Simulation of equatorial electrojet magnetic effects with the thermosphere-ionosphere-electrodynamics general circulation model

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 112,, doi: /2007ja012308, 2007 Simulation of equatorial electrojet magnetic effects with the thermosphere-ionosphere-electrodynamics general circulation model V. Doumbia, 1,2 A. Maute, 3 and A. D. Richmond 3 Received 29 January 2007; revised 9 April 2007; accepted 21 May 2007; published 28 September [1] In this work, the magnetic variations simulated by the NCAR thermosphereionosphere-electrodynamics general circulation model (TIE-GCM) in the vicinity of the magnetic equator are examined to evaluate the ability of this model to reproduce the major features of the equatorial electrojet (EEJ) as observed on the ground as well as on board low-altitude orbiting satellites. The TIE-GCM simulates electric currents of various origins and reproduces their associated magnetic perturbations. We analyze the diurnal and latitudinal variations of the EEJ magnetic effects calculated on the ground in West Africa under approximately the same solar activity condition as in 1993 for the March equinox and June and December solstices. The latitudinal and local time structures of these simulated results correspond well to those that are observed. We also compare longitudinal variations of the midday EEJ magnetic perturbations observed by the CHAMP satellite with the model predictions. Although the simulations and observations both show multiple maxima and minima in longitude, the locations of these extrema often disagree. In the model most of the longitudinal variation of the magnetic variations is associated with nondipolar structure of the geomagnetic field. We find that the modeled contributions of the thermospheric migrating diurnal and semidiurnal tides to the magnetic perturbations have large longitudinal variations, and we suggest that an increase in the amplitude of these tides in the TIE-GCM may cause them to play a major role in explaining the morphology of the EEJ longitudinal variation. Citation: Doumbia, V., A. Maute, and A. D. Richmond (2007), Simulation of equatorial electrojet magnetic effects with the thermosphere-ionosphere-electrodynamics general circulation model, J. Geophys. Res., 112,, doi: /2007ja Introduction [2] The equatorial electrojet (EEJ), a dayside current sheet flowing eastward along the geomagnetic dip equator at 105 km altitude, is associated with the regular dynamo process in the low-latitude ionosphere. This current system has been widely investigated through its magnetic signatures recorded along station chains across the dip equator [Forbush and Casaverde, 1961; Fambitakoye and Mayaud, 1976; Hesse, 1982], and on board low-altitude orbiting satellites [Cain and Sweeney, 1973; Onwumechili and Agu, 1980; Agu and Onwumechili, 1981]. Thus most of its spatial and temporal features were set forth on the basis of ground-based and satellite magnetic measurements. During the International Equatorial Electrojet Year (IEEY), a worldwide campaign of simultaneous measurements [Amory-Mazaudier et al., 1993; Arora et al., 1993; Rigoti et al., 1999] allowed to update those features [Doumouya et 1 Laboratoire de Physique de l Atmosphère, Université de Cocody, Abidjan, Ivory Coast. 2 Temporarily at High Altitude Observatory, National Center for Atmospheric Research, Boulder, Colorado, USA. 3 High Altitude Observatory, National Center for Atmospheric Research, Boulder, Colorado, USA. Copyright 2007 by the American Geophysical Union /07/2007JA al., 1998], and to undertake a global study of the EEJ using ground-based magnetic data [Doumouya et al., 2003]. The advent of Oersted, CHAMP and SAC-C satellites makes it possible to examine the longitudinal variation and other global characteristics of the EEJ through the magnetic records at various longitudes and local times [Jadhav et al., 2002; Lühr et al., 2003; Doumouya and Cohen, 2004]. Although those studies have yielded increased progress in understanding the EEJ phenomenon, most of its characteristics are still poorly interpreted. In effect, the day to day, seasonal, and solar cycle variability, the reversal known as counter-electrojet, and the unexpected behaviors in its longitudinal variation in certain longitude sectors, need more objective explanations. Furthermore, the question of how the circuit of the EEJ closes, or how it is coupled with large-scale current systems like the planetary Sq system or the meridional current evidenced by Maeda et al. [1982] in Magsat dusk data, are not yet resolved. [3] The main purpose of this work is to research the causes of some of the above EEJ features with the help of the National Center for Atmospheric Research (NCAR) thermosphere-ionosphere-electrodynamics general circulation model (TIE-GCM) [Richmond et al., 1992]. The TIE- GCM is designed to calculate the coupled dynamics, chemistry, energetic, and electrodynamics of the global thermosphere-ionosphere system between about 97 km and 500 km altitude. Its inputs include solar ultraviolet 1of16

2 radiation flux intensity, parameterized by the F10.7 index, and upward propagating atmospheric tides at the lower boundary. In particular, the TIE-GCM calculates ionospheric electric fields and currents and their associated magnetic perturbations at the ground and at low-earth-orbit altitudes, using a realistic geomagnetic main field, under various seasonal and solar activity conditions. Because of longitudinal variations of the main field, the local-time patterns of TIE-GCM ionospheric currents and geomagnetic perturbations show longitudinal variations, even when the upward propagating atmospheric tides input at the lower boundary have a fixed local time pattern, the so-called migrating tides. Longitudinal variations can also be produced by nonmigrating tides [e.g., Immel et al., 2006]. In this study we will focus on the migrating tides as first step. [4] The present work evaluates how well the TIE-GCM accounts for the EEJ observed features, using standard model inputs that include migrating tides. For that purpose, the diurnal, latitudinal and longitudinal patterns of the simulated EEJ magnetic signature are analyzed and compared with ground-based and satellite measurements of the EEJ magnetic effects. We examine the influences of the migrating tides and of the geomagnetic main field structure on the EEJ in order to determine their roles in the EEJ longitudinal variation and its seasonal variability. We do not consider here the influences of non-migrating tidal components, for which the daily cycle varies with longitude. Even though observations indicate that non-migrating tides may have substantial amplitudes in the lower thermosphere [e.g., Talaat and Lieberman, 1999; Oberheide and Gusev, 2002; Manson et al., 2004; Huang and Reber, 2004], these tidal components have not yet been carefully evaluated in the TIE-GCM. 2. TIE-GCM Ionospheric Current Configurations and Electrodynamic Parameters in Equatorial Latitudes [5] For a description of the different physical processes included in the TIE-GCM we refer to earlier publications and the references therein: Dickinson et al. [1981, 1984], Roble et al. [1982, 1988], and Richmond et al. [1992]. Although the model features are described in these references, the most relevant ones for this study are mentioned in the following. Especially the low-latitude electrodynamics and current configuration are described for a better understanding of the subsequent magnetic perturbations calculation Basic Assumptions [6] The TIE-GCM uses a realistic geomagnetic field model (International Geomagnetic Reference Field 2005), with magnetic apex coordinates according to Richmond [1995]. In the low-latitude region the geomagnetic field lines are assumed to be equipotential, which results in a symmetric electric potential about the geomagnetic equator. Within the polar cap (above/below ±75 latitude), the electric potential distribution is prescribed by the Heelis et al. [1982] model. At the lower boundary, approximately 97 km, tidal perturbations can be included. At the upper boundary, approximately km, the vertical O + flux is specified, which approximates the plasma exchange between the ionosphere and the plasmasphere. [7] In this paper, we examine the current and magnetic perturbations only at low latitudes. We focus on simulations for the March equinox and the December and June solstices, for moderate solar activity conditions (F 10.7 = 160 sfu, where 1 sfu =10 22 Wm 2 Hz 1 ), and for geomagnetic quiet time conditions. The integrated hemispheric power of auroral electrons is set to 15 GW and the cross polar cap potential is kept constant at 45 kv. For most of the model runs, we specify migrating diurnal and semidiurnal upward propagating tidal perturbations at the lower boundary as calculated by the Global Scale Wave Model (GSWM) from Hagan and Forbes [2002, 2003], but do not include the nonmigrating components. The TIE-GCM is run for several days to reach a diurnally reproducible state Current Systems in the Low-Latitude Regions [8] To determine the magnetic perturbation at the ground and above the ionosphere we need to know the current system. The TIE-GCM solves for the electric potential assuming equipotential geomagnetic field lines. Knowing the electric potential, we can calculate the electric current in the ionosphere and the geomagnetic field-aligned current at the top of the conducting ionosphere. The horizontal current density between 90 km and 600 km is integrated in height and represented as a thin current sheet at 110 km. Divergence of this horizontal current is assumed to connect to purely field-aligned current above the current sheet. [9] We treat the quasi-dipole current components [Richmond, 1995] as if they were true current components on a spherical Earth with a dipolar geomagnetic field, so that we can use spherical harmonic analysis as described by Richmond [1974] to calculate the equivalent current function and the associated magnetic perturbation at the ground. The equivalent current is a fictitious divergence-free horizontal sheet current that produces the same magnetic perturbation at the ground as the true three-dimensional current system. To calculate the equivalent current function we use spherical harmonics of degree and order up to M = 24 and N = 72, which ensures that the equatorial electrojet can be resolved. The total sheet current can be expressed by an equivalent current and a residual current. We define the residual current as the difference between the sheet current and the equivalent current. For the magnetic effects associated with currents induced within the Earth, we assume that at 600 km depth there lies a perfectly conducting layer, where the vertical component of the magnetic perturbation field vanishes. For calculating the magnetic perturbation above the ionospheric current sheet layer, we have to take into account, in addition to the equivalent current, the residual and the field-aligned current. The residual current couples with the field-aligned current to form a divergencefree current system that produces no magnetic perturbation on the ground. Shortly above the current sheet the magnetic perturbation due to the residual current is DB res where DB res = m o.b z xk res, with K res the residual current, b z the unit upward vector, and m o the permeability of free space. At higher altitudes the meridional component of the magnetic perturbation associated with the residual currents is assumed, for simplicity, to be constant in height, while the zonal component is adjusted in height to account for the 2of16

3 3.1. Local Time and Latitudinal Variations [11] Figures 1a, 1b and 1c show the daily variations of the D, H and Z components simulated at the magnetic stations of the West African meridian network [Doumouya et al., 1998; Vassal et al., 1998] in March, June and December, respectively, along with the corresponding seasonal averages of the observations during the IEEY in 1993 (see Table 1 for the coordinates of the 10 stations along the West African meridian chain across the magnetic dip equator). Figures 1a 1c show the consistency of the diurnal variations of TIE-GCM-simulated EEJ magnetic components with the observations from one station to another. Indeed, the diurnal variations of H and Z present the expected shape from the Southern Hemisphere to the Northern Hemisphere, including the amplitude variation with respect to the latitude of the stations. Note that the average value of F 10.7 for 1993 was about 110 sfu, smaller than that used in the simulations. [12] The simulated D variation is weaker than the observed one in December; and the diurnal patterns are also different. For March and June, the computed D amplitude is stronger than in December, and better capture the local time variation of the observations in the stations located to the north than in those to the south of the magnetic equator. The Figure 1a. Diurnal variation of the horizontal (magnetic east D and north H) and vertical Z components during the March equinox. The dotted red lines represent the seasonal averages of the magnetic quiet time data recorded in West Africa during the International Equatorial Electrojet Year (IEEY) (1993), and the solid blue lines represent the TIE- GCM simulation in approximately the same medium solar activity and seasonal conditions as in Only six stations of the network are represented. horizontal component of the field-aligned current flowing between the current sheet and the altitude in question. Once the magnetic perturbations have been calculated in the dipolar geometry, they are treated as though they are quasi-dipole components of the magnetic perturbation field in order to map them over the Earth. 3. Results [10] In this section, the diurnal variations of the geographic and magnetic northward components (respectively X and H), the eastward components (Y and D), and the vertical downward component Z, are analyzed as functions of magnetic local time and geographic latitude and longitude at different seasons. Figure 1b. Same as Figure 1a but for the June solstice. 3of16

4 Figure 1c. Same as Figure 1a but for the December solstice. latitudinal variation of the computed D component seen in Figure 2 is relatively small and is most of the time comparable with the observations except for stations situated to the southern side of the magnetic equator where they exhibit larger variations, especially near 8 N geographic latitude. The contour maps (Figures 3a and 3c) show that the simulated D component in the Northern Hemisphere is eastward (positive) in the morning and westward (negative) in the afternoon, while the opposite configuration is observed to the south during March and December; in June (Figure 3b) two different orientations are observed in the morning (weakly eastward to the north and westward to the south), while D is entirely westward in the evening. Table 1. Coordinates of the 10 Stations Along the West African Meridian Network Across the Magnetic Dip Equator During the International Equatorial Electrojet Year Geographic Location Stations Code Latitude, N Longitude, W Dip Latitude, deg Tombouctou TOM Mopti MOP San SAN Koutiala KOU Sikasso SIK Nielle NIE Korhogo KOR Katiola KAT Tiebissou TIE Lamto LAM of16

5 Figure 2. Latitudinal variation of the horizontal (magnetic east D and north H) and vertical Z components during equinox. The dotted red lines represent the seasonal averages of the magnetic quiet time data recorded in West Africa during the IEEY; the solid blue lines represent the TIE-GCM simulation for approximately the same solar activity and seasonal conditions. The small vertical lines indicate the position of the magnetic equator. [13] These structures of simulated D could be related to the planetary Sq current system rather than to the EEJ. Although these patterns are analogous to those observed in West Africa during the International Equatorial Electrojet Year (IEEY) by Doumouya et al. [1998], there are some differences with the observations. For the observation data, the structure of D seems to contain an additional component that closes on the edges of the EEJ, as confirmed in Figure 4 representing the D component recorded in central Africa in 1969 [Fambitakoye and Mayaud, 1976]. The 1969 experiment in central Africa was made along a meridian chain of 9 magnetic stations that expanded on a large latitude profile Figure 3. Contour map representation of the TIE-GCM simulation of the D component, (a) for the equinox, (b) for the June solstice, and (c) during the December solstice for medium solar activity conditions as observed during the IEEY. 5of16

6 Figure 4. Examples of a typical pattern of the D component diurnal variation in the EEJ influence area, observed in central Africa in 1969 on different days. The contour interval is 5 nt. of about 3000 km, an advantage which allows a comprehensive representation of the structure of the D component [Doumouya et al., 1998]. The contour map of January 31, 1969, exhibits four distinct peaks with opposite signs: a positive peak (20 nt) at 8 N dip latitude and a negative peak (± 20 nt) at 6 S for the morning hours; a negative peak ( 20 nt) at 8 N and a positive peak (25 nt) at 6 S in the early afternoon. For the other contour maps similar peaks are often, but not always present. Such a structure of the D variation suggests there may sometimes exist a current system that flows in a horizontal plane, with two vortices flowing respectively clockwise to the south, and anticlockwise to the north, focused at about 6 S and 8 N on either side of the dip equator. Such a current system flowing in the restricted equatorial area is not reproduced by our TIE-GCM simulations, and needs to be investigated further. [14] The TIE-GCM computed H and Z components present more significant amplitudes during the day time (Figure 1), arising from dawn to dusk, with the maximum around noon. The amplitude of the computed H (Figure 2) is stronger near the dip equator, while Z is maximum near the southern edge and minimum near the northern edge of the EEJ, around 3 from the magnetic equator. These shapes of H and Z are typical magnetic signatures of an eastward EEJ around noon. Nevertheless the model underestimates the amplitude of the observations and there is a shift between the times of the maxima. For the observations the maximum occurs earlier. [15] Beside the normal eastward EEJ effects, the computed H and Z components are reversed around 0700 LT in the morning (Figures 1a, 1b, and 1c), and this reversal is found at all seasons, with varying amplitude as 6of16

7 Figure 5. Diurnal variation of the H component of the TIE-GCM simulated EEJ magnetic effect in 30 increments along the dip equator between 175 and 155 geographic longitude. The solid blue line represents December solstice, the dashed green line represents June solstice, and the dotted red line represents March equinox. for the normal eastward EEJ, stronger in March (equinox) and weaker in December and June (solstices). In contrast to the model, the observations exhibit only weak reversals in March, with amplitudes and durations smaller than those of the model. The reversals of H and Z are interpreted, in terms of equivalent current, as a current system flowing westward along the dip equator. This equatorial westward current has been observed at other longitudes, and is referred to as counter-electrojet [Gouin and Mayaud, 1967] Longitudinal Variation Description of the Longitudinal Pattern of the TIE-GCM Simulated EEJ Magnetic Effect [16] We have computed the diurnal variation of the magnetic perturbations along the dip equator in different seasons. Figure 5 represents the simulated diurnal variation of the H component. The amplitude of the H component depends on season, being highest in March. This plot shows a net change in the amplitude and the shape from one 30 longitude sector to another. A 0:30 to 1:30 hour shift of the maximum is observed for June and December, depending on the longitude sector. The counter-electrojet (CEJ) effects occur in the morning, the amplitudes of which also change with longitude. The largest amplitudes of the CEJ are located between 90 W ( 90 E) and 65 E geographic longitude, covering South America, the Atlantic Ocean and Africa, with the strongest amplitudes in South America. In the afternoon there are no apparent reversals attesting the occurrence of afternoon CEJ. [17] The longitudinal profiles (Figure 6) of the H, D and Z noon values at the dip equator show that all the components vary with longitude. The H component exhibits two significant maxima respectively west of South America ( 85 geographic longitude) and in the eastern Asia-Pacific area (135 ). There is a broad minimum between 20 and 60, with a slight tertiary maximum in Africa (30 ) during March and June, and a secondary minimum in the mid- Pacific (±180 ). The D component shows a longitudinal variation that exhibits a strong amplitude (50 nt) at the dip equator in the Atlantic-South American longitude sectors, with a maximum around 80 and a minimum around 7of16

8 Figure 6. Noontime longitudinal variation of the TIE- GCM simulated EEJ magnetic effects: (top) H component, (center) D component, and (bottom) Z component, in equinox (dotted red line), June solstice (solid green line), and December solstice (dot-dash blue line). 30. The D component oscillates around 0 nt elsewhere, with weaker amplitude (less than 10 nt). [18] The existence of a nonzero D component at the magnetic equator has two causes. First, we find that the vortices of equivalent current in the Northern and Southern hemispheres can extend across the magnetic equator. Figure 7 shows the equivalent current function for two universal times near the times when the noontime D component in Figure 6 minimizes or maximizes: 1300 UT and 1700 UT, when local noon is at 15 longitude or 75 longitude, respectively. It is seen that around local noon the equivalent current flows from the Southern to the Northern Hemisphere at 1300 UT, producing a negative D component, while the noontime equivalent current flows from the northern to the Southern Hemisphere at 1700 UT, producing a positive D component. [19] A second factor that affects the D component at the magnetic equator is the nonperpendicularity of the magnetic equator to the direction of the main geomagnetic field. Since the EEJ current flows along the magnetic equator, the strongest horizontal magnetic perturbations are perpendicular to the magnetic equator. The D component is measured in the direction perpendicular to the main field, but this direction is not necessarily perpendicular to the strong EEJ magnetic perturbations, and so the D component senses a part of the EEJ magnetic perturbation. Figure 8 shows (dotted black line, labeled angle ) the angle y between the direction in which D is measured and the magnetic equator, where a positive angle means that the D direction is to the north of the equatorial direction, such that an eastward electrojet would contribute positively to D at the ground. The solid blue line shows D angle = Hsiny at local noon, which is approximately the contribution to D from the strong EEJ component of magnetic perturbation perpendicular to the magnetic equator. The total D component from Figure 5 has been replotted in Figure 8 in dashed red. D angle comprises a minor, but nonnegligible, part of the total D component. The dashed-dotted line in brown shows the difference between the total D perturbation and D angle, and is due to the flow of equivalent current across the magnetic equator, as discussed in the previous paragraph. It is interesting to note that this latter (dashed-dotted line in brown) component tends to oscillate with longitude in a similar manner as the (solid blue line) D angle component but with the peaks and valleys shifted in longitude. [20] The Z component varies with very weak amplitude at the dip equator (Figure 6), as expected. However, it is amplified in the longitude sector of South America, where a minimum of about 20 nt is noticed around 50 longitude. [21] In section 3.2.2, the TIE-GCM-simulated longitudinal variation of the H component will be compared with the CHAMP satellite borne EEJ magnetic signature Comparison of the TIE-GCM Computed EEJ Longitudinal Variation With CHAMP Satellite Observations [22] In Figure 9, the noontime magnitudes of the H component at 430 km altitude above the magnetic equator from the TIE-GCM are shown as a function of longitude (blue dotted lines) during the March equinox and the June and December solstices, along with the corresponding CHAMP satellite observed mean longitudinal variations (red solid lines) of the magnitude of the EEJ total magnetic perturbation (jdfj). The scattered dots represent the single values of jdfj at the dip equator for passes occurring between 11 and 1300 LT. The EEJ magnetic perturbation 8of16

9 Figure 7. Equivalent current function [ka] representing the Sq current system at (top) 1300 UT and (bottom) 1700 UT over geographic longitude and latitude. Equivalent current flows counterclockwise about the northern minimum and clockwise about the southern maximum. The geomagnetic equator is indicated by the dashed line. The model result is for March equinox with F 10.7 = 160 and migrating diurnal and semidiurnal tides prescribed at the lower boundary. DF has been isolated from the total satellite magnetic records by first extracting the main geomagnetic field total strength (F) using the IGRF 2000 model coefficients. The resulting residuals for each pass over the dip equator were fitted with a polynomial function, excluding the values close to the dip equator, in order to identify and remove a background signal that is superimposed on the EEJ contribution. This background signal is intended to include all the magnetic perturbations due to non-eej sources (crustal field, magnetospheric currents, Sq current, field aligned current, etc.). The remaining signal attributed to the EEJ depends on this processing technique. Doumouya and Cohen [2004] have discussed the CHAMP satellite magnetic data processing in more detail. Since the TIE-GCM values do not have a corresponding background value removed, we cannot directly compare the magnitudes of the simulations and observations, but will instead focus on the respective longitudinal variations. [23] In the equinox and the June solstice the satellite observations in Figure 9 exhibit four maxima, at about 170, 85, 10 and 100. In June the maximum around 10 is enhanced and shifted westward. Four minima at about 130, 45, 40 and 150 are also observed. Similar observations were made by Jadhav et al. [2002] during the March equinox and June solstice. As described in section 3.2.1, the TIE-GCM simulations show only two main peaks ( 85 and 135 ), a broad minimum ( 20 to 60 ) with a tertiary maximum (30 ), and a secondary minimum (±180 ). For the two results, only the maxima around 85 coincide. A shift of about 40 between the peaks in Africa and Asia is observed. In the December solstice, the satellite observation exhibits three maxima around 135, 45 and 60, two minima around 100 and 20, and an almost flat minimum between 100 and 140, whereas the TIE-GCM gives two peaks at 90 and 140, a minimum at ±180 and a flat minimum between 30 and 70. [24] The above results show poor agreement between the TIE-GCM simulation and the observed longitudinal variation of the EEJ at satellite height. The TIE-GCM longitu- 9of16

10 Figure 8. Contributions to the longitudinal variation of the D component at the magnetic equator. The dashed red line represents the TIE-GCM total D component; the black dotted line represents the angle between the direction of the D component and the magnetic equator; the solid blue line represents the D variation due to projection onto the magnetic east direction of the EEJ magnetic component perpendicular to the magnetic equator. The dashed-dotted brown line represents the effect of the equivalent current that extends across the magnetic equator. The green line represents the magnetic equator. dinal variation of the EEJ is dominated by the geomagnetic main field structure. Note that the satellite observed average longitudinal variation of the EEJ magnetic perturbation is subject to various error sources, such as the processing method and the day to day variability of the EEJ that is shown by the scatter of the dots in Figure 9. The standard deviations with respect to mean variations are respectively 6 nt for the equinox, 8 nt for the June solstice and 7 nt for the December solstice Effects of the Migrating Thermospheric Tidal Forcing on the EEJ Magnetic Field [25] The migrating diurnal and semidiurnal tidal effects on the EEJ are studied by running the TIE-GCM till we get a diurnally reproducible result for a given season. To isolate the tidal effect, we have computed the TIE-GCM EEJ magnetic effects along the dip equator, with and without the tidal forcing at the lower boundary of the model. Differencing the two results allows us to isolate the tidal forcing effect. Figure 10 shows the diurnal variation of the H component in different longitude sectors at the ground level for the March equinox. The dotted red lines represent the total H component including the tidal effect, the dashed black lines (H 0 ) show the H component without tidal forcing referred to as tidal-free component ; and the solid blue lines show the difference H t (H t =H H 0 ), representing the tidal contribution. Note that the red dotted lines in Figure 5 and Figure 10 are the same. It can be noticed that the tidal contribution is an important component of the EEJ magnetic signature. The amplitude of its contribution can exceed 35 nt on the ground around noon. Furthermore, according to its daily variation, the tidal contribution seems to play also the most dominant role in the CEJ occurrences. Indeed, the reversals are mainly carried by the tidal contribution (solid blue lines). Although the tidal contribution exhibits reversals in the afternoon as well as in the morning, the total magnetic perturbations do not show any apparent afternoon CEJ effect. This is confirmed in Figure 5 for the three seasons (March equinox, June and December Solstices) and different longitude sectors. The lack of afternoon reversals of the total magnetic perturbation is due to the fact that the amplitude of the afternoon reversals in the tidal component is much weaker than that in the morning, and is not enough to counterbalance the tidal-free component, and then to cause the reversal of the total perturbation. [26] Figure 11 shows the longitudinal profiles of the H component at local noon, in the three cases mentioned above for March equinox. The dotted red line, with tidal effect, has been described section 3.2.2; the dashed black line, without tidal effect, exhibits two maxima, one in South America and the other in eastern Asia, and a flat minimum over Africa. The tertiary maximum around Africa seen in the results that include tides is not observed when tides are removed. This feature seems to be exclusively related to the tidal effect (the solid blue line). Notice that the tidal effect 10 of 16

11 mostly amplifies the different undulations of the EEJ longitudinal variation. [27] We compare in Figure 12 the longitudinal variation of the noontime tidal effect on the H component with that of the total EEJ magnetic perturbation observed by the CHAMP satellite, for various seasons. Figure 12 shows the magnitudes of these (negative) perturbations, with different scales shown on the right and left. For the March equinox (top), the (positive) tidal effect on ground from the model is also shown (dashed red curve). The CHAMP observations (solid green line) show a similar number of peaks in longitude as the simulation of the tidal effects, but except for the maximum at 85 in all three curves, there are significant shifts between the simulated and observed minima and maxima, yielding sometimes to opposite phases. At 430 km altitude, the longitudinal profile of the tidal effects (dotted blue line) shows the same phases as on ground, except for the minimum at 140 at ground that is shifted westward to 175 at 430 km. For the June and December solstices, there is less agreement between the model and the satellite observations. In June the longitudinal changes of the tidal effect are very weak between 40 and 100, whereas the satellite observations exhibit very strong variations, more nearly out of phase than in phase with the model variations, especially between 0 and 180. In December the oppositions between the two curves are mostly between 180 and Effects of the Geomagnetic Main Field Structure in the EEJ Longitudinal Variation [28] The EEJ as an ionospheric current system is related to the ionospheric regular dynamo process and in turn depends on the local magnitude of the geomagnetic main field through the ionospheric conductivity as given by Ohm s law in the low-latitude ionosphere [Davis et al., 1967]. Both the Pedersen conductivity (s 1 ) and Hall conductivity (s 2 ) depend on the intensity of the main field (B) through the gyrofrequency [Baker and Martyn, 1953]. The longitudinal inequalities in the main field induce longitudinal changes in the EEJ intensity. In addition, longitudinal changes in the main field direction and curvature affect the EEJ. In this section, we verify the importance of the geomagnetic main field structure on the EEJ (Figure 13) by comparing the ground-level TIE-GCM result for March at noon at the dip equator (black dotted lines) with that in which the IGRF geomagnetic field (B) is replaced by a tilted dipole model field (red solid lines). The H component of the simulated EEJ magnetic effect related to the tilted dipole field is nearly constant with respect to longitude. Only the D component is changing with longitude in a sinusoidal function. These results are very different from those obtained with the IGRF model. For the dipole field model, no significant effects of the tides are noticed in the shape of the longitudinal variation. 4. Discussion [29] According to the morphological analysis of the EEJ magnetic signature simulated by the TIE-GCM, it can be 11 of 16 Figure 9. Comparison between the CHAMP satellite observed longitudinal variation of the amplitude of the (negative) EEJ magnetic effect (solid red line) and the TIE- GCM simulations (dotted blue line) at the dip equator during (top) March equinox, (center) June solstice, and (bottom) December solstice. The red dots represent the noontime amplitude of the EEJ magnetic signature at the dip equator for each single satellite pass, and the solid red line represents the average longitudinal variation.

12 Figure 10. Contribution of the thermospheric migrating diurnal and semidiurnal tides to the EEJ magnetic effects at the ground in different longitude sectors for the March equinox. The dotted red line represents the total magnetic effect including the tidal effect, the dashed black line represents the magnetic effect without tidal effect, and the difference (solid blue line) represents the exclusive tidal effect. 12 of 16

13 Figure 11. Longitudinal variation of TIE-GCM simulated EEJ magnetic effects at the dip equator during the March equinox at local noon on the ground, with tidal contribution (dotted red line), without tidal contribution (dashed black line), and the difference representing the exclusive tidal effect (solid blue line), scaled on the right side. concluded that the TIE-GCM reproduces many of the main features of the EEJ reasonably well, and also qualitatively accounts for some of its variations such as its local time, latitude, longitude, and seasonal dependence. We have shown that the H and Z components of the EEJ magnetic effect are approximately reproduced, as observed at the ground and low-earth-orbit altitudes. The TIE-GCM produces also a diurnal variation in the D component at the magnetic equator, induced mainly by meridional current flow across the dip equator. The success of the model in reproducing observed EEJ features qualifies it for addressing several questions relating to the total configuration of the EEJ, in particular how its circuit closes, how it connects to the planetary current system, and its relation to thermospheric winds. For example, this work has shown that the migrating diurnal and semidiurnal tides are the main cause of the counterelectrojet occurrence and its longitudinal dependence in the model. The fact that differences are identified between some details of the TIE-GCM predictions and the observations indicates that the model and its inputs can still be improved. For example, a different amplitude, phase, or distribution of the tides input to the TIE-GCM could affect the EEJ magnitude and time of maximum, and could alter the amount of current crossing the magnetic equator that affects the equatorial D variation. To further investigate such effects goes beyond the scope of this study. [30] The TIE-GCM reproduces less well the longitudinal variation of the EEJ as obtained through CHAMP satellite observations. We have shown that the nondipolar structure of the main field is the primary cause of the undulations in the EEJ longitudinal variation in the TIE-GCM. In particular, the migrating tides interact with this nondipolar structure in a manner that produces a tidal contribution to the longitudinal variations of the EEJ strength that shows multiple longitudinal maxima somewhat similar to those in the CHAMP observations. However, the modeled strength of these maxima is generally less than the observations indicate, and the longitude of their positions is sometimes different. This suggests the possibility that migrating thermospheric tides may play a more important role in the shape of the longitudinal profile of the EEJ strength than our TIE-GCM results predict, if these tides are underestimated in the TIE- GCM. Nonmigrating tides, though not included in our simulations, may also play a significant role. Indeed, Immel et al. [2006] have shown that nonmigrating tides may help explain the presence of four longitudinal maxima in the strength of the nighttime equatorial ionization anomaly observed by Sagawa et al. [2005]. Whether this longitudinal structure in the nighttime ionosphere is closely related to the longitudinal structure in the daytime EEJ remains to be determined. The disagreement between the TIE-GCM simulation of the EEJ longitudinal variation and the satellite observations may also be related in part to the way the EEJ signature was extracted, as well as to the nature of these data. Indeed the data used here come from CHAMP scalar measurements. Furthermore, the method used to extract the EEJ signature could have underestimated the EEJ contribu- 13 of 16

14 tion and even missed it at certain places. For instance, we have remarked that the longitudinal variations of the mean satellite data are smaller than the modeled ones, in contrast to ground-based measurements. Especially, this underestimation could happen in the areas where the dip equator is strongly tilted, as in South America-Atlantic longitude sector. Since the satellite orbit is polar, its passes in that region of strong dip equator slope askew the EEJ profiles. [31] On a quantitative basis, not only does the model underestimate the amplitude of the EEJ ground-based magnetic effects, a local time shift of the extrema of the diurnal variation of the H and Z components is observed. Most of the time, the H maximum occurs earlier in the observations than in the model results. Rastogi [2004] has also shown that the maximum of the H component occurs about one hour before noon. One possible cause of the disagreement between the observations and the TIE-GCM results can be the relative permanence and overestimation of the counterelectrojet effect by the model. [32] The strong amplitude of the D component in the longitude sector of South America is caused primarily by the structure of equivalent Sq current. A hemispheric invasion of the current vortices, characterized by a current flow across the dip equator, is observed in this sector where the geomagnetic field is weak and much distorted, with a strong declination. The Z component is also influenced by this effect. The hemispheric asymmetry of the equivalent Sq current vortices was also modeled by Takeda [2002] using ground-based magnetic observations. [33] Contributions of the tides are also noticed on the D and Z components at the dip equator, with strong amplitudes in the South American-Atlantic sector. For the D component, the tidal effect at noon is opposite to the global effect, westward to the west of the intersection of the geographic and magnetic equators, and eastward to the east. This is shown by an opposite meridional current flow with respect to the equivalent Sq current mentioned above. The tidal effect on the Z component keeps the same upward orientation at noon as for the global equivalent current, peaking around the intersection of the geographic and dip equators. [34] Notice that we did not compare the simulations of the D and Z component longitudinal variations with observations to confirm such a behavior of the current systems. Such a comparison requires vector measurements, and may be possible by using the CHAMP satellite and ground-based magnetic data. Our immediate future work consists in using satellite vector magnetic data in order to pursue this study. For that purpose, we will adopt a more objective approach to analyze CHAMP satellite data. A better adapted internal 14 of 16 Figure 12. Comparison between the amplitude of the (negative) H component of the TIE-GCM tidal effects and the corresponding CHAMP satellite-borne longitudinal profile of the EEJ scalar magnetic signature at local noon. (top) Tidal effect represented on the ground (dashed red line) and at 430 km (dotted blue line) with CHAMP observation (solid green line) during March equinox. Tidal effect represented at 430 km (dotted blue line) and CHAMP observation (solid green line) at the (center) June and (bottom) December solstices.

15 field model will be used to isolate the ionospheric contributions from the satellite magnetic measurements. The POMME model [Maus et al., 2005] could be used for this purpose. The last version of this model includes many interesting features, such as the attitude corrections, which will bring more accuracy to CHAMP vector measurements. [35] Acknowledgments. This study was partly supported by the NASA Sun-Earth Connection Theory Program. Vafi Doumbia s visit at the National Center for Atmospheric Research (NCAR) was financed by a Fulbright grant supervised by the Council for International Exchange of Scholars (CIES). Note that V. Doumbia s previous papers are authored as V. Doumouya. [36] Amitava Bhattacharjee thanks Subramanian Gurubaran and another reviewer for their assistance in evaluating this paper. Figure 13. Effect of the main geomagnetic field structure on the longitudinal pattern of the EEJ magnetic signatures at local noon for March. The solid red line represents the TIE- GCM simulation of the EEJ magnetic effects using a tilted dipole main field; the dotted black line represents the TIE- GCM simulation of the EEJ magnetic effects using the IGRF model. References Agu, C. E., and C. A. Onwumechili (1981), Comparison of the POGO satellite and ground measurements of the magnetic field of the equatorial electrojet, J. Atmos. Terr. Phys., 43, Amory-Mazaudier, C., et al. (1993), International equatorial electrojet year: The African sector, Rev. Bras. Geofis., 11, Arora, B. R., M. V. Mahashabde, and R. Kalra (1993), Indian IEEY geomagnetic observational program and some preliminary results, Braz. J. Geophys., 11(3), Cain, J. C., and R. E. Sweeney (1973), The POGO data, J. Atmos. Terr. Phys., 35, Baker, W. G., and D. F. Martyn (1953), Electric currents in the ionosphere. I. The conductivity, Philos. Trans. R. Soc. London, Ser. A, 245, Davis, T. N., K. Burrows, and J. P. Stolarik (1967), A latitude survey of the equatorial electrojet with rocket-borne magnetometers, J. Geophys. Res., 72, Dickinson, R. E., E. C. Ridley, and R. G. Roble (1981), A three-dimensional general circulation model of the thermosphere, J. Geophys. Res., 86, Dickinson, R. E., E. C. Ridley, and R. G. Roble (1984), Thermospheric general circulation with coupled dynamics and composition, J. Atmos. Sci., 41, Doumouya, V., and Y. Cohen (2004), Improving and testing the empirical equatorial electrojet model with CHAMP satellite data, Ann. Geophys., 22, Doumouya, V., J. Vassal, Y. Cohen, O. Fambitakoye, and M. Menvielle (1998), The equatorial electrojet at African longitudes: First results from magnetic measurement, Ann. Geophys., 16, Doumouya, V., Y. Cohen, B. R. Arora, and K. Yumoto (2003), Local time and longitude dependence of the equatorial electrojet magnetic effects, J. Atmos. Sol. Terr. Phys., 65, Fambitakoye, O., and P. N. Mayaud (1976), The equatorial electrojet and regular daily variation S R : - I. A determination of the equatorial electrojet parameters, J. Atmos. Terr. Phys., 38, Forbush, S. E., and M. Casaverde (1961), The equatorial electrojet in Peru, Publ. 620, Carnegie Inst. Washington, D. C. Gouin, P., and P. N. Mayaud (1967), A propos de l existence possible d un contre-électrojet aux longitudes magnétiques équatoriales, Ann. Geophys., 23, Hagan, M. E., and J. M. Forbes (2002), Migrating and nonmigrating diurnal tides in the middle and upper atmosphere excited by tropospheric latent heat release, J. Geophys. Res., 107(D24), 4754, doi: / 2001JD Hagan, M. E., and J. M. Forbes (2003), Migrating and nonmigrating semidiurnal tides in the upper atmosphere excited by tropospheric latent heat release, J. Geophys. Res., 108(A2), 1062, doi: /2002ja Heelis, R. A., J. K. Lowell, and R. W. Spiro (1982), A model of the highlatitude ionospheric convection pattern, J. Geophys. Res., 87, Hesse, D. (1982), An investigation of the equatorial electrojet by means of ground-based magnetic measurements in Brazil, Ann. Geophys., 38, Huang, F. T., and C. A. Reber (2004), Nonmigrating semidiurnal and diurnal tides at 95 km based on wind measurements from the High Resolution Doppler Imager on UARS, J. Geophys. Res., 109, D10110, doi: /2003jd Immel, T. J., E. Sagawa, S. L. England, S. B. Henderson, M. E. Hagan, S. B. Mende, H. U. Frey, C. M. Swenson, and L. J. Paxton (2006), Control of equatorial ionospheric morphology by atmospheric tides, Geophys. Res. Lett., 33, L15108, doi: /2006gl of 16

16 Jadhav, G., M. Rajaram, and R. Rajaram (2002), A detailed study of the equatorial electrojet phenomenon using Ørsted satellite observations, J. Geophys. Res., 107(A8), 1175, doi: /2001ja Lühr, H., M. Rother, S. Maus, W. Mai, and D. Cooke (2003), The diamagnetic effect of the equatorial Appleton anomaly: Its characteristics and impact on the geomagnetic field modeling, Geophys. Res. Lett., 30(17), 1906, doi: /2003gl Maeda, H., T. Iymori, T. Araki, and T. Kamie (1982), New evidence of meridional current system in the equatorial ionosphere, Geophys. Res. Lett., 9, Manson, A. H., C. Meek, M. Hagan, X. Zhang, and Y. Luo (2004), Global distributions of diurnal and semidiurnal tides: Observations from HRDI- UARS of the MLT region and comparisons with GSWM-02 (migrating, nonmigrating components), Ann. Geophys., 22, Maus, S., H. Lühr, G. Balasis, M. Rother, and M. Mandea (2005), Introducing POMME, Potsdam Magnetic Model of the Earth, in Earth Observation With CHAMP: Results From Three Years in Orbit, edited by C. Reigber et al., pp , Springer, Berlin. Oberheide, J., and O. A. Gusev (2002), Observation of migrating and nonmigrating diurnal tides in the equatorial lower thermosphere, Geophys. Res. Lett., 29(24), 2167, doi: /2002gl Onwumechili, C. A., and C. E. Agu (1980), Longitudinal variation of the equatorial electrojet parameters derived from POGO satellite observations, Planet. Space Sci., 29, Rastogi, R. G. (2004), Electromagnetic induction by the equatorial electrojet, Geophys. J. Int., 158, Richmond, A. D. (1974), The computation of magnetic effects of fieldaligned magnetospheric currents, J. Atmos. Terr. Phys., 36, Richmond, A. D. (1995), Ionospheric electrodynamics using magnetic apex coordinates, J. Geomagn. Geoelectr., 47, Richmond, A. D., E. C. Ridley, and R. G. Roble (1992), A thermosphere/ ionosphere general circulation model with coupled electrodynamics, Geophys. Res. Lett., 19, Rigoti, A., F. H. Chamalaun, N. B. Trivedi, and A. L. Padilha (1999), Characteristics of the equatorial electrojet determined from an array of magnetometers in N-NE Brazil, Earth Planets Space, 51, Roble, R. G., R. E. Dickinson, and E. C. Ridley (1982), Global circulation and temperature structure of the thermosphere with high-latitude plasma convection, J. Geophys. Res., 87, Roble, R. G., E. C. Ridley, A. D. Richmond, and R. E. Dickinson (1988), A coupled thermosphere/ionosphere general circulation model, Geophys. Res. Lett., 15, Sagawa, E., T. J. Immel, H. U. Frey, and S. B. Mende (2005), Longitudinal structure of the equatorial anomaly in the nighttime ionosphere observed by IMAGE/FUV, J. Geophys. Res., 110, A11302, doi: / 2004JA Takeda, M. (2002), Features of the global geomagnetic Sq field from 1980 to 1990, J. Geophys. Res., 107(A9), 1252, doi: /2001ja Talaat, E. R., and R. S. Lieberman (1999), Nonmigrating diurnal tides in mesospheric and lower-thermospheric winds and temperatures, J. Atmos. Sci., 56, Vassal, J., M. Menvielle, Y. Cohen, M. Dukhan, V. Doumouya, K. Boka, and O. Fambitakoye (1998), A study of transient variations in the Earth s electromagnetic field at equatorial 11 electrojet latitudes in western Africa (Mali and the Ivory Coast), Ann. Geophys., 16, V. Doumbia, Laboratoire de Physique de l Atmosphère, Université de Cocody, 22 BP 582 Abidjan 22, Côte d Ivoire. (doumouya@ucar.edu) A. Maute and A. D. Richmond, High Altitude Observatory, National Center for Atmospheric Research, Boulder, CO , USA. 16 of 16