In-situ plasma density measurements from the CHAMP satellite between about 350–420 km altitude are used to delineate intra-annual variations in the longitudinal structure of the F-region ionosphere. It is shown that the longitude structures during mid-day local times are dominated by the space-based longitudinal wavenumbers ks = 2, 4 and 3 in January, July and December, respectively. These conform to the same dominating ks-values characterizing solar thermal tide zonal winds in the dynamo region, namely the westward-propagating semidiurnal tide with planetary-fixed zonal wavenumber s = 4 (SW4), and the eastward-propagating diurnal tides with s = −3 (DE3) and s = −2 (DE2), respectively. DE3 is the dominating tide during the other months. The results presented indicate that nonmigrating tides other than DE3 may significantly impact the longitudinal structure of the low-latitude ionosphere during certain months.
 The dominance of a zonal wavenumber s = 4 (hereafter called wave-4) longitude structure of the equatorial ionization anomaly (EIA) was first observed by Sagawa et al.  using data from the Far Ultraviolet (FUV) imager onboard the IMAGE satellite. The dominant wave-4 longitudinal structure has subsequently been observed in total electron content (TEC) [Scherliess et al., 2008; Wan et al., 2008], electron density profiles [Lin et al., 2007] and in-situ electron density measurements [Kil et al., 2007; Liu et al., 2008]. Immel et al.  first proposed that the wave-4 longitudinal structure of the EIA arises due to upward propagating tides that modulate electric fields in the ionospheric E-region. The predominant origin of the wave-4 structure appears to be the eastward- propagating diurnal tide with zonal wavenumber s = −3 [Hagan et al., 2007] often referred to as DE3. This wave is generated by the longitude dependence of latent heating due to deep convection in the tropics [Hagan and Forbes, 2002; Forbes et al., 2001, 2006; Oberheide et al., 2006]. DE3 propagates vertically, and achieves large amplitudes in the ionospheric E-region [Forbes et al., 2008; Oberheide and Forbes, 2008] where electric fields are generated by the dynamo mechanism. The wave-4 longitudinal structure observed in the equatorial electrojet [Lühr et al., 2008] and E × B drift velocities [Kil et al., 2007; Hartman and Heelis, 2007] have provided strong observational evidence for the modulation of E-region electric fields by DE3. These electric fields map into the F-region, produce vertical plasma drifts and control the formation and strength of the EIA.
 We preface introduction of the problem at hand by noting that a solar tide with frequency nΩ where Ω = h−1 and n = 1, 2, 3 (i.e., diurnal, semidiurnal, terdiurnal) and zonal wavenumber s [s < 0 (s > 0) corresponding to eastward (westward) propagation] appears as a longitude variation with zonal wavenumber ks = ∣s − n∣ at a constant local time, or equivalently, from Sun-synchronous orbit. Let us adopt the convention of DWs or DEs to denote a westward- or eastward-propagating diurnal tide, respectively, with zonal wavenumber s. For semidiurnal and terdiurnal oscillations let S and T replace D. The corresponding zonally-symmetric oscillations are denoted D0, S0,T0, and stationary planetary waves (SPW) with zonal wavenumber s are expressed as SPWs. Then, it is simple to see that DE3 (n = 1, s = −3) and its various manifestations (e.g., electric fields, composition variations) produce wave-4 (ks = 4) structures in a local time frame. However, it is also simple to see that SE2 (n = 2, s = −2), DW5 (n = 1, s = 5), SW6 (n = 2, s = 6), SPW4 (n = 0, s = 4), and so on, also appear as ks = 4. By the same token, a number of other combinations of n and s can produce ks = 1, 2 and 3 signatures.
 Recent studies have shown that similar variability is observed in the wave-4 structure of the low-latitude ionosphere and the DE3 tide [Immel et al., 2008; Liu et al., 2008; Scherliess et al., 2008; Wan et al., 2008]. Using SABER temperature data, Forbes et al.  demonstrated that while DE3 dominates the nonmigrating tidal spectrum during most months, other waves (e.g., DE2, DW2, SW4, TW5) also make important contributions and moreover some waves dominate over DE3 during certain months. Similar conclusions are expected for dynamo-region winds, and Oberheide and Forbes  recently validated a method that uses tidal temperature and wind measurements to derive a full specification of tidal winds in the dynamo region. These studies raise the question of whether similar month-to-month variability exists in the longitude structure of the dynamo region and the low-latitude F-region ionosphere. The time periods when DE3 is no longer the dominant tidal component are of particular interest as it is presently unclear how this impacts the longitudinal structure of the EIA. The primary aim of this paper is to further explore the month-to-month variability of the longitudinal structure of the EIA with particular emphasis on time periods when DE3 is not the dominant tidal component. F-region electron densities from the Planar Langmuir Probe (PLP) onboard the Challenging Minisatellite Payload (CHAMP) satellite and equatorial zonal winds derived from the SABER and TIDI instruments onboard the TIMED satellite are used to further explore the connection between E-region winds and the longitudinal structure of the EIA.
2. Data and Analysis
2.1. CHAMP In-Situ Electron Densities
 The CHAMP satellite is in a nearly Sun-synchronous orbit and it precesses in local time at the rate of ∼5.44 minutes per day. In-situ electron densities can be derived from the PLP onboard the CHAMP satellite [Rother et al., 2004] and are available from the Information Systems and Data Center operated by Geo Forschungs Zentrum (GFZ) Potsdam (http://isdc.gfz-potsdam.de).
 We use CHAMP electron density data during the four time periods given in Table 1. Also given in Table 1 is the altitude of the CHAMP satellite between −30° and +30° latitude, the observed local time, and the mean F10.7 flux. These time periods were selected because geomagnetic activity was low and CHAMP sampled mid-day local times when longitudinal structures are expected to be present. DE3 is weakest during December–January and we have thus mainly focused on this time period. One time period in July is included in order to demonstrate that a wave-4 longitudinal structure is observed when DE3 is the dominant nonmigrating tide. Due to the changing local time of the CHAMP satellite, only one time period in January was suitable. Two time periods in December were suitable and both have been included to provide additional confirmation of the results obtained during this time period. Representative global ionospheric maps are created for each time period by binning the data in 10 degrees longitude and 1 degree latitude. As shown in Table 1 the sampling altitude differs for each of the time intervals analyzed which will impact the observed electron densities. It is also uncertain where any given measurement falls with respect to the F-region peak density. However, all of the measurements fall within the 300–450 km height range that Lin et al.  found to have a significant wave-4 structure indicating that the in-situ measurements are a suitable means for observing longitudinal structures in the F-region.
Table 1. Local Time, Height and Mean F10.7 for the Time Periods Considered in the Present Studya
Mean F10.7 (W/m2/Hz)
The height range corresponds to the change that occurs between −30° and +30° latitude.
where ne,n and ne,s are the electron densities of the north and south peaks and ne,t is the electron density of the equatorial trough. The electron density profile at each longitude can then be characterized by a single number. Note that the CTR is set to 1 when the EIA has not been well established at satellite altitudes. Alternatively, the latitude separation of the EIA peaks may be used to characterize the strength of the EIA. We have obtained similar results using both approaches, and have chosen to use the CTR for the present analysis.
2.2. Zonal Wind Amplitudes From TIMED
 The nonmigrating tidal amplitudes of the equatorial zonal wind at 100 km are determined using a Hough Mode Extension (HME) analysis of temperature and zonal and meridional wind measurements from the SABER and TIDI instruments onboard the TIMED satellite [Oberheide and Forbes, 2008]. An altitude of 100 km is used due to superior data quality at this altitude. Since the nonmigrating tides are vertically propagating, observations at 100 km can be viewed as representative of the nonmigrating tides entering the dynamo region. A five-year mean (2002–2006) is used in order to determine the month-to-month variability in the leading nonmigrating tidal components. The equatorial zonal wind results are representative of the broader latitude region between ±30° latitude.
3. Results and Discussion
 The month-to-month variation of the equatorial zonal winds at 100 km for the leading nonmigrating tidal components is shown in Figure 1. The leading tidal components are DE2, DE3, and SW4 which when viewed at a fixed local time appear as wave-3, wave-4, and wave-2, respectively. The amplitude of each nonmigrating tide exhibits significant month-to-month variability. In particular, it should be noted that during January the leading tidal component is SW4 followed by DE2 and DE3. During March–November DE3 is greatest followed by DE2 and SW4. In December DE2 is the primary component followed by SW4 and DE3. Given the premise that the E-region tides are expected to induce a longitudinal dependence in the dynamo electric fields, one may expect to observe a similar variability in the longitudinal structure of the EIA. Therefore, the dominant longitudinal structure of the EIA is expected to be wave-2 in January, wave-4 in July, and wave-3 in December. Other tidal components, many of which appear as wave-2, -3, and -4, are also present and we have focused here on the three leading tidal components. A detailed discussion of the various nonmigrating tides that appear as wave-2, -3, and -4 and the relative strengths of each is given in Forbes et al. .
 Representative global electron density maps for the four time periods considered are displayed in Figure 2 (left). A Fourier fit is performed using the CTR values at each longitude and the resulting wavenumber spectra are shown in the right panel. Similar to the variations seen in the zonal wind amplitudes, the primary component of the CTR amplitude exhibits intra-annual variability. In January 2003, the dominant component is wave-2 followed by wave-3 and wave-4 is the weakest component. The primary component changes to wave-4 in July 2004 and a similar amplitude is seen for waves 1–3 during this time period. In December 2004 and 2005 the dominant wavenumber is 3 followed by wave-2, -1 and -4. There are also significant changes in the amplitudes from December 2004 to December 2005 which is thought to be due to changes in the orbital altitude of the CHAMP satellite and/or changes in the level of solar activity (see Table 1).
 Although the primary focus is on SW4, DE2, and DE3 and their corresponding signatures in the ionosphere, a significant wave-1 component is also present. The offset of the geomagnetic field with respect to the geographic equator and longitudinal variation in the neutral winds create hemispheric asymmetries in the longitudinal structure of the ionosphere [McDonald et al., 2008]. The resulting hemispheric asymmetries will impact the CTR value and may be partly responsible for the observed wave-1 component. A wave-1 component can also result from the nonmigrating tides D0, DW2, SW1, and SW3. These tides arise from nonlinear interaction of SPW1 with DW1 and SPW1 with SW2 [Angelats i Coll and Forbes, 2002; Hagan and Roble, 2001] The months in question overlap with SPW1 activity in the respective winter hemispheres indicating that the observed wave-1 may partly be due to these nonmigrating tides.
 We now turn our attention to discuss the results shown in Figure 2 in the context of Figure 1. In January the leading component in the zonal wind amplitude is SW4 (wave-2) which is also the dominant wavenumber of the CTR. This demonstrates that SW4 may be capable of modulating the dynamo electric fields in order to create a wave-2 longitudinal structure in the EIA during January. During this same time period, the second leading component is observed to be wave-3 in both the zonal wind amplitude as well as the CTR. A wave-4 (DE3) structure dominates the zonal wind amplitude in July and this is also the dominate wavenumber of the CTR. In July 2004 the CTR amplitude for wave-2 and wave-3 are equivalent whereas from Figure 1 it is expected that wave-3 be greater than wave-2. The results shown in Figure 1 represent a five year mean and this discrepancy may be the result of year-to-year variability in the zonal wind. Lastly, a wave-3 structure is observed to be the primary component of both the zonal wind amplitude and the CTR during December. In both December 2004 and December 2005, wave-2 is the second major component of the CTR which is in agreement with the zonal wind amplitudes. Similar to SW4 in January, this demonstrates that DE2 is likely able to influence the dynamo electric fields in order to create a wave-3 structure in the EIA during December.
 Recent studies that have explored coupling between the dynamo region and the F-region ionosphere have emphasized the similar variability that is observed in the wave-4 structure and DE3 [Immel et al., 2008; Liu et al., 2008; Wan et al., 2008]. Through the inclusion of additional longitudinal structures and tidal components, the present paper provides additional insight into this coupling process. During January and December, DE3 is no longer the dominant tidal component in the equatorial zonal winds in the E-region and the dominant longitudinal structure of the EIA becomes wave-2 and wave-3, respectively. The change in the primary longitudinal structure of the EIA from wave-4 to wave-2 and wave-3 is thought to be due to the dominant nonmigrating tide no longer being DE3, but, being SW4 in January and DE2 in December. This demonstrates that, in addition to DE3, other nonmigrating tides are likely able to modulate the dynamo electric fields and subsequently influence the longitudinal structures in the F-region ionosphere.
 The four time periods used in the present analysis have provided observational evidence of longitudinal structures of the EIA other than wave-4 that develop when DE3 is no longer the primary nonmigrating tide in the dynamo region. It is believed that this is a recurrent feature in January and December due the dominant nonmigrating tides being SW4 and DE2, respectively. Monthly climatologies of the longitudinal structure of the low-latitude ionosphere that include a full spectrum of wavenumbers based upon a broader database of measurements (such as TEC or radio occultation measurements) would serve to further illustrate the importance of nonmigrating tides other than DE3 in creating longitudinal structures in the low-latitude ionosphere. First-principles modeling with a particular emphasis on how SW4 and DE2 influence the dynamo electric fields and F-region plasma densities are also necessary in order to further advance our understanding of this fascinating phenomenon that links together deep tropical convection in the troposphere with the temporal and spatial variability hundreds of kilometers above in the ionosphere.
 This work was supported under grant NNX07AB74G from the NASA TIMED Program and grant ATM-0719480 from the National Science Foundation as part of the Space Weather Program. J.O. is supported by the DFG through its priority program CAWSES, grant OB 299/2-2.