Mitsue Den

The plasma‐sheet dynamics leading to the substorm onset is analyzed from a global simulation. We calculated three cases, the normal (case 1), reduced magnetosphere‐ionosphere (M‐I) coupling (case 2) and reduced tail dissipation (case 3). The normal solution reproduces the thinning, reconnection, tail flow, and the dipolarization in the magnetosphere, and the AU/AL, the initial brightening, the beads, the pre‐onset arc, and the poleward expansion in the ionosphere. Case 2 shows that the M‐I coupling promotes the thinning by evacuating magnetic field from the plasma sheet. Case 3 shows that it is implausible to reproduce the AU/AL of the expansion phase only with numerical dissipation. The projection area of the growth phase plasma sheet is so narrow that the quiet arcs on the high‐latitude side are projected to the lobe. The key point to understand such narrow projection is to consider the origin of the field‐aligned current (FAC). The FAC occurs to transmit motion, and particle precipitation is the consequence. The upward FACs for the quiet arc and initial brightening, which produce the substorm sequence, are generated to transmit enhanced convection from the magnetosphere to the ionosphere. The substorm onset is the transmission of the convection transient. The onset FAC is generated by the near‐earth dynamo. Before the onset, a two‐stage reconnection occurs at the medium‐tail (MTNL) and the near‐earth (NENL) neutral lines through retreating nulls. The accompanying flow explains the model that assumes two causal areas for the onset in the near‐earth and middle tails.

We reproduced the substorm by global simulation and analyzed the development of convection, shear, the dynamo, and the field‐aligned current (FAC) at the final stage of the growth phase. From these investigations, we show that the substorm is a manifestation of reconfiguration in global flow dynamics. Ionospheric convection can be understood from two aspects. One understanding is as the potential field generated by the FAC, and the other is as the projection of magnetospheric convection. In order for the two to coincide, the FAC must be transmitted together with the motion. As a consequence, the resulting convection must be continuous from the magnetosphere to the ionosphere. We see this connection from the drawing of shear to recognize that the substorm is the projection process of transient convection. In the growth phase, convective shear that causes the quiet arc and the Harang reversal (HR) occurs on the open/closed boundary along the surface of the plasma sheet with a continuous flow structure from the magnetosphere to the ionosphere. The onset starts from a more local flow that is induced by the near‐earth neutral line (NENL). A narrow shear commences from the NENL in the mid‐tail and extends to the ionosphere to replace growth phase shear. Along a closed magnetic field line connected to the NENL, direct penetration flow, squeezing flow, the near‐earth dynamo, release of the HR, and the onset FAC occur successively and consequently cause the ground onset. Such onset mechanism is quite different from the model adopting the current wedge.

Japanese space weather research and operation began in 1896 with research on radio propagation. LF, MF, and HF radio communication used to be important before the era of communication satellites; however, they are affected by ionospheric conditions and solar activity. A center for the research and development of radio systems was established at Hiraiso, which is now also a space weather observatory for measuring solar activity, geomagnetic fields, and ionospheric conditions. The need for space weather information initially decreased in the communication satellite age, but since the beginning of the 21st century it has been increasing again with the growing use of positioning satellite systems. The role of Southeast Asian countries in space weather research has become important and the range of activities is also growing. The contribution of these countries to space weather research and operation is expected to grow in future.

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.

With the interplanetary magnetic field (IMF) that turns from northward to southward, the global simulation successively reproduces the growth phase, the onset, and the expansion phase of the substorm. The calculated ionospheric convection for the growth phase reproduces the development of the Harang reversal (HR), the upward field‐aligned current (FAC), and the upcoming onset point as observed. Magnetic field lines traced from the center of the nightside upward FAC are open, while the magnetic field line traced from the onset point is closed. These open magnetic field lines map to flow shear just outside the O/C boundary. Seen from the magnetic configuration, the growth phase proceeds as the replacement process of nulls. Two new nulls appear on the dayside under the southward IMF, while two old nulls under the northward IMF retreat tailward forming two bifurcation regions on the dawn and dusk flanks. Flow shear around the O/C boundary forms by magnetospheric convection that returns to the dayside via bifurcation regions. The expansion phase proceeds through a topological change by the near‐earth neutral line (NENL). The NENL occurs inside the thinned structure of the northward IMF remnant, on the low‐latitude side of the flow shear, and projects down to the low‐latitude edge of the upward FAC. Associated with the NENL, the convection return path changes to the center of the plasma sheet and reveals in the ionosphere as the release of the HR. By the shrinkage of projecting magnetic field line, the O/C boundary migrates poleward in the ionosphere.

Japanese space weather research and operation began in 1896 with research on radio propagation. LF, MF, and HF radio communication used to be important before the era of communication satellites; however, they are affected by ionospheric conditions and solar activity. A center for the research and development of radio systems was established at Hiraiso, which is now also a space weather observatory for measuring solar activity, geomagnetic fields, and ionospheric conditions. The need for space weather information initially decreased in the communication satellite age, but since the beginning of the 21st century it has been increasing again with the growing use of positioning satellite systems. The role of Southeast Asian countries in space weather research has become important and the range of activities is also growing. The contribution of these countries to space weather research and operation is expected to grow in future.

We developed a reliable probabilistic solar-flare forecasting model using a deep neural network, named Deep Flare Net-Reliable (DeFN-R). The model can predict the maximum classes of flares that occur in the following 24 hr after observing images, along with the event occurrence probability. We detected active regions from 3 × 10⁵ solar images taken during 2010–2015 by Solar Dynamic Observatory and extracted 79 features for each region, which we annotated with flare occurrence labels of X-, M-, and C-classes. The extracted features are the same as used by Nishizuka et al.; for example, line-of-sight/vector magnetograms in the photosphere, brightening in the corona, and the X-ray emissivity 1 and 2 hr before an image. We adopted a chronological split of the database into two for training and testing in an operational setting: the data set in 2010–2014 for training and the one in 2015 for testing. DeFN-R is composed of multilayer perceptrons formed by batch normalizations and skip connections. By tuning optimization methods, DeFN-R was trained to optimize the Brier skill score (BSS). As a result, we achieved BSS = 0.41 for ≥C-class flare predictions and 0.30 for ≥M-class flare predictions by improving the reliability diagram while keeping the relative operating characteristic curve almost the same. Note that DeFN is optimized for deterministic prediction, which is determined with a normalized threshold of 50%. On the other hand, DeFN-R is optimized for a probability forecast based on the observation event rate, whose probability threshold can be selected according to users’ purposes.

In this paper, currents causing the sudden commencement (SC), the AU/AL indices, and the positive bay during the substorm are identified from the global simulation and Biot‐Savart's law. Candidate currents assumed as causes of these ground magnetic variations are the ionospheric Hall current, the ionospheric Pedersen current, the field‐aligned current (FAC), and other magnetospheric currents than the FAC. In general, FAC effect and Pedersen current effect cancel out each other under the restriction of Fukushima's theorem. During the SC, for instance, the midlatitude preliminary positive impulse appears in the prenoon and midlatitude preliminary reverse impulse (PRI) appears in the postnoon, due to the remaining effect of the Hall current. However, violations of the Fukushima's theorem are also common such as in the cases of the equatorial PRI, the auroral electrojet, and the positive bay. The equatorial PRI caused by the Pedersen current appears both in the prenoon and postnoon regions. In the auroral region, the Hall current effect prevails over other currents so much and determines the AU/AL indices only from it regardless other currents. The midlatitude positive bay on the nightside is generated by the effect of the FAC. From these diverse reproduction of ground magnetic variations, a further verification is given for the global simulation in reproductions of the magnetosphere‐ionosphere coupling process.

Convection under the due northward interplanetary magnetic field (IMF) is reproduced by the global simulation. The resulting magnetosphere is closed except in the XZ plane and separated from the solar wind by the separatrix generated from cusp nulls. Inside the separatrix, there exist three plasma regimes of the cusp high‐pressure region, the low‐latitude boundary layer (LLBL) and the plasma sheet. In the ionosphere, the northward Bz (NBZ) current and the reverse cell occur in higher latitudes than 80°, and the fun‐shaped arc‐like field‐aligned current and the main oval occur in lower latitudes than 80°. Magnetic field lines in the antisunward flow region of the reverse cell are connected to the LLBL that is accelerated to supersonic flow by the cusp pressure. Circulation on the reverse cell in the ionosphere is as a whole constructed to the interchange cycle in the magnetosphere. Convection is looked upon as the process to discharge stress generated by the dayside cusp reconnection. Magnetic stress generated by the reconnection is first converted to thermal energy in the cusp. This thermal energy is drained through three possible routes: release of plasma downtail through the LLBL, dissipation as electromagnetic energy through formation of the dynamo, and evacuation down to the ionosphere through the plasma sheet.

The transition of magnetic topology is analyzed during the substorm. During the growth phase in the polar ionosphere, magnetic fields consist of three categories: closed magnetic field, open magnetic field leading to the northward interplanetary magnetic field (IMF), and open field leading to the southward IMF. While the open magnetic field region leading to the southward IMF expands consistently during the growth phase, the ionospheric onset starts inside the closed field line region. There should be triple points where three kinds of magnetic field coexist in the ionosphere. Connected to the triple points, complex magnetic topologies appear in the magnetosphere. There occur definitive changes in the global magnetic topology such as retreat of cusp nulls generated under the northward IMF, generation of plasma sheet magnetic fields with reverse curvature, generation of bifurcation areas on the magnetopause, formation of new dayside nulls, development of null lines on the flank, and generation of bifurcation areas (group of nulls) on the magnetopause. During the late growth phase, divergence V_x flow is excited in the central plasma sheet associated with the appearance of reverse curvature. The plasma sheet reconnection starts 5 min before the ionospheric onset from a localized area at X = −17.5 Re due to the reduction of the normal component B_z caused by divergence V_x flow. It occurs in the residual magnetic structure formed under the northward IMF as a three‐dimensional reconnection with a guide field B_y. The resulting two‐turn coil of magnetic field forms a plasmoid, which is ejected downtail.