In this paper, we demonstrate the applicability of the data-driven and self-consistent solar energetic particle model, Solar-wind with FIeld-lines and Energetic-particles (SOFIE), to simulate acceleration and transport processes of solar energetic particles. SOFIE model is built upon the Space Weather Modeling Framework (SWMF) developed at the University of Michigan. In SOFIE, the background solar wind plasma in the solar corona and interplanetary space is calculated by the Aflv\’en Wave Solar-atmosphere Model(-Realtime) (AWSoM-R) driven by the near-real-time hourly updated Global Oscillation Network Group (GONG) solar magnetograms. In the background solar wind, coronal mass ejections (CMEs) are launched by placing an imbalanced magnetic flux rope on top of the parent active region, using the Eruptive Event Generator using Gibson-Low model (EEGGL). The acceleration and transport processes are modeled by the Multiple-Field-Line Advection Model for Particle Acceleration (M-FLAMPA). In this work, nine solar energetic particle events (Solar Heliospheric and INterplanetary Environment (SHINE) challenge/campaign events) are modeled. The three modules in SOFIE are validated and evaluated by comparing with observations, including the steady-state background solar wind properties, the white-light image of the CME, and the flux of solar energetic protons, at energies of $\ge$ 10 MeV.

A document by Zhenguang Huang. Click on the document to view its contents.

We consider the flare prediction problem that distinguishes flare-imminent active regions that produce an M- or X-class flare in the future 24 hours, from quiet active regions that do not produce any flare within $\pm 24$ hours. Using line-of-sight magnetograms and parameters of active regions in two data products covering Solar Cycle 23 and 24, we train and evaluate two deep learning algorithms—CNN and LSTM—and their stacking ensembles. The decisions of CNN are explained using visual attribution methods. We have the following three main findings. (1) LSTM trained on data from two solar cycles achieves significantly higher True Skill Scores (TSS) than that trained on data from a single solar cycle with a confidence level of at least 0.95. (2) On data from Solar Cycle 23, a stacking ensemble that combines predictions from LSTM and CNN using the TSS criterion achieves significantly higher TSS than the “select-best” strategy with a confidence level of at least 0.95. (3) A visual attribution method called Integrated Gradients is able to attribute the CNN’s predictions of flares to the emerging magnetic flux in the active region. It also reveals a limitation of CNN as a flare prediction method using line-of-sight magnetograms: it treats the polarity artifact of line-of-sight magnetograms as positive evidence of flares.

We present a reduced magnetohydrodynamic (MHD) mathematical model describing the dynamical behavior of highly conducting plasmas with frozen-in magnetic fields, constrained by the assumption that, there exists a frame of reference, where the magnetic field vector, B, is aligned with the plasma velocity vector, u, at each point. We call this solution “stream-aligned MHD” (SA-MHD). Within the framework of this model, the electric field, E = −u x B ≡ 0, in the induction equation vanishes identically and so does the electromagnetic energy flux (Poynting flux), E x B ≡ 0, in the energy equation. At the same time, the force effect from the magnetic field on the plasma motion (the Ampere force) is fully taken into account in the momentum equation. Any steady-state solution of the proposed model is a legitimate solution of the full MHD system of equations. However, the converse statement is not true: in an arbitrary steady-state magnetic field the electric field does not have to vanish identically (its curl has to, though). Specifically, realistic tree-dimensional solutions for the steady-state (ambient) solar atmosphere in the form of so-called Parker (1958) spirals, can be efficiently generated within the stream-aligned MHD (SA-MHD) with no loss in generality.

To simulate solar Coronal Mass Ejections (CMEs), predict their time of arrival and geomagnetic impact, it is important to accurately model the background solar wind conditions in which CMEs propagate. We use the Alfvén Wave Solar-atmosphere Model (AWSoM) within the the Space Weather Modeling Framework (SWMF) to simulate solar maximum conditions during two Carrington rotations and produce solar wind background conditions comparable to the observations. We describe the inner boundary conditions for AWSoM using the ADAPT global magnetic maps and validate the simulated results with EUV observations in the low corona and measured plasma parameters at L1 as well as at the position of the STEREO spacecraft. This work complements our prior AWSoM validation study for solar minimum conditions Sachdeva et al. (2019), and shows that during periods of higher magnetic activity, AWSoM can reproduce the solar plasma conditions (using properly adjusted photospheric Poynting flux) suitable for providing proper initial conditions for launching CMEs.

In this work, we develop gradient boosting machines (GBMs) for forecasting the SYM-H index multiple hours ahead using different combinations of solar wind and interplanetary magnetic field (IMF) parameters, derived parameters, and past SYM-H values. Using Shapley Additive Explanation (SHAP) values to quantify the contributions from each input to predictions of the SYM-H index from GBMs, we show that our predictions are consistent with physical understanding while also providing insight into the complex relationship between the solar wind and Earth’s ring current. In particular, we found that feature contributions vary depending on the storm phase. We also perform a direct comparison between GBMs and neural networks presented in prior publications for forecasting the SYM-H index by training, validating, and testing them on the same data. We find that the GBMs have a comparable root mean squared error as the best published black-box neural network schemes.

We use machine learning methods to predict whether an active region (AR) which produces flares will lead to a solar energetic particle (SEP) event using Space-Weather Michelson Doppler Imager (MDI) Active Region Patches (SMARPs). This new data product is derived from maps of the solar surface magnetic field taken by the Michelson Doppler Imager (MDI) aboard the Solar and Heliospheric Observatory (SOHO). We survey the SMARP active regions associated with flares that appear on the solar disk between June 5, 1996 and August 14, 2010, label those that produced SEPs as positive and the rest as negative. The AR SMARP features that correspond to each flare are used to train two different types of machine learning methods, the support vector machines (SVMs) and the regression models. The results show that the SMARP data can predict whether a flare will lead to an SEP with accuracy (ACC) {less than or equal to}0.72{plus minus}0.12 while allowing for a competitive leading time of 55.3{plus minus}28.6 minutes for forecasting the SEP events.

Modeling the terrestrial impacts of the sun’s solar wind is critical to understanding geomagnetic storms. We use a database of 144 storms from 2010-2019 and showed how these storms affect magnetometers on the ground. We also extracted profiles of the magnetic field along the magnetotail. Skill scores are assigned to the individual stations on the ground based on how well they can forecast magnetic indeces like SYM-H and AL. We us our Space Weather Modeling Framework’s geospace configuration. Our model includes coupling of 3D MHD solver (BATSRUS), the Rice Convection Model, and the Ridley Ionospheric Model. We have found that all stations have a positive Heidke Skill Score which is encouraging in terms of space weather forecasting.

We present a data-driven and self-consistent SEP model, SOFIE, to simulate the acceleration and transport processes of energetic particles using the Space Weather Modeling Framework (SWMF). In this model, the background solar wind plasma in the solar corona and interplanetary space are modeled by the Alfven Wave Solar-atmosphere Model(-Realtime) (AWSoM(-R)) driven by the near-real-time hourly updated GONG (bihourly ADAPT-GONG) magnetograms. In the background solar wind, the CMEs are launched employing the Eruptive Event Generator using Gibson-Low configuration (EEGGL), by inserting a flux rope estimated from the free magnetic energy in the active region. The acceleration and transport processes are then modeled self-consistently by the multiple magnetic field line tracker (M-FLAMPA) and the Adaptive Mesh Particle Simulator (AMPS). We will demonstrate the capability of SOFIE to demystify the acceleration processes by the CME-driven shock in the low corona and the modulation of energetic particles by the solar wind structures. Besides, using selected historical SEP events, e.g. 2013 Apr 11 event, we will illustrate the progresses toward a faster-than-real-time prediction of SEPs.