SEARCH

SEARCH BY CITATION

Keywords:

  • atmospheric tide;
  • diurnal eastward-propagating with zonal wavenumber-3 (DE3);
  • semidiurnal eastward-propagating with zonal wavenumber-2 (SE2);
  • transequatorial wind;
  • wave-4 structure

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results and Conclusions
  6. Acknowledgments
  7. References

[1] Recently a so-called “wave-4 structure” was discovered in the equatorial ionosphere, which describes plasma density enhancements within four longitude sectors separated about 90° from each other. This structure was proposed to be controlled by the tide mode of DE3 (diurnal eastward-propagating with zonal wave number-3) excited by different heating in the tropical troposphere due to land-sea differences. Here, using F2-layer peak density (NmF2) and the peak height (hmF2) extracted from the COSMIC data set, we investigate the wave-4 structure by decomposing it into interhemispheric symmetric and antisymmetric components. Our results indicate that the generally accepted mechanism of DE3 modulation of E region dynamo only accounts for the daytime symmetric components, while the antisymmetric components could be explained well in the terms of the SE2 (semidiurnal eastward-propagating with zonal wave number-2) tide in transequatorial neutral wind. Surprisingly, the antisymmetric component dominates the wave-4 structure in hmF2 during nighttime, suggesting the SE2 transequatorial wind is the leading contributor to the nighttime wave-4 structure in hmF2.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results and Conclusions
  6. Acknowledgments
  7. References

[2] Longitudinal variations of the plasma density in the ionosphere are usually attributed to the differences in geomagnetic field configuration [West and Heelis, 1996; He et al., 2009, 2011]. However, recent studies suggest that longitudinal wavelike structures in the equatorial ionosphere are controlled by atmospheric tides [e.g., Sagawa et al., 2005]. Atmospheric tides refer to global-scale oscillations in the atmosphere with periods of integral fractions of a day. Among the non-sun-synchronous tidal components, the DE3 component is generally the most dynamical in the mesosphere/low thermosphere (MLT) region during most of the year [e.g., Hagan and Forbes, 2002; Pedatella et al., 2008]. In a sun-synchronous frame, four longitudinal peaks are observed from the DE3 tide due to the Doppler Effect, known as wave-4 structure. The DE3 tide or wave-4 structure is widely observed in various neutral parameters in the MLT region, such as temperature and zonal winds [Forbes et al., 2003, 2006].

[3] In the ionosphere, the wave-4 signature was first detected in the nighttime airglow intensity over equator, and was attributed to the DE3 modulation of the E region dynamo and F region fountain [Sagawa et al., 2005; Immel et al., 2006]. Evidence supporting this hypothesis includes the longitudinal wave-4 fluctuation in the vertical plasma drift and the spatiotemporal correlations between ionospheric wave-4 signatures and the DE3 tides in the MLT region [Ren et al., 2009; Wan et al., 2008, 2010]. However, recently it was argued that thermal tides could directly penetrate into the F region [Forbes et al., 2009; Häusler and Lühr, 2009; Oberheide et al., 2009, 2011; Talaat and Lieberman, 2010]. Numerical simulations [England et al., 2010; Zhang et al., 2010] also suggested that, in addition to the electrodynamic couplings, the penetration of tides into the F region could contribute directly to the ionospheric wave-4 structure. One purpose of the present work is to explore evidence for the electrodynamic coupling and the ionospheric signatures of direct penetration of tides to evaluate the contributions from the main processes.

[4] We also aim to explore the ionospheric wave-4 signatures from tides beyond the DE3. As illustrated in Figure 1, the wave-4 structure could arise from other tide modes besides the DE3, such as the DW5 (diurnal westward-propagating with zonal wave number-5), SPW4 (stationary planetary wave number-4), and SE2 (semidiurnal eastward wave number-2). An indicator supporting the connection between wave-4 structure and DE3 tide is that the wave-4 pattern shifts eastward at ∼90°/day [e.g., Wan et al., 2008; Häusler and Lühr, 2009; Lühr et al., 2008]. However, the wave-4 pattern is found to be almost stationary in GUVI-O/N2 [He et al., 2010], suggesting a potential connection to the SPW4 mode. Based on an empirical model extended from tides in the MLT region, Oberheide et al. [2011] suggested that the ionospheric wave-4 structure was also contributed by the SE2 transequatorial wind resulted from tropospheric forcing and by the SPW4 and SE2 zonal winds excited by the nonlinear interaction between the DE3 and DW1 tides. Here, we report direct evidence in ionospheric wave-4 structures indicating considerable contributions from the SE2 tide mode.

image

Figure 1. Sketch maps for four tide modes contributing potentially to the longitudinal wave-4 structure. The solid line in each panel indicates the isoline of 12 local time (LT). Pluses on each of the lines illustrate that four longitudinal peaks would be recorded in fixed LT frame as observed by (quasi-) Sun-synchronous satellites.

Download figure to PowerPoint

2. Data and Methods

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results and Conclusions
  6. Acknowledgments
  7. References

[5] The receivers on board COSMIC mission collected GPS signals, from which vertical electron density profiles were retrieved through the Radio Occultation technique. The COSMIC mission had collected more than 2.7 million profiles from May, 1, 2006 to November 29, 2010. This period corresponds to the minimum phase of solar cycle 23/24, during which the daily 10.7 cm solar radio flux (F10.7) varies from 65.1 to 103.2 (10−22·Wm−2·Hz−1) with an average of 73.4. The level-2 products of COSMIC profiles are derived from the COSMIC Data Analysis and Archive Center. We fit each profile using a cubic spline to determine the F2-layer peak density (NmF2) and its height (hmF2). To simplify data operations, we construct empirical models for NmF2, hmF2 via the following procedures.

[6] First, we pick up the data collected under geomagnetically quiet time (Kp index < 3) and then bin the NmF2 measurements by day of year (DoY) and geomagnetic longitude (MLON) with a resolution of 30.4 days by 15° longitudes (12 × 24 bins in total). In each DoY-MLON bin, we fit the coefficients (cmn, cmn, smn, smn (n = 0, 1…, 13, m = 1, 2…, n)) of the associated Legendre polynomials according to:

  • equation image

where MLAT is geomagnetic latitude, MLT is the magnetic local time, Pnm is the Legendre polynomial of degree n and order m, Cmn = cmn + cmn · F10.7 and Smn = smn + smn · F10.7 are linear functions of F10.7. Second, in each DoY-MLON bin, NmF2 is computed by summing the Legendre polynomials on the 91 × 48 mesh grid of MLAT = −90, −88… 90 versus MLT = 0.25, 0.75… 23.75, at F10.7 = 70. Thus, NmF2 is represented by a 4-D matrix (12 × 24 × 91 × 48). Third, by separating variables into two groups, 〈DoY, MLON〉 and 〈MLAT, MLT〉, we reshape the 4-D matrix into a 2-D matrix (288 × 4368). Following the methodology introduced by Jolliffe [2002], we perform a Principal Components Analysis on the 2-D matrix, and construct the empirical model of NmF2 with four input parameters (DoY, MLON, MLAT, MLT) by summing the least principal components containing 99% variance of the total. Finally, following the similar procedures, we construct the empirical model of hmF2.

3. Results and Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results and Conclusions
  6. Acknowledgments
  7. References

[7] With the NmF2 model, we are able to calculate a series of the F2-layer maximum electron density (NmF2i) at a geomagnetic longitude series (MLONi = 15(i − 1), i = 1, 2…24), at any MLAT, LT and DoY. Fourier analysis could be performed for the NmF2i according to the equation:

  • equation image

where p = equation image, ω = e2πi/24, i = equation image. The calculated AW0 is the longitudinal average of NmF2 and AW4 is the wave number-4 component. We calculate the AW0 and AW4 on the mesh grid of MLAT = −30, −27.5… 30 versus LT = 0, 1… 24 at DoY 280, shown in Figure 2a. In Figure 2a, during 10—18 LT the ionization trough is created over the geomagnetic equator and the ionization anomaly crests around ±15°. The equatorial anomaly arises from the fountain effect: the E region dynamo electric field maps into the equatorial F region, causes E × B drift, and drives the plasma to great heights from where the plasma diffuses downward along field lines to subequatorial latitudes enhancing the subequatorial plasma density [Hanson and Moffett, 1966; Immel et al., 2006]. Another striking feature in Figure 2a is the interhemispheric asymmetry of AW4, e.g., the magnitude of AW4 phasor peaks at 20° MLAT, 14 LT in the Northern Hemisphere but at 15° MLAT, 18 LT in the Southern. Figure 3a exhibits the same plot as Figure 2a but for hmF2, in which two bulges locate respectively around noon due to daytime maximum upward drift [Fejer et al., 1991] and around the sunset due to pre-reversal enhancement [Fejer et al., 1989]. The interhemispheric asymmetry of AW4 in Figure 3a is more significant. Particularly, at 00 LT the AW4 phasor at the northern subequatorial latitudes is opposite in phase to that at the southern, marked by the green arrows.

image

Figure 2. (a) Longitudinal average of NmF2 as a function of MLAT and LT at DoY 280, calculated from the COSMIC model described in the “Data and Methods” section. Arrows exhibit the Fourier-filtered phasor (also called phase vector or complex amplitude of the longitudinal wave-4 component AW4. The length of the arrow denotes the AW4 magnitude and the direction denotes the AW4 phase. For example, the rightward arrow stands for the AW4 component peaks at the four longitudes of 0° + n × 90° (n = 0, 1, 2, 3), the upward for 22.5° + n × 90°, and the leftward for 45° + n × 90°. (b) Phasor (arrow) and its magnitude (color code) of the interhemispheric symmetric wave-4 component SYMW4. (d) The diurnal cumulative variation in phase of the arrow on each of the gray lines in Figure 2b. Each red line in Figure 2d corresponds to a gray line in Figure 2b. (c, e) The same plots as Figures 2b and 2d but for the antisymmetric component ASYMW4. Dashed boxes in Figure 2b indicate that the SYMW4 at the equator is in antiphase with that at subequatorial latitudes. The arrows in the oval in Figure 2c would be compared with the corresponding arrows in Figure 3.

Download figure to PowerPoint

image

Figure 3. The same plot as Figure 2 but for hmF2. The ASYMW4 arrows in the oval in Figure 3c are largely opposite in phase to the corresponding NmF2 arrows in Figure 2c.

Download figure to PowerPoint

[8] In order to evaluate the asymmetries, we decompose the AW4 phasor in Figure 3a into interhemispheric symmetric and antisymmetric components (SYMW4 and ASYMW4, shown in Figures 3b and 3c, respectively):

  • equation image

To quantify the phase variation with LT, we calculate the diurnal cumulative variation in the SYMW4 phase (cΦ, shown in Figure 3d) along each gray line in Figure 3b:

  • equation image

where

  • equation image

During 08—18 LT, the SYMW4 phase increases at the rate of ∼2π/day (Figure 3d), i.e., the rate of an eastward diurnal tide. Different from the SYMW4 component, the ASYMW4 phase increases at the rate of ∼4π/day (Figure 3e). Similarly, we decompose the NmF2 AW4 phasor in Figure 2a into SYMW4 and ASYMW4, shown in Figures 2b and 2c. Generally, the SYMW4 phase increases at the rate of ∼2π/day (Figure 2d) and the ASYMW4 at ∼4π/day (Figure 2e), similar to those of hmF2. Among all tide modes contributing potentially to the wave-4 structure, only the DE3 mode propagates eastward at 2π/day and the SE2 at 4π/day. Thus, it is the SE2 and DE3 modes that account mainly for the ASYMW4 components and the daytime SYMW4 components, respectively.

[9] A simple consideration for the ASYMW4 components is with a SE2 tide in transequatorial neutral wind. The transequatorial wind transports plasma along the geomagnetic field lines, upwards (downward) in the upwind (downwind) hemisphere. Thus, the transport raises the F2-layer peak at the upwind side but lowers it at the downwind side [Rishbeth, 2000], resulting in an interhemispheric asymmetry in hmF2. Contrary to the asymmetry in hmF2, in NmF2 an asymmetry would be produced with higher value at the downwind side [Rishbeth, 1977, 2000]. A fact supporting this supposition is that the hmF2 ASYMW4 is largely in antiphase with the NmF2 ASYMW4 at the subequatorial latitudes (arrows in the ovals in Figures 2c and 3c). Besides, in Figure 3c, the ASYMW4 in hmF2 is generally stronger during nighttime than daytime, potentially arising from the less ion drag during nighttime for lower plasma density. The presence of the SE2 transequatorial wind in the F region is supported by the Hough mode extension empirical model [Oberheide et al., 2011]. According to the modeling, long-wavelength SE2 tides excited in troposphere could penetrate into the F region, causing transequatorial SE2 winds with amplitude in excess of 10 m/s.

[10] The SYMW4 in NmF2 could be attributed to the DE3 modulation of the E region dynamo and the F region fountain [Immel et al., 2006; Sagawa et al., 2005]. A fact supporting this proposition is marked by the green arrows in Figure 2a. At 16 LT, the AW4 phasor over the northern crest is almost opposite in phase to that over the equator. At the longitude where the dynamo electric field is enhanced most by the DE3 tide, the upward drift would be enhanced most, and so does the drainage of plasma from the equatorial region to subequatorial latitudes, resulting in a zonal minimum plasma density over the equator accompanying with a zonal maximum at subequatorial latitudes. In Figure 2b, the antiphase between the SYMW4 over the equator and that at 15—25° MLAT persists just during daytime from 9 to 18 LT. After sunset, the plasma in the E region depletes quickly and the E region conductivity decreases considerably in the absence of solar irradiation, resulting in the disappearances of the E region dynamo and the associated wave-4 structures. Meanwhile, the daytime SYMW4 in hmF2 could also be attributed to the tidal modulation of the E region dynamo. The upward drift driven by the eastward electric field starts in the morning and reaches its diurnal maximum around the noon [Fejer et al., 1991; Scherliess and Fejer, 1999]. An enhanced upward drift would transport the plasma to higher altitude to diffuse further poleward. Accordingly, the SYMW4 magnitude in Figure 3b exhibits a U-shaped pattern during daytime and peaks around noon.

[11] Generally, the NmF2 AW4 phasor in Figure 2a is dominated by the SYMW4 component while the hmF2 phasor in Figure 3a is dominated by the ASYMW4 component from 18 to 10 LT, which suggests the DE3 modulation on the electrodynamic coupling is the leading contributor to the wave-4 structure in NmF2 but the transequatorial SE2 wind is the major contributor to the nighttime wave-4 structure in hmF2.

[12] The main behaviors reported above at DoY 280 could be found during most of the year, observed from plots same as Figures 2 and 3 but for all other months (not shown). For example, around midnight the AW4 phasor at northern subequatorial latitudes is opposite in phase to that at the southern in most months (Figure 4a), and the daytime antiphase between the SYMW4 over the equator and SYMW4 at the subequatorial latitudes persists from March to November (Figure 4b).

image

Figure 4. (a) Longitudinally average hmF2 (color code) at 00 LT as a function of month and MLAT, with the AW4 phasor (arrow) in hmF2. The interhemispheric asymmetry in the AW4 is significant. (b) NmF2 SYMW4 phasor at the equator (blank arrow) and 20° latitude (green arrow) as a function of month and LT, with the angle (color code) between the green and blank arrows.

Download figure to PowerPoint

[13] In summary, considerable interhemispheric asymmetries are observed in wave-4 structures of NmF2 and hmF2. In particular, around midnight nearly antiphase wave-4 component in hmF2 is observed at the equatorial flanks. The wave-4 structures in NmF2 and hmF2 are decomposed into interhemispheric symmetric and antisymmetric components (SYMW4 and ASYMW4). The ASYMW4 components in both NmF2 and hmF2 shift eastward at the rate of 4π/day while the SYMW4 components at the rate of 2π/day during daytime. The daytime SYMW4 components are attributed to the DE3 modulation of the E region dynamo, while the ASYMW4 components are attributed to a SE2 tide in transequatorial wind. Evidence supporting the above hypotheses includes: 1) The NmF2 SYMW4 at the equator is in antiphase to that at the 15—25° during daytime, indicating the draining and diffusing effects of the fountain. 2) During daytime, the hmF2 SYMW4 enhances within a U-shape pattern, which could be ascribed to the daytime variation of eastward electric field and the associated upward drift. 3) At the subequatorial latitudes, the hmF2 ASYMW4 is largely opposite in phase to the NmF2 ASYMW4, which could be explained in the terms of SE2 tide in transequatorial wind.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results and Conclusions
  6. Acknowledgments
  7. References

[14] The authors thank Kazuo Shiokawa for his kindly help. This research is supported by National Natural Science Foundation of China (40725014, 41074112) and the China Meteorological Administration Grant (No. GYHY201106011). Y. Wei is supported by National Natural Science Foundation of China (41004072).

[15] The Editor thanks two anonymous reviewers for their assistance in evaluating this paper.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results and Conclusions
  6. Acknowledgments
  7. References