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.
 The third FRONT (F-region Radio and Optical measurement of Nighttime TID) campaign was carried out during the new-moon period of May–June 2003, in order to investigate the geomagnetic conjugacy of medium-scale and large-scale traveling ionospheric disturbances (MSTIDs/LSTIDs) at midlatitudes. Seven all-sky airglow imagers were operated in Japan and Australia. For almost all clear-sky nights, we observed MSTIDs in the 630-nm airglow images with horizontal wavelengths of 100–400 km propagating southwestward in Japan and northwestward in Australia. All of them show a one-to-one correspondence of wave structures between the Northern and Southern Hemispheres, indicating strong electrodynamic coupling between the two hemispheres through the geomagnetic field line. We also found that airglow intensity variations with a timescale longer than that of MSTIDs often show a good correlation between the two hemispheres. On 28 May, we succeeded in detecting an equatorward propagating LSTID (spatial scale ∼1000 km) as airglow enhancements at the conjugate stations. The airglow peak of the LSTID in the Northern Hemisphere was ∼20 min earlier than that in the Southern Hemisphere, indicating that the observed LSTID is caused by a wave in the neutral atmosphere rather than by an electric field structure.
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.
 Medium-scale traveling ionospheric disturbances (MSTIDs) are a common feature in the midlatitude F layer. Shiokawa et al. [2003a] showed a high occurrence (50–60%) of MSTIDs in summer in the Japanese longitudinal sector, using 2 years of airglow imaging observations. Garcia et al.  suggested a high occurrence (more than 50%) in winter in the American longitudinal sector over Arecibo. Thus the investigation of MSTIDs is very important for understanding the basic variability of the midlatitude ionosphere. In 630-nm airglow images, MSTIDs are observed as wave-like structures moving southwestward in the Northern Hemisphere with typical wavelength, velocity, period, and amplitude of 100–300 km, 50–100 m/s, 0.5–1.5 hours, and 5–15%, respectively [Shiokawa et al., 2003a]. Saito et al. [1998a] were the first to show a large-scale map of MSTID structures over Japan, as variations of total electron content (TEC) measured by more than 1000 GPS receivers.
 The first FRONT (F-region Radio and Optical measurement of Nighttime TID) campaign was carried out in May 1998 in order to obtain large-scale features of the MSTIDs in 630-nm airglow images. Southwestward moving MSTIDs traveling more than 1000 km over the Japanese islands were obtained using five airglow imagers and the multipoint GPS receivers [Kubota et al., 2000; Saito et al., 2001]. The correspondence of wave structures in the airglow images and in the GPS-TEC map was fairly good [Saito et al., 2001; Ogawa et al., 2002]. The second FRONT campaign was held in August 1999 in order to obtain the southwestern edge of the MSTID propagation. Although only limited amounts of data were obtained due to cloudy sky conditions, Shiokawa et al. [2002a] suggested, on the basis of airglow imaging observations in the southern island of Japan, that there is a possible limit of southwestward MSTID propagation around ∼18° MLAT.
 The electrodynamics of MSTIDs has been extensively studied using airglow imagers, incoherent scatter (IS) radars, and low-altitude satellites. Fukao et al.  and Kelley and Fukao  showed turbulent upwelling of plasma, which probably corresponds to the MSTID structure, on the basis of the Middle and Upper Atmosphere (MU) radar in the summer nighttime ionosphere. They explained the turbulent upwelling in the framework of ionospheric instability proposed by Perkins . Miller et al.  showed that an intense electric field exists in the mesoscale structure (∼650 km) in the nighttime F layer, on the basis of simultaneous measurements by an airglow imager and the IS radar at Arecibo. Saito et al. [1995, 1998b] reported similar electric field fluctuations at midlatitudes from the DE 2 and Freja satellites. Shiokawa et al. [2003b] showed a one-to-one correspondence between the electric field pattern detected by the ion drift meters on board the Defense Meteorological Satellite Program (DMSP) satellite and the MSTID wave structures observed by a ground airglow imager. One can easily imagine that the MSTID wave structures are symmetric between the Northern and Southern Hemispheres because the electric field in the ionosphere can map to the other hemisphere along the geomagnetic field line. Otsuka et al.  was the first to report a symmetric pattern of the MSTID images obtained in the geomagnetic conjugate points between the Northern and Southern Hemispheres. They showed that the MSTID structures observed by two airglow imagers (630.0 nm) at Sata, Japan, and Darwin, Australia, have a symmetric structure in geomagnetic field.
 Large-scale traveling ionospheric disturbances (LSTIDs) are characteristic disturbances in the storm-time midlatitude ionosphere. They have a horizontal scale of more than 1000 km and propagate equatorward with a velocity of 400–1000 m/s [Hunsucker, 1982; Hocke and Schlegel, 1996]. These disturbances are probably caused by thermospheric neutral waves generated by the auroral-zone energy input [e.g., Hooke, 1968; Francis, 1975]. Ho et al. [1996, 1998] and Tsugawa et al. [2003, 2004] have shown two-dimensional images of LSTIDs as global and regional total electron content (TEC) maps using the multipoint GPS networks. Shiokawa et al. [2002b, 2003c] reported comprehensive measurements of LSTIDs using airglow images, two-dimensional GPS-TEC maps, multipoint ionosondes, an incoherent scatter radar, and a Fabry-Perot interferometer (FPI). They concluded that the drastic increases in airglow, TEC, and ionospheric peak density, and the decrease in ionospheric height, during LSTIDs are caused by passages of poleward neutral-wind enhancements. There may be some electrodynamic coupling of LSTIDs between the hemispheres because they are prominent ionospheric disturbances. However, simultaneous measurement of LSTIDs at midlatitude conjugate points has not been done yet.
 The third FRONT campaign reported in this paper was carried out during the new-moon period of May–June 2003, in order to obtain simultaneous MSTID/LSTID images through 630-nm airglow emission at geomagnetic conjugate points in the Northern and Southern Hemispheres. Five all-sky airglow imagers were operated in Japan and two were operated in Australia. All the MSTIDs observed during the campaign show symmetric structures between the two hemispheres. A conjugate LSTID event on 28 May 2003 shows a phase difference between the conjugate hemispheres.
2. Observation Overview
 The nocturnal airglow emission at a wavelength of 630.0 nm comes from the bottomside ionosphere at altitudes of 200–300 km through the interaction between O2 and O+. Because the O2 density increases toward lower altitudes and the O+ density (nearly equal to the F layer electron density) peaks at around 400–500 km, the 630-nm airglow intensity is sensitive to the F layer height and density variations and is a good indicator of the MSTIDs/LSTIDs. The recent development of the highly sensitive cooled-CCD camera enables us to obtain two-dimensional images of 630-nm airglow emission.
Figure 1 shows a map of the station locations and fields-of-view of all-sky airglow imagers (∼500 km) for the 630-nm airglow emission. The all-sky imagers were operated at Rikubetsu (43.5°N, 143.8°E), Shigaraki (34.8°N, 136.1°E), Misato (34.1°N, 135.4°E), Nishiharima (35.0°N, 134.3°E), and Sata (31.0°N, 130.7°E) in Japan and at Darwin (12.4°S, 131.0°E) and Renner Springs (18.3°S, 133.8°E) in Australia. The imagers at Nishiharima and Renner Springs were newly installed for the campaign observation, while the imagers at the other stations take airglow images routinely. Using the IGRF-2000 magnetic field model, the geomagnetic field line was traced from a 250-km altitude in the Southern Hemisphere to a 250-km altitude in the Northern Hemisphere. From the field-line tracing, the geomagnetic conjugate points of Darwin and Renner Springs in the Northern Hemisphere were calculated to be (28.7°N, 131.4°E) and (35.0°N, 134.7°E), which were very close to Sata and Shigaraki, respectively. Thus we have two pairs of geomagnetic conjugate stations (Shigaraki–Renner Springs and Sata–Darwin) for airglow imaging observations during the FRONT3 campaign.
 In this paper we use data from Rikubetsu, Shigaraki, Sata, Darwin, and Renner Springs, where all-sky airglow imagers of the same type were installed. Details of the all-sky imagers were described by Shiokawa et al. [1999, 2000]. The 630-nm airglow images were taken every 5.5 min with an exposure time of 165 s at these stations. The clock of the personal computer, which controls the airglow imager at each station, was always corrected using GPS receivers and network connections.
Figure 2 is an overview of the sky condition and MSTID appearance at each station during the FRONT3 campaign period. Unfortunately, airglow imagers at Sata and Darwin were stopped because of instrument trouble for the periods of 25 May to 1 June and 27 May to 1 June, respectively. According to the statistical results of Shiokawa et al. [2003a], we identified the MSTIDs in the 630-nm airglow images as wave structures with horizontal scale sizes of ∼100–400 km and periods of 0.5–2 hours. A high occurrence of MSTIDs can be recognized in Figure 2. The sky conditions at Renner Springs were fairly good in the dry season of inland Australia. The MSTID waves were observed every night at Renner Springs during the campaign. It is interesting to note that the wave occurrences at Darwin and Rikubetsu were relatively low compared with those at Shigaraki and Renner Springs. This fact may suggest the preference of MSTID generation at midlatitudes around Shigaraki and Renner Springs, although the image quality at Darwin was relatively low due to the frequent passage of airplanes in the images. The occurrence rates of MSTIDs determined from Figure 2 at Rikubetsu, Shigaraki, Sata, Darwin, and Renner Springs were 63%, 80%, 100%, 52%, and 70%, respectively.
 In the following data presentations, we use 630-nm airglow images in the geographical coordinates in the Northern Hemisphere. Figure 3 shows examples of the coordinate conversion at Shigaraki and Renner Springs. As shown in Figures 3a and 3c, the 630-nm airglow images in units of Rayleighs were calculated from the raw images in count rate with subtracting the sky-background intensity, using background images taken every 30 min at a wavelength of 572.5 nm. Then, the all-sky images were converted to geographical coordinates, as shown in Figures 3b and 3d. Images at Shigaraki and Sata were converted to the geographical coordinates at (29.9–39.9°N, 131.1–141.1°E) and at (26.0–36.0°N, 125.7–135.7°E), respectively, with an assumption of airglow emission altitude of 250 km (Figure 3b). Images at Renner Springs and Darwin were converted to the same (latitude, longitude) area in the Northern Hemisphere, respectively, using the IGRF-2000 model (Figure 3d). To avoid longer timescale variations and spatial inhomogeneity of airglow intensity, we used deviation images in units of percent from 1-hour running averages. The deviations Id are calculated from raw intensities I and 1-hour running averages Ia as Id = (I − Ia)/Ia. In Figures 3b and 3d, the MSTID structures can be seen clearly as nearly vertical dark bands around the center of the images at both stations. Owing to the subtraction of 1-hour running averages, the MSTID structures became clear compared with those in the original all-sky images in Figures 3a and 3c.
Figure 4 shows NE-SW keograms of 630-nm airglow images obtained at Renner Springs during all 15 nights of the FRONT3 campaign. Because most of the MSTID structures propagate to the southwest, we took the keograms as time sequences at the NE-SW baseline from (39.9°N, 141.1°E) to (29.9°N, 131.1°E), as shown by the dashed line in Figure 3d. The length of the baseline is ∼1400 km. We took airglow intensity variations from 1-hour running averages. The Milky Way (Galaxy) can be identified as the striated structures running diagonally from the top left to the bottom right in the keograms in Figure 4. The “blocky” feature of the keogram is due to the relatively coarse time resolution (5.5 min) of airglow images.
 In Figure 4, wave structures of MSTIDs were observed all night as slanted stripes propagating from NE to SW in the keograms with periods of 0.5–2 hours. They propagate to the southwest (to the northwest in the original images in the Southern Hemisphere) with a velocity of ∼100 m/s (∼4 hours to propagate from NE to SW over 1400 km) and wavelengths of 100–400 km. One exception is at 1500–1800 UT on 26 May 2003, when the MSTIDs seem to move northeastward. It is noteworthy that the MSTID activity peaks in the premidnight sector and tends to decrease after midnight (e.g., 28 May, 1 and 3 June). It is also noted that there are vertical stripes without propagation in a longer timescale of 1–2 hours, often embedded on the slanted MSTID stripes (e.g., 1600 UT on 28 May, 1600 and 1800 UT on 29 May, and 1500 UT on 4 June). We will revisit these longer timescale variations in Figures 8 and 9.
 As shown by the values of ΣKp in Figure 4, geomagnetic activities were relatively high throughout the campaign period. Magnetic storms took place on 30 May (minimum Dst = −131 nT at 0200 UT on 30 May) and on 2 June (minimum Dst = −90 nT at 0900 UT on 2 June). The MSTIDs seem to occur irrespective of these geomagnetic activities.
3. Geomagnetic Conjugacy of MSTIDs
Figure 5 compares the MSTID structures seen in 630-nm airglow images at geomagnetic conjugate stations. The images are deviations from 1-hour running averages in geographical coordinates obtained at Shigaraki and Renner Springs (Figures 5a and 5b) and Sata and Darwin (Figure 5c). The images at Renner Springs and Darwin are converted to the Northern Hemisphere along the magnetic field line in the same area as those of Shigaraki and Sata, respectively. The white curves in the top parts represent phase structures of MSTIDs seen at Shigaraki and Sata. The curves are also drawn in the bottom parts (Renner Springs and Darwin) at the same geographic locations. Thus the white curves drawn at the top and bottom indicate what would be seen if there were one-to-one correspondence of the MSTIDs between the two hemispheres.
Figure 5a shows the MSTIDs on 28 May 2003 (every 35 min), observed simultaneously at Shigaraki and Renner Springs. The amplitude is very large, up to ∼40% from 1-hour running averages. The MSTIDs (wavelength ∼200 km) have phase surfaces in the NNW-SSE direction and propagate toward WSW at ∼80 m/s. The white curves drawn at the same locations at the top and bottom indicate one-to-one correspondence of the MSTID phase structures between Shigaraki and Renner Springs.
Figure 5b shows the MSTIDs on 2 June 2003 (every 30 min). The MSTID structures (wavelength ∼370 km) can be identified in the northwestern half of the images at both Shigaraki and Renner Springs, as NW-SE band structures moving southwestward at ∼120 m/s. Although the Galaxy overlaps with the MSTIDs at Renner Springs, fairly good correspondence can be seen in the MSTID structures at Shigaraki and Renner Springs, as guided by the white curves.
Figure 5c shows the MSTIDs on 4 June 2003 (every 30 min), observed simultaneously at Sata and Darwin. The top edge of the images at Sata and the left edge of the images at Darwin are masked by obstacles. Sata images were hazy. Passing airplanes reduced the image quality at Darwin. Nevertheless, one-to-one correspondence of MSTID structures can be recognized, as shown by the white curves.
Figure 6 shows a composite image of MSTIDs observed in the 630-nm airglow at 1543 UT on 28 May 2003 (from the right side of Figure 5a). The top half is the image at Shigaraki, and the bottom half is the image at Renner Springs, which was converted to the Northern Hemisphere along the geomagnetic field line. The MSTIDs are identified as two nearly north-south airglow depletion bands around the center of the images, with a horizontal wavelength of ∼200 km. The phase surfaces of the MSTIDs clearly coincide at the two stations with a comparable amplitude.
 To show the southwestward motion of the MSTIDs more clearly, we took cross sections (keograms) along the NE-SW baseline of the images in Figure 5. Figure 7 compares the MSTID structures in the NE-SW keogram of the 630-nm airglow images (in the same format as that in Figure 4) at geomagnetic conjugate stations. Like those in Figure 5, the white curves at the top represent MSTID structures seen at Shigaraki and Sata. The curves are drawn at the bottom at the same geographic locations.
Figure 7a shows the MSTIDs at 1200–1800 UT on 28 May 2003 at Shigaraki and Renner Springs. The MSTIDs can be identified as the slanted stripes propagating southwestward. The white curves were drawn only for some of the phase surfaces. The structures show clear one-to-one correspondence between the top and bottom. It is noteworthy that the MSTID activity ceases around 1700 UT (0200 LT) at both Shigaraki and Renner Springs simultaneously.
Figure 7b shows the MSTIDs at 1300–1600 UT on 2 June 2003. Again there are clear one-to-one correspondences between Shigaraki and Renner Springs. Figure 7c shows the MSTIDs at 1400–1700 UT on 4 June 2003 at Sata and Darwin. Despite the low image quality of these stations, fairly good conjugacy of the southwestward moving MSTIDs can be recognized between the two stations.
 The FRONT3 campaign (May–June 2003) was held in the summer season of the Northern Hemisphere. Thus the background ionospheric electron density may be different between the Northern and Southern Hemispheres. To compare the absolute values of airglow intensity between the two hemispheres, we plot the temporal variations of the 630-nm airglow intensities at Shigaraki (solid curves) and Renner Springs (dashed curves) at (34.9°N, 136.1°E; at the zenith of Shigaraki) in Figure 8. The images at Renner Springs are converted to the Northern Hemisphere along the geomagnetic field lines.
 The absolute intensities in Figure 8 are generally comparable between the two stations. The intensity variations associated with MSTIDs with a timescale of ∼0.5–1.0 hour are well correlated, with comparable amplitudes between the two stations, except for the MSTID of 1400 UT in Figure 8d. It is noteworthy that some variations at timescales longer than 1 hour also correlate between the two hemispheres, such as 1430–1700 UT in Figure 8b and 1500–1730 UT in Figure 8e. The latter is a LSTID, which will be shown in detail in the next section. The good correspondence for longer-scale variations can also be seen in the keograms in Figure 7, such as intensity enhancements at 1300, 1400, and 1600 UT in Figure 7a and at 1630 UT in Figure 7c. These “background” enhancements do not show southwestward propagation in the keograms. Similar nonpropagating variations in longer timescales can be identified in the keograms in Figure 4 (e.g., after 1600 UT on 29 May).
Figure 9 shows a comparison of the absolute intensity of 630-nm airglow between Sata and Darwin at (31.0°N, 130.7°E; around the zenith of Sata). The absolute intensity is generally higher at Sata, which may correspond to the higher ionospheric density in the summer hemisphere, particularly at lower latitudes. The intensity variations embedded on this offset are well correlated between the two stations with comparable amplitudes except for 1800–1900 UT on 4 June 2003. It is again noted that not only are the variations at a shorter timescale of ∼0.5–1.0 hour associated with MSTIDs, but those at longer timescales also show good correlation between the two stations.
4. Geomagnetic Conjugacy of a LSTID
 Among those longer timescale variations, we identified a LSTID event at 1400–1700 UT on 28 May 2003, as a southward propagating airglow structures observed at several stations. Though magnetic storm did not occur on this night, the Dst variation (provisional) was rather large (−40 nT at 1600 UT). Figure 10 shows the north-south keograms of 630-nm airglow intensities at Rikubetsu, Shigaraki, and Renner Springs (Figures 10a–10c), airglow intensities at three meridional points at the conjugate stations (Figure 10d–10f), thermospheric wind measured by a Fabry-Perot interferometer (FPI) at Shigaraki (Figure 10g), and foF2 and hpF2 obtained by four ionosondes in Japan (Figures 10h and 10i), during the LSTID event of 28 May. The airglow intensities at Renner Springs are those converted to the conjugate Northern Hemisphere in the same area as that of Shigaraki. The values of hpF2 were calculated on the basis of Shimazaki  and Berkey and Stonehocker .
 The LSTID was identified as airglow enhancements moving southward with a velocity of ∼600 m/s at Rikubetsu (1400–1530 UT), Shigaraki (1500–1630 UT), and Renner Springs (1530–1700 UT) in Figures 10a–10c. The LSTID was also identified by GPS-TEC map (not shown) over Japan as a weak TEC enhancement of ∼0.5 TECU propagating southward with a north-south scale size of ∼1000 km. In Figure 10g the southward wind velocity suddenly decreased at 1500 UT at Shigaraki. This southward wind decrease probably made ionosphere down to lower altitudes along the magnetic field line at midlatitudes (due to less support of the ionosphere) and caused the 630-nm airglow enhancement, as discussed by Shiokawa et al. [2002b]. Actually, the decrease of the ionospheric height was observed by the ionosondes as decrease of hpF2 in Figure 10i from northern station at 1400 UT at Wakkanai (45.4°N), then middle part of Japan at 1440 UT at Kokubunji (35.7°N) and Yamagawa (31.2°N), to the southern edge of Japan at 1500 UT at Okinawa (26.3°N). The values of foF2 tend to increase at the LSTID passage, except for Wakkanai.
 It is noteworthy that the airglow enhancements at Renner Springs are delayed ∼20 min from conjugate Shigaraki, as shown in Figures 10d–10f. This fact indicates that the electrodynamical coupling between the conjugate hemispheres is not the cause of the observed enhancements of 630-nm airglow emission. As shown in the FPI data in Figure 10g, northward neutral wind enhancement can be a direct cause of the F layer descent and subsequent airglow enhancement. The 20-min delay indicates that the atmospheric wave in the thermosphere, which was probably launched from the high-latitude auroral zone, propagates independently between the two hemispheres.
 As shown in Figure 4, MSTIDs were observed every night during the campaign period at Renner Springs, where the sky condition was good. This fact indicates that the MSTID is a common feature in this Northern Hemispheric summer season in the Japanese-Australian longitudinal sector. Most of them propagate southwestward/northwestward in the Northern/Southern Hemisphere. The MSTIDs become most active at premidnight local time and tend to cease after midnight. A similar local-time tendency was reported by Shiokawa et al. [2003a] in a statistical study of MSTID amplitudes observed by airglow imagers at Shigaraki and Rikubetsu.
 All the nighttime MSTIDs observed during the FRONT3 campaign show fairly good one-to-one correspondence between the Northern and Southern Hemispheres with comparable amplitudes. This fact indicates that the geomagnetic conjugacy of MSTIDs, which was first reported by Otsuka et al. , is a common feature in the midlatitude ionosphere. This conjugacy indicates strong electrodynamic coupling of MSTIDs between the two hemispheres through electric fields. The generation of MSTIDs by electric field variations has been suggested by many previous reports in the literature [Perkins, 1973; Behnke, 1979; Fukao et al., 1991; Kelley and Fukao, 1991; Miller et al., 1997; Kelley et al., 2002]. Shiokawa et al. [2003b] have shown spatially oscillating electric fields correlated with the MSTID phase structures on the basis of simultaneous ground-satellite observations of MSTIDs at Shigaraki. A spatially oscillating east-west electric field can cause upward/downward motion of the F layer through E × B drift at midlatitudes. The 630-nm airglow intensity increases when the F layer height decreases to interact with rich molecular oxygen at lower altitudes. The oscillating electric fields will be easily mapped to the other hemisphere to produce the observed symmetric structure of the MSTIDs because of the high conductivity along the geomagnetic field line.
 The mechanism of producing a wave-like electric field variation in the midlatitude ionosphere has not been clarified yet. Atmospheric gravity waves can be a seed of the wave structure of MSTIDs. Instabilities in the midlatitude F layer may develop to amplify the oscillating electric field [Perkins, 1973; Hamza, 1999]. The present result indicates that some electrodynamic coupling through magnetic field lines occurs between the summer and winter hemispheres in the MSTIDs. Such a coupling process should also be taken into account to model the production of an oscillating electric field.
 As shown in Figures 8 and 9, we also found that not only the MSTID structures but also airglow variations in longer timescales often show good correlations between the two hemispheres. It seems that the longer-scale variations are caused by neutral wind variations, such as gravity waves from lower altitudes, tidal variations, and atmospheric waves from the equatorial midnight temperature maximum (MTM) and from the auroral zone energy input (LSTIDs) [e.g., Rishbeth and Garriot, 1969; Mendillo et al., 1997; Otsuka et al., 2003; Shiokawa et al., 2003c]. The neutral wind data from FPI on 28 May in Figure 10 do show that the decrease of southward wind at 1500 UT caused the airglow enhancement. We also found similar decrease of southward wind (not shown) during the airglow enhancement at 1500–1700 UT on 4 June in Figure 8. The latitudinal motion of the equatorial anomaly, which is controlled by the zonal electric field at the magnetic equator, may also cause longer timescale variations, particularly at lower latitudes.
 The longer timescale neutral-wind variations may cause divergence of the electric field through ion-neutral collisions [Miller, 1997; Miller and Kelley, 1997]. The divergence would be mapped to the other hemisphere to produce the observed good correlation of airglow intensity variations between the two hemispheres. In such a case, the variations between the Northern and Southern Hemispheres should occur simultaneously without time differences.
 On the other hand, if the large-scale atmospheric waves from equatorial MTM or from the auroral zone (LSTID) are launched simultaneously in the Northern and Southern Hemispheres, they would produce geographically symmetric structures between the two hemispheres. For the case of the LSTID in Figure 10, similar equatorward moving airglow structures were observed at Shigaraki and Renner Springs. This suggests that the auroral energy input, which launched the observed LSTIDs toward midlatitudes, were similar in the northern and southern auroral zone. However, the observed time difference of ∼20 min between Shigaraki and Renner Springs indicates that the atmospheric waves independently propagates equatorward in the both hemispheres and the electrodynamic coupling between the hemispheres are not important to generate the observed airglow enhancements in LSTIDs.
 It should be noted that the ionospheric Perkins instability, which is a possible cause of MSTIDs, is an electromagnetic instability based on the coupling between ionospheric current and conductivity to develop electric field oscillation. This instability can be seeded by neutral wind oscillation, such as gravity waves. However, this neutral wind oscillation is just to give an initial perturbation of the electromagnetic instability and would not be necessary to have significant amplitude to produce the observed electric field oscillation through ion-neutral collision. On the other hand, the neutral wind variation that produces the LSTIDs is necessary to have sufficient amplitude to produce the ionospheric perturbation directly through ion-neutral collision.
 We have conducted the FRONT3 campaign in the period of May–June 2003 in order to investigate geomagnetic conjugacy of the midlatitude MSTIDs and LSTIDs. Seven airglow imagers were operated simultaneously in Japan and Australia. The observed results are summarized as follows.
 1. The MSTIDs were observed every night at Shigaraki and Renner Springs, indicating that they are a common feature in the midlatitude ionosphere in the Northern Hemispheric summer season in the Japanese-Australian longitudinal sector.
 2. For all the analyzed events, we found one-to-one correspondence of the MSTID structures between the Northern and Southern Hemispheres with comparable amplitudes. This fact indicates strong electrodynamic coupling between the two hemispheres along geomagnetic field lines.
 3. The MSTID activity peaks before midnight and tends to cease after midnight. This activity variation is also well correlated between the two hemispheres.
 4. Most of the observed MSTIDs propagate southwestward/northwestward in the Northern/Southern Hemisphere.
 5. The MSTID occurrences were relatively low at Rikubetsu and Darwin compared with those at Shigaraki and Renner Springs, suggesting a preference for MSTID generation at latitudes around Shigaraki and Renner Springs.
 6. For some cases, the airglow intensity variations with a timescale longer than that of MSTIDs also show good correlation between the two hemispheres.
 7. For the LSTID event of 28 May 2003, we observed equatorward moving airglow enhancements nearly simultaneously in Japan and Australia. However, the timing of the LSTID appearance was ∼20 min earlier in the Northern Hemisphere than in the conjugate Southern Hemisphere, indicating that the LSTID propagates equatorward independently in the both hemispheres as atmospheric waves in the thermosphere.
 The symmetric structures of MSTIDs between the Northern and Southern Hemispheres indicate that the electric field in the ionosphere plays an important role in the generation of MSTIDs, though the mechanism of producing a wave-like electric field pattern has not been clarified yet. During the FRONT3 campaign, not only the multipoint airglow imagers but also many other instruments were in operation, particularly in Japan, as some of them are shown in Figure 10. A Fabry-Perot interferometer and the MU radar at Shigaraki measured neutral winds and electron density profiles, respectively. The MU radar was also used to observe field-aligned irregularities in the F layer. Ionospheric measurements were carried out by using multipoint GPS receivers (more than 1000 points) and ionosondes (four stations with a time resolution of 3 min). A detailed study of MSTID electrodynamics using these multi-instruments will be published elsewhere.
 We thank Y. Katoh, M. Satoh, T. Katoh, K. Hanano, and K. Hidaka of the Solar-Terrestrial Environment Laboratory, Nagoya University, for their kind support of the development and operation of the all-sky imagers. The observation at Shigaraki was carried out in collaboration with the Radio Science Center for Space and Atmosphere, Kyoto University. The MU radar at Shigaraki belongs to and is operated by the Radio Science Center for Space and Atmosphere, Kyoto University. The all-sky imager at Darwin was operated at the ionosonde station of IPS Radio and Space Services, Australia. The imager at Nishiharima was operated in collaboration with the Nishiharima Astoronomical Observatory. This work was supported by a Grant-in-Aid for Scientific Research (11440145, 13573006, and Priority Area 764) and Dynamics of the Sun-Earth-Life Interactive System (No.G-4, the 21st Century COE Program) of the Ministry of Education, Culture, Sports, Science and Technology of Japan.
 Arthur Richmond thanks Santimay Basu and Jonathan Makela for their assistance in evaluating this paper.