Introduction

The subauroral zone is located between the projection of the position of the plasmapause along the magnetic field lines of the geomagnetic field and the equatorial boundary of the auroral oval. During increasing geomagnetic activity, the position of the subauroral zone changes rapidly. Phenomena observed in the subauroral zone are unique. For example, it is precisely in the subauroral region where narrow streams of fast westward subauroral ion drifts occur near the projection of the plasmapause at altitudes within the F-layer of the ionosphere. These streams are known as polarization jets (PJ) or subauroral ion drifts (SAID)1,2 and are most noticeable during storms/substorms against the background of large-scale plasma convection. The decrease in plasma density in the ionospheric F-layer within PJ/SAID significantly affects the conditions for the propagation of shortwave radio emissions, indicating the practical importance of studying PJ/SAID.

Various studies involving different space missions, numerical modelling, and radar research have allowed for the identification of key large-scale features in the formation and development of PJ/SAID3,4,5,6,7,8,9,10. Previous studies have also helped to determine the time and location where subauroral ion drifts are most likely to occur and establish their connection with geomagnetic disturbances. In recent studies using data from the Norwegian microsatellite NorSat-1, equipped with a high-frequency Langmuir probe system, the small-scale structure of PJ/SAID has been investigated the first time11,12,13. Thus, it is now well-known that within PJ/SAID, there exist plasma irregularities of various scales that can significantly impact the propagation of radio waves, radio communication, and cause scintillations in navigation signals.

Since it is now known that PJ/SAID can be a source of various inhomogeneous structures in the ionosphere, it became important to understand whether it is possible to trace their effects on the quality of trans-ionospheric radio signals, and hence on the positioning with the Global Navigation Satellite Systems (GNSS), and to separate these effects from those originating from irregularities caused by the energetic particle precipitation during auroras, especially in the southern hemisphere. The polar and subpolar regions in the southern hemisphere, particularly around the Antarctic, are the least studied regions to date. This is mainly because the influence of various ionospheric effects on the propagation of radio signals, including those from navigation satellites (GPS, GLONASS, Galileo, BeiDou), has been studied primarily for the northern hemisphere, where there is a dense network of ground-based receivers, radars, ionosondes, and other instruments. Recently, the pace of development of international and regional programs for the exploration of Antarctica has been increasing, along with the need to ensure the reliable functioning of radio-technical systems in the Antarctic region. It is important to note that the subauroral region is located approximately near the coast of Antarctica, where there are a significant number of polar stations (some with airstrips for delivering supplies and personnel to Antarctica) and where there is relatively dense ship traffic. Therefore, it is necessary to understand the processes occurring in the ionosphere in order to avoid or predict potential issues with radio communication and the degradation of positioning accuracy in navigation systems.

In this study, we explore for the first time how strong turbulent flows occurring in the subauroral ionosphere of the southern hemisphere during the PJ/SAID development affect slips (i.e., block of radio signals from the GNSS satellites) in satellite-based navigation systems and lead to scintillations in the received signal on the ground. To achieve this, a unique multi-instrumental approach is employed involving a large number of instruments and data sources. The uniqueness of the considered case is that at that moment of time all instruments were working, and we were able to comprehensively study the geomagnetic disturbance and subsequent scintillations and loss of signals from the GNSS satellites. We used data from three space missions (NorSat-1, DMSP, Swarm), data from two receivers installed in Antarctica, and data from magnetometers. Such a rich set of tools made it possible to study in detail the effects in the subauroral region during the presence of PJ/SAID and its impact on GNSS operations, as well as obtaining new results relevant not only to the Antarctic but also to the Arctic region since many processes occur simultaneously in the polar zones during varying geomagnetic activity.

Main

We chose conditions close to the spring equinox on March 18, 2018, during a period of low solar activity, when a moderate geomagnetic disturbance was observed with a maximum of Kp = 6, SYM-H =  − 53 nT and Auroral Electrojet Index (AE) was up to 1600 nT. In the evening local sector there was an outstanding combination of instruments available that could identify the presence of PJ/SAID and this was the determining factor for focusing on this event. A significant feature of this work is the rich amount of satellite data available, which has shed light on important details associated with structures in the subauroral region. Figure 1 presents UV snapshots of auroral emissions from the SSUSI instrument installed onboard DMSP satellite F-18 when it entered a new orbit from the south pole at 20:45 UT towards the equator. In Fig. 1, the trajectory of the F-18 satellite is shown with black line and the part with PJ/SAID is marked with a thick line. Figure 1 also shows in a similar manner the overflights of other LEO satellites: Swarm B, C and DMSP F-17, as well as NorSat-1 (colour coded). More details on how the PJ/SAID has been identified from the measurements is given in the Methods section. In the middle of the figure, we indicate locations of two GNSS Ionospheric Scintillation and TEC Monitor (GISTM) receivers at the Troll (TRL) and SANAE IV (SAN) research stations and two magnetometers at the SANAE IV and Neumayer stations (shown with the red, black and white stars respectively). Phase scintillation indices σϕ observed with the GISTM receivers are presented at locations corresponding to ionospheric piercing points at the height of 350 km (circles for TRL receiver and squares for the SNA receiver), and σϕ intensity is colour-coded according to the colour bar scale. The values of the phase scintillation index are given for the moment of time 21:25 UT, which is four minutes after the PJ/SAID was detected according to data from the NorSat-1 satellite (more images, as well as videos showring the dynamics of changes of the indices are given in the Supplementary Information). Since the SSUSI images are taken dynamically (about 20 min of shooting is required), the image of the boundaries of the precipitating particles in the UV range shows the average location of the diffuse and discrete auroras. However, even now it is clear from this figure that the area of increased values of the phase scintillation index up to 0.6 rad coincides with the area of the intense discrete aurora, as well as on the equatorial boundary of the diffuse aurora, where, according to satellite data and theory, the PJ/SAID is located. During detailed analysis of plots and results, our attention was focused on selected GNSS satellites (Troll station, GLONASS system: PRN 45 at 21:23 UT, PRN 49 at 21:50 and 21:52 UT, PRN 46 at 22:39, 22:46 and 22:50 UT; SANAE IV station, GPS system: PRN 1 at 21:30 UT, PRN 14 at 22:59 and 23:01 UT and the Galileo system at TRL, PRN 73 two hours later at 01:19 UT), that showed extreme values of the phase scintillation index. In the usual practice of scintillation data analysis such values should be removed from the data set because the corresponding lock time values are less than 240 s. However for us, that means that at some particular moment, the receiver cannot track the satellite. In our case, such satellites were united by the fact that they were: (1) lost by the receiver and the positioning was disrupted, such that a new tracking count started; (2) in the area where the PJ/SAID was located; (3) lost at low elevation angles (less than 25°). Note that in order to separate different satellite systems on one plot we use PRN + 0 for GPS, PRN + 40 for GLONASS, and PRN + 70 for Galileo systems.

Fig. 1
figure 1

Multi-instrumental view of the PJ/SAID influence on GNSS satellites signal phase scintillations at 21:25 UT. UV snapshots of auroral emissions in geomagnetic coordinates from the SSUSI instrument installed onboard DMSP satellite F-18 that flies over the area of Dronning Maud Land (Queen Maud Land), Antarctica during the geomagnetic event on 18.03.2018. The passages of LEO satellites are shown by lines in different colours, in accordance with the legend, and the thick part of the lines indicates the position of the PJ/SAID. The stars show the Antarctic research stations Neumayer (grey), SANAE IV (black), and Troll (red). The corresponding circles show the field of view of the receivers in SANAE IV (black) and Troll (red) at an altitude of 350 km for the elevation angle of 10°. The orange line shows the approximate boundary between the auroral and subauroral zones in the ionosphere. The phase scintillation index values are shown at 21:25 UT in colour according to the colour scale below the figure, with circles for the observations with the TRL receiver and squares for observations with the SNA receiver. Note that the F-17 flyby (shown with red dashed line) is a geomagnetic projection of its associated flyby in the Northern Hemisphere.

Full size image

The analysis of other times and days showed that satellite PRN 1 of the SANAE IV receiver has a constant phase tracking failure in the same place. Which, apparently, may be associated with local features of the conditions near the receiver. The same is true for the PRN 49 satellite of the Troll receiver, where approximately in the same place and at times shifted due to the orbits of the satellites, a tracking failure was also observed. This again indicates interference associated with the area around the receiver, rather than with the state of the environment along the signal path from the GNSS satellite to the receiver on the ground. The remaining four satellites (PRN 45, 46, 14 and 73) indicate a disruption associated with changes in the ionosphere. In addition, in the area where the tracking of satellite PRN 45 is disrupted, after 1.5 h the tracking of satellite PRN 14 also fails.

When we look at the dynamics in detail, we see that at 21:08 UT the phase scintillation indices σϕ begin to increase in the line parallel to the geomagnetic isoline approximately at 62–63° S MLAT above the receivers (see details in Supplementary Information). The values of the σϕ more than 0.45–0.5 rad14 are considered moderate and indicate the presence of an intense scintillation event. Values exceeding 1 rad are considered as strong scintillations. As shown in the work14, strong scintillations can lead to positioning errors of the order of 10 m. Rapid phase changes (scintillations) can induce a Doppler shift in the GPS signal, leading to a loss of phase lock15. The enhancement of σϕ above stations and more poleward is associated with the flow shear instabilities that are modulated by the auroral particle precipitation, as shown by Skjæveland et al.16 who also considered this event. It is noteworthy that the amplitude scintillation index S4 is more sensitive in the zone of discrete aurora, but we also observe increased values S4 from satellites passing through the zone of the equatorial boundary of a diffuse aurora where PJ/SAID is usually observed. At 21:17 UT, according to magnetometer data at SANAE IV and Neumayer (see Fig. 6 in Methods), a sharp increase in the northern component N of the magnetic field begins, and at the same time the values of σϕ above the receivers increase by more than 0.8 rad (PRN satellites 14 = 1.67 rad, 3 = 1.49 rad, 32 = 0.88 rad, 31 = 1.63 rad, 47 = 1.72 rad, 46 = 1.77 rad, 75 = 2.03 rad, 12 = 0.82 rad). In the next few minutes, even more polar satellites show values σϕ > 1 rad. A subsequent decrease in the N component at 21:20 UT at both magnetometers indicates an increase in the westward electrojet associated with increased electric field or conductivity. The behaviour of the S4 index is not so strongly influenced by the expansion phase of the substorm. However, we observe an increase in the S4 value from the satellites passing through the edges of the auroral oval. This generally corresponds to the statistical study of Spogli et al.17, where it was shown that the amplitude scintillation index gets enhanced in correspondence with the boundaries of the auroral oval. Thus, satellite PRN 45 already had high values σϕ and S4 even before the loss of lock at 21:23 UT (elevation angle α = 13.4°). The GPS PRN 1 satellite data obtained with TRL and SNA receivers also show an increase in S4 in the position of the PJ/SAID (α is about 13°). The PRN 46 satellite had low σϕ index values of approximately 0.25 rad and a bit higher S4 index of about 0.27 before entering the expected PJ/SAID zone. The latest satellite measurements confirming the presence of a PJ/SAID are by Swarm B at 22:20 UT. Then, in the PJ/SAID passage zone, communication with satellite 46 is lost 19 min later at 22:39 UT (α = 23.49°). The loss of lock is occurring again at 22:46 UT (α = 19.83°), and in the period from 22:50 (α = 17.76°) to 22:54 UT (α = 15.7°). Afterwards, although the minimum cut-off elevation angle of 10° chosen in the analysis had not yet been reached, communication with the satellite was no longer established. GPS satellite 14 shows an increase in the values of the S4 and σϕ indices after 22:35 UT as the satellite approaches the area where the PJ/SAID was detected. After passing through a slightly disturbed region at 22:59 UT there was a loss of lock for the L1CA signal from the SANAE IV receiver (α = 12.42°). Later at night Galileo satellite 73 experienced three times the loss of signal lock at 01:19, 01:23 and 01:28 UT, corresponding to approximately − 57°–58° MLAT (α = 15.85°, 14.40°, 12.59°). Measurements from the driftmeter as well as UV images from SSUSI onboard DMSP F-17 and F-18 satellites confirm the still stable presence of the PJ/SAID structure around − 60° MLAT at 01:56 and 02:00 UT, respectively. This indicates that the structure is long-lasting, dynamic (with poleward and equatorward movements), and also varying in thickness from 1° to 3°. All satellites considered experienced an increase in the phase and amplitude scintillation indices, as well as loss of signal lock in areas where their signal passed through the PJ/SAID structure. This again indicates the strong influence of the PJ/SAID on GNSS signals.

Discussion

In this work, we studied how the PJ/SAIDs affect slips, phase and amplitude scintillations of the received radio signal from GNSS satellites in the subauroral region in the southern hemisphere near and over Dronning Maud Land (Queen Maud Land), Antarctica. To explicitly determine whether the PJ/SAID was present at a given location, we used data from several satellites. Data received from satellites DMSP F-17, DMSP F-18, Swarm C, Swarm B and NorSat-1 confirmed the presence of the PJ/SAID in the subauroral region during the geomagnetic substorm of March 18, 2018. According to the analysis, PJ/SAID existed during a considered event at least from 19 UT March 18, 2018 to 02 UT March 19, 2018. Based on their research Anderson et al.3 concluded that a typical PJ/SAID lifetime is between 30 min and 3 h. The mechanisms behind the PJ/SAID formation allow the existence of these phenomena for a longer period during geomagnetic storms due to a continuous energy input from the magnetosphere/ionosphere4,9,18. It is also likely that the PJ/SAID can fade and reappear during one geomagnetic event, depending on geomagnetic activity variations. According to various satellite data, the PJ/SAID considered in this study is located between − 53° and − 65° MLAT in the southern hemisphere and from 16 to 02 MLT, which is in good agreement with Spiro et al.2 and Karlsson et al.7 who, based on statistics, claim that typical invariant latitudes of the PJ/SAID are between 55° and 65° and that the PJ/SAID can be stretched along latitude for more than ~ 6 h of MLT in the midnight and evening sectors. In Fig. 1 one can see that PJ/SAID detected by DMSP F-18, DMSP F-17 and Swarm C satellites is located at higher latitudes (approx. – 59°, – 60° and  – 65° MLAT respectively) than PJ/SAID detected by Swarm B (ca. − 55° MLAT at 22:20 UT and ca. – 57° MLAT at 20:48 UT) and NorSat-1 (ca. – 54° MLAT). This is because the mean magnetic latitude of PJ/SAID lowers with an increase of MLT moving from the evening to post-midnight sector according to Plate 1 in Karlsson et al.7 and Foster and Vo19 for SAPS (Subauroral Polarization Stream), which most likely has similar features to PJ/SAID. The latitudinal width of the considered PJ/SAID is around 1° which is as well the characteristic of PJ/SAID and most researchers agree that typical PJ/SAID width is between 1°and 2° 2,3,7.

We decided to check whether our case was perhaps only an isolated event. We can confirm that it is a regular phenomenon. According to the DMSP satellite data, the PJ/SAID was also present during days close to the selected event. One of the cases on the 25th of March is shown in Fig. 2. Here the geomagnetic disturbance was weaker, and the equatorial boundary of the diffuse aurora was located above the receivers. The thick areas along the satellite’s flight path show how dynamic the PJ/SAID structure is over time. The 10-min difference between the two flybys results in the PJ/SAID moving more poleward about 1° MLAT. The poleward shift is associated with the general dynamics of displacing boundaries of the auroral oval with a decrease in geomagnetic activity20. However, at the equatorward boundary of the SSUSI auroral precipitation region, where the narrow region of the PJ/SAID begins, we also see enhanced phase and amplitude scintillation indices (see Fig. 2 and Supplementary Information). There is also a loss of signal lock of some satellites observed with the Troll receiver, GPS PRN 3 at 23:04 UT (α = 15.13°, − 57.7° MLAT); GLONASS PRN 55 from 19:44–19:55 UT (α = 19.77° and 25.6°, − 59.6° and − 60.4° MLAT), PRN 46 at 23:29 UT (α = 25.84°, − 60.0° MLAT) and the receivers stops tracking the satellite after that, PRN 58 at 23:42 UT (α = 11.3°, − 63.8° MLAT). These satellites are characterized by a higher elevation angle during loss of lock than the considered March 18th event. This corresponds to the PJ/SAID position which is more poleward on March 25th than on March 18th. Thus, it is clear that the increase in scintillation indices during the development of the PJ/SAID during the geomagnetic event of March 18th, 2018, is not an “isolated” event, but it is a regular phenomenon in the high-latitude region. It is of course worth noting that not all GNSS satellites are affected.

Fig. 2
figure 2

As Fig. 1, but for 25.03.2018 at 22:50 UT. DMSP satellites flew over the southern hemisphere at 22:39 (F-17) and at 22:49 (F-18). Note that during this event there were no close flights of the NorSat-1 and Swarm satellites here and there were also gaps in their data.

Full size image

During the development of the polarization jet intense electric fields drive strong plasma flows causing frictional heating leading to plasma temperature increases. Joule heating can enhance ionospheric recombination rates leading to a depletion in electron density. Plasma heating leads not only to horizontal drift of the plasma, but also to vertical drift, which means that temperature increases strongly and the plasma is carried upwards21,22. The interaction of the neutral atmosphere and ionized plasma (ion drag) can lead to significant changes in thermospheric winds. These winds can redistribute heat and cause temperature fluctuations, which in turn affect the plasma density and create irregular structures. Temperature gradients can drive plasma instabilities, such as gradient-drift instability. Liu et al.23 using Swarm data show that the temperature gradient drift instability occurs in midlatitude ionospheric trough polarward wall.

This study was carried out during the period of minimum solar activity in the southern hemisphere close to the equinox conditions. Under these conditions, the conjugate point in the northern hemisphere is subject to higher insolation due to the difference between magnetic and geodetic poles and is expressed in a higher background electron density. The background values are crucial for occurring scintillations24. In high-latitudes both E- and F-regions are important. In each region, the relative density fluctuation is required to exceed the corresponding background for the increase in density to have an impact on GNSS signals25. Sousasantos et al.26 show a clear connection between the background F-region densities produced with GOLD data and the scintillation severity for the S4 index in the low-latitude region. In addition, higher background density increases the ionospheric conductance, which affects the electric fields and currents in the E-region which influences the strength of PJ/SAID27.

Plasma structuring, high-speed flow and irregularities in the ionosphere can cause degradation of communication and precise positioning performance, cycle slips and loss of lock28,29,30,31. It is known that rapid changes in the ionosphere can lead to scintillations and fluctuation in phase and amplitude measurements30. In addition, scattering by irregular structures in the ionosphere can cause trans-ionospheric radio signals to be subject to random amplitude and phase modulations32. The satellites considered earlier have a low elevation angle. According to the work by Vasylyev et al.33 receiver zenith angle at 85° corresponds to the zenith angle at the scattering point on plots of about 40.9° that lead to enhanced scintillation for several link configurations on the global scale (which correspond to the values of 0.16 and 0.3 rad for the S4 and σϕ over our region). At high latitudes, both phase and amplitude scintillations of radio signals interfere with communication and navigation as ionospheric irregularities usually intensify during geomagnetic activity periods17. The dynamics of the subauroral ionosphere during a substorm is driven by the interaction of the ring current and the plasmasphere/plasmapause, resulting in increased fluxes of the westward convection, that is, to the emergence of the PJ/SAID where areas of irregular plasma density troughs and enhanced electron temperature are located, leading to scintillations at subauroral latitudes34. These are among the fastest flows observed anywhere in the ionosphere, so scintillations observed in this region should have different properties35. In the region of the PJ/SAID, we observe an increase in the indices of both phase and amplitude scintillations. The phase scintillation index is sensitive to large (larger than the Fresnel scale) and small-scale irregularities along the signal path, while the S4 index is an indicator of small-scale irregularities below Fresnel scales. The Fresnel scale at n = 1 is about 365 m at a height of 350 km for L1 frequency36. Variations in these two indices indicate the presence of plasma irregularities of different scales.

The PJ/SAIDs themselves are very structured, having within them irregularities of different scales, which are clearly visible from data from the NorSat-1 satellite, equipped with a high-resolution Langmuir probe system. In this work, inside the considered PJ/SAID small-scale irregularities in plasma parameters are found (see Methods, Fig. 5). The spatial size of irregularities of electron density is up to several hundreds of meters which does not exceed the background level of subauroral plasma during disturbed geomagnetic conditions. On the contrary, irregularities in the electron temperature inside the PJ/SAID have significantly smaller minimum sizes (dozens of meters) and are more intensive inside than outside of PJ/SAID.

Although the PJ/SAID has a rather narrow structure (1°–2° in latitude), strong plasma flows and large gradients of plasma parameters can significantly impact the passage of radio waves. In the considered example of a geomagnetic event, this is manifested in increased values of both scintillation indices (but without a significant increase in the S4 values). However, it is necessary to consider the geometry of the study case. Since we are considering the Antarctic region, the navigation signal arrives at the receiver at a lower elevation angle compered to mid-latitudes, as the inclination of the GNSS satellites is relatively small. We consider in more detail three cases of the signal on the satellite-receiver path. The first case concerns the satellite with a low elevation angle (the satellite in the lower left corner in Fig. 3). The signal passes through both the PJ/SAID and diffuse aurora regions, which means that both of these regions will affect the received signal on the ground. The IPP is located somewhat higher in latitudinal coordinates than the PL/SAID projection. We see such cases in Figs. 1 and 2, as well as in the video from Supplementary Information, when the IPPs show significant values of scintillation indices equatorially above the PJ/SAID boundaries. The most interesting case is with the satellite with a higher elevation angle, the signal of which also passes through these two regions. Due to the chosen approximation of the thin ionospheric layer and the geometry of the magnetic field lines, the IPP at 350 km is located close to the diffuse aurora boundary, which creates the impression in Figs. 1 and 2 of taking into account the influence of only the right part of the signal path passing through the green region of the diffuse aurora. However, it is clearly visible that the commensurate signal path passes through the PJ/SAID region, which is located predominantly in the F region of the ionosphere and is measured by LEO satellites at altitudes greater than 450 km. Examples of such satellites are visible at the boundaries of the precipitations in Figs. 1 and 2. Therefore, the radio signal from the “equatorward” satellites (as illustrated in Fig. 3) passes not only through the PJ/SAID region with its manifestation in the ionospheric F-region but also through the region of diffuse auroras. The last satellite (on the right) shows the signal passing predominantly through the discrete aurora region. Accordingly, we do not experience difficulties here in determining the main sources affecting the signal. For polar-orbiting satellites, the auroral station may observe enhanced scintillations from field-aligned irregularities when the ray path is along the local magnetic field. However, this impact is not significant for GNSS at auroral latitudes37,38.

Fig. 3
figure 3

Schematic representation of the GNSS satellite navigation signal passage geometry in a subauroral region in MLAT-height coordinates. The signal passage trajectory is shown by a black dashed line, the GISTM ground-based receiver location is shown by a black star. Horizontal blue dashed lines display ionospheric layers E and F boundaries at 90, 180 and 350 km. The red area indicates the narrow PJ/SAID band, while the green area shows the auroral zone of particle precipitation (solid green for diffuse aurora (DifA) and hashed green for discrete aurora (DA)). Blue arrows illustrate the direction of Earth’s magnetic field lines (MFL) in the subauroral region in the Antarctic. The red squares on the satellite-receiver beam represent the IPPs at an altitude of 350 km.

Full size image

Therefore, it is quite challenging to separate one effect from another or to consider them in context from each other. Our results show that it is a combination of two effects together from the PJ/SAID and the diffuse aurora boundary. Note that satellite signals passing only through areas of diffuse aurora, as well as the boundary between diffuse and discrete aurora, were not that affected and there was no loss of tracking of the GNSS satellites. A large network of receivers in future is necessary to identify how scintillation is related to subauroral density structures.

Conclusions

In this study, we were the first to use a multi-instrumental approach to study the influence of the PJ/SAID on the occurrence of scintillations and inhomogeneous structures of various scales (from dozens of meters up to several hundreds of meters) in the ionosphere in the Antarctic subauroral region. Our multi-instrumental approach is based on the use of various satellites (NorSat-1, DMSP, Swarm) and ground-based data (GISTM receivers, magnetometers). This technique allowed for a comprehensive consideration of geomagnetic events, which made it possible to draw firm and reasonable conclusions. A multi-instrumental complex study confirmed the presence of a PJ/SAID in the region of two receivers under consideration.

We have considered two geomagnetic events where the PJ/SAID was observed by different LEO satellites in the area of two GNSS receivers. In the equatorward boundary of the auroral precipitation region, in the theoretical and observational location of the PJ/SAID, we observed an increase in the phase and amplitude indices of scintillation, as well as a loss of GNSS signals lock. We have carefully considered each of the cases to exclude repeated events of tracking failure approximately in the same place and at times shifted due to the orbits of the satellites. We highlight the importance of taking into account the geometry of the beam from the satellite to the receiver through different areas and show that in practice we have a combined effect of the PJ/SAID and the boundaries of the diffuse aurora, which are quite difficult to separate from each other due to their close proximity. It is important to note that satellites passing only through the diffuse aurora, as well as the boundary between diffuse and discrete aurora, were not that affected and there was no loss of tracking of the GNSS satellites.

Our study shows that in practical applications (such as maritime navigation) during geomagnetic disturbances, it is necessary to consider not only the precipitation of charged particles in the auroral oval but also the development of a PJ/SAID in the subauroral region in the evening and night local hours, which can lead to loss of communication with the satellite. Further studies should be carried out also for the northern hemisphere. A distinctive feature of such studies for the northern hemisphere is since most instruments at high latitudes are concentrated in the auroral zone and the polar cap since this is where the main interest is directed. We demonstrate the importance of taking into account the subauroral zone, which is where fairly stable plasma formations occur, with a strong flow and various-scale irregularities inside, which lead to serious interference in satellite communications.

Methods

This section describes instrumentations and methods used to process the observation data. To conduct this study, we had at our disposal an unprecedented set of data from different instruments in this region. Two scintillation receivers at SANAE IV (SAN, geographic coordinates ɸ = 71.7°S; ℓ = 2.8°W) and Troll (TRL, ɸ = 72.0°S; ℓ = 2.5°E) stations, which are located on the equatorial side and boundary of the quiet auroral oval. NorSat-1 satellite with a very high-resolution sampling data, flying in the evening sector in the field of view of the receivers, which allows us to evaluate the small-scale irregularities. DMSP satellites show the occurrence of particle precipitation and measure drift. Swarm satellites show the electron density variations. All data from the LEO satellites allow us to trace the dynamics of the position and development of PJ/SAID over time (represented in Fig. 4 by thick sections of the flyby trajectories of various LEO satellites). Magnetometers allow us to check the position and direction of PJ/SAID above the stations (more information about it will be at the end of this section).

Fig. 4
figure 4

LEO satellite passages and GISTM receivers’ field of view map. Different coloured lines show LEO satellite passages (NorSat-1, DMSP F-18 and F-17, Swarm C and B) over Dronning Maud Land (Queen Maud Land) in Antarctica. The dashed red line shows the geomagnetic projection of the flight of the DMSP F-17 satellite over the northern hemisphere. The position of the SAID, determined from in situ data, is shown by thick sections of the fly-by trajectories. Stars show scintillation receivers, and the projection of the field of view with an elevation angle of 10° and 20° at a height of 350 km is presented by circles (black for SANAE station, red for Troll). The grey star represents the magnetometer at Neumayer station.

Full size image

In our study, we used data from three space missions in low Earth orbits: NorSat-1, Swarm and DMSP. The Norwegian microsatellite NorSat-1 (∼16 kg, 23 × 39 × 44 cm) was launched on July 14, 2017 from the cosmodrome of Baikonur into a circular sun-synchronous orbit with an inclination of 98°, altitude of ∼600 km, and orbital period of 95 min. The satellite has a multi-needle Langmuir Probe (m-NLP) developed at the University of Oslo. It should be noted that the classic Langmuir probe makes measurements over a range of voltages, building a current–voltage characteristic, from which the plasma parameters are calculated. Since the coverage of the entire range takes time, the time resolution of the obtained plasma parameters is usually low, which makes it difficult to study small-scale plasma structures. The m-NLP consists of four cylindrical probes (needles) at different voltage biases within the electron saturation region of the current–voltage characteristic39. The absence of the need to cover the voltage range makes it possible to measure the electron current at a much higher frequency, which allows obtaining high-frequency characteristics of the plasma. To determine the electron density Ne in measurements by the system of LPs, at least two cylindrical probes operating at different fixed voltages are used40. Also, a main feature of the m-NLP approach is that it does not need information about plasma potential and electron temperature to measure electron density. While the absolute electron temperature cannot be derived without knowing the satellite potential in accordance with the Langmuir probe theory, we can estimate the change of temperature with time and latitude41. Therefore, in fact, we mean the less rigorous term “temperature parameter” instead of the formal concept of temperature here. This should be kept in mind when using the term “temperature” from the NorSat-1 satellite. The high resolution of m-NLP instrument made it possible to study for the first time the small-scale structure of PJ/SAID in the subauroral region11,12, as well as to discover a new phenomenon called Stratified Subauroral Ion Drift (SSAID)13.

On 22 November 2013, the European Space Agency launched three Swarm satellites into nearly polar orbits. These satellites eventually reached heights of 460 km (Alpha and Charlie) and 510 km (Bravo). The main purpose of the Swarm mission is to carry out accurate multipoint measurements of low-frequency magnetic and electric fields in the ionosphere of the Earth. Two spherical LPs that were placed on the satellites were used to measure the values of electron density and temperature42. Swarm satellite data has also been used to study PJ/SAID (often in conjunction with other satellite data), as well as such new phenomena as STEVE (strong thermal emission velocity enhancement), the physical mechanism of which is closely related to PJ/SAID12,43,44,45,46,47.

A large number of studies conducted by various scientific teams in different years on the investigation of PJ/SAID are associated with the use of data from the Defense Meteorological Satellite Program (DMSP) satellite series. The DMSP satellites have been collecting weather data for U.S. military operations for over five decades and provide assured, secure global weather and space weather data. The DMSP satellites are polar-orbiting spacecraft at about 835–860 km altitude with 101 min orbital period. Many studies of PJ/SAID are based on data from the DMSP satellites9,21,48,49,50,51,52. Also, using the data from these satellites, interesting cases were found when two peaks of PJ/SAID velocities were observed side by side on one satellite pass, that is, the so-called Double peak SAID (DSAID)53,54,55. Since NorSat-1 and Swarm are not equipped with ion drift measurements, data from the DMSP mission is used to indicate the existence of PJ/SAID explicitly in our work.

The Special Sensor Ultraviolet Spectrographic (SSUSI) instrument56 measures ultraviolet (UV) emissions in five different wavelengths and these instruments are flown on board the DMSP satellites. SSUSI is mounted on the nadir-looking panel of the satellite. The multicolour images from SSUSI cover the visible Earth disk from horizon to horizon and the anti-sunward limb up to an altitude of approximately 520 km. The SSUSI on board the DMSP satellites provide nearly hourly, 3000 km wide high-resolution (10 km × 10 km) UV snapshots of auroral emissions. These UV data have been converted to average energies and energy fluxes of precipitating electrons. Thus, using SSUSI it is possible to determine the boundaries of auroral precipitation and thereby clearly understand where the boundary between the auroral and subauroral zones is in the ionosphere. As is known from various works (for example by Galperin57), PJ/SAID is located inside the main ionospheric trough in the subauroral region and equatorward of the soft electron precipitation boundary, therefore, it is important to know these boundaries for a more accurate determination of the location of PJ/SAID.

The auroral energy flux derived from SSUSI images taken onboard DMSP is organised in Altitude Adjusted Corrected Geomagnetic (AACGM) coordinates for the nightside (18:00–06:00 MLT) of the southern hemisphere58. Therefore, the AACGM coordinate system is used everywhere in this study for uniformity. The AACGM coordinates59 are an extension of corrected geomagnetic (CGM) coordinates that more accurately represent the actual magnetic field. In the AACGM coordinates, points along a given magnetic field line are given the same coordinates and thus better reflect the magnetic conjugacy.

Figure 5a, b shows position of the PJ/SAID at ~ 20:56:30 UT at geomagnetic latitudes from − 58.1° to − 57.5° and at 18 MLT according to DMSP F-18 data. Despite the gap in the data, which is probably associated with instrument overscale, inside the electron density dip, which lies within the geomagnetic latitudes between − 58.5° and − 57.4°, an increase in the horizontal westward drift velocity of ions up to 2400 m/s is noticeable. According to Fig. 5b, the increase in the ion temperature by approximately twice the average level does not spatially coincide with the peak of the vertical and horizontal drift velocities and is located more poleward by approximately 2° in latitude. This may be due to a significantly lower spatial frequency of ion and electron temperature measurements (~ 0.25 Hz) than the spatial frequency of drift velocity measurements (1 Hz) on the DMSP satellite. The horizontal drift velocity of ions above 1000 m/s, coinciding with the electron density drop in the subauroral region of the ionosphere in the evening and midnight MLT sectors, indicates explicitly the presence of PJ/SAID in the considered region of the geomagnetic latitudes3. As already mentioned, the position of this PJ/SAID is marked in Figs. 1–2, 4 with thick sections of the fly-by lines.

Fig. 5
figure 5

DMSP F-18 and NorSat-1 satellites measurement. (a) Dynamics of the horizontal ion drift velocity (red line), vertical ion drift velocity (purple line) and electron density (blue line) depending on geomagnetic latitude (upper abscissa, 350 km projection) and UT time (bottom abscissa) according to DMSP F-18 at 20 UT 18.03.2018. PJ/SAID is shown with a blue area. (b) Dynamics of the ion temperature (red line), electron temperature (blue line), and MLT (purple line) according to DMSP F-18 at 20 UT 18.03.2018. (c) Dynamics of the electron temperature (red line) and electron density (blue line) depending on geomagnetic latitude (upper abscissa, 350 km projection), and UT time (bottom abscissa) according to NorSat-1 at 21 UT 18.03.2018. (d) Power spectrogram obtained using the discrete Fourier transform of the electron density data. (e) Power spectrogram obtained using the discrete Fourier transform of the electron temperature data. (f) Spectral power of electron density inside PJ/SAID as a function of spatial wave number. (g) Spectral power of electron temperature inside PJ/SAID as a function of spatial wave number.

Full size image

Figure 5c illustrates PJ/SAID detected from the high-frequency electron density and electron temperature data measured by the m-NLP onboard NorSat-1 satellite. As can be seen from Fig. 5c in an interval of the geomagnetic latitudes from − 52.0° to − 52.7°, there is an electron density dip that spatially coincides with an increase in electron temperature value and fluctuations. The median electron density within abovementioned interval is ~ 22,300 cm−3 and median electron temperature is ~ 2.33 eV. Outside of the interval the median electron density is ~ 137,600 cm−3 and 46,248 cm−3 equatorward and poleward respectively, while the median electron temperature is ~ 1.85 eV and ~ 2.04 eV. Thus, the electron density inside the density dip is ~ 6 times lower than equatorward of it and ~ 2 lower than poleward of it while the electron temperature is ~ 1.26 higher than equatorward of it and ~ 1.14 higher than poleward of it. Since NorSat-1 is not equipped with an ion drift velocity instrument, we cannot directly determine that the PJ/SAID is present in the considered time interval, however, the well-known properties of PJ/SAID (increase in temperature and decrease in ion density) indicate the presence of a PJ/SAID. In addition, thanks to DMSP F-18 horizontal ion drift velocity data it is already known that PJ/SAID existed at 20:56 UT at 18 MLT. According to the fact that average PJ/SAID spans over ~ 6 MLT7 we conclude that PJ/SAID also existed at 21:20 UT at 00 MLT where NorSat-1 took measurements. Location inside the electron density decrease that coincides with an elevation in electron temperature at the subauroral latitudes in the evening and midnight MLT sector is characteristic of PJ/SAID3. Besides that, the increase in fluctuations of the electron density and electron temperature are also characteristic of PJ/SAID11,12,34. Thus, we conclude that the considered electron density dip coinciding with an electron temperature increase detected from NorSat-1 data is PJ/SAID. Similarly, we conclude that PJ/SAID is present during the Swarm B and Swarm C passages since the Swarm Spherical LPs electron density and electron temperature data at 19:50, 20:48 and 22:20 UT show similar behaviour to the NorSat-1 data (Swarm satellites data are not shown in the figures). Despite the fact that the resolution of the Langmuir probes on Swarm and NorSat-1 is different (2 Hz and 1 kHz respectively), the smoothed values are sufficiently close.

Power spectrograms of the discrete Fourier transform of electron density and electron temperature are presented in Fig. 5d and e respectively. The spectrogram in Fig. 5d shows that fluctuations of electron density are present inside the PJ/SAID (an increase in spectral power at frequencies up to ~ 50 Hz which corresponds to a spatial size of plasma irregularities of several hundreds of meters) yet do not exceed the background level. On the contrary, the spectrogram in Fig. 5e illustrates a significant rise in fluctuations of electron temperature inside the PJ/SAID relative to the background level. The spectral power of electron temperature rises at frequencies up to 300–400 Hz which corresponds to the spatial size of plasma irregularities of several dozens of meters. The electron density spectrum illustrated in Fig. 5f also shows the presence of plasma irregularities inside the PJ/SAID with dimensions of hundreds of meters. The slope of electron density spectrum up to k = 10 km−1 is − 1.87 ± 0.02. Mishin and Blaunstein34 suggest that the power spectra of plasma density irregularities inside SAPS are well-represented by a power with a spectral index in the range − 5/3 ≥ pn ≥ − 2. Besides, slopes of the electron density spectra measured inside the PJ/SAID by NorSat-1 satellite are in the range of − 1.8 ≥ pn ≥ − 1.94 12. The slope of electron density spectrum shown in Fig. 5f falls within the above range. Note that it is different from the slope of − 1.5 which is seen in low-latitude ionospheric perturbations measured by DMSP F-1060. The electron temperature spectrum shown in Fig. 5g significantly differs from the electron density spectrum. The slope of the temperature spectrum up to k = 10 km−1 is − 1.2 ± 0.08 which is less than the density spectrum. This suggests that small-scale fluctuations of temperature are more pronounced than fluctuations of density. Thus, we can conclude that PJ/SAID includes fluctuations and plasma irregularities of various scales, which can significantly affect radio communication, radio wave propagation, and GNSS positioning accuracy. This emphasises the relevance and importance of this study.

Together with the high-speed flow (that is usually above 1 km/s for PJ/SAID), such irregularities can lead to rapid fluctuations in the phase and amplitude measurements observed by receivers on the ground. To check how strong the influence of plasma structuring in PJ/SAID can be, we used data from two GNSS Ionospheric Scintillation and TEC Monitor (GISTM) receivers. The station SANAE IV has a NovAtel GSV4004B receiver, operated by the South African National Space Agency (SANSA). In December 2017, a NovAtel GPStation-6 receiver was installed at the Norwegian research station Troll as a part of Troll Ionospheric Observatory managed by the University of Oslo. GISTM receiver records signals from the GPS, GLONASS, and Galileo satellites, and it outputs raw observational data every second, including 50 Hz phase and amplitude measurements and 1 Hz TEC measurements. The raw phase measurements detrended with a 6th-order Butterworth high-pass filter with a 0.1 Hz cutoff frequency. NovAtel GSV4004B receiver records signals only from GPS satellites. We consider two scintillation indices: S4 defined as the normalized variance of the signal intensity24 and is an indicator for small-scale irregularities at and below the Fresnel radius, and σϕ—the standard deviation of the measured phase (e.g., more in Rino32) which is a proxy of plasma structuring on large and small scales. We focus also on the lock time parameter that indicates how long the receiver follows the satellite and has been locked to the carrier phase on the signal. The thin ionospheric layer approach and the AACGM system were used to recalculate the coordinates of ionospheric piercing points (IPPs) at 350 km 61. We chose an altitude of 350 km since PJ/SAID is most pronounced in the F region of the ionosphere. For both receivers, we used 1 Hz data (for the scintillation indices) from the L1 signal type (to comply with data resolution from SANAE IV) and thresholds of 10° elevation angle to collect more data that can cover regions located equatorward of the auroral oval.

The cut-off elevation mask to a GNSS satellite (elevation angle) is an essential parameter for studying the ionosphere. The receiver starts to track the satellite radio signal after the satellite rises above the horizon at a certain angle. Usually, the cut-off elevation mask value is set between 10° and 30° for ionospheric research. The advantages of using lower elevation angles are the following: (1) Lower cut-off angle would provide better spatial data coverage which is important because in this work we are considering the Antarctic region, where GNSS satellites are visible only from a certain angle due to the low inclination of the orbits of GNSS satellites. (2) Even at elevation angles larger than 30°, the multipath effects (due to signal propagation along several paths due to scattering and reflections in a medium with inhomogeneities) can result in phase slips and pseudorange distortions due to the signal interference28. Therefore, in this study, we choose the cut-off angle of 10°. The Antarctic region is a less noisy area (lack of buildings, trees, etc.), which makes it possible to consider lower cut-off elevation angles.

In addition, data from magnetometers at SANAE IV and Neumayer (ɸ = 70.7°S; ℓ = 8.3°W) stations were used to infer the ionospheric current system (Fig. 6). Neumayer data is available at the INTERMAGNET database62. The baselines were subtracted from the measurements of the magnetic fields. Variations in the N-component during the considered period had the same appearance with a change in polarity (from positive variation to negative) which may indicate the presence of a PJ/SAID63. Similar behaviour of the N-component during the passage of PJ/SAID was observed earlier at the Yakutsk station, as reported by64. Their study specified that the rapid formation of PJ/SAID 10 min after the onset of the substorm expansion phase, was accompanied by a bipolar magnetic variation, similar to the fast passage of the Harang discontinuity over the station. Therefore, magnetic variations of the N-component over stations SANAE IV and Neumayer indicate the possible formation of PJ/SAID in this geomagnetic event.

Fig. 6
figure 6

Magnetometer data for 18–19.03.2018. Magnetometer data in the NEZ-coordinate system from SANAE IV (a) and Neumayer (b) stations on 18 and 19 March 2018. Magnetic north is presented by black lines, magnetic east—by red lines, and vertically down component—by green lines.

Full size image