RESEARCH LETTER Magnetospheric Signatures of STEVE: Implications 10.1029/2019GL082460 for the Magnetospheric Energy Source and Key Points: • Magnetosphere observations show Interhemispheric Conjugacy that STEVE corresponds to SAID, Y. Nishimura1,2 , B. Gallardo‐Lacourt3 , Y. Zou4,5 , E. Mishin6 , D. J. Knudsen3 , plasmapause, structured plasma 3 7 8 boundaries, and waves in the E. F. Donovan , V. Angelopoulos , and R. Raybell magnetosphere 1 • The picket fence is driven by Department of Electrical and Computer Engineering and Center for Space Physics, Boston University, Boston, MA, USA, electron precipitation; the red arc is 2Department of Atmospheric and Oceanic Sciences, University of California, Los Angeles, CA, USA, 3Department of fl driven by heat ux or frictional Physics and Astronomy, University of Calgary, Calgary, Alberta, Canada, 4Department of Astronomy and Center for Space heating Physics, Boston University, Boston, MA, USA, 5Cooperative Programs for the Advancement of Earth System Science, • Simultaneous conjugate 6 observations show that part of University Corporation for Atmospheric Research, Boulder, CO, USA, Space Vehicles Directorate, Air Force Research STEVE has interhemispheric Laboratory, Kirtland AFB, Albuquerque, NM, USA, 7Department of Earth, Planetary, and Space Sciences, University of conjugacy California, Los Angeles, CA, USA, 8Citizen scientist, Seattle, WA, USA

Abstract We present three STEVE (strong thermal emission velocity enhancement) events in Correspondence to: Y. Nishimura, conjunction with Time History of Events and Macroscale Interactions (THEMIS) in the magnetosphere [email protected] and Defense Meteorological Satellite Program (DMSP) and Swarm in the ionosphere, for determining equatorial and interhemispheric signatures of the STEVE purple/mauve arc and picket fence. Both types of

Citation: STEVE emissions are associated with subauroral ion drifts (SAID), electron heating, and plasma waves. The Nishimura, Y., Gallardo‐Lacourt, B., magnetosphere observations show structured electrons and flows and waves (likely kinetic Alfven, Zou, Y., Mishin, E., Knudsen, D. J., magnetosonic, or lower‐hybrid waves) just outside the plasmasphere. Interestingly, the event with the picket Donovan, E. F., et al. (2019). fi ‐ Magnetospheric signatures of STEVE: fence had a >~1 keV electron structure detached from the electron plasma sheet, upward eld aligned Implications for the magnetospheric currents (FACs), and ultraviolet emissions in the conjugate hemisphere, while the event with only the energy source and interhemispheric mauve arc did not have precipitation or ultraviolet emission. We suggest that the electron precipitation conjugacy. Geophysical Research Letters, 46, 5637–5644. https://doi.org/10.1029/ drives the picket fence, and heating drives the mauve as thermal emission. 2019GL082460 Plain Language Summary STEVE (strong thermal emission velocity enhancement) has Received 14 FEB 2019 become increasingly popular among citizen scientists due to its distinct colors and structures of Accepted 9 APR 2019 emission in the night sky and its occurrence over more populated areas than for typical aurora in the Accepted article online 16 APR 2019 auroral oval. This study addresses two major questions of STEVE: What is the energy source of the Published online 4 JUN 2019 STEVE purple or mauve colored arc and green picket fence up in space? and Does STEVE occur in the Northern and Southern Hemispheres at the same time? Using a set of imaging and satellite observations, this study found that STEVE is connected to fast plasma flows, sharp plasma boundaries, and intense waves 25,000 km (15,000 miles) up in space. Photographs taken by citizen scientists have played a key role in finding STEVE and its morphology. Plasma heating due to the fast flows and waves is suggested to drive the mauve colored arc. But this mechanism does not explain the picket fence. We found that energetic particle precipitation drives the picket fence. The picket fence is found to occur in both hemispheres at the same time, supporting that the energy source far up in space feeds energy to both hemispheres.

1. Introduction STEVE (strong thermal emission velocity enhancement) is a recently discovered upper atmospheric emis- sion in the subauroral ionosphere (MacDonald et al., 2018). Its primary feature is the purple or mauve color arc that is approximately east‐west aligned, which is distinctly different from aurora near the equatorward boundary of the auroral oval. STEVE arcs sometimes also exhibit green‐colored rays known as picket fence, which appear to form at lower altitudes than the mauve arc (MacDonald et al., 2018). STEVE arcs occur at premidnight MLTs (Gallardo‐Lacourt, Nishimura, et al., 2018) and are associated with the subauroral ion drifts (SAID), enhanced electron temperature, and downward field‐aligned currents (FACs) at the midlati- ‐ ©2019. American Geophysical Union. tude trough (MacDonald et al., 2018). The event studied by Gallardo Lacourt, Liang, et al. (2018) does not All Rights Reserved. show any substantial precipitation of energetic particles over the STEVE arc, suggesting that the arc is not

NISHIMURA ET AL. 5637 Geophysical Research Letters 10.1029/2019GL082460

aurora but thermal emission due to elevated the elevated temperature that is caused by heat flux or ion‐neutral frictional heating. Regardless of precipitation, a possible energy source is substorm injections because STEVE arcs tend to occur during the substorm recovery phase (Gallardo‐Lacourt, Nishimura, et al., 2018). This is consistent with the occurrence characteristics of SAID (Anderson et al., 1993, 2001). A type of stable auroral red (SAR) arc also shows a connection to substorms (Takagi et al., 2018), although STEVE arcs are distinct from SAR arcs because of their different spectral characteristics. Past studies of STEVE arcs have been limited to observations by ground‐based imaging and low‐altitude satellites. It is unknown what the magnetosphere counterpart of STEVE arcs (for both the purple/mauve arc and picket fence) is and whether STEVE arcs simultaneously occur in both hemispheres. Another major question is what process makes a difference between the mauve arc and green picket fence. Although the emission in the F region ionosphere could be explained by thermal emission under elevated temperature (also known as the mechanism of SAR arcs; such as Nagy et al., 1970; Sazykin et al., 2002), there have been no studies of the magnetospheric source of the purple/mauve arc. Moreover, the F region heating process does not create substantial green line emission, suggesting that additional unresolved processes exist asso- ciated with STEVE arcs. To address these questions, we examined magnetosphere signatures of STEVE arcs by using three conjunc- tion events with Time History of Events and Macroscale Interactions (THEMIS), whose footprints passed within 1‐hr magnetic local time (MLT) from the areas of STEVE arcs detected by the citizen scientist coauthors or by the THEMIS all‐sky imager (ASI) network. The northern footprints of the Defense Meteorological Satellite Program (DMSP) and Swarm satellites also passed through during the STEVE arcs analyzed here. The magnetospheric satellites provide measurements of DC and AC electric and magnetic fields and particle energy spectra in the inner magnetosphere. The ASIs and low‐altitude satellites were used to check that the arcs of interest were located equatorward of the auroral oval. DMSP measures ionospheric density, velocity, magnetic field, and precipitating particles every 1 s and temperature every 4 s at ~800‐km altitude. Radiation belt contamination in the particle data has been subtracted. The Special Sensor Ultraviolet Spectrographic Imager (SSUSI) instrument onboard DMSP detects far ultraviolet (FUV) emis- sions at 165‐ to 180‐nm wavelength along the orbits, which are sensitive to energetic electron precipitation. Swarm at 400‐ to 500‐km altitude provides a similar set of data except for energetic particles. Swarm also gives 16‐Hz flow (electric field) and 50‐Hz magnetic field for wave analysis.

2. Results 2.1. The 8 May 2016 Event (With Picket Fence) A STEVE arc was measured at the West Coast of North America near 55° magnetic latitude on 8 May 2016. The arc existed at least between 5:50 and 6:34 UT based on multiple citizen scientist reports. It occurred during the storm main phase (−55 nT Dst) and high auroral activity (~1,000 nT AE). One of the photographs taken at 6:00 UT is shown in Figure 1a. The mauve (or red to white)‐colored emission of STEVE was approxi- mately east‐west aligned and located equatorward of the premidnight auroral oval (seen as green diffuse aurora near the northern horizon, also seen as the equatorward boundary of white light aurora in Figure 1b), which are the typical properties of STEVE. This event also shows a distinct picket fence as green quasiperiodic rays. Using the star positions, the red channel of the image was projected to 250‐km altitude (see MacDonald et al. (2018) for the projection altitude) and is overlaid on Figure 1b. STEVE was located ~1.5–2° equatorward of the auroral oval equatorward boundary (dashed line). Fortunately, the northern magnetic footprints of DMSP, Swarm, and THEMIS‐E passed over the area during the lifetime of the STEVE arc. Although DMSP F17 and Swarm‐A were in the Southern Hemisphere, the SSUSI data on DMSP F17 show that the STEVE arc in this event was a conjugate phenomenon (Figure 1c, showing the detached arc at ~55° magnetic latitude and ~20 MLT). Thus, the Southern Hemisphere data can be used to examine this event. Both hemispheres in this MLT sector were in darkness. Here the Tsyganenko (2002) magnetic field model (T01) and OMNI solar wind were used for mapping, and given that STEVE is a subauroral phenomenon, the magnetic field mapping is fairly reliable. The oval and STEVE latitudes along the satellite trajectories may not be exactly the same as in the images due to the longitude and hemispheric differences, but structures can be compared by referring to the equatorward boundary of the auroral oval.

NISHIMURA ET AL. 5638 Geophysical Research Letters 10.1029/2019GL082460

Figure 1. Satellite orbits and imager data during the STEVE events on (a–c) 8 May 2016 and (d, e) 12 April 2008. (a) Citizen scientist photograph taken at 6:00 UT. (b) Image of panel (a) mapped to 250‐km altitude (red), THEMIS all‐sky imager data at Athabasca at 6:00 UT (gray scale), and northern foot- prints of THEMIS‐E (black), DMSP F17 (green), and Swarm‐A (cyan) using the T01 magnetic field model. (c) DMSP F17 SSUSI data in the Southern Hemisphere. (d) THEMIS all‐sky imager data at McGrath (bottom) and Fort Yukon (top) at 9:30 UT and northern footprints of THEMIS‐C and DMSP. (e) DMSP F17 SSUSI far ultraviolet emission data. The white spot in each imager of panel (d) is the moonlight. STEVE = strong thermal emission velocity enhancement; THEMIS = Time History of Events and Macroscale Interactions; DMSP = Defense Meteorological Satellite Program; SSUSI = Special Sensor Ultraviolet Spectrographic Imager.

Figures 2a–2f show DMSP F17 data of this event. The first line marks the equatorward boundary of plasma sheet electron precipitation, which corresponds to the diffuse auroral equatorward boundary. The notable features within 2° from this boundary are the trough, SAID, and ambient electron heating, which are the ionospheric plasma signatures of STEVE, consistent with the finding by MacDonald et al. (2018). No substantial ion precipitation was present equatorward of the electron plasma sheet, confirming that STEVE is not proton aurora. Interestingly, DMSP also detected another electron precipitation between the two SAID flow peaks, ~1–1.5° equatorward of (detached from) the electron plasma sheet precipitation (second vertical line). The plasma sheet electron energy fluxes near the equatorward boundary decreased monotonically with energy, while the detached precipitating flux peaked at ~10 keV with a comparable energy flux level. This difference and the SSUSI observation indicate that the detached precipitation is not a small‐scale perturbation of the oval equatorward boundary but is a separate population that has experienced additional acceleration. The acceleration process can also be inferred from the existence of a weak upward FAC on the poleward shoulder of the detached electron precipitation, surrounded by the region 2 downward FAC. The subauroral precipitation structure and upward FAC are strikingly different from typical SAID events (Anderson et al., 1993, 2001). Moreover, this electron precipitation was not seen for the STEVE event studied by Gallardo‐Lacourt, Liang, et al. (2018), which did not have a picket fence, and could not be examined for the MacDonald et al. (2018) event due to the lack of particle measurements. The ~2 mW/m2 (erg/cm2/s)‐ integrated energy flux and ~2‐keV mean energy of this population can drive visible emission dominated

by the 557.7‐nm wavelength peaking at 120‐km altitude (I557.7 ~ 2.5 kR and I630.0 ~ 1.5 kR based on the Global Airglow code modeling (Solomon et al., 1988), not shown), which could explain the green emission of the picket fence that appears at lower altitude of the mauve arc. The precipitation resembles suprathermal electron bursts seen in the auroral oval (Johnstone & Winningham, 1982; Lin & Hoffman, 1979), which also suggests wave‐particle interaction for accelerating the electron population.

NISHIMURA ET AL. 5639 Geophysical Research Letters 10.1029/2019GL082460

Figure 2. DMSP F17 and Swarm‐A data during the Figures 1a–1c event. (a) Ionosphere density, (b) cross‐track sunward (blue) and upward (red) velocity, (c) electron (blue) and ion (red) temperature, (d) FAC (positive upward), and (e, f) electron and ion precipitating energy fluxes. (g) Ionosphere density, (h) cross‐track sunward velocity, (i) electron temperature, (j) FAC, and (k, l) wavelet spectrograms of electric and magnetic fields. The along‐track flow component is not available. DMSP = Defense Meteorological Satellite Program; FAC = field‐aligned current; SAID = subauroral ion drifts.

Figures 2g–2l show the Swarm‐A data during this event. While Swarm detected similar signatures as DMSP, the electron heating is even more substantial (reaching >10,000 K), and the higher sampling rate observa- tions reveal high‐frequency oscillations in the satellite frame. The frequency spectrograms show locally enhanced waves reaching beyond several hertz. The E/B ratio is on average larger than 10,000 km/s, faster than the Alfven speed of ~4,000 km/s. Thus, the high‐frequency oscillations are likely propagating waves rather than small‐scale spatial structures. Those waves could be a high‐frequency extension of subauroral polarization streams (SAPS) wave structures (SAPSWS) because of the similarity to SAPSWS (high E/B ratio) and the broadband nature (Mishin et al., 2003). DMSP F18 and Swarm‐B also passed the subauroral region at slightly different times. As shown in Figure 3, these satellites detected similar signatures as DMSP F17 and Swarm‐A. Swarm‐B was in the Northern Hemisphere and detected essentially the same signatures as Swarm‐A in the Southern Hemisphere, indicat- ing that the features in Figure 2 are conjugate between the hemispheres. The exception is that DMSP F18 did not observe detached electron precipitation, which could be either because the precipitation ceased by the time of the satellite crossing or the satellite missed structured precipitation as the picket fence image suggests. The plasma structures in the ionosphere are analogous to another type of optical emission near the duskside equatorward boundary reported by Pedersen et al. (2007). However, a number of differences (the L‐shell alignment, lack of swirls, shorter duration, and noncorotating feature) indicate that STEVE is a different phenomenon. Figure 4 shows THEMIS observations for this event. THEMIS‐E was located just outside the plasmasphere (Figure 4d) when its footprint passed over the STEVE arc (~06:30 UT). The satellite detected enhanced sunward and duskward flows that were confined around the earthward edge of the electron plasma sheet between the two lines (Figures 4b, 4c, and 4e). The flow and electric field shows a double‐peak structure, which is consistent with the SAID observation by DMSP. For ~10‐keV electrons, the sharp flux gradient is present at the first vertical line and a detached flux exists at 6:28 UT. This electron structure

NISHIMURA ET AL. 5640 Geophysical Research Letters 10.1029/2019GL082460

Figure 3. DMSP F18 and Swarm‐B data during the Figures 1a–1c event. The format is same as in Figure 2. DMSP F18 elec- tron temperature data are not available. Swarm velocity data after 05:49:50 UT are not shown due to low quality. DMSP = Defense Meteorological Satellite Program; FAC = field‐aligned current.

Figure 4. THEMIS‐E and THEMIS‐A data during the Figures 1a–1c event. (a) Magnetic field deviation from the T01 quiet time field, (b) electric field, (c) ion velocity moment perpendicular to the magnetic field, (d) total plasma density from the spacecraft potential, (e, f) electron and ion omnidirectional fluxes, and (g, h) electric and magnetic field wave spectro- grams. (i–p) Same as panels (a)–(h) for THEMIS‐A. The total density data from THEMIS‐A are not available. The white lines in the wave data show the lower‐hybrid frequency. THEMIS = Time History of Events and Macroscale Interactions.

NISHIMURA ET AL. 5641 Geophysical Research Letters 10.1029/2019GL082460

Figure 5. DMSP F16 and THEMIS E data during the Figure 1d‐e event. The format is same as in Figures 2a‐f and 3a‐h. The stepwise changes in panel (g) are artificial due to shadow effects. DMSP = Defense Meteorological Satellite Program; FAC = field‐aligned current; THEMIS = Time History of Events and Macroscale Interactions; SAID = subauroral ion drifts.

also resembles the DMSP F17 observations in Figure 2e, while weaker fluxes exist until 6:36 UT and this population may not precipitate to the ionosphere. Based on the similarities between the THEMIS and DMSP observations, THEMIS‐E likely passed the L‐shells of the SAID. However, in contrast to the typical equatorial SAPS observations (Mishin & Burke, 2005; Nishimura et al., 2008), the enhanced plasma drift penetrated into portion of the electron plasma sheet where the electron flux did not monotonically decrease earthward but had substructures, including an ~10‐keV flux enhancement at 06:27–06:30 UT (Figure 4e). Enhanced ring current ion fluxes are also collocated with the enhanced flow and the structured electron earthward boundary. The sharp gradient of the ~10‐keV ion flux likely marks the earthward edge of the ring current since a sharp Bz gradient coincides with the ion flux gradient. These ions in the region of the enhanced flows seem to be trapped ions because DMSP did not detect ion precipitation. The ion flux data also show cold (~10 eV) ions at the peak of the V⊥ flow speed at 06:35–36 UT, which are likely portion of the cold ions reaching the measurable energy range due to the drift (Lee & Angelopoulos, 2014). The enhanced flows are also collocated with broadband electrostatic waves that extend up to the lower‐hybrid frequency (Figure 4g). A similar result was reported by Mishin and Burke (2005) except that the electric field is more confined in the present event. The electrostatic and broadband nature up to the lower‐hybrid frequency suggests that these are lower‐hybrid (Mishin & Burke, 2005), magnetosonic (Mishin & Sotnikov, 2017), or kinetic Alfven waves (Chaston et al., 2015), and those waves have been shown to convert injected plasma energy to electron heating. Thus, these waves may increase ionosphere electron temperature by electron precipitation or heat flux, which could be the energy source of the mauve arc of STEVE. Eight hours later, THEMIS‐A traversed the same L and MLT region well after the STEVE observation, and the inner magnetosphere environment was drastically different. The SAID electric field and waves were not present, and the electron and ion earthward edges became much more smooth and gradual. The ion and electron earthward edges were not collocated, but the electron earthward edge was located ~0.5 RE outward, which is typical in the duskside inner magnetosphere. The comparison of the two THEMIS satellite observa- tions highlights the enhanced electric field and waves in the radially narrow region between the ion and electron earthward boundaries as the unique equatorial features during the STEVE event. The intense

NISHIMURA ET AL. 5642 Geophysical Research Letters 10.1029/2019GL082460

electric field and waves would be a possible source of heating for the mauve STEVE arc, and the structured electron boundary may drive precipitation for the picket fence. 2.2. The 12 April 2008 Event (Without Picket Fence) Figures 1d and 1e show a STEVE event where an arc was measured by one of the THEMIS ASIs as a faint arc located ~1° equatorward of the diffuse aurora. The imagers detected the arc between 08:30 and 10:30 UT during moderate auroral activity (AE ~ 500 nT). Although color imaging is not available for this event, the arc did not show small‐scale structure such as a picket fence. DMSP F16 and THEMIS‐C's footprint crossed this region during the arc's presence. DMSP F16 SSUSI data in the Southern Hemisphere did not detect a corresponding arc. Within the MLT coverage of SSUSI, this arc is not associated with substantial electron precipitation, in contrast to the Figures 1a–1c event. Figures 5a–5f show DMSP F16 in situ data of this event. Similar to Figures 2a–2f, the subauroral ionosphere within a degree from the electron equatorward boundary shows SAID, electron heating, upward FAC on the poleward shoulder of SAID, and the lack of ion precipitation. Contrary to the previous event, no substantial electron precipitation above 100 eV existed in this region, but a small flux of <100‐eV electrons was seen in the region of the subauroral upward FAC. The energy is too low to drive optical emission, but this population may be a high‐energy tail of heat flux that increases the temperature. THEMIS‐C data for this event are presented in Figures 5g–5m. Similar to Figures 4a–4h, SAID were located just outside the plasmapause, in the narrow separation between the plasma sheet electron inner edge, and enhanced ring current ions. The cold ion population is more clearly seen because the flow speed is larger in this event. Broadband waves were collocated with the SAID. In contrast, the enhanced SAID flow is confined to the region earthward of the electron plasma sheet. The waves would only interact with cold electrons, and this feature may explain the lack of >100‐eV electron precipitation but with heating in the ionosphere in this event.

3. Conclusion In order to understand magnetospheric signatures of the STEVE purple/mauve arc and picket fence, we have examined three STEVE events that have coordinated observations with equatorial and low‐altitude satellites. Consistent with the event studied by MacDonald et al. (2018), the STEVE arcs are associated with SAID, elec- tron heating, and trough. Particle observations by DMSP show that the STEVE arcs are located equatorward of ion precipitation, confirming that STEVE is not proton aurora. The STEVE arcs are also located equator- ward of the bulk of the electron plasma sheet precipitation. In the second and third events, STEVE is not asso- ciated with >~1 keV precipitation, in agreement with Gallardo‐Lacourt, Liang, et al. (2018). However, our observations have shown several interesting differences from the past findings. In the event with the picket fence, the STEVE arc corresponds to substantial electron precipitation of >~ 1 keV, upward FACs, and electrostatic waves. The electron precipitation could be the possible driver of the picket fence that occurs at lower altitude than the mauve STEVE arc and shows more dynamic structures. The equatorial magnetosphere observations have also revealed structured plasma sheet electron fluxes, flows including SAID, and broadband waves around the earthward boundary of the electron plasma sheet. The structured electron fluxes are consistent with the electron precipitation over the STEVE arc seen by DMSP. Broadband waves are enhanced up to the lower‐hybrid frequency and are likely kinetic Alfven, magnetosonic, or lower‐hybrid waves. Those waves are suggested to create electron precipitation or heat flux that drives the STEVE arcs. The intense SAID could also contribute to ion‐neutral frictional heating. When the earthward boundary of the electron plasma sheet is structured by flows and waves, an additional layer of electron precipitation occurs and it could be the source population of the picket fence. For the event without the picket fence, DMSP did not detect >~1 keV electron precipitation but instead measured a localized enhancement of <100‐eV electrons. In the magnetosphere along the THEMIS‐E orbit, the SAID and waves are also located predominantly earthward of the electron plasma sheet just outside the plasmapause. In this event, it is likely that heating by low‐energy precipitation, heat flux, or ion‐neutral friction contributed to the F region ionosphere heating. Such heating processes can excite thermal emission in the ionosphere (Nagy et al., 1970; Sazykin et al., 2002) and could explain the STEVE arc without energetic electron precipitation.

NISHIMURA ET AL. 5643 Geophysical Research Letters 10.1029/2019GL082460

The DMSP SSUSI observation in the Southern Hemisphere has provided the conjugacy aspect of the STEVE arcs. For the event with the picket fence, the STEVE arc is seen in both hemispheres, indicating that the STEVE picket fence is a conjugate phenomenon. SSUSI did not detect substantial FUV emissions for the event without the picket fence. These observations support the idea that the picket fence is driven by precipitation to both hemispheres, while the purple/mauve arc is not due to precipitation but to thermal processes in the ionosphere.

Acknowledgments References This work was supported by NASA grants NNX17AL22G and Anderson, P. C., Carpenter, D. L., Tsuruda, K., Mukai, T., & Rich, F. J. (2001). Multisatellite observations of rapid subauroral ion drifts Journal of Geophysical Research 106 – 80NSSC18K0657 and NSF grants PLR‐ (SAID). , (A12), 29,585 29,599. https://doi.org/10.1029/2001JA000128 1341359, AGS‐1737823, AFOSR Anderson, P. C., Hanson, W. B., Heelis, R. A., Craven, J. D., Baker, D. N., & Frank, L. A. (1993). A proposed production model of rapid Journal of Geophysical Research 98 – FA9559‐16‐1‐0364, and subauroral ion drifts and their relationship to substorm evolution. , (A4), 6069 6078. https://doi.org/ LRIR19RVCOR038. The THEMIS 10.1029/92JA01975 Space Science Reviews 215 mission and ASIs are supported by Angelopoulos, V., Cruce, P., Drozdov, A., Grimes, E. W., Hatzigeorgiu, N., King, D. A., et al. (2019). , (1), 9. https:// ‐ ‐ ‐ NASA contract NAS5‐02099, CSA doi.org/10.1007/s11214 018 0576 4 contract 9F007‐046101, and NSF grant Chaston, C. C., Bonnell, J. W., Wygant, J. R., Kletzing, C. A., Reeves, G. D., Gerrard, A., et al. (2015). Extreme ionospheric ion energization Geophysical Research Letters 42 – AGS‐1004736. The THEMIS, DMSP, and electron heating in Alfvén waves in the storm time inner magnetosphere. , , 10,531 10,540. https:// and Swarm data were obtained through doi.org/10.1002/2015GL066674 ‐ their websites (themis.ssl.berkeley.edu, Gallardo Lacourt, B., Liang, J., Nishimura, Y., & Donovan, E. (2018). On the origin of STEVE: Particle precipitation or ionospheric sky- Geophysical Research Letters 45 – cdaweb.gsfc.nasa.gov, cedar. glow? , , 7968 7973. https://doi.org/10.1029/2018GL078509 ‐ openmadrigal.org, and swarm‐diss.eo. Gallardo Lacourt, B., Nishimura, Y., Donovan, E., Gillies, D. M., Perry, G. W., Archer, W. E., et al. (2018). A statistical analysis of STEVE. Journal of Geophysical Research: Space Physics 123 – esa.int) as daily CDF and HDF files. , , 9893 9905. https://doi.org/10.1029/2018JA025368 Journal of Geophysical Research Data access and processing were done Johnstone, A. D., & Winningham, J. D. (1982). Satellite observations of suprathermal electron bursts. , 87 – using SPEDAS V3.1; see Angelopoulos (A4), 2321 2329. https://doi.org/10.1029/JA087iA04p02321 Journal of et al. (2019). Lee, J. H., & Angelopoulos, V. (2014). On the presence and properties of cold ions near Earth's equatorial magnetosphere. Geophysical Research: Space Physics, 119, 1749–1770. https://doi.org/10.1002/2013JA019305 Lin, C., & Hoffman, R. (1979). Fluctuations of inverted V electron fluxes. Journal of Geophysical Research, 84(A11), 6547–6553. https://doi. org/10.1029/JA084iA11p06547 MacDonald, E. A., Donovan, E. F., Nishimura, Y., Case, N. A., Gillies, D. M., Gallardo‐Lacourt, B., et al. (2018). New science in plain sight: Citizen scientists lead to discovery of optical structure in the upper atmosphere. Science Advances, 3, eaaq0030. https://doi.org/10.1126/ sciadv.aaq0030 Mishin, E., & Sotnikov, V. (2017). The turbulent plasmasphere boundary layer and the outer radiation belt boundary. Plasma Physics and Controlled Fusion, 59(12), 124003. https://doi.org/10.1088/1361‐6587/aa8481 Mishin, E. V., & Burke, W. J. (2005). Stormtime coupling of the ring current, plasmasphere, and topside ionosphere: Electromagnetic and plasma disturbances. Journal of Geophysical Research, 110, A07209. https://doi.org/10.1029/2005JA011021 Mishin, E. V., Burke, W. J., Huang, C. Y., & Rich, F. J. (2003). Electromagnetic wave structures within subauroral polarization streams. Journal of Geophysical Research, 108(A8), 1309. https://doi.org/10.1029/2002JA009793 Nagy, A. F., Roble, R. G., & Hays, P. B. (1970). Stable mid‐latitude red arcs: Observations and theory. Space Science Reviews, 11,709–727. Nishimura, Y., Wygant, J., Ono, T., Iizima, M., Kumamoto, A., Brautigam, D., & Friedel, R. (2008). SAPS measurements around the magnetic equator by CRRES. Geophysical Research Letters, 35, L10104. https://doi.org/10.1029/2008GL033970 Pedersen, T., Mishin, E., & Oksavik, K. (2007). Observations of structured optical emissions and particle precipitation equatorward of the traditional auroral oval. Journal of Geophysical Research, 112, A10208. https://doi.org/10.1029/2007JA012299 Sazykin, S., Fejer, B. G., Galperin, Y. I., Zinin, L. V., Grigoriev, S. A., & Mendillo, M. (2002). Polarization jet events and excitation of weak SAR arcs. Geophysical Research Letters, 29(12), 1586. https://doi.org/10.1029/2001GL014388 Solomon, S. C., Hays, P. B., & Abreu, V. J. (1988). The auroral 6300 Å emission: Observations and modeling. Journal of Geophysical Research, 93(A9), 9867–9882. https://doi.org/10.1029/JA093iA09p09867 Takagi, Y., Shiokawa, K., Otsuka, Y., Connors, M., & Schofield, I. (2018). Statistical analysis of SAR arc detachment from the main oval based on 11‐year, all‐sky imaging observation at Athabasca, Canada. Geophysical Research Letters, 45, 11,539–11,546. https://doi.org/ 10.1029/2018GL079615 Tsyganenko, N. A. (2002). A model of the near magnetosphere with a dawn‐dusk asymmetry: 1. Mathematical structure. Journal of Geophysical Research, 107(A8), 1179. https://doi.org/10.1029/2001JA000219

NISHIMURA ET AL. 5644