Annual and semiannual variations of the midlatitude ionosphere under low solar activity

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. A8, 1166, /2001JA000267, 2002 Annual and semiannual variations of the midlatitude ionosphere under low solar activity S. Kawamura and N. Balan 1,2,3 Radio Science Center for Space and Atmosphere, Kyoto University, Kyoto, Japan Y. Otsuka Solar-Terrestrial Environment Laboratory, Nagoya University, Nagoya, Japan S. Fukao Radio Science Center for Space and Atmosphere, Kyoto University, Kyoto, Japan Received 15 August 2001; revised 15 October 2001; accepted 16 October 2001; published 6 August [1] The annual and semiannual variations of the midlatitude ionosphere under low solar activity are studied using middle and upper (MU) radar (135 E, 35 N) incoherent scatter observations and Sheffield University plasmasphere-ionosphere model (SUPIM). The variations of the daytime electron density (Ne) and electron and ion temperatures (Te and Ti) at km altitudes measured by the radar under low solar activity (F ) are satisfactorily reproduced for the first time by incorporating the radar measured values of the magnetic meridional neutral wind velocity (U q ) and northward perpendicular plasma drift velocity (V? ) into SUPIM that uses mass spectrometer incoherent scatter 1986 (MSIS-86) for neutral densities and neutral temperatures. The study shows that the annual and semiannual variations of the midlatitude ionosphere during daytime at altitudes near and above the ionospheric peak under low solar activity depend more on the direct effect of the neutral wind arising through the changes in ionospheric height than on the indirect effect arising through the changes in thermospheric composition (or atomic to molecular concentration ratio); the indirect effect, however, predominates on the variations at altitudes below the ionospheric peak. The electron temperature (Te) undergoes similar but almost opposite seasonal and semiannual variations as the electron density. The ion temperature (Ti) is closer to neutral temperature than to electron temperature at altitudes up to 400 km and shows comparatively weak annual and semiannual variations. INDEX TERMS: 2437 Ionosphere: Ionospheric dynamics; 2443 Ionosphere: Midlatitude ionosphere; 2467 Ionosphere: Plasma temperature and density; KEYWORDS: annual variation, semiannual variation, midlatitude, ionosphere, thermospheric wind, MU radar 1. Introduction [2] Earth s ionosphere, studied conventionally using peak electron density (N max ) and total electron content (TEC), is known to undergo annual and semiannual variations [Yonezawa, 1959; Torr and Torr, 1973; Evans, 1977]. The annual variation, with ionization maximum in winter and minimum in summer, has been known as seasonal anomaly as the winter/summer ionization is contrary to what is expected from the solar zenith angle variation [Croom et al., 1960]. The anomaly has been interpreted in terms of the winter/ 1 Department of Earth and Planetary Sciences, Hokkaido University, Sapporo, Japan. 2 Department of Applied Mathematics, University of Sheffield, Sheffield, England. 3 On leave from Department of Physics, University of Kerala, Trivandrum, India. Copyright 2002 by the American Geophysical Union /02/2001JA summer thermospheric composition change, including vibrationally excited nitrogen [Rishbeth and Setty, 1961; Richards and Torr, 1986; Fuller-Rowell et al., 1988]. The semiannual variation has ionization maximum at equinoxes and minimum in solstices; the variation also exhibits an equinoctial asymmetry, with the ionization in March equinox being stronger than that in September equinox [Titheridge, 1973; Essex, 1977; Titheridge and Buonsanto, 1983]. The asymmetry has also been reported to change with the 11-year solar cycle [Feichter and Leitinger, 1997]. [3] Recently, Millward et al. [1996], through model calculations using their coupled ionosphere-thermosphere model, have studied the annual and semiannual variations of N max observed at midlatitude locations in the Northern and Southern Hemispheres under high solar activity. Using the coupled model, Zou et al. [2000] and Rishbeth et al. [2000] have reasonably reproduced and interpreted the annual and semiannual variations of N max observed at several midlatitude locations under low and high solar activity conditions. Richards [2001] has reproduced the yearly and solar activity SIA 8-1

2 SIA 8-2 KAWAMURA ET AL.: ANNUAL VARIATIONS OF THE IONOSPHERE variations of N max observed at several midlatitude locations using the field line interhemispheric plasma (FLIP) model. Bailey et al. [2000] have studied the yearly variations of the electron density at 600-km altitude at low latitudes using Hinotori satellite observations and coupled thermosphere ionosphere plasmasphere model (CTIP). These studies have interpreted the annual and semiannual variations of the midlatitude ionosphere in terms of the changes in thermospheric composition [Mayr et al., 1978], which introduces the indirect effect of neutral wind on the ionosphere. The direct effect of neutral wind, which arises through the changes in ionospheric height, has not received much attention though the data include both these effects. That has mainly been due to the lack of understanding of the altitude dependence of the ionospheric variations. [4] The altitude dependence of the yearly variations of the ionosphere has been studied recently for the first time by considering the electron density (Ne), electron and ion temperatures (Te and Ti), and field-parallel and field-perpendicular plasma drift velocities (V k and V? ) measured by the middle and upper atmosphere (MU) radar at Shigaraki (35 N, 135 E; magnetic latitude 25 N) in Japan [Balan et al., 1997]. A strong equinoctial asymmetry in the ionosphere observed under high solar activity (F ) has also been explained in terms of the asymmetrical neutral wind velocities measured in the two equinoxes [Balan et al., 1998, 1999]. In the present paper, we study the relative importance of the direct and indirect effects of neutral wind leading to the annual and semiannual variations and the equinoctial asymmetry of the midlatitude ionosphere under low solar activity using MU radar observations [Sato et al., 1989; Fukao et al., 1990] and Sheffield University plasmasphere-ionosphere model (SUPIM) [Bailey and Balan, 1996]. The yearly variations of daytime Ne, Te, and Ti observed at km altitudes under low solar activity (F ) are satisfactorily reproduced for the first time by incorporating the MU radar measured magnetic meridional neutral wind velocity (U q ) and northward perpendicular plasma drift velocity into SUPIM that uses MSIS-86 for neutral densities and neutral temperatures. The observations under low solar activity are also compared with those under high solar activity [Balan et al., 1998]. As mentioned above, the neutral wind introduces its indirect effect on the ionosphere by altering the neutral composition (or atomic to molecular concentration ratio), which, in turn, changes the photochemistry of the ionosphere; the wind introduces its direct effect by dynamically shifting the ionospheric peak to altitudes of lower or higher equilibrium density. 2. MU Radar and Data [5] The capabilities of the MU radar for incoherent scatter observations have been described and compared with those of other incoherent scatter radars by Sato et al. [1989] and Fukao et al. [1990]. The electron density (Ne) profiles, electron and ion temperatures (Te and Ti), and plasma drift velocities (V k and V? ) used in the present paper have been measured using the experimental techniques described in the above papers. The meridional component of the average thermospheric neutral wind velocity (U q, averaged over km altitudes) has been derived from the field-parallel plasma drift velocity (V k ) following the procedure described by Oliver et al. [1998]. The electron density Ne has range resolution of 9.6 km and time resolution of 60 min. The corresponding values for Te and Ti are 14.4 km and 60 min and for V k are 38.4 km and 60 min. The expected errors in Ne are less than 10% and in Te and Ti are 30 and 70 K, respectively, near the F region peak; the corresponding errors in plasma velocity are 20 m s 1. These errors, expected in each pointing position of the antenna beam, could reduce by a factor of 2 when all the four beam positions are combined as is done for scalar quantities [Sato et al., 1989]. The data reduction and analysis procedure followed for the present paper are same as those described before [Balan et al., 1998]. [6] The MU radar data, started compiling since 1986, have been used to study various aspects of the ionosphere and thermosphere. The studies that are relevant to the present paper are those by Fukao et al. [1991], Oliver et al. [1991], and Su et al. [1997], who have reported on ionospheric and thermospheric temperatures and densities; Reddy et al. [1990] and Oliver et al. [1993], who have reported on F region electrodynamics; and Oliver et al. [1990, 1998], who have reported on thermospheric neutral winds. The studies reviewed by Rishbeth and Fukao [1995] have used one or more of the parameters measured for periods up to As mentioned above, recently, Balan et al. [1997, 1998, 1999] have used all the parameters measured by the radar during to study a strong equinoctial asymmetry observed under high solar activity (F ). Otsuka et al. [1998] have used the Te and Ti data during to study the altitude, seasonal, and solar activity variations of the temperatures. Kawamura et al. [1998] have used the MU radar measurements to derive H + concentration. Kawamura et al. [2000] have also studied the diurnal variations of the meridional neutral wind velocity over the radar in different seasons under low and high solar activity conditions. [7] The present paper uses all the data measured by the radar under low solar activity (F ) during However, the data with magnetic activity index Ap 25 are not considered. The data are grouped into different months. The data statistics are provided along with the figures. 3. Observations [8] The annual variations of Ne, Te, and Ti are obtained at different altitudes up to 600 km separately for daytime ( LT), nighttime ( LT), morning ( LT), and evening ( LT) periods. The selected morning and evening periods, in general, correspond to the times when the electron density builds up and decays, respectively, and the density remains more or less constant during the periods in between. Figures 1a and 1b show the variations of daytime and nighttime Ne and the corresponding peak electron density (N max ). The number of days of data and the corresponding mean solar activity index (F 10.7 ) are also shown. As Figures 1a and 1b show, the yearly variation of the ionosphere depends on altitude and local time. During daytime (Figure 1a) the variation has annual and semiannual components, with the annual component (with seasonal anomaly) being predominant at

3 KAWAMURA ET AL.: ANNUAL VARIATIONS OF THE IONOSPHERE SIA 8-3 for detailed description). As seen from a comparison of Figures 1a (350 km curve) and 2a, the variation of Te is out of phase with that of Ne and has maxima in winter and summer and minima at equinoxes, which agrees with the expected anticorrelation between daytime Ne and Te. The March equinox minimum is less than the September equinox minimum by 200 K, which is caused by the equinoctial asymmetry in Ne. The variation of Ti at 350 km (Figure 2a) is comparatively weak with a small maximum at March equinox. This maximum, which corresponds to the highest value of the product of Ne and Te, shows the existence of the equinoctial asymmetry in ion temperature. Previous studies of the electron and ion temperatures [Schunk and Nagy, 1978; Brace and Theis, 1981; Oyama and Schlegel, 1988; Farley, 1991; Watanabe and Oyama, 1996] have shown features similar to those obtained in the present study. However, these studies could not reveal the equinoctial asymmetry as the data for the two equinoxes were combined. [10] Figure 2b shows a comparison of the mean daytime electron (Te), ion (Ti) and neutral (Tn) temperature profiles at March and September equinoxes; Tn is for the same conditions as Te and Ti and is obtained from MSIS-86. The horizontal bars represent standard deviations divided by square root of the number of data points. As expected from the equinoctial asymmetry, Te is much lower and Ti is slightly higher at March equinox compared to September equinox at altitudes above 250 km (Figure 2b). However, Figure 1a. Yearly variation of mean daytime ( LT) electron density (Ne) at different altitudes between 200 and 600 km under low solar activity (F , mean F 10.7 = 86.3, mean Ap = 9.5). The variation of the peak electron density (N max ) is also shown. The histograms show the data distribution in (bottom) number of days and (top) mean F altitudes near and below the ionospheric peak and semiannual component being predominant at altitudes near and above the ionospheric peak. The equinoctial maximum has slightly higher values of Ne in September equinox compared to March equinox in the bottomside ionosphere (200 km). At higher altitudes this asymmetry reverses and becomes strong with the values of Ne in March equinox far exceeding those in September equinox, by over 100%. At night (Figure 1b) the seasonal anomaly and semiannual variation disappear. However, the equinoctial asymmetry exists near and above the ionospheric peak. The variations during morning and evening hours (not shown) are found to fall between those during daytime and nighttime. The observed variations (Figures 1a and 1b) are almost independent of solar activity because the mean value of F 10.7 are nearly equal in all months, especially in the equinoctial months of March April and September October. However, the values are high in February and low in July by about 10 units. That might have minor effects on the electron density in these months, especially at altitudes near and below the ionospheric peak. [9] Figure 2a shows the yearly variations of daytime Te and Ti at a typical altitude of 350 km (see Figure 2 caption Figure 1b. Same as Figure 1a but for nighttime.

4 SIA 8-4 KAWAMURA ET AL.: ANNUAL VARIATIONS OF THE IONOSPHERE up to 400 K from low to high solar activity, Ti increases by up to 250 K. This behavior of the ionospheric temperatures is consistent with the solar activity variation of the electronion heating process [Schunk and Nagy, 1978]. However, Te is slightly higher at high solar activity than at low solar activity during April-July (Figure 2a). [12] The important points that came out from the daytime Ne data are the following: (1) The well-known annual variation (with seasonal anomaly) is predominant at altitudes near and below the ionospheric peak, (2) the semiannual variation is predominant at altitudes near and above the ionospheric peak, and (3) the equinoctial asymmetry changes its character from a weak asymmetry in the bottomside to a strong and reverse asymmetry in the topside. These observations indicate the relative importance of the physical processes involved. [13] Figure 3 shows the yearly variations of the solar zenith angle (SZA) and thermospheric composition ([O]/[N 2 ]) that control the photochemical processes at local noon over the MU radar. The variation of [O]/[N 2 ] obtained from MSIS-86 is shown for pressure level 12, the altitude (Z) of which is also shown in Figure 3. As Figure 3 shows, the variation of [O]/[N 2 ] has a strong annual component, a Figure 2a. Yearly variation of mean daytime ( LT) electron and ion temperatures (Te and Ti) at 350-km altitude under low solar activity (F , mean F 10.7 = 86.0, mean Ap = 9.2). The hourly mean values, averaged over 100-km altitudes, are shown by different symbols; the big circles represent the daily mean values. The mean variation under low solar activity, shown by the solid running mean curve, is also compared with the variation under high solar activity (dashed curve, from Balan et al. [1998]). at high altitudes above 450 km, where the accuracy of measurements is poor, the temperatures show small fluctuations. Figure 2b also shows that Ti is closer to Tn than to Te at altitudes up to 400 km, indicating that Ti at these altitudes is controlled mainly by ion-neutral collisions. As higher altitudes, Ti gradually moves away from Tn showing the importance of electron-ion heating on Tn. [11] The behavior of the ionosphere under low solar activity (Figures 1 and 2) is found to be similar to that under high solar activity [Balan et al., 1998]. However, the ionospheric density decreases by nearly 3 times from high to low solar activity. The daytime seasonal anomaly, which is predominant in the bottomside ionosphere under all levels of solar activity, extends to 250 km altitude under low solar activity (Figure 1a) and 400-km altitude under high solar activity [Balan et al., 1998]. The semiannual variation and equinoctial asymmetry become predominant at altitudes near and above the ionospheric peak under all levels of solar activity. A comparison of the mean values of Te and Ti under low and high solar activities (solid and dashed curves, respectively, in Figure 2a) shows that while Te decreases by Figure 2b. Mean altitude profiles of daytime ( LT) electron (Te), ion (Ti), and neutral (Tn) temperatures at March equinox (March April data) and September equinox (September October data). The horizontal bars are standard deviations divided by square root of the number of data points.

5 KAWAMURA ET AL.: ANNUAL VARIATIONS OF THE IONOSPHERE SIA 8-5 (from photoionization to dissociative recombination) is short. The direct effect of the wind arises through the changes in ionospheric height. This effect is important at high altitudes where the lifetime of the ionization is long, and so the wind gets time to redistribute the ionization before it is being lost through chemical recombination. The wind changes the ionospheric height by pushing and pulling the ionization up and down along the geomagnetic field lines with vertical velocity U q sin I cos I, where I is magnetic dip angle. [15] Figure 4a shows the average magnetic meridional neutral wind velocity patterns (U q ) derived from V k for four periods, December January, March April, July August, and September October under low solar activity. The corresponding northward perpendicular E B plasma drift velocity (V? ) patterns are shown in Figure 4b. The daytime wind velocities in summer and winter (patters 3 and 1, Figure 4a) are consistent with the expected summer to winter thermospheric circulation. However, contrary to expectation, the wind velocities at equinoxes (patterns 2 and 4) are highly unequal (or asymmetric), see also Igi et al. [1999]. The plasma drift velocities (Figure 4b) are also highly asymmetric at the equinoxes. The asymmetries of the neutral wind and plasma drift under low solar activity (Figures 4a and 4b) are found to be stronger than those under high solar activity [Balan et al., 1998; see also Fejer, 1993]. The direct and indirect effects of the neutral wind on the annual and semiannual variations of the ionosphere are studied in the next section. Figure 3. Yearly variations of the solar zenith angle (SZA) and neutral composition ([O]/[N 2 ]) from MSIS-86 at pressure level 12 at local noon at the MU radar location under magnetically quiet (Ap = 4) low solar activity (F 10.7 = F 10.7A = 86) condition. The altitude (Z) of the pressure level is shown. The variation of the peak electron density (N max ) and peak height (h max ) modelled under these conditions with no neutral wind and no E B drift are also shown. 4. Model Results and Discussion [16] As mentioned above, the model values of Ne, Te, and Ti are calculated using SUPIM [Bailey and Balan, weak semiannual component, and a weak equinoctial asymmetry. The annual component has been known to arise from the summer to winter thermospheric circulation or neutral wind [Mayr et al., 1978]. The semiannual component has been qualitatively reproduced recently by Fuller-Rowell [1998], who visualized it as caused by a thermospheric spoon effect. The spoon, which corresponds to the turbulent part of the circulation, mixes the thermosphere more thoroughly in solstices and less so at equinoxes. That could result in the semiannual component of [O]/[N 2 ]. The mixing effect on the composition is in addition to the conventional circulation effect. The variation of the thermospheric composition (or indirect effect of neutral wind), for example, Figure 3, used in the present study includes both the conventional circulation effect and mixing effect. [14] As mentioned before, the indirect effect of neutral wind arises through the changes in thermospheric composition (or atomic to molecular concentration ratio), which alters the photochemistry in the ionosphere. This effect is important at low altitudes where the lifetime of ionization Figure 4a. Local time variation of the mean magnetic meridional neutral wind velocity (southward positive) at F region altitudes ( km) obtained from MU radar measurements in the (1) December January, (2) March April, (3) July August, and (4) September October periods under low solar activity (F ). The mean F 10.7 and mean Ap for each case are noted. The histograms show the data distribution in number of hours.

6 SIA 8-6 KAWAMURA ET AL.: ANNUAL VARIATIONS OF THE IONOSPHERE Figure 4b. Same as Figure 4a but for the perpendicular E B plasma drift velocity (northward positive) at 300-km altitude. 1996]. In SUPIM, coupled time-dependent equations of continuity, momentum, and energy for the O +,H +,He +, N 2 +,O 2 +, and NO + ions, and the electrons, are solved along closed dipole magnetic field lines between base altitudes of 130 km in conjugate hemispheres. The model gives values for the concentrations, field-aligned fluxes, and temperatures of the ions and electrons at a discrete set of points along the field lines. For the present study, the model equations are solved along 96 eccentric-dipole magnetic field lines in the longitude (135 E) of the MU radar. The field lines are distributed with apex altitude between 150 and 3000 km and the number of points along the field lines increase from 201 forthe lowest field line to 401 for the highest field line. [17] As mentioned above, neutral densities and neutral temperatures are obtained from MSIS-86 [Hedin, 1987]. The solar EUV fluxes ( A) that produce ionization are from the EUV94 solar EUV flux model [Tobiska, 1991]. The 10.7-cm solar flux index (F 10.7 ), which has conventionally been used as an index of solar activity, is nonlinearly related to the EUV flux though the relationship is linear for low values of F 10.7, below 150 [Balan et al., 1993]. The magnetic meridional neutral winds and northward perpendicular E B plasma drifts are from MU radar measurements (Figures 4a and 4b). To study the yearly variations of the ionosphere, the values of Ne, Te, and Ti are calculated at 15-day intervals in However, the solar activity index (F 10.7 ) and its 81-day running mean (F 10.7A ) that determine the neutral atmospheric conditions are set equal to 86 for all model days, which corresponds to the average solar activity condition of the observed data. The EUV fluxes are also set equal to those for F 10.7 = F 10.7A = 86 and the magnetic activity index Ap is set equal to 4. The neutral wind and E B drift patterns 1, 2, 3, and 4 shown in Figures 4a and 4b are used for the model days 1 (and 365), 90, 195, and 285, respectively, with interpolations for the model days in between.these wind patterns are used for Northern Hemisphere, and these patterns shifted by 6 months are used for the Southern Hemisphere. However, the same E B drift patterns described above are used for both hemispheres to meet the requirement of electrical equipotentiality between conjugate points along the field lines. [18] Using the above input conditions, the model equations are solved at 15-min time intervals for 3 iteration days. The values of the parameters Ne, Te, and Ti obtained along the field lines during the last iteration day are stored and regrided at 25-km intervals in altitude, 2 in latitude, and 15 min in local time. That gives a reasonable 24-hour distribution of model data between 150- and 650-km altitude over the MU radar. The modelling procedure followed is same as that described elsewhere [Balan and Bailey, 1995]. [19] First, to study the effect of thermospheric composition (or the indirect effect of neutral wind) on the yearly variations of the ionosphere, the model data for the whole year (1993) are obtained without neutral wind and E B drift. The variations of the noontime N max and h max obtained from these calculations are shown in Figure 3, where N max closely follows the variation of [O]/[N 2 ]. It exhibits a strong annual component with strong seasonal anomaly and a comparatively weak semiannual component with small equinoctial asymmetry. The variations of the corresponding Ne at different altitudes are shown in Figure 5, which illustrates that the variation of the thermospheric composition (or the indirect effect of the neutral wind) produces a strong annual variation in Ne (with seasonal anomaly) at altitudes near and below the ionospheric peak. However, the semiannual variation in Ne becomes dominant at altitudes above the ionospheric peak. A small equinoctial asymmetry also exists in Ne at all altitudes, consistent with the asymmetry in [O]/[N 2 ]. Figure 5. Yearly variation of the noontime electron density (Ne) at different altitudes modeled with no neutral wind and no E B drift under magnetically quiet (Ap =4) low solar activity (F 10.7 = F 10.7A = 86) condition.

7 KAWAMURA ET AL.: ANNUAL VARIATIONS OF THE IONOSPHERE SIA 8-7 Figure 6. Yearly variation of the noontime peak electron density (N max ) and peak height (h max ) modeled with no neutral wind and no E B drift (dashed curves), with neutral wind but no E B drift (solid curves) and with both neutral wind and E B drift (asterisks) under magnetically quiet (Ap = 4) low solar activity (F 10.7 = F 10.7A = 86) condition. equinoctial asymmetry in N max reverses and becomes strong due to the asymmetrical neutral winds (compare the wind velocity patterns 2 and 4 in Figure 4a and the corresponding N max and h max in Figure 6). The effect of the E B drift is small in N max though it causes 10 km change in h max (compare the asterisks and solid curve in Figure 6). That may be because at midlatitudes the drift transfers ionization from lower to higher field lines (perpendicular to the field lines) and bodily lifts the ionosphere by small heights with very small change in N max. Model electron density profiles illustrating this process have been shown elsewhere [Balan et al., 1998]. [22] The variations of the noontime electron density (Ne) modeled with neutral wind and no E B drift are shown in Figure 7. The corresponding electron and ion temperature (Te and Ti) variations are shown in Figure 8 for an altitude of 350 km. As shown, the model reasonably reproduces the general features of the observed variations. As in the observations (Figure 1a), the model values of Ne (Figure 7) shows an annual component with seasonal anomaly at altitudes near and below the ionospheric peak. The semiannual component and equinoctial asymmetry become predominant at altitudes near and above the ionospheric peak. The features observed in the variations of the measured values of Te and Ti (Figure 2) are also reasonably reproduced by the model (Figure 8). The model results (Figures 5 8) have therefore shown that the direct effect of the neutral wind arising through the changes in ionospheric height is the major controlling factor on the annual and semiannual variations of the daytime midlatitude ionosphere, including equinoctial asymmetry, at altitudes near and above the [20] Figure 6 compares the variations of noontime N max and h max calculated without neutral wind and without E B drift (dashed curves), with neutral wind but no E B drift (solid curves) and with both neutral wind and E B drift. Since E B drift has only a small effect (compared to neutral wind effect) on the ionospheric density at midlatitudes [Balan et al., 1998] and since this effect is not important for the present study, the calculations with the drift (Figure 4b) are performed only for 5 model days, and the results are shown by asterisks in Figure 6. Small fluctuations present in the model data obtained with neutral wind have been removed by applying a two-point running mean (Figure 6 and other figures); the small fluctuations arise because the neutral wind velocity used for the model days other than 1, 90, 195, 285, and 365 are obtained by interpolating the neutral wind velocity patterns shown in Figure 4a. [21] A comparison of the dashed and solid curves in Figure 6 illustrates the direct effect of neutral wind. As shown, the wind reduces the annual component, enhances the semiannual component, and reverses and enhances the equinoctial asymmetry in N max. The annual component in N max reduces because the daytime poleward wind pushes the ionosphere to much lower altitudes of heavy chemical loss in winter than in summer (compare the wind velocity patterns 1 and 3 in Figure 4a and N max and h max in Figure 6). The reduction of the annual component in N max, as seen in Figure 6, also enhances its semiannual component. The Figure 7. Yearly variation of the noontime electron density (Ne) at different altitudes modeled with neutral wind under magnetically quiet (Ap = 4) low solar activity (F 10.7 = F 10.7A = 86) condition.

8 SIA 8-8 KAWAMURA ET AL.: ANNUAL VARIATIONS OF THE IONOSPHERE Figure 8. Yearly variation of the noontime electron and ion temperatures (Te and Ti) at 350-km altitude modeled with neutral wind under magnetically quiet (Ap = 4)low solar activity (F 10.7 = F 10.7A = 86) condition. ionospheric peak. There are some quantitative differences between the measured and modelled data, which are to be understood from the facts that the measured data give monthly mean values under F while the model data are obtained at 15-day intervals for the average condition (F 10.7 = 86) of the measurements. Also, the measured data (Figures 1a and 2) correspond to the average values during daytime ( LT), while the model data (Figures 7 and 8) are for noontime. The small effects of E B drift are also not included in the model data in Figures 7 and 8. [23] As mentioned in the introduction, Richards [2001] has reproduced the annual and semiannual variations of N max, including equinoctial asymmetry, during low-high solar activity periods using the FLIP model. However, he has interpreted the variations in terms of the changes in thermospheric composition. Using the coupled ionospherethermosphere model, Zou et al. [2000] have reasonably reproduced the annual and semiannual variations of N max observed at several midlatitude locations under low and high solar activity conditions. In a companion paper, Rishbeth et al. [2000] have interpreted that the observed variations of noontime N max are closely related to the ambient atomic/molecular concentration ratio and suggested that the variations of N max with geographic and magnetic longitudes are largely due to the geometry of the auroral ovals. [24] The annual and semiannual variations of the ionosphere studied using N max and TEC were reported earlier by Titheridge [1973], Essex [1977], and Titheridge and Buonsanto [1983]. They found large TEC and N max in March equinox compared to September equinox at all stations that include Northern and Southern Hemispheres. They also found stronger asymmetry in TEC than in N max, which agrees with the asymmetry in Ne (Figures 1a and 1b). However, they could not fully explain the asymmetry as the direct dynamical effects of the neutral winds were neglected owing to the lack of understanding of the altitude dependence of the asymmetry [Titheridge and Buonsanto, 1983]. [25] Recently, Feichter and Leitinger [1997] have brought out an interesting aspect of the equinoctial asymmetry in N max and TEC. They showed that the asymmetry changes from one 11-year solar cycle to the next 11-year solar cycle. They have attributed these changes, which indicate the presence of a 22-year cycle in the ionosphere, to similar changes in the thermospheric composition (or [O]/[N 2 ] ratio) that they expect could arise from the well-known 22-year cycle in geomagnetic activity. However, the asymmetry in the thermospheric composition shown in the present study and given by MSIS-86 (Figure 3) is contrary to the asymmetry in N max obtained when neutral wind is included in the model calculations (Figure 6). In this context, it will be interesting to study the annual variation of the thermospheric composition and ionospheric density by considering data for two consecutive 11-year solar cycles. [26] The present study has shown that the direct effect of the neutral wind is a major controlling factor on the variations of the midlatitude ionosphere. The direct effect becomes more important than the indirect at locations of magnetic dip angle I between 30 and 60 where the factor sin I cosi has maximum value (0.43 to 0.50), especially under low solar activity when dynamical processes become most effective in the distribution of ionization. A study similar to the present one, but for high solar activity, is being carried out. The wind velocity (Figure 4a) is also found to be largely unequal in the two equinoxes, which accounts for the strong equinoctial asymmetry (Figures 1 and 2) in the ionosphere and a weak equinoctial asymmetry in the thermospheric composition (Figure 3) [Balan et al., 1998; see also Mayr et al., 1978]. However, the asymmetry in the wind velocity remains puzzling. Equinoctial asymmetry in the wind velocity has also been reported from high-latitude locations (Kiruna, 20 E), which has been explained in terms of the possible asymmetry in high-latitude forcing [Aruliah et al., 1996a, 1996b]. However, the high-latitude forcing cannot explain the equinoctial asymmetry at other locations because the forcing depends on latitude and longitude while the asymmetry observed in the ionosphere at midlatitudes is similar at all locations [Titheridge, 1973; Essex, 1977; Titheridge and Buonsanto, 1983]. A possible explanation for the equinoctial asymmetry in the thermospheric wind velocity could be the possible existence of a similar asymmetry in the momentum/energy being transmitted from lower to upper atmosphere through tides and waves. Such possibilities are being studied. Another possibility is ion drag. Higher ion density (equal to electron density) can more effectively slow down the neutrals through ion drag. 5. Summary [27] The annual and semiannual variations of the midlatitude ionosphere under low solar activity are studied using MU radar incoherent scatter observations and Sheffield University plasmasphere-ionosphere model (SUPIM). The variations of the daytime electron density (Ne) and electron and ion temperatures (Te and Ti) at km altitudes measured by the MU radar under low solar activity (F ) are satisfactorily reproduced for the first time by incorporating the radar measured values of the meri-

9 KAWAMURA ET AL.: ANNUAL VARIATIONS OF THE IONOSPHERE SIA 8-9 dional neutral wind velocity and northward perpendicular plasma drift velocity into SUPIM that uses MSIS-86 for neutral densities and neutral temperatures. [28] The study shows that the annual and semiannual variations of the midlatitude ionosphere during daytime at altitudes near and above the ionospheric peak depend more on the direct effect of the neutral wind arising through the changes in ionospheric height than on the indirect effect of the wind arising through the changes in thermospheric composition (or atomic to molecular concentration ratio). The indirect effect, however, predominates on the variations at altitudes below the ionospheric peak. The electron temperature (Te) undergoes similar but almost opposite seasonal and semiannual variations as the electron density. The ion temperature (Ti) is found to be closer to neutral temperature than to electron temperature at altitudes up to 400 km and shows comparatively weak annual and semiannual variations. The neutral wind velocities in the two equinoxes are found to be highly unequal or asymmetric, the direct effect of which, which is opposite to the indirect effect, causes a strong equinoctial asymmetry in the ionosphere with the electron density near and above the ionospheric peak being much stronger in March equinox compared to September equinox, by over 100%. The asymmetry in the wind velocity could also account for a weak asymmetry observed in the thermospheric composition. It is suggested that the equinoctial asymmetry in the neutral wind velocity could be due to the possible differences in the momentum/energy being transmitted from lower to upper atmosphere through tides and waves. The direct effect of the neutral wind studied for the ionosphere over the MU radar is valid in general for midlatitudes of magnetic dip angle I between 30 and 60, especially under low solar activity when dynamical processes become most effective in the distribution of ionization. [29] Acknowledgments. S. Kawamura was supported by a grant of the Japan Society for the Promotion of Science (JSPS) under the Fellowships for Japanese Junior Scientists. N. Balan thanks the Radio Science Center for Space and Atmosphere (RASC) of Kyoto University for awarding a COE visiting professor fellowship. The MU radar belongs to and is operated by RASC. [30] Janet G. Luhmann thanks Henry Rishbeth and another referee for their assistance in evaluating this paper. References Aruliah, A. L., A. D. Farmer, D. Rees, and U. Brandstrom, The seasonal behavior of high-latitude thermospheric winds and ion velocities observed over one solar cycle, J. Geophys. Res., 101, 15,701, 1996a. Aruliah, A. L., A. D. Farmer, T. J. Fuller-Rowell, M. N. Wild, M. Hapgood, and D. Rees, An equinoctial asymmetry in the high-latitude thermosphere and ionosphere, J. Geophys. Res., 101, 15,713, 1996b. Bailey, G. J., and N. Balan, A low-latitude ionosphere-plasmasphere model, in STEP Handbook, edited by R. W. Schunk, p. 173, Utah State Univ., Logan, Bailey, G. J., Y. Z. Su, and K.-I. Oyama, Yearly variations in the lowlatitude topside ionosphere, Ann. Geophys., 18, 789, Balan, N., and G. J. Bailey, Equatorial plasma fountain and its effects: Possibility of an additional layer, J. Geophys. Res., 100, 21,421, Balan, N., G. J. Bailey, and B. Jayachandran, Ionospheric evidence for a non-linear relationship between the solar EUV and 10.7 cm fluxes during an intense solar cycle, Planet. Space Sci., 41, 141, Balan, N., Y. Otsuka, and S. Fukao, New aspects in the annual variations of the ionosphere observed by the MU radar, Geophys. Res. Let., 24, 2287, Balan, N., Y. Otsuka, G. J. Bailey, and S. Fukao, Equinoctial asymmetries in the ionosphere and thermosphere observed by the MU radar, J. Geophys. Res., 103, 9481, Balan, N., Y. Otsuka, S. Fukao, M. A. Abdu, and G. J. Bailey, Annual variations of the ionosphere: A review based on MU radar observations, Adv. Space Res., 25, 153, Brace, L. H., and R. F. Theis, Global empirical models of ionospheric electron temperature in the upper F region and plasmashere based on in situ measurements from the Atmosphere Explorer-C, ISIS-1 and ISIS-2 satellites, J. Atmos. Terr. Phys., 43, 1317, Croom, S. A., A. R. Robbins, and J. O. Thomas, Variation of electron density in the ionosphere with magnetic dip, Nature, 185, 902, Essex, E. A., Equinoctial variations in the total electron content of the ionosphere at northern and southern hemisphere stations, J. Atmos. Terr. Phys., 39, 645, Evans, J. V., Satellite beacon contributions to studies of the structure of the ionosphere, Rev. Geophys., 15, 325, Farley, D. T., Early incoherent scatter observations at Jicamarca, J. Atmos. Terr. Phys., 53, 665, Feichter, E., and R. Leitinger, A 22-year cycle in the F layer ionization of the ionosphere, Ann. Geophys., 15, 1015, Fejer, B. G., F region plasma drifts over Arecibo: Solar cycle, seasonal, and magnetic activity effects, J. Geophys. Res., 98, 13,645, Fukao, S., T. Sato, T. Tsuda, M. Yamamoto, M. D. Yamanaka, and S. Kato, MU radar: New capabilities and system calibrations, Radio Sci., 25, 477, Fukao, S., W. L. Oliver, Y. Onishi, T. Takami, T. Sato, T. Tsuda, M. Yamamoto, and S. Kato, F-region seasonal behavior as measured by the MU radar, J. Atmos. Terr. Phys., 53, 599, Fuller-Rowell, T. J., The thermospheric spoon : A mechanism for the semiannual density variation, J. Geophys. Res., 103, 3951, Fuller-Rowell, T. J., D. Rees, S. Quegan, R. J. Moffett, and G. J. Bailey, Simulation of seasonal and universal time variations of the high-latitude thermosphere and ionosphere using a coupled, three-dimensional model, Pure and Appl. Geophys., 127, 189, Hedin, A. E., MSIS-86 thermospheric model, J. Geophys. Res., 92, 4649, Igi, S., W. L. Oliver, and T. Ogawa, Solar cycle variations of the thermospheric meridional wind over Japan derived from measurements of h m F 2, J. Geophys. Res., 104, 22,427, Kawamura, S., Y. Otsuka, W. L. Oliver, N. Balan, and S. Fukao, MU radar observations of H + in the topside ionosphere, J. Geophys. Res., 103, 20,697, Kawamura, S., Y. Otsuka, S-.R. Zhang, S. Fukao, and W. L. Oliver, A climatology of middle and upper atmosphere radar observations of thermospheric winds, J. Geophys. Res., 105, 12,777, Mayr, H. G., I. Harris, and N. W. Spencer, Some properties of upper atmosphere dynamics, Rev. Geophys., 16, 539, Millward,G.H.,H.Rishbeth,T.J.Fuller-Rowell,A.D.Aylward, S. Quegan, and R. J. Moffett, Ionospheric F 2 layer seasonal and semiannual variations, J. Geophys. Res., 101, 5149, Oliver, W. L., S. Fukao, T. Takami, M. Yamamoto, T. Tsuda, T. Nakamura, and S. Kato, Thermospheric meridional winds measured by the Middle and Upper Atmosphere Radar, J. Geophys. Res., 95, 7683, Oliver,W.L.,T.Takami,S.Fukao,T.Sato,M.Yamamoto,T.Tsuda, T. Nakamura, and S. Kato, Measurements of ionospheric and thermospheric temperatures and densities with the Middle and Upper Atmosphere Radar, J. Geophys. Res., 96, 17,827, Oliver, W. L., M. Yamamoto, T. Takami, S. Fukao, M. Yamamoto, and T. Tsuda, Middle and Upper Atmosphere Radar observations of ionospheric electric fields, J. Geophys. Res., 98, 11,615, Oliver, W. L., Y. Otsuka, and S. Fukao, MST radar measurement of ionospheric F region winds: The layer-wind technique, Radio Sci., 33, 941, Otsuka, Y., S. Kawamura, N. Balan, S. Fukao, and G. J. Bailey, Temperature variations in the ionosphere over the MU radar, J. Geophys. Res., 103, 20,705, Oyama, K.-I., and K. Schlegel, Observations of electron temperature anisotropy in the ionosphere: A review, Ann. Geophys., 6, 389, Reddy, C. A., S. Fukao, T. Takami, M. Yamamoto, T. Tsuda, T. Nakamura, and S. Kato, A MU Radar-based study of mid-latitude F region response to a geomagnetic disturbance, J. Geophys. Res., 95, 21,077, Richards, P. G., Seasonal and solar cycle variations of the ionospheric peak electron density: Comparison of measurement and models, J. Geophys. Res., 106, 12,803, Richards, P. G., and D. G. Torr, A factor of 2 reduction in the theoretical F 2 peak electron density due to enhanced vibrational excitation of N 2 in summer at solar maximum, J. Geophys. Res., 91, 11,331, Rishbeth, H., and C. S. G. K. Setty, The F-layer at sunrise, J. Atmos. Terr. Phys., 20, 263, 1961.

10 SIA 8-10 KAWAMURA ET AL.: ANNUAL VARIATIONS OF THE IONOSPHERE Rishbeth, H., and S. Fukao, A review of MU radar observations of the thermosphere and ionosphere, J. Geomagn. Geoelectr., 47, 621, Rishbeth, H., I. C. F. Muller-Wodarg, L. Zou, T. J. Fuller-Rowell, G. H. Millward, R. J. Moffett, D. W. Idenden, and A. D. Aylward, Annual and semiannual variations in the ionospheric F2-layer: II Physical discussion, Ann. Geophys., 18, 945, Sato, T., A. Ito, W. L. Oliver, S. Fukao, T. Tsuda, S. Kato, and I. Kimura, Ionospheric incoherent scatter measurements with the Middle and Upper Atmosphere Radar: Techniques and capabilities, Radio Sci., 24, 85, Schunk, R. W., and A. F. Nagy, Electron temperatures in the F region of the ionosphere: Theory and observations, Rev. Geophys., 16, 355, Su, Y. Z., S. Fukao, and G. J. Bailey, Modeling studies of the middle and upper atmosphere radar observations of the ionospheric F layer, J. Geophys. Res., 102, 319, Titheridge, J. E., The electron content of the southern mid-latitude ionosphere, , J. Atmos. Terr. Phys., 35, 981, Titheridge, J. E., and M. J. Buonsanto, Annual variations in the electron content and height of the F layer in the northern and southern hemispheres, related to neutral composition, J. Atmos. Terr. Phys., 45, 683, Tobiska, W. K., Revised solar extreme ultraviolet flux model, J. Atmos. Terr. Phys., 53, 1005, Torr, M. R., and D. G. Torr, The seasonal behavior of the F 2 layer of the ionosphere, J. Atmos. Terr. Phys., 35, 2237, Watanabe, S., and K.-I. Oyama, Effects of neutral wind on the electron temperature at a height of 600 km in the low latitude region, Ann. Geophys., 14, 290, Yonezawa, T., On the seasonal and non-seasonal annual variations and semi-annual variations in the noon and midnight densities of the F2 layer in middle latitudes, II, J. Radio Res. Lab. Jpn., 6, 651, Zou, L., H. Rishbeth, I. C. F. Muller-Wodarg, A. D. Aylward, G. H. Millward, T. J. Fuller-Rowell, D. W. Idenden, and R. J. Moffett, Annual and semiannual variations in the ionospheirc F2-layer, I. Modelling, Ann. Geophys., 18, 927, S. Fukao, N. Balan, and S. Kawamura, Radio Science Center for Space and Atmosphere, Kyoto University, Uji, Kyoto , Japan. (fukao@ kurasc.kyoto-u.ac.jp; balan@kurasc.kyoto-u.ac.jp; kawamura@kurasc. kyoto-u.ac.jp) Y. Otsuka, Solar-Terrestrial Environment Laboratory, Nagoya University, 3-13 Honohara, Toyokawa, Aichi , Japan. (otsuka@stelab. nagoya-u.ac.jp)