Amplified Westward SAPS Flows near Magnetic Midnight in the Vicinity of the Harang Region

Abstract

Rare (only 10) observations, made in the southern topside ionosphere during 2015–2016, demonstrate the amplification of westward subauroral polarization streams (SAPS) up to 3000 m/s near the Harang region. The observed amplified SAPS flows were streaming antisunward after midnight and sunward at midnight, where the dusk convection cell intruded dawnward. One SAPS event illustrates the elevated electron temperature (Te; ~5500 K) and the stable auroral red arc developed over Rothera. Three inner-magnetosphere SAPS events depict the Harang region’s earthward edge within the plasmasheet’s earthward edge, where the outward SAPS electric (E) field (within the downward Region 2 currents) and inward convection E field (within the upward Region 2 currents) converged. Under isotropic or weak anisotropic conditions, the hot zone was fueled by the interaction of auroral kilometric radiation waves and electron diamagnetic currents. Generated for the conjugate topside ionosphere, the SAMI3 simulations reproduced the westward SAPS flow in the deep electron density trough, where Te became elevated, and the dawnward-intruding westward convection flows. We conclude that the near-midnight westward SAPS flow became amplified because of the favorable conditions created near the Harang region by the convection E field reaching subauroral latitudes and the positive feedback mechanisms in the SAPS channel.

1. Introduction

1.1. Magnetosphere–Ionosphere (M-I) Conjugate SAPS Phenomenon

Subauroral polarization streams (SAPS ) [1] are westward plasma flows driven by a poleward subauroral electric (E) field (SAEF) developed during storms/substorms at the auroral electron precipitation zone’s equatorward edge. The latitudinally broad (≥3°–5°) and longitudinally extended SAPS channel is predominantly located in the dusk–premidnight magnetic local time (MLT) sector (16–24 MLT) [2,3] and can extend to the early morning (3 MLT) sector [3,4,5,6].

The poleward SAEF is the ionospheric component of the magnetosphere–ionosphere (M-I) conjugate SAPS E field. The inner-magnetospheric component is the outward SAPS E field, which develops in the magnetic equatorial plane and maps down along the plasmapause to the subauroral ionosphere as a poleward E field. Both the SAPS flow and the poleward SAPS E field impact the thermosphere, which creates a coupled system with the ionosphere, and thus regulates the dynamics in the coupled ionosphere–thermosphere (I-T) system (see details in Section 1.4).

Based on experimental observations, the westward SAPS flow develops soon (>30 min) after substorm onset on a short timescale [7,8], negating the inherently slow processes (taking 1–2 h) [9] invoked by the traditional voltage [10] and current [11] generator theories and explaining the outward SAPS E field’s development on a short timescale in the inner magnetosphere.

According to the fast-time SAPS development theory, the outward SAPS E field develops in the inner magnetosphere at the leading edge of the expanding substorm current wedge (SCW), forming a two-loop (2L) system (SCW2L) [12] characterized by the lower-latitude/earthward loop of Region 2 (R2) and the higher-latitude/tailward loop of R1 field-aligned currents (FACs). SCW expansion is created by the magnetotail-reconnection-related hot mesoscale plasma flows (MPFs; cross-tail width of 2–3 R_E) [13] or bursty bulk flows (BBFs) [14,15], which build up an azimuthal pressure gradient at the leading edge of the SCW. Thus, the newly formed outward SAPS E field, which is an integral part of the SCW2L system [7,8,16], maps down from the inner magnetosphere to the subauroral ionosphere, near the leading edge of the westward expanding auroral bulge, forming the westward traveling surge (WTS) [17]. Fast-time SAPS development takes place (>30 min) after substorm injections and dipolarization lasting ~5–20 min and unfolding after substorm onset [7,18]. The SAPS’ quick response to the arrival of the WTS [19] often leads to structured SAPS flows called SAPS–wave structures (SAPS-WS) [19,20,21], while auroral arcs also develop via dispersive mesoscale Alfven waves. Overall, these findings demonstrate the causal connections of SAPS/SAPS-WS and WTS events [7,8,16]. Furthermore, the M-I conjugate SAPS phenomenon is closely related to the M-I conjugate Harang phenomenon (see details in Section 1.2).

1.2. M-I Conjugate Harang Phenomenon

To describe the M-I conjugate Harang phenomenon [22,23], we start with its ionospheric component, shown near magnetic midnight in Figure 1a,b, which is modified after Figure 1 presented by Koskinen and Pulkkinen in [23].

In Figure 1a, the schematic diagram shows the dominating duskside auroral eastward electrojet (EEJ) located between the duskside downward (↓) R2 and upward (↑) R1 FACs. Meanwhile, the dominating dawnside auroral westward electrojet (WEJ) is located between the dawnside ↑R2 and ↓R1 FACs. Relevant to our study, we show the Harang region in the near-midnight MLT sector. Overall, the Harang region is characterized by an auroral electrojet reversal (i.e., Harang reversal) [24] and by a discontinuity (i.e., Harang Discontinuity) [25]. In the midnight sector, the dominating auroral EEJ reverses to an auroral WEJ across the discontinuity (HD; marked in green). We also marked here the plasmapause (PP; continuous line in cyan). At the Harang region’s equatorward edge (marked as shaded region in light magenta), the oppositely directed R2 FACs are located at the same MLT (or longitude) but at different latitudes. Oppositely, at the Harang region’s poleward edge (marked as shaded region in light cyan), the oppositely directed R1 FACs are located at the same MLT (or longitude) but at different latitudes.

In Figure 1b, the schematic diagram shows the near-midnight Harang region in the ionospheric two-cell E × B polar convection [26] depicted by the lowest equipotential contour of the dusk cell (in blue) and dawn cell (in red). Here, the Harang region’s equatorward edge is located at the tip of the dawnward-intruding dusk cell (shaded region in light magenta), where the oppositely directed R2 FACs are separated by the discontinuity (HD; marked in green). Within the dawnward-intruding dusk cell, the convection flow reverses across the discontinuity (HD; in green) from sunward (at higher latitudes) to antisunward (at lower latitudes). In the near-midnight MLT sector, the strong lower-latitude auroral flows are accompanied by strong subauroral flows located equatorward of the plasmapause (PP; in cyan). Flowing in concert with the dusk cell convection flows, these westward subauroral flows stream antisunward (marked in light blue) after magnetic midnight and sunward (marked in orange) before magnetic midnight.

In Figure 1c, the schematic diagram is modified after Figure 5 of Gkioulidou et al. [27] and is shown in an X versus Y graph in Geocentric Solar Magnetospheric (GSM) coordinates. Here, the Y axis is inverted and indicates the dawnside (Y < 0) and duskside (Y > 0). This diagram illustrates the mapped-up dawnward-intruding dusk cell. At the earthward edge (marked in light magenta), the discontinuity (HD; in green) separates the earthward-located ↓R2 and tailward-located ↑R2 FACs. Meanwhile, the Harang region’s equatorward edge is located near the inner-magnetosphere plasmasheet’s earthward edge (marked as continuous line in dark magenta).

1.3. Features Associated with the M-I Conjugate Harang and SAPS Phenomena

In the inner magnetosphere, the Harang region’s earthward edge is located just tailward of the outward SAPS E field, which is associated with various inner-magnetosphere phenomena. These include the hot zone [28,29], auroral kilometric radiation (AKR) waves [30], and the Harang region’s earthward edge [24,25]. The hot zone is a region, located near the inner-magnetosphere plasmapause, where the ion temperature (Ti) peaks and the electron temperature (Te) becomes also elevated. Heating is provided near the plasmapause by the collisionless damping of electromagnetic ion–cyclotron (EMIC) waves developed due to ring current instability [28,31]. But EMIC waves and MPFs are often associated with AKR waves, which are electromagnetic waves (20–800 kHz) [30] generated during substorm onset in the 20–24 MLT sector [32] by the free energy provided by the magnetotail-reconnection-related suprathermal electrons that are injected earthward from the plasmasheet [33]. AKR waves grow into surges during the fast expansion of the auroral bulge with the WTS located at its leading edge [34]. Often, near magnetic midnight, an intense outward SAPS E field appears near the Harang region’s earthward edge, where the oppositely directed R2 FACs are located. As an inherent part of the SCW2L system, via the meridional current connecting the ↓R2-↑R1 FACs, the outward SAPS E field develops on a short timescale due to the emerging ↓R2 FACs (occurring in response to the ↑R1 FACs via the meridional current closure) and is associated with localized ring current injections [8].

In the ionosphere, the Harang region’s equatorward edge is located just poleward of the subauroral flows, where the westward SAPS flows become intense [35,36]. The westward SAPS flow is associated with various interrelated features [37], including the different types of plasma-density troughs, the stable auroral red (SAR) arc [38], and the elevated Te [39]. These troughs are the ring-current-related ionospheric trough (RIT) [40,41] (that develops at middle latitudes due to the F region plasma’s O+ content depletion created by chemical processes [42]) and the flow-stagnation-related main ionospheric trough (MIT), also called stagnation trough [43] (that develops at subauroral latitudes due to the counteracting westward convection flows and eastward co-rotation flows). When these troughs appear separately, the mid-latitude RIT is located equatorward of the subauroral MIT. But these troughs can also merge into a single trough formation [44]. Within the RIT or coinciding RIT-MIT, the SAR arc develops as the ring current decays (most intensively in the recovery phase and at the plasmapause) and Te becomes elevated due to the downward heat flux or electron precipitation (from the inner magnetosphere to the ionosphere). The SAR arc is a well-known, east–west-oriented, 680 nm monochromatic, subvisual auroral emission associated with RIT [40] and strong SAPSs [45,46,47,48].

1.4. SAPS Effects Impacting the Coupled I-T System

In the coupled I-T system, thermospheric neutral winds are directly affected by ionospheric E fields and current systems. These include the dynamic and longer-lasting SAPS and underlying Pedersen and Hall currents, since the rapid SAPS flows can last up to 3 h [49] and can also be influenced by both the convection E field and the subauroral ionospheric conductivity [50]. H. Wang et al. [51] first reported the various significant effects of SAPS on the thermosphere. These include the strong westward disturbance zonal neutral winds driven by the poleward SAPS E field, as evidenced by the westward zonal neutral winds peaking at the same latitude as the westward SAPS flow. Moreover, the heating of the thermosphere in the SAPS channel by ion-neural friction, where the heating power depends on the ion drift velocity, results in neutral air density bulges. Furthermore, they highlight the peak of the auroral EEJ (driven by the antisunward Hall currents) just poleward of the SAPS channel. Later studies investigated these SAPS effects’ spatial and temporal variations in different MLT sectors [52] and in different universal time intervals [53]. Experimental results revealed that the ion-neutral frictional heating in the SAPS channel can further deplete the plasma-density trough and increase both the ion temperature and the neutral temperature [52]. Further studies by Wang and Lühr [54] documented the resultant ion upflow in the SAPS channel and its seasonal variation. Successful model simulations were produced by the Thermospheric Ionosphere Electrodynamics General Circulation Model (TIEGCM) when an empirical SAPS model was imposed on TIEGCM called SAPS-TIEGCM [55]. The SAPS-TIEGCM model simulations not only confirmed these experimental results but also revealed their hemispheric and seasonal asymmetries [55] and the nighttime meridional wind surges’ universal time effects [56]. The study by H. Wang et al. [50] also documented that, when the dawn-to-dusk convection E field played a significant role in the development of SAPS, the equatorial EEJ also became enhanced because of the duskward convection E field’s equatorward penetration. Later, Ebihara et al. [57] used magnetohydrodynamic (MHD) simulations in their study. These revealed that the mapped-up poleward SAPS E field, located in the magnetic equatorial plane, can cause the development of equatorial counter electrojet (CEJ) by the reversal of the equatorial EEJ to the westward direction in the dayside magnetic equatorial region. Experimental observational evidence was provided by the recent study by Wang et al. [58], who explained that the mapped-up poleward SAPS E field can create an overshielding effect in the inner magnetosphere, leading to the development of dayside equatorial CEJ (i.e., reversal of equatorial EEJ to WEJ).

1.5. Previous Studies on Intense SAPS Flows in the Harang Region’s Vicinity

Previous studies state that westward SAPS flows are quite intense near magnetic midnight. According to the simple analytical model of Lyatsky et al. [59], intense SAPS flows occur in the broader midnight MLT sector. This is because of the enhanced dusk cell convection flows that developed due to the equatorward propagation of the duskward convection E field, reaching middle latitudes. Consequently, the dusk cell expands both dawnward, by crossing the magnetic midnight meridian, and equatorward. Such dawn–dusk asymmetry is introduced by a non-zero interplanetary magnetic field (IMF) B_Y component through its imposition of a torque on the magnetic field flux tubes [60]; these, in turn, twist the open magnetic flux and skew the convection pattern [61]. Thus, within the dawnward-intruding dusk cell, the flow reversal (from higher-latitude sunward to lower-latitude antisunward) occurs close to the equatorward edge of the aural oval, as reported by previous studies [4,62,63,64,65,66]. But, according to the experimental study by Zou et al. [36], such dusk cell expansion (both equatorward and dawnward) is also accompanied by the expansion of the Harang region (both equatorward and dawnward). Furthermore, the modeling study by Gkioulidou et al. [27], investigating the formation of the inner-magnetospheric Harang region’s earthward edge, relates also the intense ionospheric SAPS flows to the nearby Harang region’s earthward edge. While Zou et al. [36] documented only weak ionospheric SAPS flows (~1000 m/s) developed in the evening sector (at ~19.6 MLT and ~22 MLT), Gkioulidou et al. [27] provided no experimental observational evidence at all supporting their model simulations. Thus, the midnight MLT sector remained undocumented by experimental ionospheric and inner-magnetospheric SAPS observations.

1.6. Motivations of This Study

This study was motivated by the lack of experimental observational evidence provided by the above-mentioned previous studies regarding the intense SAPS (near and after magnetic midnight) in the Harang region’s vicinity in the coupled M-I system. Significantly, we provide observational evidence of SAPS flows that became amplified (up to 3000 m/s in magnitude) in the vicinity of the Harang region, which developed when the dusk cell expanded equatorward and dawnward. Thus, we focus on the westward SAPS flows that developed in the Harang region’s vicinity near and after magnetic midnight in the topside ionosphere and that became locally amplified.

2. Materials and Methods

For this study, we compiled a multi-satellite database augmented with various imageries and SAMI3 model simulations. The satellite data are provided by the satellites of Defense Meteorological Satellite Program (DMSP), Time History of Events and Macroscale Interactions during Substorms (THEMIS), and Geostationary Operational Environmental Satellites (GOES).

The polar orbiting DMSP satellites circle the Earth at ~840 km altitude in the topside ionosphere with an orbital period of ~101 min and inclination angle of ~98.7° [67]. From the Special Sensor for Ion and Electron Scintillation (SSIES) instrument, we used in situ measurements, mostly provided by F15 and occasionally by F17 and F19. As listed in

Table 1. DMSP instrumentations and observables used for this study.

DMSP
Instrumentation	Observables	Cadence
SSIES SLP	Ne (1/cm3)	1 s
SSIES SLP	Te (K)	4 s
SSIES RPA	Ti (K)	4 s
SSIES IDM	VHOR (m/s)	1 s
SSIES IDM	VVER (m/s)	1 s

, these measurements include electron density (Ne; 1/cm³) and electron temperature (Te; K) from the Spherical Langmuir Probe (SLP), ion temperature (Ti; K) from the Retarding Potential Analyzer (RPA), and cross-track ion drifts in the horizontal (V_HOR, m/s) and vertical (V_VER, m/s) directions from the Ion Drift Meter (IDM).

The five THEMIS satellites (TH-A—TH-E) orbit the Earth (TH-A, TH-D, and TH-E) on their highly elliptical orbits in the magnetosphere [68] and the Moon (TH-B and TH-C) since 2011. For this study, we used TH-A and TH-D data from various instruments and systems. As listed in

Table 2. THEMIS instrumentations and observables used for this study.

THEMIS
Instrumentation	Observables	Cadence
EFI	EX, EY, and EZ (mV/m)	3 s
ESA	Spacecraft potential (V)	3 s
ESA	Ion and electron energy flux (eV/cm2-s-sr-eV)	3 s
ESA	Ti⊥ (eV) and Ti‖ (eV)	3 s
GMOM	Pi (eV/cm3)	3 s
GMOM	VeX, VeY, and VeZ (km/s)	3 s
MOM	Pi (eV/cm3)	3 s
HF FBK	HF Vpeak (V)	4 s

, these include the Electric Field Instrument (EFI) providing E field measurements (E_X, E_Y, E_Z; mV/m), the Electrostatic Analyzer (ESA) providing the data of spacecraft potential (SC Pot; V), ion and electron energy flux (eV/(cm²-s-sr-eV)) at various energy channels, and field-aligned hot (up to 25 keV) ion temperature (Ti, eV) in the perpendicular (Ti_⊥) and parallel (Ti_‖) directions, permitting the computation of Ti_⊥/Ti_‖ characteristic to ion anisotropy, and the systems of ground calculated ion moments (GMOM) for 5 eV–4 Me ions and onboard moments (MOM) for 30–300 keV ions and providing ion pressure (Pi; eV/cm³) measured as a six-component symmetric tensor, and the GMOM electron drift velocity components (Ve_X, Ve_Y, Ve_Z; km/s) in GSM coordinates. From the high-frequency (HF) filter bank (FBK) suite, we used the peak voltage values (measured at 130 kHz) of HF filter channel covering the AKR emission band (100–400 kHz) [69].

The GOES satellites are on geostationary orbits (6.6 R_E) and circle the Earth once a day [70]. We used average electron flux values measured by GOES-13 at various energy channels (40–475 keV) in nine directions in order to specify the particle injections.

Orbit data of THEMIS and GOES include MLT (Hr), dipole L (R_E), spacecraft location (X, Y, Z; R_E) in GSM coordinates, and footprints in geographic longitude (GLON; °E) and latitude (GLAT); °N) and in geomagnetic latitude (MLAT; °S).

We used various types of images. These include the DMSP Special Sensor Ultraviolet Spectrographic Imager (SSUSI) [71] auroral images, depicting the various auroral features; the Super Dual Auroral Radar Network (SuperDARN) convection maps [72], showing the pattern of polar convection; the Active Magnetosphere and Planetary Electrodynamics Response Experiment (AMPERE) current density polar map images, which illustrate the distribution of large-scale FACs such as R1 and R2 [73]; finally, the 630 nm All Sky Imager (ASI) imageries for observing the SAR arc over Rothera (291.9°E; 67.5°S; geographic) located in the South American sector [74].

We also used the interplanetary magnetic field (IMF B_Y and B_Z; nT) and electric field (IEF E_Y; mV/m) data along with the various indices (AE, nT; Kp; SYM-H, nT) for specifying the underlying geomagnetic conditions. Furthermore, we used the substorm lists by Forsyth et al. [75] and Newell and Gjerloev [76], listing substorm onset times and locations, published by SuperMAG.

For obtaining SAMI3 model simulations from the Community-Coordinated Modeling Centre (CCMC), we requested SAMI3 runs that are available online. For the SAMI3 simulations requested, CCMC used the 3.22 version of SAMI3 along with the NRLMSIS2.1/HWM14 neutral atmosphere model.

3. Results

3.1. Topside-Ionosphere Amplified Westward SAPS: Selection Criteria for Specification

In order to observe the intense westward SAPS flows located in the vicinity of the Harang region in the topside ionosphere (~840 km altitude), we surveyed the calendar years of 2015 and 2016 and identified 10 rare SAPS events with westward flows (from 1200 m/s up to 3000 m/s in magnitude) in the 0–1 MLT sector, depicted by good quality observations, as listed in

Table 3. Topside-ionosphere SAPS events investigated and DMSP F15 observations made.

DMSP F15 Observed SAPS Event					DMSP F15 Observables
Number	Date	UT (Hr:Mn)	MLAT (°S)	MLT (Hr:Mn)	Ne 103 (cm−3)	VHOR (m/s)	Te (K)
1	22 February 2015	22:08	61.37	00:05	2.5	−1200	6250
2	9 September 2015	6:13	49.55	01:46	2.8	−2900	5500
3	20 September 2015	1:35	57.34	00:07	1.4	+3000	6500
4	1 January 2016	21:24	59.33	00:38	3.5	−1200	5500
5	9 January 2016	20:57	60.62	00:36	3.0	−2500	6500
6	12 January 2016	21:50	57.11	00:39	3.5	−2400	6000
7	19 January 2016	23:20	52.83	00:44	3.0	−2400	6000
8	23 January 2016	20:35	60.89	00:37	3.1	−2200	5750
9	18 February 2016	1:46	51.97	00:18	2.1	−2800	7000
10	20 March 2016	5:38	59.09	00:56	2.8	−1700	5500

.

During the calendar years of 2015 and 2016, the DMSP F15 satellite was in a nominal midnight–dusk polar orbit over the Southern Hemisphere and midday–dawn polar orbit over the Northern Hemisphere. Based on the near-midnight observations made only in the Southern Hemisphere, we could investigate the westward SAPS flows in the Harang region’s vicinity, where the dusk cell intruded dawnward. Here, the cross-track V_HOR drift velocity measurements indicate sunward drift (V_HOR > 0) before and at magnetic midnight and antisunward drift (V_HOR < 0) after midnight. Because of the lack of electron and ion differential number flux images and magnetic data, as the DMSP F15 auroral particle sensor and magnetometer were not working, we considered eight selection criteria (CR1-CR8), which had to be met simultaneously, in order to specify correctly both the subauroral region and the westward SAPS flow channel. According to these criteria, the following observations were considered: (CR1) the plasmapause (PP) appeared mostly as a steep Ne gradient and was located poleward of the SAPS channel; (CR2) the ring-current-related plasma-density trough (RIT) was present in the SAPS event and appeared on the equatorward side of the plasmapause; (CR3) Te became elevated (Te > 4000 K) and the Te peak marked the location of RIT; (CR4) the magnitude of zonal cross-track drift was large (3000 m/s > |V_HOR| ≥ 1200 m/s) in the SAPS flow channel and peaked equatorward of the plasmapause and within the RIT; (CR5) the zonal cross-track V_HOR drift direction in the SAPS flow channel was sunward (V_HOR > 0) before and at magnetic midnight or antisunward (V_HOR < 0) after magnetic midnight; (CR6) the underlying two-cell polar convection pattern, depicted by the SuperDARN convection map, showed the dusk cell and its dawnward intrusion; (CR7) the location of equatorward oval boundary, depicted by the DMSP/SSUSI imagery, demonstrated the SAPS flow channel’s location on the oval’s equatorward side; (CR8) the signature of the dusk cell, appearing in the DMSP/SSUSI imaged auroral particle precipitation, confirmed the SuperDARN observation.

We note here that all the DMSP F15 survey plots, their respective underlying SuperDARN convection maps, and the AMPERE and DMSP/SSUSI imageries are shown in Figures S1–S10 (see the Supplementary Materials).

3.2. Topside-Ionosphere Amplified Westward SAPS: Observed by DMSP F15 in the 0–1 MLT Sector over the Southern Hemisphere

In Figure 2a, the southern-hemisphere geographic map illustrates the DMSP F15 observed SAPS locations (dot symbols in colors) along with the modeled dip equator (in light magenta) and magnetic meridians (in blue). These SAPS detections were made in two longitude sectors: (1) mostly near Africa, where the magnetic meridians are closely spaced; (2) sometimes in the South American longitude sector’s wider region, where the magnetic meridians are widely spaced. These (1–2) are the two preferred longitude sectors, where the ring-current-related trough (RIT) tends to develop [37,44] because of the more intense ring current precipitation into the ionosphere due to the weak magnetic field of the South Atlantic Magnetic Anomaly (SAMA) [77] that also increases the plasma drift (v = E × B/B²) in the SAPS channel.

In Figure 2b, the MLT versus MLAT southern polar plot shows the SAPS locations (dot symbols in colors). In good agreement with the model of Lyatsky et al. [59], the amplified westward SAPS flows (related to the dawnward-intruding dusk cell) were detected near magnetic midnight. This was mostly the case in the postmidnight MLT sector, because of the duskward convection E field’s equatorward propagation, which led to the expansion of the dusk cell in the equatorward and dawnward directions.

In Figure 3a–c, we illustrate each westward SAPS event with a DMSP F15 line-plot set oriented in the equatorward direction. Here, we marked the auroral zone located on the left side. Progressing from top to bottom, the Ne plot depicts the coinciding troughs of RIT (circle in red) and MIT (circle in blue), where the plasmapause (PP) appeared on the poleward side of RIT-MIT as a Ne drop-off. As shown by the Te plot, Te became elevated. The elevated Te peaked (≥4500 K) within the coinciding RIT-MIT and reached sometimes ~7000 K. We consider these Te maximum values elevated; according to the statistical DMSP SAPS study by Wang and Lühr [54], Te increases up to ~4000 K in the SAPS channel under Kp ≥ 4 conditions. Meanwhile, the cross-track V_HOR drift plot shows the near-midnight westward SAPS flow (dot symbol in color; shaded interval in yellow) that became amplified (3000 m/s > |V_HOR| ≥ 1200 m/s in magnitude). We consider these maximum zonal drifts amplified, since the statistical study by Wang and Lühr [54] documented average SAPS flows of ~1400 m/s in magnitude under Kp ≥ 4 conditions.

Overall, these observations demonstrate the close association of the amplified SAPS flow (in the deep coinciding RIT-MIT) and elevated Te, implying that these SAPS flows became amplified because of the combined effects of the underlying (i) positive feedback mechanisms [78] and (ii) dusk cell’s expansion [62,63,64,65,66] (equatorward and dawnward because of the equatorward propagating duskward convection E field). Since these amplified SAPS flows were located equatorward of the plasmapause and therefore on closed magnetic field lines, while the plasmapause was located on open magnetic field lines, the amplified SAPS flows had a significant role in moving and re-distributing the ionosphere plasma [18] within the coinciding RIT-MIT.

3.3. Topside-Ionosphere Amplified Westward SAPS: Underlying Interplanetary, Geophysical, and Auroral Conditions

In Figure 4, each line-plot set covers a larger time-period of the SAPS event of interest and depicts the underlying interplanetary (IMF B_Y and B_Z and IEF E_Y), ring current (SYM-H), geophysical (Kp), and auroral (AE) conditions. In each set, we marked the SAPS event (dot symbol in color; shaded interval in yellow) and the preceding substorm onset (star symbol in red). While all the SAPS events occurred soon after their respective substorm onsets (see

Table 4. SAPS events investigated, their underlying (interplanetary, geophysical, and auroral) conditions, and the preceding substorm onset specified (marked as * from substorm list of Forsyth et al. [75] and/or # from the substorm list of Newell and Gjerloev [76]).

SAPS Event	UT (Hr:Mn)	IMF		SYM-H (nT)	Kp	AE (nT)	Substorm Onset UT (Hr:Mn)
		BY (nT)	BZ (nT)
22 February 2015	22:08	−2.71	−6.45	2	2+	37	17:11 *
9 September 2015	6:13	4.81	−9.63	−90	6	760	6:01 *; 6:05 #
20 September 2015	1:35	3.39	3.07	−3	3	47	1:45 *
20 September 2015	3:36	4.75	4.08	−15	5−	612	1:45 *
21 September 2015	3:21	2.29	0.76	−34	3−	279	3:35 *#
7 October 2015	3:02	9.41	2.73	−12	6	384	2:27 *
1 January 2016	21:24	−1.15	−12.96	−30	2	132	18:25 *#
9 January 2016	20:57	4.84	2.3	0	1+	90	17:03 *#
12 January 2016	21:50	−7.20	−1.40	−24	3	296	20:19 *
19 January 2016	23:20	−4.51	10.22	10	1	24	-
23 January 2016	20:35	−2.12	−1.26	−10	3−	54	19:33 *
18 February 2016	1:46	−1.25	4.30	−55	4+	70	(17 Feb) 22:50 *#
20 March 2016	5:38	−4.73	1.85	−26	3−	68	3:06 #*

), only some of the SAPS events occurred during storms, in the initial phase (SAPS Event 7), main phase (SAPS Events 2 and 9), or recovery phase (SAPS Event 4). But, commonly, all the SAPS flow amplifications occurred when the Kp index was increasing or temporarily maximizing. These imply that the polar convection was increasing and the plasmapause (PP) was moving equatorward to lower latitudes. In the topside ionosphere, the SAPS channel developed within the coinciding RIT-MIT, where Te maximized. These ionospheric features (PP, SAPS, RIT-MIT, and Te peak) are coupled dynamically in the wider midnight MLT sector [79].

3.4. Topside-Ionosphere Amplified Westward SAPS: Plasma Environment

Figure 5 and Figure 6 illustrate the plasma environment of the amplified (V_HOR ≈ 3000 m/s in magnitude) westward SAPS flow with two SAPS events occurring near magnetic midnight on 20 September 2015 (shown in Figure 5) and 18 February 2016 (shown in Figure 6). In the same style, these figures are constructed with the DMSP F15 data, and with the AMPERE and DMSP/SSUSI imageries matched (as close as possible) with the F15 SAPS observation. We plotted, over each imagery, the DMSP F15 pass (in color) and the SAPS location (dot symbol in color). Based on their respective underlying polar convection patterns, these figures demonstrate that the amplified westward SAPS flow (flowing in concert with the dusk cell) was streaming sunward at midnight and antisunward after midnight.

In Figure 5a and Figure 6a, the southern-hemisphere geographic map shows the descending DMSP F15 pass, providing the SAPS observation made over the South Atlantic, and the ascending DMSP F15 pass over the South Pacific.

In Figure 5b and Figure 6b, the previously shown (in Figure 3a,c) line-plot set is augmented with the cross-track vertical drift (V_VER) data and covers a larger section of the polar region where the midnight–duskside auroral zone is marked. In both figures, the line-plot sets depict the deep plasma-density trough (i.e., coinciding RIT-MIT), where Te became enhanced quite intensively (up to 6500 K and 7000 K) and was possibly efficient enough to trigger SAR arc development within the SAPS channel/deep trough [45] that could not be documented due to the lack of ASI stations nearby. These line-plot sets show also the amplified westward SAPS flow (symbol dot in color; shaded interval in yellow) and the underlying enhanced upward drift (reaching commonly ~3000 m/s) produced by heating via collisions [80]. Both the deep trough and the low conductivity conditions along with the strong dusk-cell-related convections created favorable settings for the poleward SAPS E field’s enhancement, via the combination of positive ionospheric feedback mechanisms [19,20] and equatorward-propagating duskward convection E field [59]. These combined phenomena commonly contributed to the intensification of the westward SAPS flows. Due to the underlying polar convection (see more details below), the westward SAPS flow (flowing in concert with the dusk cell) was streaming sunward (~3000 m/s; Figure 5b) at around magnetic midnight and antisunward (~−2800 m/s; Figure 6b) in the postmidnight MLT sector. Without passing through the polar cap region, the descending DMSP F15 pass tracked the midnight–duskside auroral zone’s equatorward edge (as marked), where the cross-track plasma drifts (V_HOR and V_VER) were weak and varied little.

Overall, in Figure 5c and Figure 6c, the AMPERE imagery shows the radial current density [72] with a range of ±0.5 µA/m² and depicts the underlying large-scale FACs (↑ currents in red and ↓ currents in blue). Illustrating the underlying SCW2L system, the duskside and dawnside oppositely directed FACs appear at auroral latitudes in each SAPS event.

In Figure 5c, the AMPERE imagery shows the previously formed SCW2L system operational during 01:16–01:26 UT near the time of DMSP F15 SAPS detection (01:35 UT). Then, there was no ↓R2 FAC detected at SAPS latitudes. This implies that the SAPS channel was still newly formed in the topside ionosphere and developed in a voltage generator, since the ↓R2 FACs need time in the magnetospheric voltage generator to start flowing [80].

In Figure 6c, the AMPERE imagery shows the previously formed SCW2L system that was still operational during 00:30–00:40 UT and was characterized by some weaker (in light blue) ↓R2 FACs, underlying the intense near-midnight SAPS flow and thus documenting the flow of ↓R2 FACs to subauroral latitudes: into the old (i.e., previously formed) SAPS channel observed by DMSP later on at 01:46 UT.

Figure 5d and Figure 6d show the DMSP/SSUSI imagery depicting the underlying auroral precipitations wherein the signature of dawnward-intruding dusk cell appears. Here, the auroral boundary defined by the SSUSI model is also marked (dotted line in dark red). Based on the SSUSI Google Environmental Data Records (EDRs) auroral specifications, the highest energy auroral precipitations shown are discrete auroral arcs associated with increased electron mean energy and flux. These discrete auroral arcs were generated by the magnetospheric generator and their precipitated electrons (creating the auroral emissions observed) carried some of the ↑FACs [81]. However, in Figure 5d, the discrete auroral arc appeared significantly less developed than the discrete auroral arc appearing on the equatorward part of the auroral oval shown in Figure 6d. With the SAPS location (plotted over), the DMSP/SSUSI imagery also illustrates that the amplified westward SAPS flow was observed equatorward of the auroral precipitation regime in the midnight MLT sector.

In Figure 5e and Figure 6e, the MLT versus MLAT polar plot is constructed based on the relevant empirical SuperDARN convection map [82] and illustrates the two-cell polar convection [26] underlying the SAPS event. We matched the convection map (in UT) with the previously shown DMSP/SUSSI imagery. We marked the lowest equipotential contour of the dusk cell (in blue) and dawn cell (in red) along with the Heppner–Maynard (H-M) Boundary (in dark green), representing the modeled low-latitude auroral oval boundary [83] and the SAPS location (dot symbol in color). These provide observational evidence that the location of SAPS in MLT defined which direction the westward SAPS flow was streaming in concert with the dusk cell. As shown in Figure 5e, the westward SAPS flow at midnight was streaming sunward, in concert with the dusk cell at the magnetic midnight meridian. Oppositely, as shown in Figure 6e, the westward SAPS flow after magnetic midnight was streaming antisunward, in concert with the dawnward-intruding dusk cell. We note here that these findings, demonstrating the amplification of westward SAPS flow near and after magnetic midnight, are in good agreement with the results of previous studies [62,63,64,65,66] (see details in Section 1.5).

3.5. Topside-Ionosphere Amplified Westward SAPS: Association with SAR Arc

Most of the DMSP F15 detections (shown in Figure 3) imply the intense near-midnight westward SAPS’ association with the SAR arc (based on the underlying elevated Te reaching sometimes ~7000 K). But we cannot provide observational evidence due to the absence of nearby ASI stations. Only one SAPS event (observed over the Rothera ASI station) could accurately be correlated (in space and time) with the SAR arc developed (as shown in Figure 7).

In Figure 7, the 9 September 2015 near-midnight SAPS event (observed over Rothera) by F15 is illustrated along with two amplified nighttime westward SAPS flows observed over the South Pacific by F17 and F19.

In Figure 7a, the southern-hemisphere geographic map shows the SAPS locations over the South Atlantic (dot symbol in dark green) and over the South Pacific (dot symbols in dark cyan and magenta) along with their respective descending DMSP passes, the 630 nm ASI station (dot symbol in yellow) at Rothera [74], and the footprints of GOES-13 (in blue) near Rothera but at higher latitudes.

In Figure 7b, the three sets of Ne, Te, and Ti, and cross-track zonal V_HOR line plots illustrate the amplified westward SAPS flows in the nighttime sector (dot symbols in dark cyan and magenta) streaming sunward (~5000 m/s and ~2700 m/s) and in the postmidnight sector (dot symbol in green) streaming antisunward (−) (~−2900 m/s), in concert with the dawnward-intruding dusk cell over Rothera. Within the RIT, Te peaked at ~5500 K at nighttime (20.27 MLT and 20.73 MLT) and at ~5000 K after midnight (at 1.77 MLT). These elevated Te peaks imply SAR arc development that we could document only for the postmidnight event unfolding over Rothera.

Figure 7c shows the GOES-13 electron flux values measured during a 3 h time-period (4–7 UT). Here, we marked the almost simultaneous SAPS detections taken at 5.85 UT (dot symbol in dark cyan), at 5.91 UT (dot symbol in magenta), and at 6.23 UT (dot symbol in green). As the electron flux line plots show, a series of substorms (marked as star symbols in red) occurred according to the substorm list of Forsyth et al. [75]. Then, GOES-13 observed a series of significant electron flux depletions starting just before their respective preceding substorm onsets (marked as star symbol in red; shaded interval in light gray) characteristic to a tail-like magnetic field. Also, a series of significant dispersionless particle injections (shaded interval in yellow) characteristic to a dipole-like magnetic field and to fast BBFs/MPFs [12]. These imply a series of magnetic field reconfigurations (from tail-like to dipole-like) and dipolarization events that occurred just before and during the SAPS detections (shaded interval in yellow). All these provide observational evidence of (1) the magnetotail-reconnection-related BBFs/MPFs traveling earthward and (2) the SAPS development in the nighttime and postmidnight MLT sectors (observed by DMSP F15, F17, and F19) after dipolarization and on a short timescale (see details in Section 3.7).

In Figure 7d, the 630 nm ASI imagery depicts the east–west-oriented SAR arc over Rothera at 6.43 UT, soon after DMSP F15 observed the elevated Te at 06:14 UT.

Figure 7e–g shows three types of MLT versus MLAT polar plots. We plotted over each polar plot the DMSP F15 pass (in green) along with the westward SAPS locations in the postmidnight sector (dot symbol in green) and nighttime sector (dot symbols in dark cyan and magenta), and the GOES-13 (dot symbol in blue) location.

In Figure 7e, the polar plot depicts the two-cell polar convection operational earlier, during 05:45–05:48 UT, based on the matched SuperDARN convection map [82], with the lowest equipotential contour of the dusk cell (in blue) and dawn cell (in red). Characterized by a larger dusk cell and smaller dawn cell, the dusk cell intruded dawnward (up to 2 MLT) where the Harang reversal developed. Favorable inner-magnetospheric conditions were created by the negative IMF B_Z and positive IMF B_Y (see Table 4), with their comparable magnitudes inducing flow shear along the lobe-plasmasheet boundary [84]. Under these IMF conditions, the ionospheric polar convection axis was slightly tilted in the prenoon–premidnight direction and the enhanced dusk cell intruded dawnward. At 1.77 MLT, the postmidnight SAPS channel (dot symbol in green) was observed in the regime of antisunward flows, related to both the dawnward-intruding dusk cell and the Harang region. Thus, the intense westward SAPS flow was streaming in the postmidnight sector antisunward (~−2800 m/s), in concert with the dawnward-intruding dusk cell. Because of the tilted convection axes, the nightside magnetotail reconnection occurred after midnight (at ~1 MLT), where GOES-13 (symbol dot in blue) was located and was able to observe the reconnection-related dispersionless particle injections (shown in Figure 7c). In the nighttime MLT sector, the amplified westward SAPS flows were streaming sunward (~5000 m/s and ~2700 m/s), in concert with the dusk cell.

In Figure 7f, the AMPERE imagery (5:41–5:51 UT) is matched (in UT) with the two-cell polar convection map (05:45–05:48 UT) and depicts the underlying SCW2L system operational earlier, during 05:41–05:51 UT, with the equally strong duskside ↓R2-↑R1 FACs and dawnside ↑R2-↓R1 FACs. We also note the strong ↑R1 FACs (in red) marking the dawnward intruded dusk cell and the weak ↓R2 FACs (in light blue) at subauroral latitudes, near the intense nighttime and postmidnight westward SAPS flows detected by DMSP latter on; thus, we were able to document the flow of ↓R2 FACs to subauroral latitudes into the old (i.e., previously formed) SAPS flow channel.

In Figure 7g, the DMSP/SSUSI imagery taken at 05:43 UT depicts the dawnward-intruding dusk cell’s signatures appearing in the auroral precipitation pattern observed. Based on the SSUSI Google EDR auroral specifications, the highest energy auroral precipitations are discrete auroral arcs. Their precipitated electrons, creating the auroral emissions observed, were associated with ↑FACs including the ↑R1 FACs shown by the AMPERE imagery as strong upward currents (in red), of which intensifications were caused by the cross-tail currents in the near-earth plasmasheet [85]. We marked (as dot symbols in colors) the westward SAPS flow channel locations situated equatorward of the auroral precipitation regime on the duskside and dawnside along with the GOES-13 footprints (in blue) situated within the polar cap.

3.6. Inner-Magnetosphere SAPS E Field and the Harang Region’s Earthward Edge

We also searched for inner-magnetosphere SAPS and Harang region observations made by THEMIS that could be matched with the DMSP F15 observations. Because of the Harang region’s rare observations by THEMIS, we cannot show any correlated M-I conjugate THEMIS-DMSP observations. However, we identified three Harang events observed by THEMIS in 2015 (listed in

Table 5. Inner-magnetosphere SAPS events investigated and THEMIS observations made.

THEMIS Observed SAPS Event					THEMIS Observables
Number	Date	UT (Hr:Mn)	MLAT (°S)	MLT (Hr:Mn)	EX (mV/m)	EY (mV/m)	EZ (mV/m)
1	20 September 2015	3:36	63.48	00:47	−10	12	9
2	21 September 2015	3:21	62.05	00:28	−13	16	8
3	7 October 2015	3:02	54.57	00:47	15	8	11

and shown in Figure 8 and Figure 9); here, only one event, the 20 September 2015 THEMIS Harang event, is loosely correlated with the 20 September 2015 DMSP F15 SAPS event.

Figure 8 is constructed with three sets of inner-magnetosphere THEMIS data depicting the inner-magnetosphere SAPS E field and the Harang region’s earthward edge observed by THEMIS in three events: on 20 and 21 September 2015 by TH-A and on 7 October 2015 by TH-D.

In Figure 8a–c, the three sets of orbit plots (X versus Z and X versus Y; in GSM) are shown. Each set depicts the THEMIS orbit sections of interest and the locations of SAPS E field (E_SAPS; dot symbol in color) and convection E field (E_C; star symbol in color) observed within the Harang region’s earthward edge. While traveling tailward, THEMIS observed these E fields above the magnetic equator and on the dawnside. In these regions, outward and duskward are positive, and antisunward is negative.

In Figure 8d–f, the three sets of THEMIS line-plot series are constructed with the data of spacecraft potential (SC Pot), E field components (E_X, E_Y, E_Z; in GSM), field-aligned ion pressure (Pi) tensors measured in different directions, suprathermal electron flux sets and high-energy hot ion flux sets measured at various energy channels, and orbit parameters (MLT; L).

In Figure 8d–f, the spacecraft potential plot shows that the tailward traveling THEMIS satellite crossed the plasmasheet’s earthward edge bounded by earthward and tailward potential drops. While the smaller earthward potential drop depicts the plasmapause (PP) that was newly formed (see details below), the tailward potential drop is larger. We marked the Harang region’s earthward edge (as dotted line in red) observed by THEMIS within the plasmasheet’s earthward edge. It represents the most earthward transported plasma (by injection and E × B convection) from the nightside magnetotail because of the magnetotail-reconnection-related particle injections and dipolarization-related increased earthward E × B convection. Through the suddenly enhanced earthward plasma transport (by particle injection and E × B convection), the plasmasheet’s earthward edge formed and became terminated (on the earthward side) at the newly formed plasmapause, where the domination of co-rotation (on the earthward side) and E × B convection (on the tailward side) abruptly changed [86,87]. Within the plasmasheet’s earthward edge, the Harang region’s earthward edge was observed by THEMIS in the regime of the oppositely directed R2 FACs [27,88] marked here (as ↓R2 and ↑R2; in red).

In Figure 8d–f, the E_X, E_Y, and E_Z line plots illustrate the outward SAPS E field (E_SAPS; shaded interval in yellow) associated with both the ↓R2 FACs and the plasmapause (PP), and the inward convection E field (E_C; shaded interval in gray) associated with the ↑R2 FACs. Within the regime of ↓R2-↑R2 FACs, which is the Harang region’s earthward edge, all three E field components (E_X, E_Y, and E_Z) show that the E_C field on the tailward side (of the PP) and the SAPS E field on the earthward side (of the PP) converged: the E_C field (directed sunward, dawnward, and inward) converged into the SAPS E field (directed antisunward, duskward, and outward).

Figure 8d–f show that THEMIS observed the oppositely directed R2 FACs in their generation region, which is the plasmasheet’s earthward edge [89,90] in all three events. In good agreement with the studies of Gkioulidou et al. [27,87], investigating the Harang reversal’s inner-magnetospheric signatures, the Pi plot (in magenta) depicts the ↓R2 FAC signature as a well-defined positive ion pressure gradient (in the tailward direction) and the ↑R2 FAC signature as a shallow negative gradient (also in the tailward direction). Appearing between the ↓R2-↑R2 FACs, the ion pressure plateau implies only weak currents flowing [91] between the oppositely directed R2 FACs. In the third event, shown in Figure 8f, TH-D continuously traveled along the magnetic equator and observed the signature of shear flow appearing as ion pressure fluctuations, depicted by the off-diagonal Pi_XY component (in orange) in the Harang region’s earthward edge.

According to the orbit parameters, THEMIS observed the Harang region’s earthward edge in the inner magnetosphere (i.e., in the regime of ↓R2-↑R2 FACs) near magnetic midnight (0–1 MLT) and close to Earth (4–5 R_E).

3.7. Inner-Magnetosphere Outward SAPS E Field Developed on a Short Timescale

In Figure 8d–f, the line-plot sets of suprathermal electron flux and high-energy hot ion flux measured at various energy channels illustrate the response of the MPFs’ electron and ion populations to the azimuthally expanding SCW, as was explained by Mishin [7] and Mishin et al. [8]. The electron flux line plots illustrate the increase in suprathermal electrons within the plasmasheet’s earthward edge, as the magnetotail-reconnection-related electron injections (i.e., the MPFs’ electron population) arrived at their stopping point located near the plasmapause (PP) on the tailward side. The 1.36–12.29 keV ion flux line plots, covering the range of ring current (~10 keV) ions, show the localized dispersionless injections associated with the SAPS E field: not only across the PP, but also earthward of the PP.

3.8. Inner-Magnetosphere Hot Zone

Figure 9a–c shows the continuation of the data presented in Figure 8d–f and depict the hot zone developed within the plasmasheet’s earthward edge, where the Harang region’s earthward edge developed. Here, we investigate the underlying heating mechanisms. As shown, the three sets of THEMIS line-plot series are constructed with the data of spacecraft potential (SC Pot), hot ion flux, ion temperature (Ti), Ti_⊥/Ti_‖, high-frequency (HF) filter bank (FBK) peak voltage characteristic to AKR waves, electron velocity components (Ve_X; Ve_Y; Ve_Z), and orbit parameters (MLT; L). In order to place the hot zone in the context of the observed Harang region’s earthward edge, we show again the SC potential plot and shaded intervals, marking the locations of the outward SAPS E field (shaded interval in yellow) and the inward convection E field (shaded interval in gray).

In Figure 9a–c, the line plot of hot ion flux (measured at ~16 keV; in red) depicts a localized peak at the plasmapause (PP), marking its location. Meanwhile, the line plot of Ti (in magenta) depicts the hot zone (shaded section in light magenta), appearing as a localized Ti increase near the PP and within the plasmasheet’s earthward edge.

In order to investigate the underlying heating mechanisms, we analyze the ion flux (~16 keV), which is characteristic to the ring current (RC) ions (10–50 keV) [92], and the Ti_⊥/Ti_‖ ratio. Within the hot zone, the Ti and ion flux line plots show strong similarities implying the earthward intrusion of RC ions (10–50 keV) [92], which peaked locally near the PP (at L ≈ 4 R_E) where the RC energy density peaked [93]. Here, in the postmidnight sector, the RC ions intruded earthward into the plasmasheet’s earthward edge during the unfolding magnetotail-reconnection-related particle injections and dipolarization (see details in Section 3.6). Then, the plasmapause was newly formed (as was explained in Section 3.6) and coincided with the earthward boundary of the plasmasheet‘s earthward edge, where the plasmasheet’s earthward edge became terminated. Observational evidence is provided by the unfolding earthward hot ion injections (shown in Figure 8d,e) and by the nature of the plasmasheet ions in the plasmasheet‘s earthward edge (shown in Figure 9). Regarding their nature, these plasmasheet ions were mostly isotropic (|Ti_⊥/Ti_‖| ≤ 1) in the first two events shown in Figure 9a,b or weakly anisotropic (|Ti_⊥/Ti_‖| > 1) in the last event shown in Figure 9c. Thus, the plasmasheet material (transported by injection and E × B convection from the tail region) was still fresh (i.e., isotropic or weakly anisotropic), since the ion population becomes more anisotropic as it spends more time within the plasmasheet’s earthward edge [94].

Under these isotropic or weakly anisotropic Ti conditions, ion ring instability cannot develop. Thus, the heating generated was due to the various waves excited by the locally developed azimuthal electron diamagnetic currents [8,95,96], as implied by the locally increased electron drift (Ve; shaded interval in cyan; see details below).

As the Ve line plots show in Figure 9a–c, the electron drift increased locally (shaded interval in cyan) near the PP. Based on the studies of Mishin et al. [8,95] and Mishin and Streltsov [96], we specified the locally increased Ve drift as the locally increased diamagnetic electron drift and its location as the stopping point (of the earthward traveling MPFs’ hot electron population), also called entry layer, situated near the PP. At the stopping point/entry layer, the locally increased diamagnetic electron drift drove electron diamagnetic currents that excited various waves such as electrostatic ion–cyclotron (EIC) waves, fast magnetosonic (MS) waves and EMIC waves [8,95,96], generating heat in the hot zone through wave–particle interactions.

Shown by the plot of high frequency (HF) filter bank (FBK) peak voltage, where the HF range matches the frequency range of AKR waves [69], THEMIS observed the most intense AKR waves in the regime of outward SAPS E field (shaded interval in yellow), partially overlapping the stopping point (shaded interval in cyan). Such localized AKR wave increase (across the outward SAPS E field) is due to a set of common underlying phenomena. These include the magnetotail-reconnection-related and earthward-directed suprathermal electron injections comprising the MPFs’ hot electron population [13], where the suprathermal electron injections provide free energy to AKR development [33]. Appearing near the outward SAPS E field and within the hot zone, the intense AKR waves developed imply that the hot zone’s underlying heating was also generated by the wave–particle interactions occurring between the AKR waves and thermal electrons/ions and unfolding within the stopping point (shaded interval in cyan). These interactions occurred additionally to the above-described heating generated by the various (EIC, MS, EMIC) waves excited by the locally developed azimuthal electron diamagnetic currents within the stopping point [8,95,96].

3.9. SAMI3 Simulations Generated for the THEMIS-Observed Harang Events Mapped Down to the Topside Ionosphere

Our investigation also includes model simulations generated by SAMI3 [97], which is a Naval Research Laboratory (NRL) ionosphere–plasmasphere (I-P) model that computes the E × B drift transporting the plasma [98]. For this study, we used the CCMC-provided version 3.22 of SAMI3 that was coupled with the NRLMSIS2.1/HWM14 neutral atmosphere model. We requested SAMI3 simulation runs for the three THEMIS SAPS events (shown in Figure 10, Figure 11 and Figure 12) and for two DMSP SAPS events occurring on 22 February 2015 and 23 January 2016 (shown in Figure 13 and Figure 14).

For the three inner-magnetosphere Harang events observed by THEMIS, we generated SAMI3 simulations (shown in Figure 10, Figure 11 and Figure 12) for the conjugate southern-hemisphere topside ionosphere (altitude of 840 km) and for the closest UT intervals of the THEMIS SAPS detections.

In Figure 10a, Figure 11a and Figure 12a, a section of the southern geographic map shows the footprints of THEMIS and the mapped-down SAPS E field (E_SAPS; dot symbol in color) and convection E field (E_C; star symbol in color) locations situated in the South Atlantic sector. For this geographic region, the mapped SAMI3 simulations of Ne (in Figure 10b, Figure 11b and Figure 12b) and Te (in Figure 10c, Figure 11c and Figure 12c) are shown along with the mapped-down THEMIS-observed E_SAPS and E_C locations. These SAMI3 Ne and Te maps depict the coinciding Ne trough and Te peak over the South Atlantic where the Harang region’s earthward edge was observed by THEMIS in the conjugate inner magnetosphere.

In Figure 10d, Figure 11d and Figure 12d, the MLT versus MLAT polar plot shows the underlying two-cell polar convection pattern (based on the matched SuperDARN convection map), depicting the lowest equipotential contour of the dusk cell (in blue) and dawn cell (in red) and the dawnward-intruding dusk cell. Here, we plotted the THEMIS footprints with the mapped-down E_SAPS and E_C locations situated at the equatorward edge of the dawnward-intruding dusk cell in the postmidnight MLT sector. Due to the lack of MLT and MLAT simulation data, we graphed the SAMI3 zonal E × B drift as GLON versus GLAT polar map (show in Figure 10e, Figure 11e and Figure 12e) and geographic map (shown in Figure 10f, Figure 11f and Figure 12f). Both maps show the dawnward-intruding dusk cell with the increased westward (−) drift (in blue) and the dawn cell with the increased eastward (+) drift (in red). We note here that the THEMIS-observed outward E_SAPS mapped down to the SAMI3-simulated increased westward drift regime, as the resultant poleward SAPS E field drove the plasma westward (or duskward) at the equatorward edge of the dawnward-intruding dusk cell. Meanwhile, the THEMIS-observed inward E_C mapped down to the SAMI3-simulated weak eastward drift regime as the resultant equatorward convection E field drove the plasma eastward (or dawnward) in the poleward region of the dawnward-intruding dusk cell.

Overall, in Figure 10, Figure 11 and Figure 12, the mapped SAMI3-simulated variables (Ne, Te, zonal E × B drift) show good correlations with the mapped-down THEMIS observed E fields (E_SAPS and E_C). Furthermore, these SAMI3 maps are also in good agreement with the topside-ionosphere DMSP F15 observations, as both depict the following aspects:

(1) The deep Ne trough ((5–15) × 10³ cm⁻³; shades of dark purple), where the simulated Ne minimum (5 × 10³ cm⁻³) is similar to the Ne values measured by DMSP F15 (shown in Figure 3) within the plasma-density trough.

(2) The elevated Te (3000–4000 K; shades of red) where the simulated Te maximum (~4000 K) is close to the lower range of Te values measured by DMSP F15 within the (coinciding RIT-MIT) trough.

(3) The dusk-cell-related westward (−) zonal E × B drift {−(712–915) m/s; shades of blue; reproducing the intense westward SAPS flow} along with the dusk-cell-related eastward (+) convection (E_C) flow (305–508 m/s; shades of light red) in the Harang region. But the SAMI3-simulated westward drift (712–915 m/s) is lower than the cross-track V_HOR (1200–3000 m/s in magnitude) measured by DMSP F15.

For two DMSP SAPS events, on 22 February 2015 (shown in Figure 3b) and 23 January 2016 (shown in Figure 3c), we generated SAMI3 simulations (shown in Figure 13 and Figure 14). These simulations cover not only the geographic region of interest, but also a (25–65)°S geographic latitude section of the descending DMSP F15 pass. These SAMI3 simulations were generated for the topside ionosphere (at 840 km altitude) and for the closest time (in UT) of the SAPS observations made by DMSP F15. These SAMI3 simulations are shown in Figure 13 and Figure 14, which are constructed in the same style. By adding the latitudinal cross sections of the SAMI3-generated Ne, Te, and zonal E × B, we could directly compare the SAMI3-simulated and F15-measured variables for the SAPS channel.

In Figure 13a and Figure 14a, the southern-hemisphere geographic map covers the African and Australian longitude sectors, and shows the ground track of the descending F15 pass over the Indian Ocean. Here, we highlighted (in magenta) the (25–65)°S geographic latitude section of interest and marked the SAPS location (symbol dot in color).

In Figure 13b and Figure 14b, the line-plot sets cover the southern geographic latitude, in the equatorward direction, from 25°S to 65°S, and are constructed with the SAMI3-simulated variables of Ne, Te, and E × B at 840 km altitude. These line plots show that, for the time of DMSP F15 SAPS detection, SAMI3 was able to reproduce the plasma-density trough where Te peaked, the increased zonal E × B drift in the SAPS channel located on the equatorward side of the coinciding Ne trough–Te peak in Figure 13b, and within the coinciding Ne trough–Te peak in Figure 14b. In

Table 6. DMSP and SAMI3 variables at trough and SAPS latitudes.

Variables	22 February 2015		23 January 2016
	DMSP	SAMI3	DMSP	SAMI3
Ne (1000/cm3)	2.4700	4.0131	2.9300	2.6415
Te (K)	6400.0	2245.4	5720.0	2251.6
zonal E × B (m/s)	−1200.800	−146.400	−2242.300	−79.208

, we compare the observed and modeled variables. These show that the SAMI3 simulations most accurately reproduced the electron density (Ne), but the values of Te and zonal E × B (i.e., V_HOR) are largely underestimated.

In Figure 13c and Figure 14c, these variables are mapped along with the F15 ground track and SAPS location. As shown, the trough was reproduced over the Indian Ocean wherein Te locally peaked and the westward zonal drift also increased locally.

4. Discussion

This study is based on the rare (only 10) observations of amplified SAPS flow that DMSP F15 made in the topside ionosphere, over the Southern Hemisphere near Africa and over the South Atlantic, during the 0–1 MLT hours in 2015–2016 (as shown in Figure 3). These 10 SAPS events depict the amplified westward SAPS flow: streaming antisunward (−) in the postmidnight sector and sunward (+) at midnight (i.e., in concert with the dawnward-intruding dusk cell). In these longitude sectors, the geomagnetic field is weak (due to the SAMA) and therefore the polar E × B convection flows (v = E × B/B²) and associated return flows are stronger than in other longitude sectors. Based on their respective underlying polar convections, these amplified SAPS flows occurred near the Harang region, as implied by the dawnward-intruding dusk cell (see Figure 5 and Figure 6). Near the Harang region, the westward SAPS flows appeared stronger (up to ~3000 m/s in magnitude) near magnetic midnight than usually in the dusk/evening sector (~1000–2000 m/s) reported by previous statistical studies [2,54]. Also, within a deep plasma-density trough (i.e., coinciding RIT-MIT), Te increased up to 7000 K, implying the development of the SAR arc [45]; we demonstrated this with the 9 September 2015 SAPS event (shown in Figure 7).

In three events, THEMIS observed the inner-magnetosphere Harang region’s earthward edge. There, the oppositely directed R2 FACs were located at similar MLT [27] and developed within the plasmasheet’s earthward edge, where the outward SAPS E field (in the regime of ↓R2 FACs) and the inward convection E field (in the regime of ↑R2 FACs) converged (see Figure 8d–f). THEMIS also observed the signatures of fast-time SAPS E field development [7,8,16] and the signatures of increased waves generated by both the electron diamagnetic currents [8,95,96] and the AKR waves developed due to the free energy provided by the earthward injected MPFs [33]. Consequently, the resultant enhanced wave–particle interactions fueled the hot zone (see Figure 9a–c).

Based on recent theories and models, we explained the M-I conjugate SAPS phenomenon in the 0–1 MLT sector. By invoking the novel fast-time SAPS generation theory of Mishin [7,8,16], we demonstrated the characteristic signatures of the newly formed SAPS E field (observed by THEMIS), providing observational evidence that the outward SAPS E field developed on a short timescale across the newly formed plasmapause. By invoking the analytical SAPS model of Lyatsky et al. [59], connecting stronger SAPS flows (in the broader midnight MLT sector) to the enhanced polar convection (resulting in the dawnward and equatorward expansion of the dusk cell), we demonstrated the development of amplified westward SAPS flows near magnetic midnight. These amplified SAPS flows were observed along with the dawnward-intruding dusk cell, where the Harang region develops. Although the model of Lyatsky et al. [59] does not require particle injections, which are physically interconnected with the development of outward SAPS E field on a short timescale [7,8,16] (that we demonstrated with the THEMIS observations covering three events), Lyatsky’s model [59] places the amplified westward SAPS flow in the Harang region, where the dusk-cell-related westward SAPS flow is better developed [35,36,87]. Traditionally, the SAPS flow is considered separate from the convection pattern [4,97]; however, recent studies consider a direct connection between the polar convection and the subauroral flow. Sangha et al. [98] provided observational evidence of the bifurcation of R2 FACs into subauroral latitudes and explained it with the development of a second R2 loop (located more earthward/equatorward that the first R2 loop in the SCW2L system). The authors put forward a modified version of the SCW2L system that allows the flow of R2 FACs at subauroral latitudes because of the second R2 loop [98]. In this scenario, the direct connection of SAPS with the R2 FACs implies that the westward SAPS flow is directly connected to the dusk convection cell. Maimaiti et al. [99] investigated the morphology of polar convection and observed the dawn–dusk convection asymmetry—created by the IMF B_Y component—also in the subauroral flow pattern. Thus, the IMF B_Y dependence of subauroral flow pattern provides observational evidence of the direct connection between the polar convection and subauroral flow via the convection E field’s propagation to subauroral latitudes [99].

We also analyzed the simulation results produced by SAMI3 along with the NRLMSIS2.1/HWM14 neutral atmosphere model. SAMI3 is a first-principles physics model that uses a self-consistent E field to compute the E × B plasma drifts approximating plasma motions [100,101]. Our SAMI3 simulations demonstrate that SAMI3 reproduced quite well the enhanced zonal westward E × B plasma drift at subauroral latitudes in the SAPS channel located in the magnetic midnight sector of the topside ionosphere (at 840 km altitude; see Figure 10, Figure 11, Figure 12, Figure 13 and Figure 14), near the Harang region. SAMI3 was also able to reproduce the deep Ne trough and elevated Te in the SAPS channel, where the reduction of Ne occurred by convection and by heating due to the locally increased Te (enhancing ion recombination via increased neutral-ion collisions [102]). These results add to the findings presented in the recent study by Huba et al. [103], reporting that, along with the Rice Convection Model (RCM) [104], the coupled SAMI3-RCM model was able to reproduce the SAPS channel with realistic westward zonal E × B values (~1500 m/s in magnitude) and the underlying electron-density trough developed over each hemisphere at 20 MLT in the ionosphere (at 315 km altitude) during the 17 March 2015 geomagnetic storm. Possibly, in our SAMI3 simulations, the NRLMSIS2.1/HWM14 neutral atmosphere model created inherent limitations in the accurate reproduction of elevated Te and amplified westward zonal E × B drift in the SAPS channel.

5. Summary of New Results

Our experimental findings have provided the following new results documenting:

(a)

The amplified near-midnight westward SAPS flows at the Harang region’s equatorward edge (in 10 events) and the associated SAR arc (in 1 event).

(b)

The convergence of the large-scale inner-magnetosphere E fields (in three events).

(c)

The inner-magnetosphere hot zone (in three events).

(d)

The AKR waves (in three events).

We generated SAMI3 simulations for the following aspects:

• The conjugate topside ionosphere (at 840 km altitude) of the three inner-magnetosphere Harang events observed by THEMIS.
• Two topside-ionosphere SAPS events observed by DMSP F15.

(e)

The demonstrated ability of SAMI3 to reproduce the Harang region’s equatorward edge with an intense westward E × B drift in the SAPS channel, located within the deep Ne trough, where Te became locally elevated.

6. Conclusions

From these new results (a–e), we conclude that the near-midnight westward SAPS flow became amplified for the following reasons:

(1)

Because of the strong inner-magnetospheric outward SAPS E field’s development in the inner-magnetosphere Harang region’s earthward edge.

(2)

As a result of the combined effects of the following aspects:

• The convection E field’s propagation to subauroral latitudes, resulting in the equatorward and dawnward expansion of the dusk convection cell near the Harang region.
• Positive feedback mechanisms in the deep plasma-density trough (i.e., coinciding RIT-MIT) where Te became elevated.

Finally, we note here that, in future studies, we plan to investigate closely correlated M-I conjugate observations provided by a complete DMSP dataset (i.e., including particle and magnetic data). Although our current study could not present observations with a complete DMSP dataset and could not provide any correlated M-I conjugate observations—these being the major shortfalls of this study—our results are still significant because they enhance our understanding of the M-I conjugate SAPS and Harang phenomena that remain poorly understood.