Tomographic imaging provides a powerful technique for obtaining images of the spatial distribution of ionospheric electron density in an altitude versus geographic latitude grid at a specified longitudinal sector. The method, which involves monitoring the total electron content (TEC) of the ionosphere using signals transmitted from the global positioning system (GPS), has the ability to reveal the detailed structure of the ionosphere including the mid-latitude trough. The tomographic reconstruction of ionospheric electron density structures are performed at different specific longitudinal sectors both in the northern and southern hemisphere. At the same time the plasmapause position is estimated from IMAGE EUV images and mapped onto tomographic images. The simultaneous location of the altitude extension of the mid-latitude trough and the plasmapause position are essentially co-located.
 The ionospheric trough is one of the most prominent features of the subauroral ionosphere. It appears within a longitudinally elongated narrow strip equatorwards of the auroral zone. The mid-latitude trough, reviewed by Moffett and Quegan , has been observed over several decades using various techniques [Lemaire, 2001; Pryse et al., 1998, and references therein]. Aladjev et al. , by combining model simulation and experimental tomographic observations, demonstrated the possible factors that contribute to the formation of the mid-latitude trough. They conclude that the meridional and zonal component of external (magnetospheric) electric fields and precipitation of soft electrons at the poleward boundary of the trough are the cause of trough formation.
 In contrast, earlier work [see, e.g., Thomas and Dufour, 1965] believed that the mid-latitude trough represents the termination of the mid-latitude ionosphere and can be regarded as a boundary between the mid- and high-latitude regions produced by different mechanisms. Others believed that the mid-latitude trough has a direct link with the equatorial plasmapause [Moffett and Quegan, 1983] and that the two features lie on the same field line. However, less progress has been made in obtaining observational proof for this long standing conjecture of magnetosphere-ionosphere (MI) coupling- namely that the mid-latitude trough is the ionospheric signature of the plasmapause. These two features have some common properties that support the conjecture that they are co-located. First, since the ionosphere is the major source of plasma within the plasmasphere, the trough structure is expected to be reflected in the plasmasphere, suggesting that the ionospheric mid-latitude trough and plasmapause should be on the same field lines. Secondly, since the cross-tail magnetospheric electric field plays a major part in the formation of both features [Moffett and Quegan, 1983], they are expected to be directly related. However, except for a few case studies of single satellite passes performed in the last decades (that usually used indirect methods to infer the plasmapause location) [Grebowsky et al., 1976; Sivtseva and Ershova, 1978; Rodger and Pinnock, 1982; Smith et al., 1987], there have not been definitive comparison studies of the two features. Yizengaw et al. , using the ground-based GPS TEC data and IMAGE EUV images of the plasmasphere, have compared the locations of the plasmapause and mid-latitude trough on a global scale. These observations, which are the first study to use global simultaneous observations of the two features, clearly demonstrated that GPS TEC maps can be used to accurately estimate the location of the plasmapause on a routine basis. However, there are still gaps of understanding in the two features' relation. For example, does the signature of the plasmapause on the ionosphere have only a direct link with the F region density depletion or with the topside ionospheric density depletion as well? Does the altitude extension of the mid-latitude trough follow the field line where the plasmapause is expected to be?
 The light ion trough is observed in the topside ionosphere as a pronounced mid-latitude depletion of H+ and He+ within a 5–10° latitude range. The upward flow of H+ ions along the magnetic field lines is believed to be the main cause of the light ion trough in the topside ionosphere where lighter ions (H+) are more abundant than heavy ions (O+) [Moffett and Quegan, 1983]. Due to charge exchange with the major ion of the ionosphere (O+), the light ion trough (depletions of H+) may occur in the absence of a significant variation of the major ion O+ [Lemaire, 2001].
 This paper uses the combination of IMAGE EUV and GPS datasets to unequivocally settle this long standing conjecture of MI coupling. The tomographic reconstruction technique is used to identify the altitude extension of the mid-latitude trough. At the same time the plasmapause location (Lpp) is estimated [Goldstein et al., 2004] from IMAGE EUV images and mapped over the two-dimensional image of the ionospheric electron density. This allows, for the first time, to track the mid-latitude trough and plasmapause as a function of altitude and latitude.
 Signals from the satellites in the Global Positioning System (GPS) are used for tomographic observations. The satellites, which orbit at a height of about 20200 km (4.2 Re), transmit coherent signals. Using a standard extraction technique [Blewitt, 1990], measurements of the TEC along a large number of intersection satellite-to-receiver ray paths are determined. During the preprocessing of the GPS data, outliers and cycle slips are detected and removed or corrected, respectively. The GPS derived TEC is also calibrated from the receiver and satellite biases [Blewitt, 1990].
 Since Austen et al.  first introduced ionospheric tomography to the scientific community, it has become one of the reliable techniques used to image the ionosphere [see, e.g., Andreeva et al., 1990; Pryse et al., 1998] and plasmasphere [Heise et al., 2002] remotely using ground- and space-based instruments, respectively. Ionospheric tomography, which uses TEC measurements made at a linear chain of stations that are aligned approximately along a common geographic meridian, is particularly adept at the imaging of horizontal structures in ionization density. Thus, one of the main features that can be replicated by tomographic approach is the mid-latitude trough.
 The absolute TEC observations are then reconstructed in an inversion algorithm to create an image of the spatial distribution of the electron density as a function of latitude and altitude. One of the most commonly used inversion techniques is called the algebraic reconstruction technique (ART). The ART algorithm, which can converge quickly in an iterative fashion compared to other reconstruction algorithms [Austen et al., 1986], is the preferable algorithm to use for ionospheric reconstruction when a limited widely spaced number of receivers are available [Austen et al., 1986; Yizengaw et al., 2004, and references therein].
 Depicted in the top panel of Figure 1 are the geographic locations of the four stations over the east cost of Australia. They span a range of ∼35° (from 19.63°S to 54.50°S) in latitude and all are within 6° of 153°E longitude. Since the ionosphere varies much more rapidly with latitude than longitude, the longitudinal TEC gradients are relatively small and can be ignored [Jee et al., 2004]. This assumption is valid especially for our case where the longitudinal separations are <6° or <620 km and the observations are made within a 25 minute time window. Thus the local time differences of all TEC measurements recorded at these four stations are adjusted to provide information along the 153°E meridian. Similarly, the GPS stations used for tomographic observation over Europe spans a latitude range of ∼38° (from 40.65°N to 78.93°N) and all stations are within <3° of 15°E longitude. A similar local time difference adjustment of the TEC recorded over Europe is performed to obtain information along the 15°E meridian.
 Since 2000 the extreme ultraviolet (EUV) imager on the IMAGE satellite routinely provides global snapshots of the plasmasphere by detecting 30.4-nm sunlight resonantly scattered by the He+ ion population with spatial and temporal resolution of about 0.1 RE and ∼10 min [Goldstein et al., 2004]. The plasmapause position is estimated from IMAGE EUV images using the technique described by Goldstein et al. . The locations of the plasmapause are manually selected by clicking on an EUV image with a computer mouse along the plasmapause with an average azimuthal spacing of about an hour of magnetic local time (MLT) [Goldstein et al., 2004]. The bottom panel of Figure 1 shows an example of IMAGE EUV snapshot with computer mouse click (white boxes) on it, representing estimated positions of the plasmapause.
 In the present study four months of data (March through July 2001), representing disturbed and quiet periods, are used to study the altitude correlation of the ionospheric mid-latitude trough and the plasmapause projection onto the ionosphere. Two regions of interests, one in the north (at 15°E longitude) and the other one in the southern hemisphere (at 153°E longitude), are selected to perform this correlation study. We have performed a series of correlation studies during both magnetically disturbed and quiet times. Two typical examples, representing disturbed and quiet period observations, are presented. Figure 2 shows examples of tomographically reconstructed images of the ionosphere during the 31 March 2001 severe storm, on an altitude-versus-geographic-latitude grid (along with geomagnetic latitudes at the top) at 15°E longitude in the northern hemisphere. The corresponding time interval is given at the top of each panel. The bottom panel of Figure 2 contains the tomographic image that shows the bottom side and F layer ionospheric density structure, showing solar-produced ionization at the lower latitudes with decreasing densities to the north ending in a clear trough and smaller-scale structuring at the northernmost latitudes. The corresponding time topside ionospheric electron density is also shown in the top panel of Figure 2. Note that the maximum densities as represented with the color bars of the two panels are different. Similarly, the solar-produced density is increasing toward lower latitudes with a density bite out in the mid-latitude region. Note that the density bite out in the top-side ionosphere appeared to be at lower latitudes compared to its position at the F layer altitude.
 The altitude extension of the plasmapause position, extracted from IMAGE EUV images, at 15°E longitude is tracked from the bottom altitude to the ceiling height of the image grid using the IGRF model. Then the plasmapause position, which is tilted equatorward (white dots), is mapped onto the tomographically imaged density structure as shown in Figure 2. As can be seen in Figure 2, the plasmapause is located equatorward of the trough minimum at the topside and F-layer ionosphere. The two features have about ∼1.4° offset. The plasmapause position as determined from IMAGE EUV has an uncertainty of ∼0.1L (∼0.6° latitude) on the nightside, where the plasmapause gradient is steep. On the dayside the error can be larger ≤0.4L (<2.1° latitude) [Goldstein et al., 2004].
Figure 3 shows an example of tomographic images of the ionosphere performed at 153°E longitude during a magnetically quiet period, on 8 July 2001. The reconstruction shown in the bottom panel of Figure 3, which shows the bottom and F layer structure of the ionosphere, is performed using ground-based GPS data recorded between 07:00 and 07:25 UT on the same day. As in the previous example, the solar-produced ionization is dominant in the lower latitude region with decreasing density towards south terminating in a clear trough structure centered at about 57°S geographic (∼69°S geomagnetic) and smaller-scale structuring at the southernmost latitudes. The density structure of the topside ionosphere is also presented in the top panel of Figure 3. Like the disturbed day correlation observation, the altitude extension of the plasmapause position (the white dots) is projected onto the tomographic images in Figure 3. The mid-latitude trough appeared at higher latitudes during this quiet day observation (Figure 3) compared to its position during the storm period (Figure 2). Unlike Figure 2, the plasmapause is at the trough minimum in the topside ionosphere as shown in the top panel of Figure 3. However, the plasmapause is equatorward of the F-layer trough minimum with an offset of ∼1.8° latitude (bottom panel of Figure 3).
 Direct comparison between F-region mid-latitude trough and light-ion (topside ionosphere) trough behavior is notably scarce in the literature. The F-region and topside ionospheric trough comparison with the plasmapause is also another unexamined physical phenomenon. Although earlier studies found that the relationship of the topside light-ion and F-region peak density may be obscured by the electron temperature effect [Moffett and Quegan, 1983], the absence of the topside light-ion trough usually corresponds to the absence of F-region mid-latitude trough on the dayside. However, on the nightside, the H+ depletion occurs in the vicinity of a much shallower depression in the electron concentration, which is presumably the diffusive equilibrium extension of the F-region mid-latitude trough into topside ionosphere.
 The altitude versus latitude tomographic imaging of the ionospheric density structure at specific longitude sectors combined with the IMAGE EUV derived plasmapause position as a function of latitude and altitude gives us a new opportunity to track the two features directly and simultaneously and prove if they are on the same field lines. Although our correlation study campaign was during a limited number of months, it provides, for the first time, observational proof of the long standing conjecture of MI coupling, namely the mid-latitude trough and plasmapause position are on the same field line. Using four months of data, we have performed about 40 simultaneous observations. The number of simultaneous observations is limited due to the limited data available from IMAGE EUV. Since the mid-latitude trough is more abundant at night than daytime, most of our observations are performed during nighttime at different local times depending on IMAGE EUV data availability. On 90% of the observations, the plasmapause positions agreed very well (with the uncertainty of the IMAGE EUV plasmapause estimates) with both the F layer and light ion trough and are equatorward of the trough minimum. In the remaining observations, mainly during quiet times, the plasmapause is at about the light-ion trough minimum and at the equatorward edge of the F-layer trough minimum. In typical storm time tomographic images, performed in the north at 15°E (Figure 2), the plasmapause is equatorward of the trough minimum both at the topside (top panel) and F-layer (bottom panel) ionosphere. Note the minimum L value was calculated from IMAGE EUV snapshots to estimate the positions of the plasmapause. However, in Figure 3, which shows the quiet day observations, in the southern hemisphere at 153°E, the plasmapause is equatorward of the F-layer trough minimum (bottom panel) and exactly at the light-ion trough minimum (top panel).
 Interestingly, the altitude extension of both mid-latitude trough and plasmapause positions are tilted to lower latitudes with increasing altitude. Since the H+ is the dominant ion at the topside ionosphere, light-ion depletion occurs due to the upward flow of H+ to the plasmasphere, which becomes stronger in the mid-latitude region where the magnetic inclination is about 45°. The equatorward inclination of the magnetic field may cause the light-ion depletion to be shifted to lower latitude at higher altitude (see top panels of Figures 2 and 3) compared to the F-region mid-latitude position (see bottom panels of Figures 2 and 3). At the same time, this confirms the long standing hypothesis, i.e., the plasmapause and mid-latitude trough are on the same field line. Both features tilt equatorward following the magnetic field lines.
 In conclusion, the altitude extension comparison of the two features observationally confirms that the plasmapause position has a close link with the mid-latitude trough at both the F layer and topside ionosphere. The equatorward tilting behavior of the two features confirms that they are closely aligned on the same field line. These excellent correlation results reflect the power of combining the tomographic conversion technique, which detects the altitude extension of the mid-latitude trough, with IMAGE EUV observations, which provides global images of the plasmasphere.
 This work has been sponsored by a NASA Guest Investigator grant (NNGO4GG343G), NSF grant (ATM-0348398), and a JPL/UCLA partnership grant. The authors are grateful to the IGS for providing GPS data and B. R. Sandel for IMAGE EUV data. Thanks are also due to David Galvan for assistance with the estimation of plasmapause positions from IMAGE EUV images.