JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 116, A01316, doi: /2010ja015925, 2011

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 116,, doi: /2010ja015925, 2011 Vertical connection from the tropospheric activities to the ionospheric longitudinal structure simulated by a new Earth s whole atmosphere ionosphere coupled model H. Jin, 1 Y. Miyoshi, 2 H. Fujiwara, 3 H. Shinagawa, 1 K. Terada, 3 N. Terada, 3 M. Ishii, 1 Y. Otsuka, 4 and A. Saito 5 Received 12 July 2010; revised 5 October 2010; accepted 23 November 2010; published 27 January [1] This paper introduces a new Earth s atmosphere ionosphere coupled model that treats seamlessly the neutral atmospheric region from the troposphere to the thermosphere as well as the thermosphere ionosphere interaction including the electrodynamics selfconsistently. The model is especially useful for the study of vertical connection between the meteorological phenomena and the upper atmospheric behaviors. As an initial simulation using the coupled model, we have carried out a 30 day consecutive run in September. The result reveals that the longitudinal structure of the F region ionosphere varies on a day to day basis in a highly complex way and that a four peak structure of the daytime equatorial ionization anomaly (EIA) similar to the recent observations appears as an averaged feature. The simulation reproduces and thus confirms the vertical coupling processes proposed so far with respect to the formation of the averaged EIA longitudinal structure; the excitation of solar nonmigrating tides in the troposphere, their propagation through the middle atmosphere, and the modulation of ionospheric dynamo, which in turn affects EIA generation. The simulation result indicates that not only the ionospheric averaged longitudinal structure but also the day to day variation can be modulated significantly by the lower atmospheric effect. Citation: Jin, H., Y. Miyoshi, H. Fujiwara, H. Shinagawa, K. Terada, N. Terada, M. Ishii, Y. Otsuka, and A. Saito (2011), Vertical connection from the tropospheric activities to the ionospheric longitudinal structure simulated by a new Earth s whole atmosphere ionosphere coupled model, J. Geophys. Res., 116,, doi: /2010ja Introduction [2] Recent studies have suggested that the upper atmosphere is significantly affected by the lower atmosphere. A prominent four peak longitudinal structure of the equatorial ionization anomaly (EIA) has been discovered by global satellite observations [e.g., Sagawa et al., 2005; England et al., 2006a; Immel et al., 2006; Lin et al., 2007]. In addition, according to modeling works, the longitudinal dependence of tropospheric convection has been shown to be an important source for the excitation of solar nonmigrating (non Sun synchronous) tides that propagate upward to the thermosphere, including the diurnal eastward propagating 1 National Institute of Information and Communications Technology, Koganei, Japan. 2 Department of Earth and Planetary Sciences, Kyushu University, Fukuoka, Japan. 3 Department of Geophysics, Graduate School of Science, Tohoku University, Sendai, Japan. 4 Solar Terrestrial Environment Laboratory, Nagoya University, Nagoya, Japan. 5 Graduate School of Science, Kyoto University, Kyoto, Japan. Copyright 2011 by the American Geophysical Union /11/2010JA tide with zonal wave number 3 (DE3), which has four peaks in the local time fixed frame [e.g., Hagan and Forbes, 2002, 2003; Miyoshi, 2006]. Therefore, the four peak structure of the EIA can be considered as a result of atmospheric coupling from the bottomside (troposphere) to the topside (ionosphere) through the nonmigrating tides, which can modulate plasma fountain via E region dynamo. This idea was supported by a number of subsequent studies. The nonmigrating tides, with their climatology, have been observed directly in the lower thermosphere [e.g., Oberheide et al., 2006; Forbes et al., 2008], which confirmed the dominant existence of the DE3 tide during many months. As evidence of modulation in the dynamo, a four peak structure has been observed in E B drift [Hartman and Heelis, 2007; Kil et al., 2007, 2008; Ren et al., 2009] and in the equatorial electrojet (EEJ) [England et al., 2006b; Luhr et al., 2008] and has been shown in electrodynamic simulations [Jin et al., 2008; Ren et al., 2010]. Furthermore, the four peak structure of the EIA has been reproduced by a simulation using the Thermosphere Ionosphere Mesosphere Electrodynamics General Circulation Model (TIME GCM), in which the nonmigrating tidal forcing derived from the Global Scale Wave Model was incorporated at its lower boundary [Hagan et al., 2007]. 1of9

2 Figure 1. Method of self consistent atmosphere ionosphere coupling. Output variables from the whole atmospheric General Circulation Model (GCM; neutral densities N n, velocity V n, and temperature T n ), the ionospheric model (ion densities N n, ion temperature T i, electron temperature T e, and electric conductivities s), and the electrodynamics model (electric fields E) are exchanged mutually via the coupler module and used as the input to another model component. [3] For understanding the atmospheric vertical coupling, global first principle models treating both atmospheric and ionospheric regions, like TIME GCM [Roble, 1996], are quite useful as demonstrated by Hagan et al. [2007]. Although the tropospheric effects on the upper atmosphere have been shown to be significant in the aforementioned studies, there are few examples for the model including the entire atmospheric regions from the troposphere to the ionosphere. One example is TIME GCM/CCM3, which couples TIME GCM and a lower atmospheric model of the Climate Community Model version 3 (CCM3) [Roble, 2000], and other models are being developed such as the Integrated Dynamics through Earth s Atmosphere model, which consists of a Whole Atmosphere Model and a Global Ionosphere Plasmasphere model [Fuller Rowell et al., 2008]. [4] In this paper, we introduce another coupled model, which is called the Ground to Topside Model of Atmosphere and Ionosphere for Aeronomy (GAIA). One of the major objectives in developing the coupled model is to study the effects of meteorological phenomena on upper atmospheric behaviors at various temporal and spatial scales, including the formation of ionospheric longitudinal structure as already introduced. As an initial result of GAIA, we show a vertical connection between the tropospheric moist convection and the ionospheric longitudinal structure. Reproducing the vertical connection by an integrated model is important for confirming the coupling mechanisms proposed so far, since most of them were based on the separate climatological studies for separate regions. We also show, as a new feature, the day to day variability in the longitudinal structure of ionospheric parameters due to the lower atmospheric activity. 2. Model Description [5] An Earth s whole atmospheric model from the troposphere to the ionosphere, called GAIA, has been developed recently, by coupling three models, that is, a whole atmospheric GCM, an ionospheric model, and an electrodynamics model, each of which has been developed independently so far. [6] The whole atmospheric GCM covers the neutral atmospheric region seamlessly from the ground to the exobase and treats a full set of physical processes appropriate for the troposphere, stratosphere, mesosphere, and thermosphere including schemes for hydrology, a boundary layer, a radiative process, eddy diffusion, and moist convection under the assumption of hydrostatic equilibrium [Miyoshi and Fujiwara, 2003, 2006, 2008; Fujiwara and Miyoshi, 2006, 2009, 2010]. The GCM is a global spectral model, and in the present coupled model we used a version of T21 (maximum horizontal wave number of 21), corresponding to a grid spacing of 5.6 longitude by 5.6 latitude. The vertical resolution is 0.4 scale height above the tropopause. The time resolution is 60 s (under the condition of moderate solar flux and quiet geomagnetic activity). [7] The ionospheric model is the ionospheric part of a global thermosphere ionosphere model (three dimensional upgrade of Shinagawa and Oyama [2006], Shinagawa et al. [2007], and Shinagawa [2011]) developed as part of an operational real time space weather simulation system ( html), which solves equations of mass, momentum, and energy for several major ion species and electrons, including production and loss due to photochemical reactions as well as ion diffusion and convection processes. The horizontal grid spacing is 5 longitude by 1 latitude, and the vertical spacing is 10 km below 600 km height, with larger spacing at higher altitudes until the upper boundary at about 3000 km height. The time resolution is 0.5 s. [8] The electrodynamics model treats the closure of global ionospheric currents induced both by the neutral wind dynamo and by the setup of the polarization electric field [Jin et al., 2008]. In solving the distribution of the electrostatic potential, the model assumed that geomagnetic field lines are equipotential and, also, that the current does not connect to the magnetospheric current or to the current below the ionosphere. The latter corresponds to the boundary conditions that the current vanishes at the low altitude boundary (70 km) and that the field aligned current is preserved between the north and the south high altitude boundaries (720 km) along the same field line. The present model assumes a tilted dipole for the geomagnetic field configuration. In order to incorporate the effect of magnetospheric convection in the polar region, the electric field distribution from an empirical model [Volland, 1975] is simply added to the solution of the electrodynamics model. The horizontal grid spacing of the electrodynamics model is 3.7 longitude by latitude at a 70 km height level. [9] The method of model coupling is described as follows. In the coding aspect the coupling among the three model components was achieved by introducing a main module that we call the coupler and by arranging each model component in the form of a FORTRAN subroutine or module so as to be called in the coupler program. The coupler handles general procedures for running the integrated model, such as setting common parameters and conditions, controlling the time increment, and exchanging physical variables among the model components as summarized in Figure 1. The exchanged variables are neutral densities of major thermospheric species (O, O 2, and N 2 ), ion densities (O +,O 2 +,N 2 +, and NO + ), neutral velocity, temperatures (neutral, ion, and electron), conductivities (Pedersen and Hall components), and 2of9

3 while the drift turns downward and the EIA decays during nighttime. [11] First, we show the simulation results of the F region ionospheric distribution for the entire period of the present simulation. Figures 3a 3c exhibit the 30 day variations of the longitudinal structure for nmf2 at both the north and the south crests of the EIA and for upward E B drift, each of which is organized in the local time fixed frame after removing short temporal variations (period less than 2 h) with a low pass filter. The local time fixed here is 11 LT for E B drift and 15 LT for nmf2, when each variable peaks in its average local time variation. The result indicates quite complex dayto day variation of the longitudinal structures. The number of longitudinal peaks as well as the peak locations and widths in longitude change from day to day. For example, the nmf2 at the south EIA has four peaks during days 2 to 7 but three Figure 2. Diurnal variations of (a) nmf2 at all geographic latitudes and (b) upward E B velocity at the magnetic equator and a 300 km height for an arbitrarily selected longitude (55 E) and day (26 September) from the Ground to Topside Model of Atmosphere and Ionosphere for Aeronomy (GAIA) simulation. electric field. The interval of the variable exchange is 180 s in the present simulation, although each model component has a finer time resolution as already described. The three models have different grid distributions, thus coordinate conversion is necessary in exchanging the variables, which is also carried out in the coupler module. Unification of spatial resolution among model components is desirable and is intended for use in the future versions. 3. Results [10] Using the present version of GAIA, described in section 2, we have carried out a 30 day consecutive run in September, which is one of the months when the ionospheric four peak longitudinal structure becomes dominant accordingtotheobservations[e.g.,kiletal., 2008; Liu and Watanabe, 2008; Scherliess et al., 2008; Wan et al., 2008]. As the initial condition, we used results from noncoupled simulations using each atmospheric GCM and ionospheric model. We assumed constant solar UV and EUV fluxes (the value of F 10.7 is set at 135.0) and a quiet geomagnetic activity (magnetospheric inputs such as the polar cap potential were set constant) during the run so that the day to day variation in the upper atmosphere is modulated mainly owing to the lower atmospheric variability. Figure 2 shows diurnal variations of F region peak electron density and equatorial vertical E B velocity for an arbitrarily selected longitude and day to indicate typical ionospheric behaviors at the low to middle latitude; the upward E B drift increases after sunset, resulting in growth of the EIA during the daytime, Figure 3. Day to day variation of F region ionospheric longitudinal structure obtained by GAIA. Plotted variables are (a) nmf2 at the north crest of EIA, (b) nmf2 at the south crest of EIA, and (c) upward E B velocity at the magnetic equator and a 300 km height during 30 days in September. The longitudinal structure plotted for each day is local time fixed at 1100 LT for E B velocity and at 1500 LT for nmf2. 3of9

4 Figure 4. Longitude latitude distribution of 30 day averaged variables in September. (a) nmf2, (b) upward E B velocity at 300 km height, and (c) neutral temperature at a 110 km height. Local time fixed frame is 1500 LT for Figure 4a and 1100 LT for Figures 4b and 4c. (d) Diurnal amplitude of the rainfall rate on the ground. Only diurnal components with a longitudinal wave number of 7 to +5 are extracted and plotted. In Figures 4a 4c, the location of the magnetic equator is shown by a dashed line (a tilted dipole field is assumed in the present model). Vertical gray thin lines between Figures 4a and 4c denote the location of the E B velocity peak. peaks during days 13 to 15 (Figure 3b). Periodic behaviors, which are different for different longitudes, are also evident. For example, 2 and 3 day oscillations are dominant between 10 and 50 longitudes at the south EIA (Figure 3b). Periodic behaviors can also be found at the corresponding longitudes for equatorial E B drift (Figure 3c). The correlation between the daily variation of the equatorial E B drift and that of the nmf2 at the EIA crests depends on the longitude and also has a north south asymmetry; there is a good correlation (correlation coefficient of ) for the north crest at 60 to 0 longitudes and for the south crest at 50 to 180 longitudes and an especially bad correlation (correlation coefficient <0) for the north crest at 80 to 120 longitudes and for the south crest at 110 to 50 longitudes. There is a tendency for a good (bad) correlation to occur at longitudes where the EIA crest is located at a lower (higher) geographic latitude (see Figure 4a for the crest latitudes), and therefore the effect of meridional wind becomes weaker (stronger). The meridional neutral wind is another driver that affects F region plasma density variation, moving the plasma up and down along the magnetic field line via ion neutral collision and, thus, changing the recombination rate of the plasma [e.g., Prölss, 1995]. It is interesting to note that the reproduced day to day variations of the F region ionosphere originate purely from the lower atmospheric variability, since other sources, such as the daily variations of UV and EUV fluxes and the magnetospheric inputs, are not considered here. Day to day variations may be caused by various kinds of atmospheric waves through thermosphere ionosphere interactions (chemical reactions, ion neutral drag, and dynamo process); this requires further analysis. [12] Most observational studies have reported averaged ionospheric longitudinal structures. In the following, we focus our attention on the averaged ionospheric structure and its relation to the lower atmosphere. Figures 4a 4d show the 30 day averaged longitude latitude distribution of the nmf2, the upward E B drift velocity at a 300 km height, the neutral temperature at 110 km, and the rainfall rate on the ground obtained by GAIA, respectively, while Figures 5a 5d are the results from a spectrum analysis for the variable and altitude corresponding to Figures 4a 4d, respectively, but only for the low latitude region. In Figure 4a the local time fixed nmf2 distribution (1500 LT) shows clear northern and southern crests of EIA away from the magnetic equator. The south crest has four peaks, around 165, 65, 15, and 105 longitudes, while such a four peak structure is less clear in the Northern Hemisphere. This feature is shown further in Figure 5a, which indicates that wave number 4 is the most significant for the south crest, but wave number 1 is dominant for the north crest at this local time. The peak of wave 1 is located in the American region for the Northern Hemisphere but in the Asian region for the Southern Hemisphere. [13] Compared with observations of the F region plasma density, the wave 4 structure becomes dominant only in the Southern Hemisphere in our simulation, while the wave 4 structure is observed in both hemispheres for a similar situation (September equinox, daytime, moderate solar activity) [Lin et al., 2007; Kil et al., 2008; Liu and Watanabe, 2008; Scherliess et al., 2008]. The amplitude of wave 4 is about 4% in our simulation, which is a degree similar to the total electron content observation by TOPEX (e.g., Figure 8 of Scherliess et al. [2008]) but significantly smaller than the electron content integrated over km obtained by the FORMOSAT 3/COSMIC observation (e.g., Figure 1 of Lin et al. [2007], which shows an amplitude of more than 10%). As for the wave 1 structure, most observations do not show a clear wave 1, except for the plasma density observation at a 400 km height made by the CHAMP satellite (e.g., Figure 1 of Liu and Watanabe [2008]). In their observation an enhancement of plasma density is shown in the American longitudinal sector for the Northern Hemisphere as well as in the Asian sector for the Southern Hemisphere; these locations 4 of 9

5 agree with our simulation. However, the north south asymmetry in the amplitudes of the wave numbers 1 and 4 components seems too large for the present simulation. [14] The F region zonal electric field (upward E B drift) is known as the driver of EIA formation, and its longitudinal distribution should contribute to the EIA structure. In Figure 4b the local time fixed distribution (1100 LT) of the upward E B drift at a 300 km height shows a four peak structure along the magnetic equator. Figure 5b also indicates the significant existence of the wave 4 structure along with the dominant wave 1 structure. As denoted by the thin lines vertical between Figures 4a and 4b, the peak locations of upward E B drift are close to those of nmf2, at least for the southern EIA crest, while the longitude of the wave 1 peak does not match between the upward E B drift and the nmf2 at the northern EIA crest; this is discussed later. Compared with observations of vertical drift, the significant existence of the wave 4 structure agrees with existing observations [Hartman and Heelis, 2007; Kil et al., 2007, 2008; Ren et al., 2009]. However, the amplitude of wave 4 is considerably lower in our simulation (about 4%) than in the observations (e.g., about 10% is shown in Figure 2 of Ren et al. [2009] for a similar local time and season). The wave 1 structure may possibly exist also in the observed distribution, although it has not been featured so far. [15] The daytime F region electric field is known to be generated mainly in the E region and transmitted via nearly equipotential geomagnetic field lines. In Figure 4c the localtime fixed distribution (1100 LT) of the neutral temperature at a 110 km height, where the Hall conductivity is highest in its altitude profile, has a dominant wave 4 longitudinal structure at latitudes of about 20 to 20. The longitudinal structure can be made up by solar nonmigrating tidal components, because their tidal phases vary with longitude at a fixed local time. In Figure 5c, other than the migrating tides (wave number 1 for diurnal and 2 for semidiurnal; hereafter, a plus sign indicates eastward propagating, while a minus sign indicates westward), the most prominent nonmigrating tide is the diurnal tide with a zonal wave number of +3 (or DE3), and other nonmigrating tides are also seen, such as diurnal tides with a wave number of 5 to 2 (DW5 DW2) and semidiurnal tides with a zonal wave number of 3 (SW3). According to the climatological temperature observation made by the TIMED/SABER instrument, diurnal tides with a wave number of 3to 1 and +2 to +4 and semidiurnal tides with a wave number of 4 to 1 are significant at nearly the same altitude in the equatorial region for the corresponding month (September) [Forbes et al., 2008]. Our simulation successfully reproduces the major tidal components in this month except for some quantitative differences. The amplitudes of dominant components such as DE3 and SW3 are in good agreement, while there are some differences for other tides (e.g., the amplitude of DW3 tide was not observed to be as high). 5of9 Figure 5. Ionospheric longitudinal structures and tidal components for low latitude and 30 day averaged variables. Amplitude of zonal wave number spectrum (a) for nmf2 at equatorial ionization anomaly (EIA) crests (at 1500 LT) and (b) for upward E B velocity at the magnetic equator (at 1100 LT). They are normalized by the background values (wave number 0 component). (c, d) Amplitudes of diurnal and semidiurnal zonally propagating components for the neutral temperature at a 110 km height averaged over 10 to 10 latitudes and those for rainfall rate on the ground averaged over 5 to 15 latitudes, where the average rainfall rate is high as shown in Figure 4d. Note that positive and negative signs of wave number indicate eastward and westward propagating, respectively.

6 Figure 6. Same as Figure 4d, but for the amplitude of the diurnal component combined from DW1 and DE3. [16] The longitudinal distribution of tropospheric moist convection is considered to be one of the main causes for the generation of nonmigrating tides that propagate upward to the thermosphere [e.g., Forbes et al., 1997; Hagan and Forbes, 2002, 2003]. In order to show the distribution of tropospheric moist convection as a tidal source, we carried out a Fourier filtering on the 30 day average rainfall rate to exhibit the amplitude of diurnal aggregate composed from the zonal wave numbers of 7 to +5 (Figure 4d). The semidiurnal component has a similar distribution (not shown). The distribution of rainfall that occurs internally in the model apparently depends on the land sea distribution, indicating that active moist convection occurs over the American continent, the African continent, and the southeast Asian to west Pacific regions. As shown in Figure 5d the rainfall rate at low latitudes can be Fourier decomposed into the prominent diurnal propagating component with a wave number of +3 and other diurnal components, such as of wave numbers 5 to 0 and +5, and semidiurnal components, such as of wave number Discussion and Summary [17] In the following we discuss the interpretations and implications of the simulation results on atmospheric vertical connection shown in Figures 4 and 5. First, the sources of dominant nonmigrating tides shown in Figure 4c are discussed. Tropospheric moist convection is one of the sources. In Figure 4d the rainfall rate increases in the equatorial region at separate longitude sectors, at about 70 W, 25 E, 110 E, and 150 E, where the vertical motion of atmosphere is enhanced, and the water vapor also increases, in the upper troposphere (not shown), thus indicating the region of active moist convection. In these regions atmospheric waves can be excited owing to latent heat release and insolation absorption by water vapor and clouds [e.g., Forbes et al., 1997; Hagan and Forbes, 2002, 2003]. Figure 6 shows a plot similar to Figure 4d, but for the amplitude of rainfall rate combined from the diurnal migrating component (DW1) and the dominant nonmigrating component (DE3). The heating source for the DE3 tide exists at longitudes around 160 W, 70 W, 20 E, and 110 E for the equatorial region; the latter three sectors, over the American continent, the African continent, and the Southeast Asian region, coincide with the region of active moist convection shown in Figure 4d. The DE3 tide propagates upward and becomes the dominant nonmigrating tidal component in the lower thermosphere, as shown in Figure 5c. Miyoshi [2006] used the same GCM to analyze the propagation of DE3 and its impact on the upper atmosphere and suggested that the DE3 tide dominantly appearing at a 110 km height originates in the tropospheric moist convection. In contrast, nonlinear interaction between the atmospheric waves is another source for the generation of nonmigrating atmospheric tides. The dominant semidiurnal nonmigrating tide, SW3, shown in Figure 5c could possibly be generated by a nonlinear interaction between the migrating SW2 tide and a stationary planetary wave of zonal wave number 1 [Forbes et al., 1995; AngelatsiCollandForbes,2002;Yamashita et al., 2002]. However, the contribution of the tropospheric activity and the nonlinear wave interaction as the source for each tidal component in the thermosphere requires further analysis. [18] Regarding the possible tidal effect on the ionosphere, we showed only the longitudinal spectrum for a local time fixed distribution (Figures 5a and 5b) rather than a space time spectrum as shown in Figures 5c and 5d because the effective tidal altitude for the dynamo process is strongly dependent on the local time (especially drastic change occurs at sunrise and sunset) [Jin et al., 2008]. In addition, the dominant drivers affecting F region plasma density (i.e., photochemical reactions, dynamo electric field, and neutral wind drag along the geomagnetic field line) also change with local time and geographic latitude, as already discussed. Instead, we have examined the local time phase shift for several major longitudinal wave components, shown in Figures 5a and 5b. Figure 7 shows the results for upward E B drift (plus symbol) and nmf2 at the northern EIA (triangle) for wave 1 components and results for E B drift (diamond) and nmf2 at the southern EIA (asterisk) for wave 4 components, where the phase is defined as the peak longitude. The DE3 signature (eastward phase shift at a speed of 3.75 /h for wave 4 components) can be found in the phase variation of E B drift between 1130 and 1700 LT and, also, in that of nmf2 at the southern EIA, with a delay of about 3 h from the E B peak, which reflects the response time of EIA density to the equatorial electric field. The wave 4 phase of E B drift is rather stationary in the earlier morning, from about 0600 to 0900 LT. The DE3 effect is also indicated by the thin lines drawn between Figures 4b and 4c. The location of the enhanced eastward electric field roughly coincides with the region of the westward gradient of neutral temperature along Figure 7. Local time variation of the phase for the several major longitudinal wave components shown in Figures 5a (nmf2 at EIA crests) and 5b (upward E B velocity at the magnetic equator and a 300 km height). Blue and red curves denote the phase speeds of SW3 and DE3, respectively. 6of9

7 Figure 8. Longitudinal distribution of zonal neutral wind along the magnetic equator at a 110 km height. The wind represents the 30 day averaged value and is shown in the local time fixed frame (1100 LT). As references, the neutral temperature along the magnetic equator at the same height is shown by the gray curve, and locations of upward E B velocity peaks at 300 km are also indicated by vertical dashed lines. the magnetic equator (except around 150 W), where the neutral wind is enhanced in the westward direction as shown by Figure 8. The opposite direction between the F region electric field and the neutral wind at the dynamo altitude is consistent with the dynamo simulation by Jin et al. [2008]. They showed that the charge accumulation by the convergence or divergence of the zonal Hall dynamo current produces the zonal polarization electric field in the direction opposite the neutral wind. [19] As for the wave 1 structure in the ionosphere, Figure 7 indicates a signature of SW3 (the westward phase shift at a speed of 30 /h for wave 1 components) for the E B drift phase variation during almost the whole daytime and, also, during some nighttime hours. The dynamo efficiency of the SW3 wind is relatively good, since the tidal wind has a symmetric about the equator at low latitudes and E region height with a vertical wavelength slightly shorter than that of the DE3 tide. The possible importance of SW3 for the dynamo process was also discussed by Forbes et al. [2008] based on the TIMED/SABER observation. [20] In contrast, the dominant wave 1 structure of nmf2 at the northern EIA is not correlated with that of E B drift, and rather it has a stationary phase variation. In the case of nmf2, the offset of geomagnetic equator from the geographic equator, which varies with longitude (thus stationary to the longitude), is likely responsible for the formation of wave 1 structure. Generally, the difference of EIA crest in the geographic latitude leads to the difference in the background poleward thermospheric wind typical for the daytime equinox season as well as the difference in the EUV flux intensity even at the same solar local time (due to the change in the solar zenith angle). Figure 9 shows the longitudinal distribution of daytime meridional wind at EIA crest latitudes in the present simulation, along with its migrating component. Since the migrating component does not vary with longitude at a fixed local time if the geographic latitude is fixed, the dominant wave 1 profile in Figure 9 can be attributed to variation in the geographic latitude of EIA crest locations. The nmf2 at the EIA crest decreases (increases) at the longitude where the crest is farther from (closer to) the geographic equator, and the poleward wind becomes stronger (weaker), since the ionospheric plasma is moved (not moved) to a lower altitude along the magnetic field line where the recombination rate becomes higher [e.g., Prölss, 1995]. Figure 9 also explains the north south asymmetry of wave 1 structure in our simulation; the meridional wind is larger in the north than in the south, thus the effect of longitudinal variation in the equator offset tends to become larger in the Northern Hemisphere. [21] Although not shown in this paper, through implementation of our simulation results (neutral density, velocity, and electric field) into the SAMI2 model [Huba et al., 2000], our model was found to overestimate the effect of meridional wind while underestimating the effect of the electric field. This error in the present model may contribute to the major dissimilarities between the present simulation results and existing observations. As discussed in section 3, the northsouth asymmetry in the longitudinal variation of EIA crest density (specifically, the asymmetry in the amplitudes of wave numbers 1 and 4 components) is overestimated in our simulation, although the phase of wave 1 is coincident with observation. Since the wave 1 structure of EIA plasma density is mainly caused by longitudinal variation of the meridional wind at the EIA crest due to magnetic equator offset, the asymmetry of wave 1 should be smaller if the effect of meridional wind is smaller in reality. We plan to improve the model in this respect in the near future. However, the qualitative features of the averaged longitudinal distribution in our simulation are similar to existing observations as discussed in the previous section. [22] In summary, we have introduced a new Earth s atmosphere ionosphere coupled model, and shown that it can be a useful tool for study of vertical coupling processes from the troposphere to the ionosphere. The model can successfully reproduce the four peak longitudinal structure of the EIA similar to the recent observations, as well as the key processes leading to its formation proposed in previous studies, that is, the excitation of nonmigrating tides in tro- Figure 9. Longitudinal distribution of meridional neutral wind along the magnetic latitudes of 15 S and 15 N (i.e., near the EIA crests) at a 400 km height. The wind represents the 30 day averaged value and is shown in the local time fixed frame (1100 LT). Dashed curves represent migrating components. 7of9

8 pospheric moist convection, their upward propagation, generation of a dynamo electric field, and the fountain effect. These processes are connected by one coupled model. The simulation also revealed several interesting features. The four peak structure is seen in the averaged distribution of the F region ionosphere, but in fact the simulated longitudinal structure varies greatly on a day to day basis. Since the solar EUV and UV fluxes and magnetospheric inputs were assumed to be constant, the reproduced ionospheric day today variation originates from the lower atmosphere, and further studies with respect to their relation to specific atmospheric waves and mechanisms will be necessary. As an average feature, not only the DE3 tide, but also other nonmigrating tides such as SW3, may also have an impact on the ionospheric structure. [23] Acknowledgments. This work was supported by a MEXT Grant in Aid for Scientific Research on Innovative Areas ( ). Computations in this study were performed on the NEC SX 8R at the National Institute of Information and Communications Technology, Japan. We used the SAMI2 ionospheric model written and developed by the Naval Research Laboratory for comparison purposes. We also thank Huixin Liu for helping us determine the model name. [24] Robert Lysak thanks the reviewers for their assistance in evaluating this paper. References Angelats i Coll, M., and J. M. Forbes (2002), Nonlinear interactions in the upper atmosphere: The s = 1 and s = 3 nonmigrating semidiurnal tides, J. Geophys. Res., 107(A8), 1157, doi: /2001ja England,S.L.,T.J.Immel,E.Sagawa,S.B.Henderson,M.E.Hagan, S. B. Mende, H. U. Frey, C. M. Swenson, and L. J. Paxton (2006a), The effect of atmospheric tides on the morphology of the quiet time, postsunset equatorial ionospheric anomaly, J. Geophys. Res., 111, A10S19, doi: /2006ja England, S. L., S. Maus, T. J. Immel, and S. B. Mende (2006b), Longitudinal variation of the E region electric fields caused by atmospheric tides, Geophys. Res. Lett., 33, L21105, doi: /2006gl Fejer,B.G.,J.W.Jensen,andS. Y. Su (2008), Quiet time equatorial F region vertical plasma drift model derived from ROCSAT 1 observations, J. Geophys. Res., 113, A05304, doi: /2007ja Forbes, J. M., N. A. Makarov, and Y. I. Portnyagin (1995), First results from the meteor radar at South Pole: A large 12 hour oscillation with zonal wavenumber one, Geophys. Res. Lett., 22(23), , doi: /95. Forbes, J. M., M. E. Hagan, X. Zhang, and K. Hamilton (1997), Upper atmosphere tidal oscillations due to latent heat release in the tropical troposphere, Ann. Geophys., 15, Forbes, J. M., X. Zhang, S. Palo, J. Russell, C. J. Mertens, and M. Mlynczak (2008), Tidal variability in the ionospheric dynamo region, J. Geophys. Res., 113, A02310, doi: /2007ja Fujiwara, H., and Y. Miyoshi (2006), Characteristics of the large scale traveling atmospheric disturbances during geomagnetically quiet and disturbed periods simulated by a whole atmosphere general circulation model, Geophys. Res. Lett., 33, L20108, doi: /2006gl Fujiwara, H., and Y. Miyoshi (2009), Global structure of large scale disturbances in the thermosphere produced by effects from the upper and lower regions: Simulations by a whole atmosphere GCM, Earth Planets Space, 61, Fujiwara, H., and Y. Miyoshi (2010), Morphological features and variations of temperature in the upper thermosphere simulated by a whole atmosphere GCM, Ann. Geophys., 28, Fuller Rowell, T. J., et al. (2008), Impact of terrestrial weather on the upper atmosphere, Geophys. Res. Lett., 35, L09808, doi: / 2007GL 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 Hagan, M. E., A. Maute, R. G. Roble, A. D. Richmond, T. J. Immel, and S. L. England (2007), Connections between deep tropical clouds and the Earth s ionosphere, Geophys. Res. Lett., 34, L20109, doi: / 2007GL Hartman, W. A., and R. A. Heelis (2007), Longitudinal variations in the equatorial vertical drift in the topside ionosphere, J. Geophys. Res., 112, A03305, doi: /2006ja Huba, J. D., G. Joyce, and J. A. Fedder (2000), Sami2 is another model of the ionosphere (SAMI2): A new low latitude ionosphere model, J. Geophys. Res., 105(A10), 23,035 23,053. 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 Jin, H., Y. Miyoshi, H. Fujiwara, and H. Shinagawa (2008), Electrodynamics of the formation of ionospheric wave number 4 longitudinal structure, J. Geophys. Res., 113, A09307, doi: /2008ja Kil, H., S. J.Oh,M.C.Kelley,L.J.Paxton,S.L.England,E.Talaat, K. W. Min, and S. Y. Su (2007), Longitudinal structure of the vertical EXB drift and ion density seen from ROCSAT 1, Geophys. Res. Lett., 34, L14110, doi: /2007gl Kil, H., E. R. Talaat, S. J. Oh, L. J. Paxton, S. L. England, and S. J. Su (2008), Wave structures of the plasma density and vertical EXB drift in low latitude F region, J. Geophys. Res., 113, A09312, doi: / 2008JA Lin, C. H., C. C. Hsiao, J. Y. Liu, and C. H. Liu (2007), Longitudinal structure of the equatorial ionosphere: Time evolution of the four peaked EIA structure, J. Geophys. Res., 112, A12305, doi: /2007ja Liu, H., and S. Watanabe (2008), Seasonal variation of the longitudinal structure of the equatorial ionosphere: Does it reflect tidal influences from below?, J. Geophys. Res., 113, A08315, doi: / 2008JA Luhr, H., M. Rother, K. Hausler, P. Alken, and S. Maus (2008), Influence of nonmigrating tides on the longitudinal variation of the equatorial electrojet, J. Geophys. Res., 113, A08313, doi: /2008ja Miyoshi, Y. (2006), Temporal variation of nonmigrating diurnal tide and its relation with the moist convective activity, Geophys. Res. Lett., 33, L11815, doi: /2006gl Miyoshi, Y., and H. Fujiwara (2003), Day to day variations of migrating diurnal tide simulated by a GCM from the ground surface to the exobase, Geophys. Res. Lett., 30(15), 1789, doi: /2003gl Miyoshi, Y., and H. Fujiwara (2006), Excitation mechanism of intraseasonal oscillation in the equatorial mesosphere and lower thermosphere, J. Geophys. Res., 111, D14108, doi: /2005jd Miyoshi, Y., and H. Fujiwara (2008), Gravity waves in the thermosphere simulated by a general circulation model, J. Geophys. Res., 113, D01101, doi: /2007jd Prölss, G. W. (1995), Ionospheric F region storms, in Handbook of Atmospheric Electrodynamics, edited by H. Volland, pp , CRC Press, Boca Raton, Fla. Oberheide, J., Q. Wu, T. L. Killeen, M. E. Hagan, and R. G. Roble (2006), Diurnal nonmigrating tides from TIMED Doppler Interferometer wind data: Monthly climatologies and seasonal variations, J. Geophys. Res., 111, A10S03, doi: /2005ja Ren, Z., W. Wan, L. Liu, and J. Xiong (2009), Intra annual variation of wave number 4 structure of vertical E B drifts in the equatorial ionosphere seen from ROCSAT 1, J. Geophys. Res., 114, A05308, doi: /2009ja Ren, Z., W. Wan, J. Xiong, and L. Liu (2010), Simulated wave number 4 structure in equatorial F region vertical plasma drifts, J. Geophys. Res., 115, A05301, doi: /2009ja Roble, R. G. (1996), The NCAR thermosphere ionosphere mesosphereelectrodynamics general circulation model, in STEP Handbook on Ionospheric Models, edited by R. W. Schunk, pp , Utah State Univ., Logan. Roble, R. G. (2000), On the feasibility of developing a global atmospheric model extending from the ground to the exosphere, Atmospheric Science Across the Stratopause, Geophysical Monogr. Ser., vol. 123, edited by D. E. Siskind, S. D. Eckermann, and M. E. Summers, pp , AGU, Washington, D. C. 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 Scherliess, L., D. C. Thompson, and R. W. Schunk (2008), Longitudinal variability of low latitude total electron content: Tidal influences, J. Geophys. Res., 113, A01311, doi: /2007ja Shinagawa, H. (2011), Ionosphere simulation, J. Natl. Inst. Inf. Comm. Technol., in press. 8of9

9 Shinagawa, H., and S. Ohyama (2006), A two dimensional simulation of thermospheric vertical winds in the vicinity of an auroral arc, Earth Planets Space, 58, Shinagawa, H., T. Iyemori, S. Saito, and T. Maruyama (2007), A numerical simulation of ionospheric and atmospheric variations associated with the Sumatra earthquake on December 26, 2004, Earth Planets Space, 59, Volland, H. (1975), Models of the global electric fields within the magnetosphere, Ann. Geophys., 31, Wan,W.,L.Liu,X.Pi,M. L. Zhang, B. Ning, J. Xiong, and F. Ding (2008), Wavenumber 4 patterns of the total electron content over the low latitude ionosphere, Geophys. Res. Lett., 35, L12104, doi: / 2008GL Yamashita,K.,S.Miyahara,Y.Miyoshi,K.Kawano,andJ.Ninomiya (2002), seasonal variation of non migrating semidiurnal tide in the polar MLT region in a general circulation model, J. Atmos. Sol. Terr. Phys., 64, H. Fujiwara, K. Terada, and N. Terada, Department of Geophysics, Graduate School of Science, Tohoku University, Aramaki Aza Aoba 6 3, Aoba ku, Sendai, Miyagi , Japan. M. Ishii, H. Jin, and H. Shinagawa, National Institute of Information and Communications Technology, Nukui Kitamachi, Koganei, Tokyo , Japan. (jin@nict.go.jp) Y. Miyoshi, Department of Earth and Planetary Sciences, Kyushu University, Hakozaki, Fukuoka , Japan. Y. Otsuka, Solar Terrestrial Environment Laboratory, Nagoya University, Furo-cho, Chikusa ku, Nagoya , Japan. A. Saito, Graduate School of Science, Kyoto University, Kyoto , Japan. 9of9