis (TDA) of the solar magnetic field, and a com- bination of these two sets. A systematic hyperparameter tuning framework is employed in order to assure fair comparisons of the models across different feature sets. Finally, principal component analysis is used to study the effects of dimensionality reduction on these feature sets.
Results. It is shown that simpler models with fewer free parameters perform better than the more complicated models on the canonical 24-h flare forecasting problem. In other words, more complex machine-learning architectures do not necessarily guarantee better prediction performance. In addition, it is found that shape-based feature sets contain just as much useful information as physics-based feature sets for the purpose of flare prediction, and that the dimension of these feature sets – particularly the shape-based one – can be greatly reduced without impacting predictive accuracy.
more »
« less
Identifying Flare-indicative Photospheric Magnetic Field Parameters from Multivariate Time-series Data of Solar Active Regions
Photospheric magnetic field parameters are frequently used to analyze and predict solar events. Observation of these parameters over time, i.e., representing solar events by multivariate time-series (MVTS) data, can determine relationships between magnetic field states in active regions and extreme solar events, e.g., solar flares. We can improve our understanding of these events by selecting the most relevant parameters that give the highest predictive performance. In this study, we propose a two-step incremental feature selection method for MVTS data using a deep-learning model based on long short-term memory (LSTM) networks. First, each MVTS feature (magnetic field parameter) is evaluated individually by a univariate sequence classifier utilizing an LSTM network. Then, the top performing features are combined to produce input for an LSTM-based multivariate sequence classifier. Finally, we tested the discrimination ability of the selected features by training downstream classifiers, e.g., Minimally Random Convolutional Kernel Transform and support vector machine. We performed our experiments using a benchmark data set for flare prediction known as Space Weather Analytics for Solar Flares. We compared our proposed method with three other baseline feature selection methods and demonstrated that our method selects more discriminatory features compared to other methods. Due to the imbalanced nature of the data, primarily caused by the rarity of minority flare classes (e.g., the X and M classes), we used the true skill statistic as the evaluation metric. Finally, we reported the set of photospheric magnetic field parameters that give the highest discrimination performance in predicting flare classes.
more » « less- NSF-PAR ID:
- 10496084
- Publisher / Repository:
- DOI PREFIX: 10.3847
- Date Published:
- Journal Name:
- The Astrophysical Journal Supplement Series
- Volume:
- 271
- Issue:
- 2
- ISSN:
- 0067-0049
- Format(s):
- Medium: X Size: Article No. 39
- Size(s):
- ["Article No. 39"]
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
more » « less
Abstract We propose a new “helicity-pumping” method for energizing coronal equilibria that contain a magnetic flux rope (MFR) toward an eruption. We achieve this in a sequence of magnetohydrodynamics relaxations of small line-tied pulses of magnetic helicity, each of which is simulated by a suitable rescaling of the current-carrying part of the field. The whole procedure is “magnetogram-matching” because it involves no changes to the normal component of the field at the photospheric boundary. The method is illustrated by applying it to an observed force-free configuration whose MFR is modeled with our regularized Biot–Savart law method. We find that, in spite of the bipolar character of the external field, the MFR eruption is sustained by two reconnection processes. The first, which we refer to as breakthrough reconnection, is analogous to breakout reconnection in quadrupolar configurations. It occurs at a quasi-separator inside a current layer that wraps around the erupting MFR and is caused by the photospheric line-tying effect. The second process is the classical flare reconnection, which develops at the second quasi-separator inside a vertical current layer that is formed below the erupting MFR. Both reconnection processes work in tandem with the magnetic forces of the unstable MFR to propel it through the overlying ambient field, and their interplay may also be relevant for the thermal processes occurring in the plasma of solar flares. The considered example suggests that our method will be beneficial for both the modeling of observed eruptive events and theoretical studies of eruptions in idealized magnetic configurations.
-
more » « less
Abstract Solar flares are explosions on the Sun. They happen when energy stored in magnetic fields around solar active regions (ARs) is suddenly released. Solar flares and accompanied coronal mass ejections are sources of space weather, which negatively affects a variety of technologies at or near Earth, ranging from blocking high-frequency radio waves used for radio communication to degrading power grid operations. Monitoring and providing early and accurate prediction of solar flares is therefore crucial for preparedness and disaster risk management. In this article, we present a transformer-based framework, named SolarFlareNet, for predicting whether an AR would produce a
$$\gamma$$ $$\gamma$$ $$\ge$$ $$\ge$$ $$\ge$$ $$\gamma$$ $$\gamma$$ -
more » « less
Abstract Magnetic reconnection is regarded as the mechanism for the rapid release of magnetic energy stored in active regions during solar flares, and quantitative measurements of the magnetic reconnection rate are essential for understanding solar flares. In the context of the standard two-ribbon flare model, we derive the coronal magnetic reconnection rate of the M6.5 flare on 2015 June 22 in two terms, reconnection flux change rate and reconnection electric field, both of which can be obtained from observations of the flare morphology. Data used include a sequence of chromospheric H
α images with unprecedented resolution during the flare from the Visual Imaging Spectrometer of the Goode Solar Telescope (GST) at the Big Bear Solar Observatory and a preflare line-of-sight photospheric magnetogram from the GST Near-InfraRed Imaging Spectropolarimeter along with hard X-ray data from the Ramaty High Energy Solar Spectroscopic Imager. The temporal correlation between the magnetic reconnection rate and nonthermal emission is found, and the variation of the reconnection electric field is mainly determined by the ribbon speed, not by the local magnetic field encountered by the ribbon front. Spatially, the hard X-ray source overlaps with the location of the strongest electric field obtained at the same time. The ribbon motion shows abundant fine structures, including a local acceleration at the location of a light bridge with a weaker magnetic field. -
more » « less
Abstract Magnetic reconnection is understood to be the main physical process that facilitates the transformation of magnetic energy into heat, motion, and particle acceleration during solar eruptions. Yet, observational constraints on reconnection region properties and dynamics are limited due to a lack of high-cadence and high-spatial-resolution observations. By studying the evolution and morphology of postreconnected field-lines footpoints, or flare ribbons and vector photospheric magnetic field, we estimate the magnetic reconnection flux and its rate of change with time to study the flare reconnection process and dynamics of the current sheet above. We compare high-resolution imaging data to study the evolution of the fine structure in flare ribbons as ribbons spread away from the polarity inversion line. Using data from two illustrative events (one M- and X-class flare), we explore the relationship between the ribbon-front fine structure and the temporal development of bursts in the reconnection region. Additionally, we use the
RibbonDB database to perform statistical analysis of 73 (C- to X-class) flares and identify quasiperiodic pulsation (QPP) properties using the Wavelet Transform. Our main finding is the discovery of QPP signatures in the derived magnetic reconnection rates in both example events and the large flare sample. We find that the oscillation periods range from 1 to 4 minutes. Furthermore, we find nearly cotemporal bursts in Hard X-ray (HXR) emission profiles. We discuss how dynamical processes in the current sheet involving plasmoids can explain the nearly cotemporal signatures of quasiperiodicity in the reconnection rates and HXR emission.
