Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
 Interaction with EMIC (electromagnetic ion cyclotron) waves is thought to be a key component contributing to the very rapid loss of both ring current and radiation belt particles into the atmosphere. Estimated loss rates are heavily dependent on the assumed spatial distribution of the EMIC wave. Statistical maps of the spatial distribution have been produced using in-situ satellite data. However, with limited satellite data it is impossible to deduce the true spatial distribution. In this study, we present ground-based observations using all-sky imager and search coil magnetometer networks, which provide the large-scale distribution and motion of the EMIC wave-particle interaction regions. We observed several spots of isolated proton auroras simultaneously with Pc1/EMIC waves at subauroral latitudes during the expansion phase of a storm-time substorm on 9 March 2008. The isolated auroras were distributed over ∼4-hours MLT preceding midnight. The POES-17 satellite confirmed enhancements of 30-keV proton precipitations over the isolated auroras. The equatorward motion of the auroras and frequency drift of the wave were consistent with the plasmasphere eroding due to a polar cap potential enhancement modeled by a numerical simulation. We also found that relativistic electron precipitation was not always associated with the isolated aurora, depending strongly on the plasma density profile near the plasmapause. This study shows that the specific distribution of ring current proton precipitation can be visualized through the ground network observations. By combining with upcoming inner-magnetosphere satellite missions, these remote-sensing observations are very important for quantitative understanding of the particle loss in the inner magnetosphere.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 The electromagnetic ion cyclotron (EMIC) wave is one of the loss sources for ring current protons and radiation belt electrons due to pitch angle scattering through the resonant wave-particle interactions [Cornwall et al., 1970; Thorne and Kennel, 1971]. The resonant protons scattered by the wave into the loss cone precipitate into the atmosphere and cause proton auroras at subauroral latitudes. Such proton auroral emissions have been observed by global FUV imaging observations from the satellite [e.g., Immel et al., 2002; Burch et al., 2002; Frey et al., 2004; Yahnin et al., 2007, 2009], and also ground-based Hβ imaging observations [Sakaguchi et al., 2007]. The subauroral proton arcs on the dusk side, so-called detached proton auroral arcs, were found to be caused by EMIC waves and be linked to the plasmaspheric plume [Spasojevic et al., 2004; Spasojevic and Fuselier, 2009] where the wave growth rate maximizes [Jordanova et al., 2007]. On a large scale, the EMIC wave-particle interaction region on the dusk side is concentrated within the density structures of the plasmaspheric plume. On the other hand, fine-scale structures of proton auroras related to the EMIC wave have been found from ground-based imaging and low-altitude satellite observations. These observations show that the luminous/precipitation width is limited to less than a few degrees of latitude, and the distribution is intermittent in longitude [Sakaguchi et al., 2008].
 To reveal the global distribution of the EMIC waves in the inner magnetosphere is important for understanding the dynamic variations of energetic particles. The proton aurora imaging observations could potentially be used to visualize the ion cyclotron wave and particle interaction region in the field of view. However, the coverage of a single ground-based imager is limited while the global imaging has limited spatial resolution. In this paper, we show the THEMIS-ASI array [Donovan et al., 2006; Mende et al., 2008] observations of the isolated proton aurora event related to the EMIC wave throughout North America. The mosaic semi-global views obtained from THEMIS-ASIs demonstrated the distribution and motion of the wave-particle interaction region during a particular event in the substorm expansion phase, and concurrent variations with the plasmapause.
2.1. Aurora and Pc1 Observations on the Ground
 The event occurred during the expansion phase of an auroral substorm that occurred during a small geomagnetic storm on March 9, 2008. The AL index fell below −500 nT at 02–06 UT, and the Dst index reached a minimum of −86 nT at 06 UT. Five THEMIS ASIs observed isolated auroral emissions at subauroral latitudes equatorward of intense electron and proton auroral zone aurora over North America. Figure 1shows a seven-ASI mosaic at 0323 UT. The imagers at Gillam (AACGM latitude: 67.1°) and Sanikiluaq (66.3°) observed intense substorm aurora activities. The five subauroral imagers at Prince George (59.8°), Athabasca (62.6°), The Pas (64.7°), Pinawa (61.1°), and Kapuskasing (60.9°) observed the equatorward boundary of the auroral oval, and several isolated aurora spots indicated inFigure 1 by the green arrows, which were intermittently distributed across the four hours of MLT preceding midnight (meaning 20 to 24 MLT). Animation 1 shows the formation and disappearance of the isolated auroras over four hours. The isolated aurora moved to lower latitudes concurrent with the equatorward expansion of the substorm auroral activity. Looking in detail, we can find that the individual spots drift and/or oscillate longitudinally and pulsate with periods of a few minutes. Note that the gray spot and white portion indicated by the two cross marks are snow and its reflection on the dome.
Figure 2shows the north–south keograms obtained from the NORSTAR Meridian Scanning Photometer (MSP) at Pinawa. The panels presented from top to bottom are the 558-nm oxygen (Figure 2a), 428-nm nitrogen molecular ion (Figure 2b), 630-nm oxygen aurora (Figure 2c), and Hβ proton intensities (Figure 2d) along geographic latitudes, assuming the emission heights of 110 km except for the 630-nm aurora at a height of 230 km. The equatorward-moving isolated auroras were found at geographic latitudes of 48–51°N (which in this sector corresponds to roughly 59–62° magnetic latitude) at 0315–0410 UT for all four emission lines. The Hβ keogram indicates that the isolated auroras include intense proton emissions with intensities of ∼500 Rayleigh. Note that the NORSTAR Pinawa MSP calibration is being reassessed at the time of writing, and that the quote brightnesses may be too large by as much as a factor of two. We confirmed that the OMTI Hβ proton imager at Athabasca also detected Hβemissions from the isolated aurora with intensities of ∼200 Rayleigh after the observation starting time at 0320 UT until 0350 UT (not shown). These filtered observations verify that the isolated auroras observed by white-light ASI were created from the strong proton precipitation from the magnetosphere.
Figures 2e and 2findicate the dynamic spectra of the north-south (Figure 2e) and east-west (Figure 2f) components of ground magnetic field up to 1 Hz observed by the CARISMA fluxgate magnetometer at Pinawa [Mann et al., 2008]. The Pc1 geomagnetic pulsations were observed at 0315–0415 UT simultaneously with the isolated aurora appearance. The connection of these two phenomena was found because the Pc1 frequency increased from ∼0.2 to ∼0.5 Hz, as the latitude of the isolated aurora moved equatorward. Pc1 geomagnetic pulsations were also observed by the STEL induction magnetometer at Athabasca at 0230–0350 UT (not shown), simultaneously with the isolated aurora appearances in the field of view of the Hβ imager at Athabasca.
2.2. Particle Observations by Satellite
 As shown in Figure 1, the POES-17 satellite passed across the isolated aurora at around 0326 UT. The trajectory of POES-17 is indicated by the red-white dashed line on the ASI mosaic image. The satellite footprints have been traced from its ∼800 km orbit down to 110 km altitude using the IGRF magnetic field model. From the lower to higher latitudes, POES-17 passed over two spots of a weak and some intense isolated auroras, and then entered into the region activated by the substorm. The POES satellite measures ions from 30 keV to more than 6900 keV, as well as electrons >30 keV, >100 keV, and >300 keV [Evans and Greer, 2000]. Although the POES satellite does not have sensors specifically designed to observe MeV electrons, the ion telescopes respond to relativistic electrons. A case of significant count rates in the nominal >6900 keV and >16 MeV proton energy channels of the ion telescopes but no response in the 2400–6900 keV proton energy channel would represent an unphysical proton energy spectrum, and a detector response of that character is interpreted as being due to the presences of >830 keV and >∼3 MeV electrons, respectively [Miyoshi et al., 2008; Rodger et al., 2010; Yando et al., 2011; J. Green, private communication, 2012]. Although it is difficult to determine the exact electron energy from proton data, as the response of the proton sensors depend on the input electron energy spectrum, the data can be used for a proxy of relativistic electrons.
Figure 3shows the POES-17 satellite observations of trapped (black) and precipitating (red) ions and electrons at 0325–0828 UT. From top to bottom, each panel shows the 30 keV ion (Figure 3a), >830 keV electron (Figure 3b), and >∼3 MeV electron (Figure 3c) count rates. In this event, zero counts were detected for the 2400–6900 keV proton energy channel. POES-17 observed localized enhancements of precipitating ions at two separated latitudes equatorward of the ion isotropic boundary. When the profiles of the ion precipitation (Figure 3a) are examined along the trajectory of POES-17 (Figure 1), we found that the two latitudes of the localized precipitating ion enhancements before and after 0326 UT corresponded exactly to the location of two isolated auroral emissions. For the higher-latitude enhancement, POES-17 passed over the intense arc-like isolated aurora, while for the lower-latitude enhancement, the satellite passed the edge of a weak isolated auroral spot. Although the count rate of the lower-latitude precipitation enhancement was 1-order smaller than the higher latitude one, it was possible to identify corresponding isolated aurora in the ground auroral images.
 In addition, Figures 3b and 3c show the enhancements of the >830 keV and >∼3 MeV precipitating electron counts. The outer radiation belt was observed from 0325:30–0327:00 UT. Similar relativistic electron precipitation into the isolated proton aurora has been reported by Miyoshi et al. . Interestingly, for this case, we found that the >830 keV electron precipitation was detected only in the equatorward half latitudes of >30 keV ion precipitation region in the lower latitudes enhancement. On the other hand, we found that the trapped >830 keV electrons were detected in a wide latitude range up to the 30-keV ion isotropic boundary. The omni-directional >∼3 MeV electron count was also detected in the lower latitude ion enhancement, but within a wider latitude range to the equator side than that of the >830 keV electron and >30 keV ion precipitations. These relativistic electron precipitations were also interpreted as a resulting from the pitch angle scattering by the EMIC waves [Miyoshi et al., 2008], as well as the proton precipitation. The possible reasons that the relativistic electron precipitation was not fully coincident with the proton precipitation are discussed in the following section.
 The THEMIS-ASI array observed several isolated aurora spots distributed over approximately four hours of MLT on night side at subauroral latitudes. The Pc1/EMIC waves were simultaneously observed by ground-based magnetometers at the sites near the appearances of isolated proton aurora. Simultaneous isolated aurora and Pc1/EMIC wave measurement events have been previously reported bySakaguchi et al.  using a proton Hβimager and an induction magnetometer. Their comprehensive analyses of these events using a low-altitude particle observation data set suggested that such auroras are generated by proton precipitation from the ring current due to pitch angle scattering of the EMIC waves near the plasmapause. We believe that the present event observed by THEMIS-ASI was generated by the same EMIC wave-particle interaction process in the inner magnetosphere.
3.1. Intensity Ratio of Emission Lines in Isolated Aurora
 For the event presented in this paper, the isolated auroras were identified by the white-light THEMIS-ASI observations, and the presence of significant Hβ emissions was confirmed by the MSP and filtered ASI observations. The Hβemission was typically dim and a diffuse, so it is essentially hard to confirm the emission only by white-light observations. We considered that the luminosity of the isolated aurora would be created by the secondary electrons induced when the protons precipitate into the atmosphere. The MSP data (Figure 2) shows that a visibly intense 558-nm line was emitted from the isolated aurora as well as the Hβ and other emission lines. The maximum intensities of the observed emission lines were 12 kR from 558 nm, 300 R from 428 nm, 500 R from Hβ, and 1.6 kR from 630 nm, respectively. We found that these intensity ratios; (Hβ: 558 nm : 630 nm) = (500 : 12,000 : 1,600) were consistent with the intensity ratios induced by a 20-keV pure proton precipitation; (Hβ : 558 nm : 630 nm) = (1 : 9.0–12.5 : 2.5–3.3) [Eather, 1968]. Thus, the isolated proton aurora was visualized by concurrent intense secondary electron auroras, which must be created from an almost pure proton precipitation impact on the atmosphere. Indeed, 30 keV and higher energy proton precipitations have been confirmed by POES-17 over the isolated auroras (Figure 3).
3.2. Polar Cap Potential Enhancement
 The isolated auroras did not remain still after their appearance, but moved equatorward as a whole (see Animation 1). Looking at them in detail, we found that the individual spots showed swing/drift motions longitudinally, and some of them pulsated with periods of a few minutes. These motions should be projections of the fine structure of the EMIC wave-particle interaction region at the equatorial plane in the inner magnetosphere. Here, we focus on the cause of the overall equatorward motion of the isolated auroras. At the beginning of the appearance, the isolated aurora was within the vicinity of the equatorward boundary of the high-latitude active aurora region. Then gradually, it separated from the boundary toward the equator by 1–2 degrees in latitude. These sequences suggest that energetic protons were injected by convection enhancement and/or the substorm activities into the inner magnetosphere, which results in EMIC wave excitations near the plasmapause.
Figure 4 shows a numerical simulation of the time variations of the equatorial plasma density [Rasmussen et al., 1993] during this event. The polar cap potential (the right axis of Figure 4a) derived from the PC index (the left axis of Figure 4a) [Troshichev et al., 1996] and the Volland-Stern electric field model [Volland, 1973; Stern, 1973]. Figure 4b shows the calculated equatorial plasma density at the MLT corresponding to the longitude of Pinawa. The radial profile was converted into dipole magnetic latitude (note that we used a dipole field for this calculation) as shown on the vertical axis. It must be noted here that we used the dipole field for this calculation. The red lines in Figure 4b indicate contours of the equatorial plasma densities of 10, 100, and 300 cm−3 around the plasmapause. We found that the plasmasphere began to shrink soon after the polar cap potential enhancement at 0130 UT, and consequently, the plasmapause reformed more sharply after ∼0230 UT. Based on Figure 4, we interpreted that the EMIC waves were generated by the injection of energetic protons, probably along with a temperature anisotropy (T⊥ > T∥), due to the convection enhancement and/or substorm activities near the steep plasmapause structure, and the interaction region moved Earthward along with the plasmapause motion.
Figure 5shows semi-global combination images of the calculated plasmaspheric density and EMIC wave distribution visualized by THEMIS-ASIs at 0235 UT (Figure 5a) and 0335 UT (Figure 5b). The blue contour indicates the mapped equatorial plasma density, (1, 10, 30, 100, 300, 1000 cm3), mapped in the ionosphere along the dipole magnetic field. We found that the isolated proton auroral emissions moved equatorward along with the steep plasmapause motion.
 In Figure 2c, the MSP observation of the 630-nm oxygen red aurora emissions shows a somewhat longer emission existence similar to a stable aurora red (SAR) arc at the isolated proton aurora location. The SAR arc is caused by thermal electron precipitation mainly resulting from the Coulomb collision between cold plasmaspheric electrons and hot ring current particles [Kozyra et al., 1997]. The red aurora observation suggested that the isolated auroras were co-located with the SAR arc emission suggesting overlapping latitudes of the plasmasphere and ring current in the magnetospheric equator.
3.3. Difference in Scattering Rate Between Proton and Electron
 As shown in Figure 3, we identified an interesting difference between the precipitating particles into the isolated auroras. During the event, the POES-17 satellite crossed two latitudinally separated isolated emission regions during a pass. However, the relativistic electron precipitations at the energies of >830 keV and >∼3 MeV were observed only when the satellite crossed the weak isolated aurora located at the lower latitude. Over the higher latitude isolated aurora, where the trapped >830 keV electron population were detected nevertheless, the >830 keV precipitating and >∼3 MeV component were not detected.
 We consider this is caused by the difference in pitch angle scattering rate between protons and electrons. In particular, the scattering rate of relativistic electrons is highly sensitive to changes in the resonant energy, ambient plasma density, magnetic field intensity, and ion compositions [Summers and Thorne, 2003; Albert, 2003; Miyoshi et al., 2008; Jordanova et al., 2008]. The energy dependence of the pitch angle scattering rate of the equatorial protons by the EMIC waves were calculated for two equatorial plasma density conditions, 10 cm−3 and 100 cm−3 [see Miyoshi et al., 2008, Figure 4]. The results show that the scattering rate of relativistic electrons is higher for a background density of 100 cm−3 than that for 10 cm−3. For the event presented in this paper, the lower latitude isolated aurora could be located inside of the plasmasphere where the equatorial cold plasma density is high compared to that at the higher latitude. Owing to such dependencies on the background conditions, it is expected that a luminous isolated proton aurora created at higher latitudes by the intense ion precipitation is not accompanied by relativistic electron precipitation, but a faint aurora at a lower latitudes is accompanied by the loss of relativistic electrons from the outer belt.
 Several spots of isolated proton auroras were observed by using a wide-coverage and high-resolution THEMIS-ASI array at subauroral latitudes on March 8, 2009 during the expansion phase of a storm-time substorm. We studied this event by conducting ground-based monochromatic optical (MSP and OMTI), magnetic field (CARISMA), and low-altitude satellite (POES-17) observations and through a numerical simulation of the equatorial plasma density. The findings obtained from this observation are as follows.
 1. The isolated proton auroral spots were intermittently distributed over ∼4-hour MLT in the pre-midnight sector equatorward of the auroral zone substorm auroral activity in the auroral latitude.
 2. The Pc1/EMIC waves were simultaneously observed by GMAG near the appearances locations of the isolated aurora.
 3. The MSP and OMTI observations showed that the isolated proton auroras were visualized from the electron auroral emissions that would be excited by secondary electrons created by the high-energy protons incident into the atmosphere.
 4. The high-resolution THEMIS-ASI showed that the individual isolated auroral spots swing/drifted longitudinally, and some of them pulsated with periods of a few minutes. The isolated auroras moved equatorward together with increase of the Pc1 frequency during enhancement of polar cap potential.
 5. Our numerical simulation of the plasmasphere indicated that the plasmasphere shrank Earthward and a steep plasmapause formed during the event due to polar cap potential enhancement. The EMIC wave was excited by the injection/convection of energetic protons into the inner magnetosphere related to the substorm activities. The wave-particle interaction region exists near the steep plasmapause structure, and drifted Earthward together with the plasmapause shift due to the enhanced convection.
 6. The localized enhancements of the 30 keV ring current proton precipitation were identified by the POES-17 satellite over two isolated auroras.
 7. The relativistic >830 keV and >∼3 MeV radiation belt electron precipitations were identified by POES-17 at the lower latitude isolated aurora, but not at the higher latitude aurora. This is probably because only the denser cold plasma population at lower latitudes satisfied the resonance condition enough to cause the relativistic electron precipitation.
 The THEMIS-ASI white-light observations, augmented by more local multispectral observations that can confirm the specific nature of the precipitation, are a powerful remote-sensing tool to monitor EMIC wave-particle interactions in the inner magnetosphere. We suggest that the fine scale structure and motion of the interaction region are projected in the ionospheric screen through the intermediary of the isolated proton aurora, improving our understanding of the EMIC wave distribution and related ring current and radiation belt particle dynamics in the inner magnetosphere.
 The authors would like to thank S. Mende for the use of the THEMIS-ASI data, and the CSA for their logistical support in fielding and data retrieval from the GBO stations. The authors would also like to thank D.K. Milling and the rest of the CARISMA team for providing the CARISMA magnetometer data. CARISMA is operated by the University of Alberta and funded by the Canadian Space Agency. The POES data is provided from NGDC.
 Masaki Fujimoto thanks the reviewers for their assistance in evaluating this paper.