Volume 45, Issue 23 p. 12,720-12,729
Research Letter
Free Access

Rapid Loss of Relativistic Electrons by EMIC Waves in the Outer Radiation Belt Observed by Arase, Van Allen Probes, and the PWING Ground Stations

S. Kurita

Corresponding Author

S. Kurita

Insitute for Space-Earth Environmental Research, Nagoya University, Nagoya, Japan

Correspondence to: S. Kurita,

[email protected]

Search for more papers by this author
Y. Miyoshi

Y. Miyoshi

Insitute for Space-Earth Environmental Research, Nagoya University, Nagoya, Japan

Search for more papers by this author
K. Shiokawa

K. Shiokawa

Insitute for Space-Earth Environmental Research, Nagoya University, Nagoya, Japan

Search for more papers by this author
N. Higashio

N. Higashio

Research and Development Directorate, Japan Aerospace Exploration Agency, Sagamihara, Japan

Search for more papers by this author
T. Mitani

T. Mitani

Institute for Space and Astronautical Science, Japan Aerospace Exploration Agency, Sagamihara, Japan

Search for more papers by this author
T. Takashima

T. Takashima

Institute for Space and Astronautical Science, Japan Aerospace Exploration Agency, Sagamihara, Japan

Search for more papers by this author
A. Matsuoka

A. Matsuoka

Institute for Space and Astronautical Science, Japan Aerospace Exploration Agency, Sagamihara, Japan

Search for more papers by this author
I. Shinohara

I. Shinohara

Institute for Space and Astronautical Science, Japan Aerospace Exploration Agency, Sagamihara, Japan

Search for more papers by this author
C. A. Kletzing

C. A. Kletzing

Department of Physics and Astronomy, University of Iowa, Iowa City, IA, USA

Search for more papers by this author
J. B. Blake

J. B. Blake

Space Science Applications Laboratory, The Aerospace Corporation, Los Angeles, CA, USA

Search for more papers by this author
S. G. Claudepierre

S. G. Claudepierre

Space Science Applications Laboratory, The Aerospace Corporation, Los Angeles, CA, USA

Search for more papers by this author
M. Connors

M. Connors

Observatories, Athabasca University, Edmonton, Alberta, Canada

Search for more papers by this author
S. Oyama

S. Oyama

Insitute for Space-Earth Environmental Research, Nagoya University, Nagoya, Japan

Ionospheric Research Unit, The University of Oulu, Oulu, Finland

National Institute of Polar Research, Tachikawa, Japan

Search for more papers by this author
T. Nagatsuma

T. Nagatsuma

National Institute of Information and Communications Technology, Koganei, Japan

Search for more papers by this author
K. Sakaguchi

K. Sakaguchi

National Institute of Information and Communications Technology, Koganei, Japan

Search for more papers by this author
D. Baishev

D. Baishev

SHICRA, Siberian Branch of the Russian Academy of Sciences, Yakutsk, Russia

Search for more papers by this author
Y. Otsuka

Y. Otsuka

Insitute for Space-Earth Environmental Research, Nagoya University, Nagoya, Japan

Search for more papers by this author
First published: 30 November 2018
Citations: 25

Abstract

There has been increasing evidence for pitch angle scattering of relativistic electrons by electromagnetic ion cyclotron (EMIC) waves. Theoretical studies have predicted that the loss time scale of MeV electrons by EMIC waves can be very fast, suggesting that MeV electron fluxes rapidly decrease in association with the EMIC wave activity. This study reports on a unique event of MeV electron loss induced by EMIC waves based on Arase, Van Allen Probes, and ground-based network observations. Arase observed a signature of MeV electron loss by EMIC waves, and the satellite and ground-based observations constrained spatial-temporal variations of the EMIC wave activity during the loss event. Multisatellite observation of MeV electron fluxes showed that ~2.5-MeV electron fluxes substantially decreased within a few tens of minutes where the EMIC waves were present. The present study provides an observational estimate of the loss time scale of MeV electrons by EMIC waves.

Key Points

  • The Arase satellite observed loss of MeV electrons in association with the activation of EMIC waves
  • Spatial-temporal variations of the EMIC wave activity are constrained by ground-based and satellite observations
  • Multi-instrument observations show loss of ~2.5-MeV electrons due to EMIC waves occurred within 10 min

Plain Language Summary

This paper shows an observational evidence that electromagnetic ion cyclotron (EMIC) waves cause rapid loss of high-energy electrons in the Earth's radiation belt. The radiation belt contains highly energetic electrons with energies above 1 Mega-electron volt (MeV), and the number of the electrons dramatically changes during geomagnetic storms. Radiation belt electron fluxes decrease orders of magnitude during the early stage of storms, and it is of interest to clarify mechanisms which are responsible for the loss of radiation belt electrons. EMIC waves are considered as an important driver of loss of radiation belt, and theoretical estimates suggest that radiation belt electrons are rapidly lost immediately after generation of EMIC waves. This paper gives observational support for the theoretical prediction using the satellite and ground-based network observations. The Arase satellite and Van Allen Probes measure variations in radiation belt electrons. The excitation timing and spatial-temporal evolution of EMIC waves are determined by using the worldwide ground-based network measurements in concert with the Van Allen Probes and the Arase satellite. We demonstrate that ~2.5-MeV electrons are lost within 10 min after the generation of the EMIC waves, which is much faster than that reported previous observational studies.

1 Introduction

The Earth's outer radiation belt exhibits dramatic changes in relativistic electron fluxes in association with geomagnetic storms. The typical flux change is characterized by decrease in electron fluxes during the main phase of storms, and the decrease is often followed by an increase in electron fluxes during the recovery phase of the storms (e.g., Baker et al., 1994; Miyoshi et al., 2003; Turner et al., 2012). However, previous studies have reported that the degree of the change in flux is different from one storm to the next (e.g., Miyoshi & Kataoka, 2005, 2008; O'Brien et al., 2001; Reeves et al., 2003) and suggested that net flux change of the radiation belt electrons during storms is determined by a delicate balance of acceleration, loss, and transport processes (Reeves et al., 2003).

There are two major loss mechanisms of radiation belt electrons (Turner et al., 2012, and references therein). One mechanism is escape from the magnetosphere through the magnetopause during the enhanced solar wind pressure (so-called magnetopause shadowing; e.g., Li et al., 1997; Matsumura et al., 2011). The magnetopause is compressed earthward by the enhanced solar wind pressure, and electrons drifting across the magnetopause are directly lost from the magnetosphere. The subsequent outward radial diffusion leads to enhanced loss of electrons far from the magnetopause (e.g., Turner et al., 2012). Turner et al. (2014) showed that this process is important for electron loss at higher L-shells, while another loss process is required to explain loss at lower L-shells (L < ~4) during a geomagnetic storm.

The other primary mechanism is precipitation into the atmosphere via pitch angle scattering by plasma waves (e.g., Millan & Thorne, 2007; Thorne, 2010). Whistler mode waves and electromagnetic ion cyclotron (EMIC) waves have been considered as primary candidates to cause precipitation loss of MeV electrons (e.g., Horne & Thorne, 1998; Thorne & Kennel, 1971). The precipitation loss of electrons by plasma waves has been mainly revealed by observations of precipitating electrons associated with plasma wave activity (e.g., Breneman et al., 2017; Lorentzen et al., 2001; Miyoshi et al., 2008). Recent Van Allen Probes observations have shown the effect of EMIC wave-driven loss on trapped relativistic electron populations; Turner et al. (2014) showed that localized loss of MeV electrons at low L-shell can be associated with EMIC wave activity. Usanova et al. (2014) showed that pitch angle distributions of relativistic electrons tend to have a peak at 90° in association with EMIC wave activity, which is consistent with the theoretical prediction that the EMIC waves scatter MeV electrons at lower pitch angles (e.g., Summers & Thorne, 2003). Shprits et al. (2017) suggested that localized loss by EMIC waves results in a deepening minimum in the radial profiles of phase space density, and the minimum associated with EMIC wave activity is revealed by the Van All Probes observations.

There is increasing observational evidence that EMIC waves cause precipitation loss of MeV electrons via pitch angle scattering. However, the rapid nature of EMIC wave-driven loss has not been well demonstrated. Previous studies have shown changes in MeV electron fluxes associated with EMIC wave activity by comparing satellite passes at different times, with the observations separated by several hours (e.g., Shprits et al., 2017; Su et al., 2017; Turner et al., 2014; Usanova et al., 2014). Theory predicts that the time scale for pitch angle scattering by EMIC waves can be as fast as a few minutes (Summers & Thorne, 2003). It is expected that trapped MeV electron fluxes change immediately in association with the activation of EMIC waves.

On 20 December 2016, the Exploration of energization and Radiation in Geospace (ERG, also called “Arase”) satellite was launched and began observations in March 2017 (Miyoshi et al., 2018). In this paper, using Arase measurements along with those from the Van Allen Probes and ground-based network observations, we show that the loss event was characterized by a short time scale for MeV electron flux variations of under 10 min, substantially faster than previously reported.

2 Observations

Observations shown in this paper were made during a weak geomagnetic disturbance associated with the arrival of a high-speed solar wind stream. Solar wind parameters and geomagnetic indices are shown in the supporting information. Prior to the arrival of high-speed solar wind stream, a geomagnetically quiet period had prevailed for a week, and weak substorm activities took place in association with the passage of the stream interface. Loss of MeV electrons was observed in association with the weak substorm activity around 07 UT on 21 March.

Figure 1 shows spin-averaged electron fluxes in the energy range from 700 to 3150 keV measured by the eXtremely high Energy Particle experiment (XEP; Higashio et al., 2018) onboard the Arase satellite, together with the local pitch angle distributions of 700- to 3,150-keV electrons. The flux time series and pitch angle distributions obtained at L < 2.5 are obscured in Figure 1 since the XEP measurements seem to be substantially affected by energetic protons in the inner radiation belt.

Details are in the caption following the image
(a) Spin-averaged electron fluxes in the energy range from 700 to 3,150 keV measured by XEP. (b–i) Pitch angle distributions of 700- to 3,150-keV electrons observed by XEP. (j) L-value of the Arase satellite calculated from the OP77Q magnetic field model. (k) Magnetic latitude of the Arase satellite. The gray shaded area in the plots corresponds to the XEP measurement at L-shell below 2.5, where the measurements seem to be affected by energetic protons in the inner radiation belt.

During the inbound and outbound passes of the Arase satellite from 18 UT on 20 March to 03 UT on 21 March, XEP measured smooth radial profiles of radiation belt electrons. Pitch angle distributions of the electrons were almost isotropic during this time interval in the entire outer radiation belt. During the third traverse of the outer radiation belt from 07 to 08 UT, the XEP measurement shows the sudden decrease in electron fluxes in all energy channels, and pitch angle distributions show peaks at 90°. Comparison of the electron pitch angle distributions between the first (20–23 UT on 20 March) and third traverses indicates that the 90° peaks are formed by the decrease in electron fluxes at lower pitch angles, and the peaks are more pronounced at the higher energy channels. The change in the electron fluxes was observed in the L range from 4.2 to 4.9. As the Arase satellite moved toward L-shell below 4.2 after ~0805 UT, radial profiles of the electron fluxes became smooth again, and pitch angle distributions were almost isotropic, as observed during the first traverse. The change in the electron fluxes is quite similar to those predicted by pitch angle scattering by EMIC waves (Summers & Thorne, 2003). We also computed the equatorial pitch angle distributions during this time, and we found that decrease in electron fluxes at lower pitch angles did result in formation of the 90° peaks observed during the third traverse. An example of the equatorial pitch angle distributions appears in the supporting information.

The ground-based observation network was developed by the “study of dynamical variation of Particles and Waves in the INner magnetosphere using Ground-based network observations” (PWING) project (Shiokawa et al., 2017). Induction magnetometers were installed in ground-based stations, and the EMIC wave activity could be monitored in a global scale. The locations of the PWING stations at Kapuskasing (KAP), Athabasca (ATH), Gakona (GAK), and Zhigansk (ZGN) are represented in Figure 2a. During the analyzed time interval, the apogees of Arase and Van Allen Probes (RBSP in Figures 2a and 2b) were located on the dawnside and duskside, respectively, as shown in Figure 2b. The satellite footprints at 120-km altitude mapped by the Olson and Pfitzer quiet magnetic field model (Olson & Pfitzer, 1977) at 08 UT are shown in Figure 2a. The stations and satellites were located at L ~ 4 at 8 UT on 21 March, and longitudinal distributions of EMIC waves could be constrained from the multipoint observations.

Details are in the caption following the image
(a) Geographic map of the observatory locations and the satellite footprint evaluated from the OP77Q magnetic field model at 08 UT on 21 March 2017. The star and circle symbols indicate the stations and satellite footprints where the Pc1/EMIC waves were observed and not observed, respectively, from 07 to 08 UT on 21 March. (b) Locations of Arase and Van Allen Probes in the solar magnetic coordinate system at 08 UT on 21 March. (c–f) Frequency-time spectrograms of magnetic fields observed by ground-based induction magnetometers at Kapuskasing (KAP), Athabasca (ATH), Gakona (GAK), and Zhigansk (ZGN), respectively. The vertical lines in the spectrograms indicate local midnight. (g) Frequency-time spectrogram of wave magnetic field obtained by the magnetometer in a part of the Electric and Magnetic Field Instrument and Integrated Science (EMFISIS; Kletzing et al., 2013) onboard Van Allen Probe-B. The solid and dotted lines in the spectrogram indicate local gyrofrequency of helium (He+) and oxygen (O+) ions, respectively.

Frequency-time spectrograms of the induction magnetometers at KAP, ATH, GAK, and ZGN are shown in Figures 2c–2f. EMIC waves excited in the magnetosphere can be monitored on the ground as wave activity at the Pc1 frequency range (0.2–5 Hz). Several Pc1 wave activities were observed at the ground stations from 20 UT on 20 March to 12 UT on 21 March. When the change in electron fluxes was observed by XEP from 07 to 08 UT on 21 March, Pc1 wave activity with increasing frequency was observed at KAP, ATH, and GAK. The EMIC wave activity continued from 0710 to 0840 UT at the stations. Similar Pc1 wave activity was also observed at ZGN from 08 to 10 UT. This type of Pc1 wave is called intervals of pulsations of diminishing periods (IPDP) pulsations (Troitskaya, 1961) and is assumed to be caused by the earthward motion of the generation region of EMIC waves in the magnetosphere (Baishev et al., 2000). The substorm activity around 07 UT would be responsible for excitation of the IPDP pulsation since the earthward motion is theorized to be a result of the direct proton injection into the source region of EMIC waves and/or enhanced convection during substorms (Baishev et al., 2000, and reference therein). The filled star (circle) symbols in Figure 2a indicate the locations where the EMIC wave activity was observed (not observed) from 07 to 08 UT on 21 March. The EMIC wave activity is estimated to extend from 20 to 02 MLT through midnight during the time interval. The EMIC wave activity subsequently observed at ZGN suggests that the generation region moved westward as protons, the free energy source for EMIC waves, drifted westward. The frequency-time spectrogram of EMIC waves observed by Van Allen Probe-B is shown in Figure 2g. The solid and dashed lines represent the local gyrofrequency of He+ (fcHe+) and O+ (fcO+), respectively. The EMIC waves first appeared below fcO+ on ~0730 UT. The main wave power was observed between fcHe+ and fcO+ from 0750 UT to 0840 UT, when Van Allen Probe-B traversed the L-shell range from 5.2 to 4.5. Comparison of the EMIC/Pc1 wave frequency at the satellite and the ground-based magnetometers suggests that the Pc1 waves observed on the ground would correspond to the He+ band EMIC waves.

Use of optical measurements is a complementary way to deduce the spatial extent of EMIC wave activity. Energetic proton precipitations are associated with generation of EMIC waves, and precipitating protons are responsible for auroral emissions at 486.1 nm (Hβ), 557.7 nm, and 630.0 nm (Eather, 1968). Proton auroras associated with the EMIC wave activity are identified as detached auroral emissions from the auroral arc at these wavelengths with concurrent Pc1 wave activity (Sakaguchi et al., 2007). Using optical images at these wavelengths, we can constrain the spatial-temporal variation of EMIC wave activity (e.g., Sakaguchi et al., 2012). The PWING stations are equipped with Optical Mesosphere Thermosphere Imagers (OMTIs; Shiokawa et al., 1999; Shiokawa et al., 2009) to measure these emissions, and we can investigate the spatial extent of EMIC waves using images obtained by OMTI.

Figures 3a–3c show keograms in a magnetic north-south meridian at wavelengths of 557.7, 630.0, and 486.1 nm, respectively, derived from the images taken by OMTI at ATH. The images at 557, 630.0, and 486.1 nm are taken every 2 min with exposure times of 5, 30, and 40 s, respectively. The vertical cyan line represents the time at the Pc1 wave onset at ATH (~710 UT). The auroral arc located at the magnetic latitude (MLAT) of 67–68° gradually moved to low latitudes from 0650 UT. The clear detached aurora appeared from ~0720 UT at all wavelengths near MLAT of ~62°, and the detached aurora then moved to lower latitudes (~60° MLAT). This detached aurora associated with the Pc1 wave activity is consistent with proton auroras excited by EMIC-driven proton precipitation (Sakaguchi et al., 2007). The motion of the detached proton aurora corresponds to the motion of the source region of proton precipitations, indicating that the generation region of the EMIC waves moved radially inward during the time interval. The McIlwain L-shell range of the EMIC wave activity is estimated to be 4.3–4.8 based on the Olson and Pfitzer quiet magnetic field model, considering the location of the detached proton aurora.

Details are in the caption following the image
Auroral keograms in a magnetic north-south meridian derived from images taken by Optical Mesosphere Thermosphere Imagers (OMTI) installed at ATH. Keograms of (a) 557.7-nm, (b) 630.0-nm, and (c) 486.1-nm emissions. The emission height is assumed at 120 km for 557.7 and 486.1 nm and 250 km for 630.0 nm. (d) A snapshot image at 486.1 nm obtained by OMTI at ATH at 07:44:14 UT on 21 March, which is mapped on the geographic coordinate assuming the emission height of 120 km. The cyan lines represent geographic latitudes and longitudes, and the magenta lines indicate magnetic latitude.

The longitudinal extent of the proton aurora cannot be evaluated from the keogram. Figure 3d shows a snapshot of the auroral image at 486.1 nm obtained by OMTI at ATH at 07:44:14 UT, which is mapped onto the geographic coordinate. Two bands of auroral emissions can be seen in the auroral image. The band located at the higher latitude corresponds to the auroral oval, and the detached proton aurora is located at latitudes lower than the auroral oval. The proton aurora observed during the event is longitudinally elongated, which is consistent with the longitudinally broad extent of the EMIC wave source region revealed from the magnetometer observations. An auroral movie of 486.1 nm emissions observed at ATH are also available in the supporting information.

The emission at 630.0 nm persistently appeared after 0800 UT without corresponding emissions at 557.7 and 486.1 nm. This emission is equivalent to the Stable Auroral Red (SAR) arc (e.g., Foster et al., 1994) and is related to low-energy (<10 eV) electron precipitations (heat conduction) produced by interactions between ring current ions and plasmaspheric cold electrons along the plasmapause (e.g., Kozyra et al., 1987). Thus, the location of the SAR arc can be a measure of the plasmapause location. It appears that the EMIC waves during this event are present near the plasmapause, where pitch angle scattering of MeV electrons by EMIC waves becomes effective (e.g., Summers et al., 1998 ; Thorne & Kennel, 1971).

As shown in Figure 4a, Arase and Van Allen Probes traversed the outer radiation belt during the activation of EMIC waves. Arase and Van Allen Probe-A were on their inbound and outbound passes during the EMIC wave event, respectively. The Arase observation was followed by the Van Allen Probe-B observation, and their time difference was ~50 min.

Details are in the caption following the image
(a) Location of Van Allen Probe-A (magenta), Van Allen Probe-B (blue), and Arase (black) around the time of interest. (b) Spin-averaged fluxes of ~2.5-MeV electrons observed by Van Allen Probe-A/MagEIS (magenta), Van Allen Probe-B/MagEIS (blue), and Arase/XEP (black) as a function of L-value. (c) Time after the Pc1 wave onset at ATH (0710 UT on 21 March) as a function of location of Van Allen Probe-A (magenta), Van Allen Probe-B (blue), and Arase (black) in L.

Figure 4b shows L-shell profiles of ~2.5-MeV electron fluxes measured by XEP (black) and Magnetic Electron and Ion Spectrometer (MagEIS; Blake et al., 2013) onboard both Van Allen Probe-A (magenta) and -B (blue). Figure 4c represents the relationship between the satellite locations in L-shell and time since the Pc1 wave onset determined by the induction magnetometer. The Van Allen Probe-A observation shows a smooth L-shell profile of ~2.5 MeV electron fluxes throughout the outer radiation belt. On the other hand, the L-shell profile measured by XEP is significantly deformed compared to that observed by Van Allen Probe-A in the L-shell range of 4.2–4.9. The L-shell range where the flux depletion is observed roughly corresponds to the region of the enhanced EMIC wave activity. Following the inbound pass of the Arase satellite, Van Allen Probe-B observed further decrease in ~2.5-MeV electron fluxes. The L-shell range where the flux decrease was observed corresponds to the region of enhanced He+ band EMIC wave activity in the Van Allen Probe-B spectrogram (Figure 2g). The Arase and Van Allen Probes showed that ~2.5-MeV electron fluxes substantially decreased at L > 4.2 within 1.5 hr in association with the enhanced EMIC wave activity.

3 Discussion and Summary

We demonstrated that relativistic electron loss observed by XEP onboard the Arase satellite was likely to be caused by the concurrent EMIC waves. The timing of the EMIC wave activation and longitudinal extent were determined by the Van Allen Probes and the PWING induction magnetometer observations. We further constrained the spatial-temporal variation of the EMIC waves by using the ground-based optical measurements. The spatial extent of the EMIC waves expected from the multipoint and multi-instrument measurements agrees well with the L-shell range where relativistic electron loss is revealed by Arase and Van Allen Probes observations. The observation reported in this paper implies that MeV electrons are being scattered by the EMIC waves, allowing us to make an observational estimate of time scale for loss of relativistic electrons through pitch angle scattering by EMIC waves based on the trapped electron measurements.

We showed that the IPDP pulsations were closely associated with the MeV electron loss observed by XEP, while EMIC wave activities were present during different time intervals. For example, Pc1 waves were observed at GAK and ZGN from 01 to 04 UT on 21 March, and the waves also appeared at KAP, ATH, GAK, and ZGN from ~0430 to ~0700 UT. These waves were observed by the induction magnetometers located at L~4.0. If the EMIC waves are responsible for the loss of MeV electrons, some signature should be identified by Van Allen Probe-A, since the satellite traversed the corresponding L-shell during the EMIC wave activity. However, decreases in MeV electron fluxes are not at an observable level. Theoretical studies have indicated that effective pitch angle scattering of MeV electrons by EMIC waves is expected to occur near the plasmapause and in the plasmasphere (Summers et al., 1998; Thorne & Kennel, 1971). Considering the fact that the EMIC waves can propagate long distances when the waves reach the ionosphere and trapped in the ionospheric duct, it is natural to consider that Pc1 pulsations are not present at the location where the waves can cause effective pitch angle scattering of relativistic electrons.

In Figure 4c, we used the time at the Pc1 wave onset (0710 UT) as a reference to show the time at which the satellite measurements were made at each L-shell. Considering the discussion above, if we instead use the time of the appearance of the detached proton aurora (~0720 UT) as a reference, loss of ~2.5 MeV electrons could occur 10 min faster than estimated from the Pc1 wave activity. Furthermore, considering the motion of the proton aurora to the lower latitude, the activation of the EMIC wave activity at lower L-shell was delayed from that at higher L-shell. Let us roughly estimate time difference between the EMIC wave activation and the Arase traverse. The detached proton aurora reached most equatorward (~60° MLAT, L~4.3) at ~0750 UT, and Arase observed ~2.5 MeV electron fluxes at corresponding L-shell at ~0800 UT. Thus, the observed flux change at L = 4.3 would occur within 10 min, if ~2.5 MeV electron fluxes did not change before the activation of the EMIC waves at L = 4.3. Considering the longitudinal difference of ~180° between the Arase and Van Allen Probe-A measurements and the drift period of an ~2.5-MeV electron at L = 4.3 (~250 s), the estimated time scale would have an ambiguity of ~2 min. The estimated time scale for the flux change is much faster than previously reported based on the trapped electron measurements, and the time scale is consistent with the theoretical prediction.

It is important to investigate the loss time scale of relativistic electrons quantitatively to test theoretical predictions of its value. Loss time scale for pitch angle scattering by EMIC waves can be evaluated using the quasi-linear diffusion theory. On the other hand, quantitative estimates of loss time scale from particle instruments onboard different satellites require sufficient intercalibration between the instruments. So far, calibration of XEP is in an initial stage, and the team has continuously made the best possible effort for that calibration. Comparison between fluxes measured by XEP and MagEIS shows that the electron flux measured by XEP agrees with that measured by MagEIS within a factor of 2 at ~2.5 MeV during the quiet period prior to the EMIC wave event (not shown). The result shown in this paper is thus reliable, while we leave quantitative analysis and comparison between theory and observation as a future work.

Acknowledgments

Science data of the ERG (Arase) satellite used in this study were obtained from the ERG Science Center operated by ISAS/JAXA and ISEE/Nagoya University (https://ergsc.isee.nagoya-u.ac.jp/index.shtml.en). The XEP (L1c v003) and MGF Level 2 v1.00 data were analyzed to generate the electron flux time series and pitch angle distributions. These data will be publicly available from the ERG Science Center based on a project-agreed schedule. The PWING induction magnetometer and OMTI data are available from the ERG Science Center. The Van Allen Probes EMFISIS and MagEIS data can be available from https://emfisis.physics.uiowa.edu/and https://www.rbsp-ect.lanl.gov/rbsp_ect.php, respectively. This study has been supported by JSPS Grant-in-Aid for Scientific Research (15H05815, 15H05747, 16H06286, and 17H06140). This study has also been supported by JSPS Bilateral Open Partnership Joint Research Projects, and the Ministry of Science and Higher Education of the Russian Federation and the Siberian Branch of the Russian Academy of Sciences (project II.16.2.1, registration number AAAA-A17-117021450059-3) and is partially funded by Russian Foundation for Basic Research (18-45-140037; D. B.). Observations made at Athabasca were facilitated by Canada Foundation for Innovation support for observatories there. Work at The Aerospace Corporation was supported by RBSP-ECT funding provided by JHU/APL contract 967399 under NASA's Prime contract NAS5-01072. The database construction for the PWING ground-based instruments is supported by the ERG Science Center (http://ergsc.isee.nagoya-u.ac.jp/) and the IUGONET (Inter-university Upper atmosphere Global Observation NETwork) project (http://www.iugonet.org/).