Dynamics of the terrestrial radiation belts: a review of recent results during the VarSITI (Variability of the Sun and Its Terrestrial Impact) era, 2014–2018

The Earth’s magnetosphere is region that is carved out by the solar wind as it flows past and interacts with the terrestrial magnetic field. The inner magnetosphere is the region that contains the plasmasphere, ring current, and the radiation belts all co-located within about 6.6 Re, nominally taken to be bounding this region. This region is highly dynamic and is home to a variety of plasma waves and particle populations ranging in energy from a few eV to relativistic and ultra-relativistic electrons and ions. The interplanetary magnetic field (IMF) embedded in the solar wind via the process of magnetic reconnection at the sub-solar point sets up plasma convection and creates the magnetotail. Magnetic reconnection also occurs in the tail and is responsible for explosive phenomena known as substorms. Substorms inject low-energy particles into the inner magnetosphere and help generate and sustain plasma waves. Transients in the solar wind such as coronal mass ejections (CMEs), co-rotating interaction regions (CIRs), and interplanetary shocks compress the magnetosphere resulting in geomagnetic storms, energization, and loss of energetic electrons in the outer radiation belt nad enhance the ring current, thereby driving the geomagnetic dynamics. The Specification and Prediction of the Coupled Inner-Magnetospheric Environment (SPeCIMEN) is one of the four elements of VarSITI (Variability of the Sun and Its Terrestrial Impact) program which seeks to quantitatively predict and specify the inner magnetospheric environment based on Sun/solar wind driving inputs. During the past 4 years, the SPeCIMEN project has brought together scientists and researchers from across the world and facilitated their efforts to achieve the project goal. This review provides an overview of some of the significant scientific advances in understanding the dynamical processes and their interconnectedness during the VarSITI era. Major space missions, with instrument suites providing in situ measurements, ground-based programs, progress in theory, and modeling are briefly discussed. Open outstanding questions and future directions of inner magnetospheric research are explored.


Introduction
The term "magnetosphere" was introduced by Gold in 1959 (Gold 1959) to describe the region above the ionosphere and coincided with James Van Allen's discovery of the radiation belts (Van Allen et al. 1959). The solar wind (Parker 1958), flows past the Earth forming the magnetopause which is the outer boundary of the magnetosphere. The solar wind is supersonic and forms the bow shock just ahead of the magnetopause. Co-located with the radiation belts is a region containing low-energy plasma which is home to a variety of plasma waves. The radiation belts comprise an outer belt and an inner belt separated by the so-called slot region. The outer belt is populated mostly by electrons ranging in energy from a few hundred keV to tens of MeV and is highly dynamical in nature varying on time scales ranging from minutes to years (Baker and Kanekal 2008). The inner belt comprises mostly of protons and is relatively stable varying slowly over time scales of years and is presumed to be predominantly the result cosmic ray albedo neutron decay (Hess 1959).
In the recent past major missions such as NASA's Van Allen Probes (Mauk et al. 2014;Sibeck et al. 2012), JAXA's Arase , THEMIS (Angelopoulos 2008), and others vastly improved our understanding of the magnetosphere, particularly the inner magnetosphere which includes the radiation belts and the plasmasphere. New observations such as the discovery of a long-term storage ring or the "third belt" (Baker et al. 2013) to the first direct observation of wave-particle interaction (Fennell et al. 2014;Kurita et al. 2018; are prime examples of new observational in situ measurements enabled by advances in instrumentation. Theoretical understanding of the importance of waveparticle interactions have also enabled understanding the physics behind the particle energization and loss (see, for example, chapter by Bortnik et al. in Balasis et al. (2016)).
The Variability of the Sun and Its Terrestrial Impact (VarSITI) program (Shiokawa and Georgieva 2020) commenced in 2014 with an aim to promote international collaboration in data analysis, modeling, and theory to understand how the solar variability affects the Earth. The program comprises four scientific elements. The SPeCI-MEN (Specification and Prediction of the Coupled Inner-Magnetospheric Environment) element's main goal is "the quantitative prediction and specification of the Earth's inner magnetospheric environment based on Sun/solar wind driving inputs".

Solar drivers of magnetospheric dynamics
It is by now well established that the Sun is the ultimate driver of magnetospheric phenomena; indeed, it is the steady solar wind that carves out the magnetosphere and delineates it from the surrounding interplanetary space. The Sun produces many transient eruptions, of which coronal mass ejections (CMEs), co-rotating interaction regions (CIRs), and interplanetary shocks affect the terrestrial magnetosphere. The consequences, i.e., magnetospheric responses, encompass phenomena such as substorms, geomagnetic storms, and aurore and affect plasma waves and energetic particle populations. The time scales of magnetospheric response, particularly the inner magnetosphere, range from minutes to years (Baker and Kanekal 2008).

Coronal mass ejections
Coronal mass ejections are large eruptions of plasma together with the embedded magnetic field from the Sun that are now understood as major drivers of geomagnetic activity and were first observed as late as the 1970s (Gopalswamy 2016). CMEs propagate at different speeds through the solar wind and often are preceded by an interplanetary shock. They may contain well-ordered magnetic fields, e.g., magnetic flux ropes and clouds (Klein and Burlaga 1982), and the interplanetary shock ahead of the CME is a prime acceleration site for solar energetic particles SEP (Desai and Giacalone 2016). A schematic illustrating the topology of a CME is shown in Fig. 1 (reproduced from Zurbuchen and Richardson (2006)).

Corotating Interaction Regions and High-Speed Streams
A stream interaction region is formed when slow solar wind is overtaken by the fast solar wind which originates from coronal holes. The coronal holes move down to lower heliospheric latitudes during the descending phase of the solar cycle and usually persist for periods longer than the solar rotation period of about 27 days (as seen from the Earth). During these times, the stream interaction regions are somewhat stable and rotate with the Sun and are termed CIRs. Figure 2 shows a conceptual schematic of a CIR (adapted from Owens and Forsyth (2013)).

Interplanetary shocks
Interplanetary shocks are known to accelerate charged particles (Jones and Ellison 1991) and are also responsible for accelerating electrons in the outer zone on very short time scales of the order of a few drift times (Blake et al. 1992). The acceleration and injection are quite distinct ) from either radial transport or wave-particle interactions (see the Section 2.5 section).

The inner magnetosphere 2.2.1 Morphology and structure
Until recently, observations of the Van Allen belts indicated that there are two belts, an outer belt comprised mostly of electrons and an inner belt mostly comprised  Zurbuchen and Richardson (2006)). The interplanetary shock ahead of the CME is followed by a turbulent sheath region. Closed magnetic loop lines have counter-streaming particles and may have amagnetic cloud topology of protons. These two regions were separated by the socalled slot region which was a region that was normally bereft of energetic electrons. Other particle populations include transiently trapped solar particles and somewhat more long-lasting trapped anomalous cosmic rays (Cummings et al. 1993). However, after the launch of the Van Allen Probes, new observations by the REPT (Relativistic Electron Proton Telescope) (Baker et al. 2014a) showed that a third belt or "storage ring" was formed soon after the passage of a CME effectively "splitting" the outer Fig. 2 A conceptual schematic of a CIR (adapted from Owens and Forsyth (2013)). Forward (reverse) shock is shown as a dashed blue (red) line Kanekal and Miyoshi Progress in Earth and Planetary Science (2021)  belt. Figure 3 reproduced from Baker et al. (2013) shows the logarithm of relativistic differential electron fluxes at 3 distinct energy ranges, viz., 3.2 to 4.0 MeV, 4.0 to 5.0 MeV, and 5.0 to 6.2 MeV in panels a, b, and c respectively. The time period is from 1 September to 4 October 2012 with the ordinate showing the L* parameter (for a detailed discussion of L*, McIlwain L value, and radiation belt coordinates, see Roederer and Lejosne (2018)) with the fluxes are color coded according to the color bar at right. After about September 4, when the belts were enhanced, the plasmapause was located at very high L* values (> 4.5) which may have accounted for the stable existence of the newly formed ring. It has been suggested that the electron decay lifetimes which are due to interaction with plasmaspheric hiss inside the plasmapause are long (Thorne et al. 2018). Other competing theories have also been proposed and include rapid storm-time outward ULF wave transport ) and lack of resonant scattering of ultra-relativistic electrons by chorus and hiss (Shprits et al. 2013).
Another hitherto unnoticed morphological feature of the outer electron belt at relativistic and at ultrarelativistic energies is that of the so-called impenetrable barrier (Baker et al. 2014a). The observations of highenergy electron fluxes by the REPT instrument onboard Van Allen Probes showed that over prolonged periods of many years. Figure 4 reproduced from Baker et al. (2014a) shows in a format similar to Fig. 3 color-coded logarithm of differential electron fluxes of energies ranging from 2.0 to 8.8 MeV (panels a through e) for a period ranging from 1 September 2012 to 1 May 2014. It is seen that the electrons do not "penetrate" to L shells lower than 2.8 across all energies for this entire period.
This remarkable feature may be due to extremely slow inward diffusion coupled with pitch angle scattering removing those electrons that reach these low L values. However, there has been some speculation that the "barrier" could also potentially arise from particle scattering by human produced waves by radio transmitters.
Thus, in the recent past, measurements have revealed hitherto unobserved morphological features of the radiation belt in the inner magnetosphere made possible by new missions with state-of-the-art instrumentation

Inter-connectedness and cross scale/cross-energy coupling
The influence of the plasmasphere on radiation belts via wave-particle interactions is well known (see the "Wave-particle interactions" section) and arises due to the plasmasphere being home to plasmaspheric hiss, while outside the plasmapause injection of low-energy electrons and ions from the magnetotail and their anisotropy results in EMIC and chorus waves ; Magnetosphere-Ionosphere Coupling in the Solar System, 2016; See Chapter 9 by Thorne. R., et al. 2016). The plasma waves generated by the low-energy particles (so-called source populations; see Jaynes et al. (2015)) transfer energy to intermediate energy particles ("seed" populations) energizing them. The waves also pitch angle scatter electrons de-trapping them and ultimately leading to particle loss. Thus, a crossscale coupling exists that connects the lowest energy to the highest energy particle populations in the inner magnetosphere, via the intermediary plasma waves. Figure 5 reproduced from  shows schematically the inter-connectedness and cross-energy coupling. The "impenetrable barrier" for relativistic electrons as observed by the REPT instrument on Van Allen Probes (reproduced from Baker et al. (2014a)). At all energies, inward radially diffusing electrons do not penetrate below L ≈ 2.8 That the plasmasphere itself may be affected by solar influences was demonstrated recently by observations from Van Allen Probes of the cold plasma density and electric field in the inner magnetosphere (Thaller et al. 2019). Using long-term observations from Van Allen Probes, Thaller et al. (2019) demonstrated that the outer plasmasphere and the plasmapause boundary were driven by modulations in the solar wind. Furthermore, their findings indicated that these modulations were stronger during the declining phase of the solar cycle, possibly due to the prevalence of CIRs, and were substantial in amplitude. As mentioned earlier, the processes in plasmasphere are intimately connected to the electron populations in the outer radiation belts. Thus, these modulations of the plasmasphere potentially could be important in understanding the driver dependence of outer radiation belt dynamics

Inner magnetosphere observatories
The VarSITI era saw several major space missions that were either directly dedicated to the study of the inner magnetosphere, e.g., NASA's Van Allen Probes (Mauk et al. 2014;Sibeck et al. 2012) and JAXA's Arase , or were able to contribute significantly to scientific questions pertaining to the inner magnetosphere. These include THEMIS (Angelopoulos 2008) and MMS (Burch et al. 2016). The VarSITI era also witnessed the maturation of the paradigm shifting technology of Cube-Sats and associated inexpensive access to space.
In the following, we provide a very brief overview of these major missions and illustrate the CubeSat approach using as an example the CeREs mission .
The Van Allen Probes mission launched in 2012 comprises two identically instrumented spacecraft each (2021) 8:35 Page 6 of 22

Fig. 5
Diagram illustrating the interconnectedness and cross-scale coupling between particles, waves, and underlying physical processes (reproduced from ). The energy and spatial scales are indicated on the ordinate and abscissa, respectively. The processes and coupling are shown as colored arrows carrying a comprehensive suite of instruments measuring plasma waves, magnetic fields, and energetic particles (electrons, ions, and ion composition) over a wide energy range. The goal of the mission was to "provide understanding, ideally to the point of predictability, of how populations of relativistic electrons and penetrating ions in space form or change in response to variable inputs of energy from the Sun. " The mission and its science goals are fully described by Mauk et al. (2014). The ERG project is a Japanese geospace exploration project, which consists of three different research teams: Arase (ERG) satellite, ground-based observations, and modeling simulation. The main science target of the project is understanding of transportation, acceleration, and loss of energetic particles in radiation belts and dynamics of geospace during space storms. As a key observation of the ERG project, Arase was launched in 2016 and has started the regular observations since March 2017. Arase has nine science instruments to measure plasma/particles and field and waves. The orbit of Arase is similar to Van Allen Probes, but Arase measures higher magnetic latitudes up to 40 degrees and higher L-shell up to 10. Overview of the project and science goals are fully described by .
The THEMIS mission was launched in 2007 and comprises five identical spacecraft whose main science goal was the study of substorms in the magnetotail. Each spacecraft carries onboard particle instruments that measure low-energy electrons and ions as well as electric and magnetic fields. While the mission's main focus is on the magnetotail, since the spacecraft travers through the radiation belts, the mission also contributes to the study of the inner magnetosphere. The details of the mission are described by Angelopoulos (2008). Since 2010, two of THEMIS spacecraft orbit around the Moon while the remaining three remain in Earth orbits.

The CubeSat/SmallSat paradigm
In the recent past, CubeSats, which comprise 10 cm cubes called 1U, have moved from being a teaching and academic tool to becoming contenders for doing significant science (National Academies of Sciences, Engineering, and et al. 2016). In particular, low earth orbit (LEO) is well suited for CubeSats, since currently due to their power and volume limitations, data transmission rates are somewhat of an issue. However, innovative new instrumentation utilizing advances in detector technology, electronics, and rad-hard processors have brought the capabilities of instruments onboard major missions in the past to these compact space platforms. For example, the MERiT ) instrument onboard the CeREs ) combines avalanche photodiodes (APD) and solid-state detectors (SSD) to provide electron measurements from a few keV up to ten MeV. Figure 6 shows the CeREs CubeSat with the MERiT instrument at the top. The CubeSat, which was launched December 2018, however soon lost contact with the ground station only about a week into the mission. CubeSats remain a risky proposition while affording the possibility of good science at a low cost. Another 3U CubeSat, the Colorado Student Space Weather Experiment CubeSat, CSSWE has however been successful and made some significant science contributions , such as establishing that energetic electrons from cosmic ray neutrons get trapped in the radiation belts . Other CubeSat missions that are studying terrestrial radiation belts include ELFIN (Shprits et al. 2017), AC6 (Blake andFireBird (Spence et al. 2012). The latter two have studied electron microburst precipitation and have provided information regarding microburst spatial extent  and spectra (Crew et al. 2016). CubeSat missions are being considered for interplanetary regions as well, for example, the CUSP ) CubeSat expected to launch early 2020 will measure supra-thermal and solar energetic protons. Figure 7 shows a synoptic view measurements made by the FireBird (Crew et al. 2016) together with those made by the MagEIS ) instrument onboard Van Allen Probes illustrating the complementary nature of contributions to radiation belt science by CubeSats and major missions.
One of the main advantages that CubeSats bring to the study of inner magnetosphere is the capability to do multipoint measurements. Detailed in situ measurements of particles and waves in the inner magnetosphere have been mostly confined to single point measurements or at most missions comprising two spacecraft. Large major missions are expensive and take many person-years of effort. Cube-Sats, taking advantage of recent developments in both inexpensive launch vehicle systems and instrumentation, can fill this lacuna and provide valuable scientific insights that arise from multi-point measurements, e.g., resolving spatio-temporal ambiguities.

Electron dynamics in the outer belt
The Earth's outer radiation belt comprising mostly of electrons is a very dynamic region with electron intensities varying by orders of magnitude in time scales ranging from minutes to days (see, for example, Kanekal (2006); ) which has been well established by observations going back to the nineties . Figure 8 shows the measurements of energetic electrons spanning nearly a decade by SAMPEX and Polar spacecraft. The dynamic nature of the outer zone is evident from the figure, which shows many electron enhancements as well as loss over a nearly a solar cycle. Furthermore, the global coherent nature (Kanekal et al. 2001) of electron energization is also well illustrated in Fig. 8. Note that SAMPEX was in low Earth orbit while Polar was in a high-inclination, high-altitude orbit during this period. It has long been established that the causative agents for electron energization are high solar wind speeds (Paulikas and Blake 1979;Baker et al. 1989) and a persistent southward component of the interplanetary magnetic field . Electron losses are important as it is currently known that the net flux of energetic electrons is a "delicate balance" between energization and loss processes. Figure 9 illustrates this by showing observations of three storms of similar geomagnetic intensities wherein net energization, loss, and no change in pre-and post-storm electron fluxes are seen.
In the following sections below, we discuss and review the current understanding of electron energization and loss processes in the context of focusing on recent results during the SPECiMEN era.

Electron energization processes
Radial transport Electrons moving radially inward gain energy due to the conservation of the first adiabatic invariant; this process is termed radial diffusion and has been established several decades ago (Falthammar 1965;Schulz and Lanzerotti 1974); similarly outward radial motion tends to decrease particle energy and more importantly to loss via escape through magnetopause. Radial diffusion is caused by magnetic and electric field perturbations (Falthammar 1965), and more recent work has shown that  Kanekal (2006) enhanced ULF wave activity speeds up this process considerably (Elkington et al. 1999;Shprits et al. 2008;Su and et al. 2015). Jaynes et al. (2018) Zhao et al. (2018) demonstrated that inward radial diffusion was responsible for energizing electrons to ultra-relativistic energies, again using data from the REPT instrument.
Another study by Ozeke et al. (2020) used ground-based measurements of ULF waves to characterize the global distribution of the ULF wave power and simulated radial diffusion effects during March 2015 and March 2013 on electron fluxes in the outer zone. They compared their simulation results with observations made by the REPT instrument onboard Van Allen Probes and concluded that ULF wave-driven radial diffusion was an important mechanism for both energization and loss via magnetopause shadowing (see Fig. 4 of Ozeke et al. (2020)).
It is clear from these and other studies that enhanced radial diffusion plays an important role in radiation belt dynamics while energization and pitch angle scattering by wave particle are also an equally significant mechanism.
Wave-particle interactions Wave-particle interactions play an important role in energizing radiation belt electrons. There are two different processes in the wave-particle interactions: (1) drift resonance between drifted electrons and MHD fast mode waves and (2) (2021) 8:35 Page 10 of 22 cyclotron resonance between gyrating electrons and Whistler mode/magnetosonic mode waves. The first process has been recognized as an adiabatic process due to conservation of the first and second invariants, while the second process has been recognized as non-adiabatic process through violation of all invariants.
Since the late 1990s, theories have suggested that interactions through cyclotron resonance causes accelerations of relativistic electrons (Summers et al. 1998), and several observational studies indicate that interactions with Whistler mode chorus contribute to flux enhancement of relativistic electrons. Measurements on the radial profile of the phase space density are essential to discriminate the processes (Green and Kivelson 2004). Several studies carried out prior to the Van Allen Probes mission reported on the observation of a peak in the phase space density inside the outer belt likely arising from internal electron energization. However, detailed and comprehensive observations by Van Allen Probes provided definitive evidence for local electron energization as distinct from energization due to radial transport (Reeves et al. 2013). Reeves et al. (2013) showed growing peak of the phase space density, which indicates that generation of MeV electrons occur inside the outer belt. Thorne et al. (2013) showed variations of the plasma wave data and electron data from Van Allen Probes and compared with the quasi-linear simulation. They concluded that Whistler mode waves actually cause acceleration of MeV electrons.
During the Van Allen Probes era, there are a number of reports that whistler mode waves accelerate electrons and growing peak of the phase space density inside the outer belt (Reeves et al. 2013), and it has been established that internal acceleration through resonance with whistler mode waves should be a key process on the formation of MeV electrons of the outer belt. The elementary process of accelerations has also been identified by both Van Allen Probes and Arase (Fennell et al. 2014;Kurita et al. 2018).
Another important aspect of the wave-particle interactions is understanding the non-linear wave-particle interactions. Whistler mode chorus waves are generated through non-linear interactions with electrons through formation of the electron hole/hill in the phase space (Omura et al. 2009). Evidence of non-linear interactions has been reported by comparing theories (Foster et al. 2017;Kubota and Omura 2018).
Generation of chorus waves and acceleration of relativistic electrons are the result of an interplay among different plasma/particle populations via wave-particle interactions. Cross-energy coupling process ) is a key concept to understand enhancement of the outer belt electrons. Figure 11 (Jaynes et al. 2015) shows a schematic diagram of how different plasma/particle populations contribute to generate of plasma waves and subsequent enhancement of outer belt electrons. These processes occur associated with substorm activities. Sustained substorm activities for a few days driven by the small amplitude of southward IMF are essential to cause the large flux enhancement through this process (Miyoshi et al. 2013).

Rapid injection by IP shocks
Interplanetary shocks can inject energetic electrons into the magnetosphere very rapidly on time scales of a few minutes, i.e., on electron time drift time scales; the most dramatic such an event was first observed by the CRRES spacecraft (Blake et al. 1992). The observations showed very high energy electrons with energies > 13 MeV were injected deep into the magnetospheric slot region ( 2 < L 3). Blake et al. point out that the spectrum continues up to almost 50 MeV. The injected electrons were stably trapped in the inner zone and lingered for many years as observed by the SAMPEX spacecraft (Looper et al. 1994). The process of injection is fairly well understood and has been modeled (Li et al. 1993); (Hudson et al. 1997); Wygant et al. (Wygant and et al. 1994)) and results from the electrons "surfing" a magnetosonic pulse electric field while conserving their first invariant. During the VarSITI era, multiple IP shocks have resulted in rapid injections of electrons energized during their transport from outer regions of the magnetosphere (Foster et al. 2015;Kanekal and et al. 2016;Schiller et al. 2016). The injected electrons often lead to enhanced bunches electrons as they drift around (drift echos). Figure 12 shows the radiation belt electron response to the IP shock event of March 2015 ). The figure shows pitch angle-resolved electrons of several energies as measured by the REPT instrument. Rapid injection followed by velocity dispersed drift echos are seen, with the arrival of the shock indicated by a vertical black line. Electron enhancements are seen to occur within about 5 min of the shock arrival. Due largely to the paucity of measurements, studies relating various shock properties to characteristics of electron injections have not been well established. For example, the March 2015 shock event showed a more harder spectrum (as compared to electron spectrum just prior to injection) than the September 2017 Page 13 of 22 event which was a much stronger shock (Kanekal 2020). A study by Schiller et al. (2016) studied the relationship between shock characteristics and the highest energy of the injected electrons and found that shock mach number correlated best and suggest that shock strength plays an important role. The studies mentioned in the previous paragraph have established underlying physical processes that result in these injections; however, we emphasize no conclusive relationship between IP shock parameters and electron properties such as spectral hardness has been established.
The space weather implications of rapid energization due to IP shocks are significant, since, as the March 1991 event demonstrated, very high-energy electrons can be injected deep unto the magnetosphere and persist for several years. Persistent high fluxes of electrons pose a critical threat to spacecraft.

Solar driver dependence
Relativistic electrons of the outer belt show different responses associated with the solar wind disturbances. As mentioned previously, largescale solar wind structures such as CMEs and CIRs cause geomagnetic storms, and the outer belt electron flux and spatial distributions also change during the period. For geomagnetic storms, intense southward IMF is essential to cause severe magnetic storm through the enhancement of the cross-polar cap potential. On the other hand, magnetic storms do not always cause large flux enhancement of the outer belt electrons (Reeves et al. 2013). Several studies showed some cases of strong outer belt relativistic electron enhancements during non-storm times , and these studies pointed out increasing of chorus wave power during the flux enhancement event. It has been indicated that the high-speed coronal hole stream following CIR is more effective for the large flux enhancement at GEO rather than CME-driven storms and the southward IMF embedd in the high-speed coronal hole stream is important for the large flux enhancement (Miyoshi and Kataoka (2005), (2008); Kilpua et al. 2015). Li et al. (2011) has shown that the solar wind speed is not a necessary condition for the flux enhancements at GEO. These studies indicate the importance of the southward IMF besides the solar wind speed that has been considered as a primary parameter for the outer belt electrons, and prolonged substorm activities are essential to cause the flux enhancement.
From the analysis of 200 stream interface events in solar cycle 23, Miyoshi et al. (2013) have shown that the southward IMF in the high-speed stream that is controlled by the Russell-McPherron effect (Russell and McPherron 1973) affects key parameters of electron acceleration by chorus waves. These parameters include source electron population, thermal plasma density, and chorus wave activity itself. They showed that prolonged substorm activity is essential for the large flux enhancement. Their model as depicted in Fig. 13 (Miyoshi et al. 2013) shows how solar wind speed and the prolonged southward IMF control relativistic electron flux enhancement through wave-particle interactions. More recently,  showed increase of electron PSD associated with solar wind/geomagnetic parameters as well as the important role played by substorms.
It is generally accepted that CMEs predominantly occur during the ascending phase and CIRS during the descending phase of the solar cycle; however, that is not always the case. Kanekal et al. (2015) studied an interesting period during November 2013 when the radiation belts were driven concurrently by weak CMEs during an ongoing HSS event. They used Van Allen Probes measurements of electron intensities to calculate phase space densities (PSD) and found that the lower-energy electrons were accelerated by radial diffusion while the higher-energy electrons were energized by wave-particle interactions. Figure 14 shows the radial profile of PSD obtained using MagEIS and REPT measurements onboard Van Allen Probe A (similar results were found for Probe B). The profiles for two values of K, the second adiabatic invariant and for multiple values of the first, i.e., μ show that at lower values of the latter, the profiles are monotonic whereas at higher values exhibit a peak.
Another recent study by Pandya et al. (2019) compared radiation belt responses to 28(27) CME(CIR) events. They examined the dependence of electron flux variability upon interplanetary parameters such as solar wind speed, IMF B Z . Their findings emphasize the different nature of electron responses to these two types of solar drivers. The distinct response of the radiation belts to solar drivers has also been demonstrated by Baker and et al. (2014c) who studied an event during March 2013 when a high-speed stream was followed by a coronal mass ejection. They showed that this resulted in energization due to prolonged gradual radial diffusion which was followed by a much more rapid enhancement. More recently, Baker et al. (2019) used relativistic electron measurements spanning 6 years from the REPT instrument onboard Van Allen Probes. Their long-term study confirmed the important role played by solar wind speed in energizing electrons, specifically a threshold speed of 500 km/s being required to enhance relativistic electron intensities. Long periods when the solar wind speed did not exceed the threshold showed no detectable high-energy electrons. As the authors point, this fact, despite being known for decades (see, for example, Baker et al. 1979;Paulikas and Blake 1979), still remains unexplained as to the physical processes leading to the existence of such a threshold value.
From these studies, prolonged southward IMF, highspeed solar wind, and low solar wind dynamic pressure The numerical values are listed in the legend accompanying the figures (reproduced from Kanekal et al. (2015) have been recognized as the main contributions to cause the large flux enhancement. These situations are readily found in trailing edge of the stream interface with the toward/away sector polarities in spring/fall through Russell-McPherron effect. During the period, intense chorus waves are observed (Miyoshi et al. 2013;Li et al. 2014).
It is important to note that prolonged substorm activities with/without magnetic storms are essential to cause the large flux enhancements. This is an important paradigm to indicate the need for more accurate space weather forecasting models.

Electron loss processes
Magnetopause shadowing An important mechanism of loss of electrons from the radiation belts is the so-called magnetopause shadowing, where drifting electrons are on trajectories that encounter the magnetopause, and as a consequence, they escape the magnetosphere into interplanetary space. During disturbed times, as for example when a CME compresses the magnetosphere, the magnetopause moves inward and electrons which were on closed drift shells now encounter the magnetopause and escape. Studies suggest electron loss occurs associated with enhancements of the solar wind dynamic pressure, suggesting that the magnetopause shadowing leads to electron loss (Turner and Ukhorskiy 2020). If the phase space density of electrons in the outer part of the outer belt decreases, negative spatial gradient of the phase space density is expected, and subsequent outward radial diffusion takes place (Miyoshi et al. 2003;Shprits et al. 2006), which contributes to electron loss of the whole outer belt. Simulations using the Fokker-Planck equation have assumed that electron phase space density decreases if the last-closed drift L-shell is larger than the dayside magnetopause.
On the other hand, actual trajectories of electrons in the dayside magnetopause are complicated, and drift-shell bifurcation causes the pitch angle scattering. Several test particle simulations have investigated these electron trajectories in detail (Kim et al. 2008, 2010, Saito et al. 2010. As mentioned by Kim and Lee (2014) and Mauk et al. (2016), it is difficult for electrons to cross the field line of the dayside magnetopause because gyro-radius of electrons are too small, and further detailed analysis lb Kanekal and Miyoshi Progress in Earth and Planetary Science (2021) 8:35 Page 16 of 22 considering actual structure of the dayside magnetopause is necessary.
Particle precipitation Pitch angle scattering is another important process to cause loss of relativistic electrons from the outer belt. Whistler mode waves and EMIC waves are important for the pitch angle scattering of relativistic electrons (Millan and Thorne 2007). From late of the 2000s, observational evidence has shown that EMIC waves actually cause the pitch angle scattering of MeV electrons Rodger et al. 2015;Usanova et al. 2014), which has been theoretically predicted in 1971 (Thorne and Kennel 1971). Since EMIC waves are generated by a temperature anisotropy of hot ions, ring current ions affect MeV electrons through wave-particle interactions. Several studies (Kubota and Omura 2017) indicated that rapid precipitations are possible through non-linear interactions between EMIC waves and MeV electrons. Whistler mode waves also contribute to the loss of electrons. Hiss waves cause the pitch angle scattering of MeV electrons, and the flux of MeV electrons gradually decreases. Interactions with hiss waves are a primary loss process of MeV electrons in the slot region. Recent studies  showed that interactions with hiss waves produce reversed energy spectra in the slot region.
Chorus waves also cause precipitation of electrons. The Arase observations identified for the first time that flux enhancements of tens keV electrons inside loss cone correspond to enhancement of chorus waves , which is definite evidence that pitch angle scattering and subsequent precipitation by wave-particle interactions actually occur. Figure 15 reproduced from Kasahara et al. (2018) illustrates schematically the interaction between chorus waves at the equator and bouncing electrons which are pitch angle scattered ultimately into the loss cone and precipitate. Figure 16 also reproduced from Kasahara et al. (2018) shows ERG/Arase measurements illustrating the simultaneous enhancement of wave power (top panel) and increased intensities of nearly fieldaligned electrons clearly indicating the causal connection between the waves and precipitation of these electrons. If chorus waves propagate to higher latitudes, chorus waves can resonate with high-energy electrons and cause the pitch angle scattering of sub-relativistic/relativistic electrons (Miyoshi et al. 2015;Miyoshi et al. 2020), which may contribute tothe loss of MeV electrons. One of the significant phenomena about particle precipitation is the microbursts of relativistic electrons. Combined observations by balloon, FIREBIRD-II, AC-6 Cubesat identified the spatial and temporal evolution of a microburst region (Anderson et al. 2017). The relationship between microbursts and chorus waves as observed by FIREBIRD-II and Van Allen Probes (Breneman et al. 2017) indicates that microbursts of MeV electrons result from non-linear wave-particle interactions between electrons and chorus waves (Horne and Thorne 2003;Saito et al. 2012;Wang and Shprits;Miyoshi et al. 2020).
While it has been confirmed that both Whistler mode waves and EMIC waves cause significant precipitation of energetic electrons from the outer belt, quantitative estimation to discriminate dominant loss process of the outer belt has still been open problem.

Ground-based observations and conjunction studies
During solar cycle 24, multi-spacecraft have observed Geospace, which enable us to realize multi-point coordinated observations. Many conjunction studies among Van Allen Probes, Arase, THEMIS, MMS, and other satellites are achieved. One of the coordinated observations from multi-spacecraft is the radial profile of the phase space density. Boyd et al. (2018) investigated the phase space density profile up to L = 7.5 using both Van Allen Probes and THEMIS satellite, and they showed the peak of PSD beyond apogee altitude of Van Allen Probes. Multi-spacecraft observations could discriminate spatialtemporal variations of particle and field/waves. For example, Kurita et al. (2018) showed loss of MeV electron flux on timescales 30 min using both Van Allen Probes and Arase satellites and argued that EMIC waves cause the pitch angle scattering. Teramoto et al. (2019) identified modulations of MeV electrons by the MHD-fast mode waves occurring over the limited azimuthal longitudes using Van Allen Probes and Arase satellite.
Another great advantage for the inner magnetosphere research is the development of ground-based network observations. Several multi-point network observations have been developed, SuperDARN radars, magnetometers, optical imagers, riometers, and VLF/LF radio wave receivers. There have been many opportunities for conjugate observations between satellites and ground-based observations, which identified latitudinal/local time distributions.

Open questions and future directions
During the VarSITI era, new discoveries (e.g., multiple belt structure) and new observations (e.g., direct observations of wave-particle interactions) advanced the current understanding of the dynamics of the inner magnetosphere, in particular, that of energization and loss of radiation belt electrons. This was made possible by improved instrumentation and dedicated missions such as Arase, Van Allen Probes, MMS, and THEMIS. For example, Arase and van Allen Probes carried comprehensive instrument suites that characterized electric and magnetic fields, plasma waves, and energetic particles.
Nevertheless, many open questions remain, and indeed, some of the recent discoveries may have opened up new frontiers. For example, while the role of wave-particle  Kasahara et al. (2018) interactions in energization and loss has long been recognized (for a review see for example Thorne (2010)), the importance of non-linear wave-wave and wave-particle interactions remains an open question. Recent theoretical and modeling (e.g., see Omura et al. 2015) suggest that they play a crucial role. Another important open question is in regard to electron loss from the radiation belts, viz., the relative importance of magnetopause shadowing and wave-particle precipitation (Turner et al. 2012). Of critical importance is the role of energetic particle precipitation in affecting the chemistry of the upper atmosphere (Turunen et al. 2016), an important open question that needs to be answered regarding climate change. The discovery of the so-called impenetrable barrier (Baker et al. 2014a The need for multi-point measurements in advancing our quantitative understanding of the inner magnetospheric dynamics has long been recognized. This is a key direction in future scientific endeavors. The advent of CubeSats has made this a real and tangible goal. Innovations in space instrumentation and low-cost access to space have further spurred interest in the community, and proposed missions include constellations of Cube-Sats to study radiation belts, ring current, and plasma waves that drive particle dynamics (Kanekal et al. , (2021 Page 18 of 22 2019; Li et al. 2013). These CubeSat missions also play a complementary role to major "flag ship" type missions as the latter often have capabilities that are comprehensive (Fox and . Increasing societal reliance upon space-based assets for communication, Earth observation, and global climate issues are also driving the need for future mission to comprehensively understand the near-Earth environment. Flagship type missions that address the Sun-Earth, inner heliosphere as an integrated system are another important future trend that is being recognized as necessary to understand not only the system-as-a-whole but also the constituent parts such as the radiation belts and ring current.

Conclusions
During the varSITI era, much progress has been made in understanding the dynamics of the inner magnetosphere, and this review has focused particularly on the dynamics of the outer zone electrons. Major missions such as Arase, THEMIS, Van Allen Probes, and CubeSats and high-altitude balloons have contributed to the progress. New discoveries, such a multi-belt morphology, existence of a putative prevention of continued inward motion of energetic electrons (so-called impenetrable barrier mentioned earlier), and direct observation of wave-particle interactions, are some of the significant observational findings during the VarSITI era. In addition, our understanding of the underlying physical processes of electron energization and loss has also advanced considerably, e.g., in delineation of the importance of source, seed populations that lead to energization. Other areas of advances in understanding electron energization include the complementary role played by wave-particle and radial diffusion in producing ultrarelativistic electrons. Conjunctions between spacecraft and ground-based observations have been used to understand particle precipitation and the role of various plasma waves in general in regard to electron energization as well as loss. Contribution to space weather aspects, such as detailed observations of rapid injection of multi-MeV electrons by interplanetary shocks, has re-emphasized the need for continuous monitoring of radiation in geospace.
Yet, despite the impressive achievements during the VarSITI era, not unexpectedly, the new findings and advancements have opened up new vistas with many new scientific questions. This era has also seen the coming-of-age of CubeSats and advances in the miniaturization of space instrumentation, which have for the first time made multi-point in situ measurements possible. We believe that the future is bright for further progress in our understanding of the inner magnetosphere, especially in the light of technological innovations regarding space platforms such as CubeSats, low-cost launches and sophisticated instrumentation.