The First Pi2 Pulsation Observed by China Seismo-Electromagnetic Satellite

: On 2 February 2018, the China Seismo-Electromagnetic Satellite (CSES) ZhangHeng 01 (ZH-01) was successfully launched, carrying on board, in addition to a suite of plasma and particle physics instruments, a high precision magnetometer package (HPM), able to observe the ultra-low frequency (ULF) waves. In this paper, a night time Pi2 pulsation observed by CSES is reported for the ﬁrst time. This Pi2 event occurred on 3 September 2018, and began at 14:30 UT (02:37 magnetic local time), when the satellite was in the southern hemisphere between − 49 and − 13 magnetic latitude (MLAT). Kakioka (KAK) ground station in Japan detected the same Pi2 between 14:30–14:42 UT (23:30–23:42 local time). The Pi2 oscillations in the compressional, toroidal, and poloidal components at the CSES satellite and the H-component at the KAK station are investigated by estimating coherence, amplitude, and cross-phase. We noticed a high degree of similarity between the Pi2 event in the horizontal component at KAK and the ionospheric ﬂuctuations in the compressional component at CSES. This high correlation indicated the magnetospheric source of the Pi2 event. In addition, Pi2 is exhibited clearly in the δ By component at CSES, which is highly correlated with the ground H component, so the Pi2 event could be explained by the Substorm Current Wedge (SCW). This interpretation is further conﬁrmed by checking the compressional component of Van Allen Probe (VAP) B satellite inside the plasmasphere, which, for the ﬁrst time, gives observational support for an earlier model. This ULF wave observation shows the consistency and reliability of the high precision magnetometer (HPM) equipped by two ﬂuxgate magnetometers (FGM1 and FGM2) onboard CSES. both at the ground and in the ionosphere indicates that these oscillations are caused by a common source mechanism. At CSES, δ Bz and δ Bx oscillate exactly in phase for the entire Pi2 interval and their amplitudes are similar. The ﬂuctuations in CSES δ By and their waveform / period seem to match δ H with small phase delay.


Introduction
Pi2 magnetic pulsation is an impulsive nightside ultra-low frequency (ULF) wave and it is considered a sensitive indicator to the onset of the geomagnetic substorms within low-and mid-latitudinal regions. The presence of Pi2s with the substorm onset makes it as an imperative intermedia for understanding the electrodynamic processes occurring in the coupling between magnetospheric and ionospheric regions. The frequency range of this wave is 6.6-25 mHz, i.e., a period of 40-150 s, with an amplitude of 0.2-2.3 nT. Pi2 pulsations have a common period and can be monitor of the geomagnetic field and its variations and pulsations (as it will be investigated in this work) to better understand the physics of the ionosphere, the Van Allen belts, and more [28].
To the best of authors' knowledge, this is the first investigation of Pi2 pulsation made using the CSES satellite. Few studies have addressed the observation of global high-correlation Pi2 on the ground and in the ionosphere and magnetosphere. The authors of [29] show the formation and time evolution of a three-dimensional (3-D) oscillatory current system in the SCW. Our results strongly support the result of [29] using in situ observation of Pi2 in the magnetosphere. Pi2 analysis confirms the consistency and reliability of the high precision magnetometer (HPM) equipped by two fluxgate magnetometers (FGM1 and FGM2) onboard CSES. We use the KAK low-latitude ground station as the reference signal relative to those spectral intensities, coherence, and phase of the ionospheric magnetic field observed at the CSES and Van Allen Probe (VAP) satellites. The paper can be organized as follows.
In Section 2, all data sets are described. The observations are presented in Section 3. In Section 4 we discuss the results. Section 5 presents the conclusions of this paper.

CSES Satellite
The first CSES is also called Zhangheng (ZH-1) in honor of Zhang Heng, a Chinese scientist who lived from AD 78 to 139 and who invented the first seismoscope. The satellite, managed by the Chinese Space Agency, China Earthquake Administration, and Italian Space Agency, has been working for over two years since its launch on 2 February 2018, and it is still in orbit and in nominal working conditions. The main scope of the mission is to search for perturbations of the ionosphere and to study their possible correlation with the occurrence of seismic events. To achieve this goal, the satellite has been equipped with nine payloads that monitor the ionospheric magnetic and electric field, two-particle detectors, and other instruments to analyze the plasma, like LAP (Langmuir Probes) and PAP (Plasma Analyzer Package) to determine the electron density, temperature, and which ionized elements are present in the orbital environment.
CSES-01 is the first satellite of a constellation of at least five satellites, and China and Italy are presently building the second one, i.e., CSES-02. CSES-01 works in two modes: burst and survey, with different parameters and sampling rates. The satellite is set to work in burst mode over China and in a seismic belt around the world, and in survey mode in other regions. The data are downloaded from China when the satellite passes over the country. The orbit is quasi polar and Sun-synchronous, with a coverage return time of exactly five days.
To measure the magnetic field, CSES has been equipped with two fluxgate magnetometers (HPM1 and HPM2), placed on two orthogonal booms, which provide the orientation of the magnetic field while an absolute scalar magnetometer allows us to determine the intensity of the same field and calibrate the fluxgates. These instruments form the High Precision Magnetometer (HPM) instrument. The data are sampled at 60 Hz (however, we analyzed them at 1 Hz). The orientation of the spacecraft is checked by two star-cameras placed on the satellite body. The electron density and electron temperature are measured by a couple of Langmuir probes. An Electric Field Detector (EFD) placed at the end of another boom is dedicated to measuring the electric field vector. To detect possible particles emitted from the active faults, the satellite has been provided with two-particle detectors: The Low Energy Particle Detector and the High Energy Particle Detector; the latter detector has been designed and assembled in Italy by the Italian National Institute of Nuclear Physics (INFN).
The data are elaborated by the CSES team in four levels, starting from the raw level. In this study we processed the Level2, which is the calibrated data reported in the Earth Reference System. The data have been downloaded from the CSES portal (www.leos.ac.cn) where the data are available after a free registration.
CSES data are provided in the WSG84 reference system and coordinates. This reference system is defined in the Earth center with the x-y plane as the equatorial one, where X points toward the Greenwich meridian, Z is parallel to the Earth's rotational axis, and Y is orthogonal forming a right-hand reference system. The origin of the reference system is then translated to the satellite position. To report the magnetic signal to the commonly used NEC frame (i.e., the one used by Swarm satellites), a rotation has been applied. The new coordinates have been calculated by two rotation matrices following the next equation: where λ and ϕ are the geographical latitude and longitude of the CSES satellite at the acquisition time.
As the CSES measures the total magnetic field (including the internal field, crustal field, and external field), we have used the CHAOS-6 model [30] of Earth's geomagnetic field, which was developed using more than 6.5 years of high-precision geomagnetic measurements from the Swarm three satellites, to retrieve the external ionospheric component. The residual fields, obtained from this model, are then transformed to Mean-Field Aligned (MFA) coordinates, poloidal, and toroidal components in the topside ionosphere.

VAP Mission
The twin VAP mission was launched on 30 August 2012, by NASA for its Living with a Star program. The VAP mission has two near-equatorial orbit spacecrafts (VAP-A and VAP-B). Both VAPs, with apogee near L~6, contain several sensors prepared to investigate the Earth's radiation medium [31]. The triaxial fluxgate magnetometers onboard the two satellites are qualified to measure 64 vectors per second [32].
We used Electric and Magnetic Field Instrument Suite and Integrated Science (EMFISIS) [32] to get 1Hz data of the magnetic field. The EMFISIS magnetic data are rotated from the GSM coordinates into mean field-aligned (MFA) coordinates. This decomposition allows the dominant magnetic field wave polarization to be determined as toroidal (azimuthal), poloidal (radial), or compressional (parallel). In this system, the Bx is outward, perpendicular to the averaged magnetic field; By is eastward, perpendicular to the averaged magnetic field; and Bz is northward, along the averaged magnetic field.

KAK Geomagnetic Observatory
We used one-second resolution data from the low-latitude KAK geomagnetic observatory in Japan to examine Pi2 magnetic pulsations as detected on the ground. The geographic latitude, geographic longitude, and magnetic latitude of the KAK is 36.23 • N, 140.18 • E, and 27.18 • , respectively. The KAK L-value = 1.31 and its local time = universal time +9 h.

Supermag Auroral Electrojet Index (SML)
We used 1 min data from the SML index to identify the onset of the substorm. The SML index is the equivalent of the AL index as identified from SuperMAG derived auroral electrojet indices. SuperMAG is a worldwide collaboration of organizations, and the derived indices are determined from approximately 110 stations. The authors of [33,34] described the substorm identification technique.  . Both the KAK observatory and the CSES spacecraft were near and within 2.5 h of local time from each other. The magnetic data from the simultaneous availability of CSES and KAK give an opportunity to study, for the first time, the amplitude and phase of the Pi2 pulsation in the ionospheric magnetic field fluctuations at the CSES altitude. We examine the relationship between the Pi2 magnetic oscillation in the topside ionosphere and the Pi2 pulsations at a low-latitude region on the ground. Figure 2a shows the time series of KAK H while Figure 2b shows the Bx, By, and Bz oscillations observed at the CSES satellite. The zero-phase band pass filter (6.6-25 mHz) time series demonstrates that the Pi2 oscillates in all three components at the CSES satellite. The vertical dashed lines point out the crests of the H component. Figure 2c shows the magnitude scalogram of the compressional component of the CSES. give an opportunity to study, for the first time, the amplitude and phase of the Pi2 pulsation in the ionospheric magnetic field fluctuations at the CSES altitude. We examine the relationship between the Pi2 magnetic oscillation in the topside ionosphere and the Pi2 pulsations at a low-latitude region on the ground. Figure 2a shows the time series of KAK δH while Figure 2b shows the δBx, δBy, and δBz oscillations observed at the CSES satellite. The zero-phase band pass filter (6.6-25 mHz) time series demonstrates that the Pi2 oscillates in all three components at the CSES satellite. The vertical dashed lines point out the crests of the δH component. Figure 2c shows the magnitude scalogram of the compressional component of the CSES.

Pi2 Event at KAK and CSES
First, we note that the KAK δH displayed seven cycles of a quasiperiodic fluctuation starting at 14:30 UT. CSES sees the same seven cycles in all magnetic field components. The waveform of the magnetic field in δH at KAK is very similar to δBz at CSES without significant phase delay; however, the amplitude of KAK is larger than δBz by a factor of ∼3. This identical waveform reordered both at the ground and in the ionosphere indicates that these oscillations are caused by a common source mechanism. At CSES, δBz and δBx oscillate exactly in phase for the entire Pi2 interval and their amplitudes are similar. The fluctuations in CSES δBy and their waveform/period seem to match δH with small phase delay. Remote Sens. 2020, 12, x FOR PEER REVIEW 6 of 17 First, we note that the KAK H displayed seven cycles of a quasiperiodic fluctuation starting at 14:30 UT. CSES sees the same seven cycles in all magnetic field components. The waveform of the magnetic field in H at KAK is very similar to Bz at CSES without significant phase delay; however, To confirm the similarity between KAK on the ground and CSES in the topside ionosphere, spectral dissection in the frequency domain was performed. Figure 3 shows the time series and spectral Remote Sens. 2020, 12, 2300 7 of 16 properties of Pi2 observed by KAK and CSES during the 10 min interval of 14:30-14:40 UT. Note that when the coherence was higher than 0.6, then the cross phase was plotted.
shape, but the Bz power spectrum is smaller than that of H. The Bz-H coherence (Figure 3g) is more than 0.6, and the cross phase is near to 0° (Figure 3j).
The Bx spectral dissection (Figure 3e) also points out an enhancement at ∼6.5 mHz, and the intended Bx− H coherence (Figure 3h), estimated for 10 minutes, is more than 0.6 with a phase delay of 180° (Figure 3k). Figure 3f shows that the power spectrum of By is similar to that of Bz near 6.5 mHz, and the By power is quite similar to the H power. The By− H calculated coherence ( Figure  3i) shows a higher spectrum than 0.6 at ∼6-7.5 mHz, but quickly reduces to less than 0.2 at frequencies of ∼8-25 mHz.

VAP Observations
Here we will examine the Pi2 event using data from the VAP magnetospheric mission. During the time interval of the Pi2 pulsation (14:30-14:40 UT), VAP-A and VAP-B satellites were located in the night side (01:49 and 22:06 MLT, respectively). The radial distances at the onset of the substorm were 4.62 RE and 5.35 RE for VAP-A and VAP-B, respectively. Both the VAP-A and VAP-B orbits were visualized onto the solar magnetic x-y plane for the 6 h intervals, and they are shown in Figure  4a and 4b, respectively. To examine if both satellites are inside or outside the plasmasphere, we checked the electron number density (Ne). Figure 4c and 4d show the Ne as being calculated from As expected from Figure 2, the spectral dissection for the δH and δBz components (Figure 3d) shows a dominant spectral enhancement centered at around 6.5-7.5 mHz with a similar spectral shape, but the δBz power spectrum is smaller than that of δH. The δBz-δH coherence (Figure 3g) is more than 0.6, and the cross phase is near to 0 • (Figure 3j).
The δBx spectral dissection (Figure 3e) also points out an enhancement at ∼6.5 mHz, and the intended δBx−δH coherence (Figure 3h), estimated for 10 minutes, is more than 0.6 with a phase delay of 180 • (Figure 3k). Figure 3f shows that the power spectrum of δBy is similar to that of δBz near 6.5 mHz, and the δBy power is quite similar to the δH power. The δBy−δH calculated coherence ( Figure 3i) shows a higher spectrum than 0.6 at ∼6-7.5 mHz, but quickly reduces to less than 0.2 at frequencies of ∼8-25 mHz.

VAP Observations
Here we will examine the Pi2 event using data from the VAP magnetospheric mission. During the time interval of the Pi2 pulsation (14:30-14:40 UT), VAP-A and VAP-B satellites were located in the night side (01:49 and 22:06 MLT, respectively). The radial distances at the onset of the substorm were 4.62 RE and 5.35 RE for VAP-A and VAP-B, respectively. Both the VAP-A and VAP-B orbits were visualized onto the solar magnetic x-y plane for the 6 h intervals, and they are shown in Figure 4a  Modifications in Acknowledgments by red color: E.G. would like to express his gratitude to the Egyptian Government for providing him a fellowship for visiting ICSWSE, Kyushu University, Japan. The geomagnetic field data from KAK were provided by Kakioka Magnetic Observatory (http://www.kakioka-jma.go.jp/). We gratefully acknowledge SuperMAG for SML index data (http://supermag.jhuapl.edu/) and Van Allen Probes Science Operations Centers located at University of Iowa (http://emfisis.physics.uiowa.edu) for Van Allen Probes magnetic data. W.K. is acknowledged for deriving electron density detected by Van Allen Probes. This work made use of the data from CSES mission, a project funded by China National Space Administration (CNSA) and China Earthquake Administration (CEA) in collaboration with Italian Space Agency (ASI) and Istituto Nazionale di Fisica Nucleare (INFN). We acknowledge also LIMADOU-Science Project funded by ASI and the project "The electromagnetic data validation and scientific application research based on CSES satellite" funded by ISSI-BJ under which some of the coauthors make active research. This work was supported by JSPS KAKENHI Grand Number JP20H01961 and JP15H05815. At the onset time of the Pi2, VAP-A was near the inner edge of the plasmapause. Thus, the Pi2 event occurred when VAP-A was inside the plasmasphere. VAP-B was at first in the plasmasphere from 13:00 UT to ∼15:50 UT, and after that it ran across a severe Ne decrease from ∼15 cm −3 to ∼2 cm −3 at L = ∼6.2 before midnight (∼23:00 MLT). This decrease in density is attributed to an outbound plasmapause crossing. VAP-B ran across an abrupt Ne increase from ∼2 cm −3 to ∼13 cm −3 , which is due to an inbound plasmapause crossing, at about 17:03 UT (L = ∼5.9 and MLT = 23:55). From this Ne information, we can emphasize that VAP-B was inside the plasmasphere when the Pi2 pulsation was observed at the ground station.

Discussion
This paper reports, for the first time, the observation of a Pi2 event from the CSES satellite in the topside ionosphere. In previous studies and observations [35][36][37], the 3-D fluctuation current system of SCWs have been investigated. The authors of [38] reported the distribution of the two-dimensional

Discussion
This paper reports, for the first time, the observation of a Pi2 event from the CSES satellite in the topside ionosphere. In previous studies and observations [35][36][37], the 3-D fluctuation current system of SCWs have been investigated. The authors of [38] reported the distribution of the two-dimensional (2-D) substorm system structure using AMPERE satellite data. According to [10,39], the first deflection in a Pi2 wave could be a proxy for the signature of the substorm current wedge, and the Pi2 wave displays deflections at the sub-storm onset [40]. This Pi2 pulsation initial deflection could be the signature of the onset time of the substorm current wedge [29].
The oscillations of the KAK δH in a low latitude and the δBy component in the topside ionosphere are in agreement with the SCW model, but at the same time, the δBy component displayed some phase shifts exhibited after the Pi2 onset (see Figure 2b). On the other hand, a 90 s phase shift during the Pi2 event is shown between the δBz component oscillations at VAP-B in the plasmasphere with both the δBz component at CSES and at the KAK δH (see Figures 7 and 8). As shown in Figures 3, 7 and 8, the Pi2 wave in the KAK δH component has the same spectral peak at~7 mHz, with few phase differences, and with both δBz components at CSES and VAP-B. The FLR model is not likely considered as a source of a Pi2 event because the frequency of a spectral peak was not different at the magnetosphere, ionosphere, and on the ground, so the driven Alfvénic waves generated by the fast wave mode is likely to be considered as valid. This behavior could be interpreted by the SCW, and we assume that the fast mode wave is the reason for the exhibition of the Pi2 wave at the KAK δH component.
The SCW source mode and cavity mode have not been verified by several studies, while a few of them showed a coherent relationship. The authors of [41] showed that the behavior of SCW was detected in conjugate stations at a high latitude with plasma flow vortices of up to 900 km/s. These observations were in the premidnight region at the radial distance between 9 and 12 RE and generated the FAC of the SCW at the substorm expansion phase onset time. The authors of [42] showed that the compressional pulse could modulate the SCW. Returning to Figure 1a, we notice that the substorm activity associated with our Pi2 pulsation was quite low. In the onset time of the Pi2 event, SML was −25 nT. At the same time, the positive bay at KAK H was detected (see Figure 1b). It suggests that the SCW could be sufficient for exhibiting the magnetic bay signature in our Pi2 study. Figure 2a,b show that the compressional oscillation detected at CSES in the topside ionosphere was observed at the same time with the horizontal component identified on the ground without time delay. Figure 2b suggests that the first two cycles of the δBy component fluctuations were due to the SCW pulses. Power spectrum density almost has a similar pattern between δH and δBx/δBz, but not for δH and δBy (Figure 3a-c). It suggests a different generation mechanism between δBx/δBz and δBy.
During the Pi2 event, CSES moved rapidly from middle-latitude to the near-equatorial region ( Figure 5). Magnetic perturbations of northern and southern FACs of SCW are almost canceled out in the equatorial latitude. The decay of CSES δBy amplitudes was more rapid than those of δBx/δBz (Figures 2 and 3). The higher frequency part of the δBy power was more attenuated than those of δBx/δBz. It could be due to the integrated effect of the fine-structured FACs [29]. This is consistent with what has been schematically illustrated by [11]. This scenario in the magnetosphere-ionosphere coupling system proposed that the Pi2 wave current system generated an AC circuit in this system [43].
The authors of [11] developed the magnetohydrodynamics (MHD) model to describe wave propagation. This model estimates the time delay from the satellite to the conjugate ionosphere as 50 s. In our event, no time of flight is observed between the low latitude station and the ionosphere (Figure 2a,b). Observation shows that the periodicity characteristics of the dominant pulses between δH and δBy corresponded to one another, although δH and δBy have different decay patterns and spectrum signatures. This is a manifestation of global high-correlation Pi2 waves [11,29].
The same model of [11] also estimated the time of flight of the fast wave propagation from the near-Earth plasma sheet to Earth. They found that it was~60 s for the paths between the THEMIS satellites and equatorial ground station, Ancon (ANC, geomagnetic latitude = −2.1 • , geomagnetic longitude = 355.6 • ). The time of flight of the fast wave, identified in our event between the VAP-B satellite and KAK station, is 90 s (Figure 7). However, the displayed wave at the CSES was delayed by 90 s compared with those displayed in the magnetosphere (Figure 8). This lateness could be referring to the Alfvén wave propagation along the SCW [29]. Therefore, it is well known that the Pi2 oscillation is triggered first from the magnetosphere to the ionosphere by the Alfvén wave, which is considered responsible for FACs evolution [44,45], along with the initial SCW signal [29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44][45][46][47].

Conclusions
In this work, we report a Pi2 event observed by CSES satellite on 3 September 2018, at 14:30 UT. To the best of the authors' knowledge, this is the first investigation of a Pi2 pulsation made using the CSES satellite. The notable key findings of this study are the following: (1) Pi2 observation and analysis confirm the consistency and reliability of the high precision magnetometer (HPM) equipped by two fluxgate magnetometers (FGM1 and FGM2) onboard CSES. Van Allen Probe B satellite inside the plasmasphere, which, for the first time, gives observational support for an earlier model of the three-dimension oscillations of the SCW associated with the Pi2 event as studied by [29].
With the interpretation of magnetic field data from different space sources in the magnetosphere, the ionosphere can help to better understand the propagation of Pi2 magnetic pulsation. Simultaneous Pi2 events recorded in several locations of the geospace are necessary, and LEO missions such as the CSES and Swarm satellites in the topside ionosphere, integrated with VAP, MMS, ERG (Arase), and GOES observations in the inner magnetosphere, as well as THEMIS observations in the plasma sheet, are expected to resolving many Pi2 issues that are still unclear. Moreover, the integration of space-based satellites with ground-based measurements can explain how Pi2 pulsation spread spatially.