F-region vertical drifts were made using ionosonde for the Africa equatorial station Ouagadougou (12°N, 1.5°W; 5.9°N dip) from 3 a of data during January 1987 to December 1989 for solar cycle minimum, medium, and maximum conditions (F10.7 = 85, 141, and 214, respectively) under geomagnetic quiet-time. The variations are found to be dominated by the characteristics morning peak and evening prereversal enhancement (PRE) velocities. Seasonal and solar cycle effects are prominent near the dusk sector with an increase of PRE from solar minimum to maximum. The average equinoctial evening prereversal enhancement increases by almost a factor of three from low to high flux. On the average, the values of daytime and nighttime ionosonde-derived vertical drifts are smaller by about a factor of four than the magnitude often mentioned in publication for equatorial regions from other experimental techniques. In addition, the morning peak velocities maximize between about 0830–0930 LT with typical values of 12–18 m/s. In contrast, PRE have largest amplitude between about 1700–2200 LT with typical values of 5–10 m/s. The morning reversal times do not reveal any dependence on season and solar cycle; but the most probable time of occurrence is around 0530 LT. The evening reversal times are in excellent agreements for the three levels of solar activity periods apart from June solstice that exhibit considerable variations. The most likely time of occurrence is near 2000 LT. Of importance, the link between the onset / inhibition parameters of postsunset irregularities over Ouagadougou indicate anti-correlations with solar variability.
 The ionosphere plays a unique role in the Earth's environment because of strong coupling processes to regions below and above. The dynamics of the ionosphere, however, are controlled by a number of coupled processes involving production, recombination, diffusion, electrodynamic plasma drifts, and thermospheric neutral winds.
 At equatorial regions, the zonal electric field is known to play a significant role in essentially all electrodynamic processes of the region. The plasma drifts, particularly the vertical drift at the equatorial F region during the evening period is primarily driven by a zonal electric field formed as a part of the vertical polarization field developed due to enhanced thermospheric winds and decay of E region conductivity. These changes can sometimes trigger the plasma instabilities that lead to the onset of spread F irregularities [e.g., Fejer and Scherliess, 1997] and scintillations often observed at VHF and L-band frequencies.
 Since almost two decades, extensive new studies of low-latitude plasma drifts have been carried out using ion drift meter (IDM) and vector electric field measurements on the polar orbiting Dynamics Explorer 2 (DE-2), and on the low inclination Atmospheric Explorer E (AE-E) and San Marco Satellites, in particular, their longitudinal variability [e.g., Coley and Heelis, 1989; Heelis and Coley, 1992; Fejer et al., 1995]. Moreover, several theoretical and numerical models developed have contributed to continuing improvement in our understanding of the behavior of ionospheric. F region plasma drifts [e.g., Rishbeth, 1971; Heelis et al., 1974; Richmond et al., 1976; Farley et al., 1986; Crain et al., 1993a; Fesen et al., 2000]. Scherliess and Fejer  combined the AE-E data set of Fejer et al.  with extensive Jicamarca incoherent scatter radar vertical drift observations from 1968 to 1992 to present the first detailed global empirical model for the equatorial F region vertical drifts that take into account their diurnal seasonal, solar cycle, and longitudinal variations.
 The equatorial F region vertical plasma drift is upward by day and downward at night with the transition occurring around sunset and sunrise [e.g., Fejer, 1981]. The daytime upward drift undergoes an enhancement in the dusk period before reversing to downward direction. A unique feature of the diurnal pattern of the vertical drift at the low latitude ionosphere is the pre-reversal enhancement (PRE) and the peak value of the increase exhibits a strong positive dependence on solar activity, as represented by the 10.7 cm radio flux or sunspot number [e.g., Fejer et al., 1991; Ramesh and Sastri, 1995; Fejer et al., 1995].
 However, experimental data of plasma drifts therefore constitute a vita input for regional and global ionospheric models and also for assessment of the predictions of global atmospheric models [e.g., Anderson et al., 1987a, b; Heelis et al., 1990; Bailey et al., 1993]. By contrast, there is dearth of observational data of F region E × B drifts at the African sector of the equatorial and low latitude ionosphere due to scarcity of ground and space-based scientific instruments.
 In this paper we first describe our measurement techniques and ionospheric data used (section 2), and then examine the diurnal, seasonal, and solar activity effects on the F region vertical plasma drifts at Ouagadougou (section 3) before proceeding to discuss our results in section 4. The major results are summarized in section 5.
2. Measurement Techniques and Data Analysis
 The database used in the present study consists of hmF2 (the peak height of the F2 layer) for Ouagadougou (12.4°N, 1.5°W, dip angle, 5.9°N; magnetic declination, 3°21′W), Burkina Faso, an equatorial station in Africa region. These data were taken for 1987, 1988 and 1989 corresponding to low, moderate, and high solar activity conditions, respectively. The solar activity as represented by the F10.7 flux for the observational periods of 1987, 1988, and 1989 were confined to the ranges 71–101, 104–200, and 182–240 with mean values of 85, 141, and 214 (in units of 10−22W/m2/Hz), respectively. The corresponding average values of sunspot number (Rz) were 29, 100, and 162 for 1987, 1988 and 1989, in that order. At Ouagadougou, hmF2 values are not given and only M(3000)F2 values are given, hmF2 values are obtained from M(3000)F2 using the formula of Shimazaki . Hourly values of ionosonde hmF2 data were collected during geomagnetically quiet days of each month of the year were averaged. The 12 months were classified under equinoxes (March–April, September–October), June Solstice (May, June, July, and August), and December Solstice (November–December, January–February). Furthermore, the observed vertical drift velocity is obtained from 4-month seasonal averages of hmF2 by taking ΔhmF2/Δt, where ΔhmF2 is the difference between the hmF2 values at two consecutive 1-h intervals. Ionosonde observations have been used for several years to estimate the vertical motion of the equatorial ionosphere in the Brazilian and Indian sector but by measuring the time rate of change of h′F, the virtual height of the bottomside of F region, Δh′F /Δt [e.g., Abdu et al., 1981; Batista et al., 1986; Hari and Krishna Murthy, 1995]. This method is often limited to the sunset and evening hours when the F region bottom height rises beyond 300 km [e.g., Jayachandran et al., 1993; Subbarao and Krishna Murthy, 1994; Batista et al., 1996], the apparent F layer vertical displacement velocity (Δh′F/Δt), inferred from ionosonde measurements, represent the true vertical drift. Below 300 km the apparent vertical velocity starts to depart significantly from the true vertical drift, owing to the increasing dominance of the recombination process at these lower heights [e.g., Bittencourt and Abdu, 1981; Batista et al., 1996]. It is important, however, to note that the ionosonde drift derived from Δh′F/Δt reflects the vertical movement of the based of the layer, while ionosonde measurements inferred from ΔhmF2/Δt reflects the vertical motion of the peak of the layer. In addition, except for the study of Balan et al.  that derived the F region vertical E × B drift velocity from the hourly variation of hmF2 for the magnetic equator of Trivandrum (India) and Fortaleza (Brazil) and compared to that measured at Jicamarca (Peru), estimation of vertical E × B drift velocities from hmF2 has not serious received any attention since then. Nonetheless, for the observations presented now, the peak height of F layer remains mostly above 300 km, except for a short duration in the midnight to early morning (0300 and 0700 LT) during 1987 low solar activity period for the three seasonal periods. The layer changes would not, therefore, alter the basic pattern of the velocity variation. Our data sets are very reliable and suited for studying diurnal, seasonal, and solar cycle effects on the F region vertical plasma drifts over Ouagadougou.
 In this section we will examine in detail the characteristics of the seasonal and solar activity of the Ouagadougou average F region vertical plasma drifts. Figures 1a–1c illustrate the F region quiet time average upward drifts for three levels of the 10.7 cm solar flux index during June solstice, December solstice, and equinox seasons in 1987, 1988, and 1989, corresponding to low, moderate, and high solar activity periods, respectively. The average values of the solar decimetric index are indicated in the figures. In order to know the heights where ionosonde measurements were actually made, variations of hmF2, the mean peak height of F layer for Ouagadougou, were equally reported with the seasonally averaged vertical drift results on the bottom panels in Figures 1a–1c. The main results noted from Figures 1a–1c are summarized below.
 The variation patterns for each season are found to be more or less similar; drifts are upward during the day and downward during the night. The characteristics morning peak velocities dominate the daytime variations in the early morning hours and with pre-reversal enhancement during evening hours. For the greater part of the daytime, the variations do not reveal any such strong dependence on season or on solar activity; except for the midnight-dawn periods, which show small variation. The morning reversal times are independent of season and solar activity effects; the evening reversal times are similarly consistent for the three levels of solar cycle and season except for June solstice months, which exhibits significant variation.
 Furthermore Figures 1a–1c also indicate that the data differ primarily in terms of solar activity, the effect that seem to be most striking on the pre-reversal enhancement. The peak velocities of the pre-reversal enhancement found in the present study are generally smaller in values than what is commonly reported in the literature for equatorial regions. The pre-reversal enhancement is absent during the June solstice of 1987 year of low solar activity (Figure 1a), while the peak velocity of the pre-reversal enhancement increase by about a factor of two from moderate to high solar flux. During December solstice, the peak velocity of the pre-reversal enhancement is lowest during the low solar activity year of 1987 with a value of about 5.1 m/s, intermediate during the moderate solar activity year of 1988 with magnitude of about 13.1 m/s, and highest during the high solar activity year of 1989 with a value of approximately 14.1 m/s (Figure 1b). Moreover, the solar activity dependence also is found to be equally dramatic with equinoctial peak velocities being in the order of about 6–19 m/s (Figure 1c), with the value for the high solar activity year a little lower than the value for the moderate solar activity period. The magnitude of the daytime and nighttime seasonal averaged velocities are roughly similar, independent of season and solar activity, with the values much less than about 10 m/s. The nighttime downward velocities are found to be appreciably greater for the June solstice than for the other season during the year of high solar activity. The vertical ion drifts shown in Figures 1a–1c are largely in agreement with previous observations of vertical drifts estimated for other equatorial sites employing time variation of ionosonde h′F measurements [Batista et al., 1986; Namboothiri et al., 1989].
 The bottom panels in Figures 1a–1c give typical characteristics of the Ouagadougou average local time variation of the F2 region peak height (hmF2) for low, moderate, and high solar activity periods during June solstice, December solstice, and equinoxes. Notice that the hmF2 variations were used to obtain the vertical plasma drifts in the top panels in Figures 1a–1c. Bottom panel of Figure 1a shows a comparison of diurnal hmF2 variation for June solstice season for different solar activity periods. Generally, diurnal variation pattern of hmF2 observed at Ouagadougou during winter season at three average levels of the 10.7 cm solar flux indices are quite similar. Between 0600 and 1100 LT, hmF2 rises sharply and fairly linearly within the time interval 1200 and 1800 LT. This is followed by a sudden increase, reaching a peak at 2000 LT with a value of ∼460 km, ∼545 km, respectively, for 1988 moderate and 1989 high solar activity periods. Such peak is absent during 1987 low solar epoch. Notice that the postsunset peaks and the occurrence times of hmF2 are indeed coincident with the amplitude of PRE upward velocity, showing similar solar activity trends. Furthermore, as can be seen in Figures 1a–1c, the discrepancies between hmF2 values for the three seasons at the three solar activity levels start between local noon and last until around local midnight, whereas the variations follow each other closely within the time interval 1000–1200 LT. The values of hmF2, also increases with increasing solar activity. A minimum value of hmF2 is observed at about 0600 LT in all the seasons at the three different solar flux values.
 The morphology of hmF2 in December solstice (bottom panel in Figure 1b) is similar to that of June solstice (bottom panel in Figure 1a). Here, the post-sunset peaks are ∼374 km at 1900 LT, ∼470 km at 1900 LT, and ∼533 km at 1900 LT, respectively, for 1987 low, 1988 moderate, and 1989 high solar activity periods. On the other hand, post-sunset peaks in equinox season (bottom panel of Figure 1c) are ∼394 km at 2000 LT, ∼499 km at 2000 LT, and ∼535 km at 1900 LT, respectively, for average decimetric solar indices 85, 141, and 214 (in units of flux).
 Pre-reversal enhancement upward velocity lifts the F2 layer by about 85 km from medium to high solar activity (June solstice), by about 160 km from low to high flux (December solstice), and by about 140 km from low to high solar activity during the equinoxes. The dependence of the morning reversal times and morning peak velocities on season and solar activity are brought out in detail in Figures 2a–2b. Figure 2a gives the variation of the monthly value of the morning reversal times, separately for the 3a. Figure 2a indicates that morning reversal times during high solar activity are most variable. On the other hand, morning crossover times during low and moderate solar flux conditions are found to be perfectly similar except between September and November months. Statistical analysis shows that the occurrence time of morning crossover from downward to upward drift can take place anytime during 0445–0745 LT, with most probable time of occurrence around 0530 LT. A contrast among morning peak drift velocities for low, moderate, and high solar activity periods is shown in Figure 2b. One may note that the fluctuations between the months of January and May are not as large as that between June and December months. The morning peak vertical velocities are found to distribute over 10–50 m/s with typical most probable value of about 12–18 m/s. The occurrence time of peak velocity range from 0600–1200 LT with most likely time of occurrence at about 0830–0930 LT.
Figures 3a–3b show the dependence of evening reversal times and evening peak pre-reversal enhancement vertical velocities on season and solar activity. Figure 3a compares the seasonal variation of evening reversal times for three levels of F10.7 solar flux mean values. The variations show similar solar and solar activity trends except for low solar activity during the months of January and March, and April-May moderate solar activity periods. The evening crossover from upward to downward drifts can take place anytime during 1800–2200 LT, with most probable time of occurrence around 2000 LT.
Figure 3b presents the variation of the magnitude of PRE upward velocity for low, moderate, and high solar activity periods, respectively. Figure 3b indicates that there are large fluctuations in the magnitude of the peak PRE upward drift velocities during moderate and high solar activity periods. The evening peak velocity is found to distribute over 5–25 m/s with most probable value occurring in the range of 5–10 m/s. The occurrence time of the peak ranges from 1700–2200 LT, with the most likely occurrence time around 1900 LT.
 In order to test the relationship between the characteristics morning peak (VZPM) and evening PRE velocities (VZPE), as well as postsunset maximum in F2 peak height (hmF2p) and the solar variability, we use 36 monthly values of VZPM, VZPE, and hmF2p from January 1987 to December 1989. Corresponding 17 values of the F10.7 indices were used as a proxy for solar activity and the correlation of F10.7 with monthly values of morning and evening peak velocities, and maximum in ionospheric F2 peak altitude was investigated. Simple linear fits were used to study any dependence of VZPM, VZPE, and hmF2p on F10.7. The overall values of VZPM are presented in Figure 4 where VZPM values are plotted versus F10.7. The best fit straight line is also shown in Figure 4. The 17 months averaged value of F10.7 = 147, representative of moderate solar activity is also indicated in the plot. From the results given in Figure 4, it can be concluded that the VZPM and F10.7 are related as
with correlation coefficient R of 0.17013. One notes that there is no significant relation between F10.7 and VZPM, this relation must, at best, be regarded with caution.
 In Figure 5, the evening pre-reversal enhancement in vertical plasma drift, VZPE values are plotted versus F10.7 indices. Again, VZPE and F10.7 are correlated as
with correlation coefficient R of 0.6938. From the results given in Figure 5, it is clearly shown that VZPE has a reasonably good correspondence with solar activity. Results for the postsunset maximum in F2 peak height, hmF2p dependence on solar activity is presented in Figure 6. As before the best fit line to the data is given by
with regression coefficient of 0.9176. Further from this relation we conclude that solar effect is much more pronounced on the variation of hmF2p.
Figure 7 shows the dependence of pre-reversal maximum in F2 peak height, hmF2p. on postsunset vertical E × B peak velocity between 1700–2200 LT for 1987 and between 1900–2000 LT for both 1988 and 1989. There are three panels of plots in Figure 7, representing solar cycle minimum (top panel), moderate (middle panel), and maximum (bottom panel) conditions, with averaged monthly solar flux intensity F10.7 values of 85, 141, and 214, respectively. Linear regression parameters are shown in each plot. There are a couple of interesting features noticed in the results presented in Figure 7. First, the slopes connecting hmF2p and VZPE are found to decrease considerably with increasing phase of solar cycle. Detailed analysis shows that the percentage change in slope is about 36 and 8 from low to medium, and from medium to high solar activity, respectively. Second, the correlation coefficients, R values in the linear regression analysis are very good for low and moderate solar activity year of 1987 and 1988, in that order; but very poor during high solar flux conditions of 1989. In addition, correlation coefficient values decrease appreciably with increase in the epoch of solar cycle. From low to moderate solar activity the change in R value is just about 3%, while ∼76% change in R is observed from moderate to high solar activity conditions. Three, VZPE and hmF2p. are confined to the range ∼1.5–14.1 m/s, ∼326–445 km (low); ∼4.1–26.3 m/s, ∼406–544 km (moderate); and ∼9.4–26.2 m/s, ∼514–547 km (high). We observe immediately that the range of occurrence of post-sunset peak in hmF2p. decreases with increasing solar activity, although hmF2p. increases with increasing solar flux intensity. The connection between VZPE and hmF2p. is further examined by plotting 36 monthly values of hmF2p against the corresponding values of VZPE. A summary of the results is provided in Figure 8. From this data, one may note that while the value of the slope is in agreement with the value of the slope obtained for low solar activity year of 1987 (cf. upper panel in Figure 7), the correlation coefficient value is consistent with that of moderate solar activity period of 1988 (cf. middle panel in Figure 7). However, we can state that hmF2p. correlate well with VZPE at low and moderate solar activity periods for a given longitude near the dip equator.
 There is a vast published literature showing the occurrence of the plasma instabilities that leads to the development (or inhibition) of postsunset spread F irregularities during quiet geomagnetic periods in the bottomside low-latitude ionosphere. However, it is generally accepted that upward E × B drift in the F region and altitude are the main governing parameters for the uplift that leads to the irregularities during this period. Although threshold parameters correlate well with solar activity, but they have not been correlated previously with each other at different levels of solar activity. A clear and unique relation exists between the magnitude of postsunset maximum in F2 peak height and the magnitudes of the prereversal enhancement. Since the dependence of hmF2p. on VZPE showed an exceptional solar effect, an overall value (in Figure 8) could be misleading, if viewed in isolation. Consider along with the solar effect in Figure 7, the observation in Figure 8 will help emphasize the significant of solar influence on the threshold parameters required to trigger postsunset spread F irregularities. Thus we suppose that the results in the top and middle panels of Figure 7 for low and moderate solar activity conditions, respectively, postsunset ESF can likely occur, while the data in the bottom panel of Figure 7 suggests that even when meaningful threshold values needed to ignite postsunset equatorial irregularities are met, plasma instabilities can still be suppressed at high solar flux conditions.
 Indeed it is of note that the results presented in this work are extremely encouraging, it is apparent that the values of vertical drift during moderate solar activity year 1988 are found to be comparable or even larger than the maximum solar activity year 1989 (Figures 1a–1c and Figures 2 and 3). Additionally, we find excellent variability patterns and trends in F region vertical plasma drift during moderate and high solar activity years near the sunset periods in December solstice and equinoxes seasons. A clear avenue of investigation was therefore to look at the daily and monthly solar flux intensity F10.7 in 1988 versus that in 1989 since it is known that solar activity is an important factor in the variability of F region equatorial electrodynamics. Nonetheless, the observed results are connected to the anti-solar activity correlation between the solar flux intensity F10.7 during a moderate solar activity year of 1988 and a high solar activity year of 1989. The cause of anti-solar activity correlation in turn seems to be linked to the transequatorial meridional wind that exhibits a more pronounced instability effect during a high solar activity year than during a low solar activity year. Then we present in Figure 9a (top panel) the daily and Figure 9b (bottom panel) the monthly solar flux intensity F10.7 in 1988 versus that in 1989. Four main points are obvious. First, the lowest daily solar flux intensity in 1988 is slightly higher than 100, but significantly larger than 130 in 1989. Second, the daily solar flux intensities in 1989 is seen at all times higher than in 1988 except for about 3 days (day 183–185) in June, 1 day (day 211) in July, 3 days (day 266–268) in August, and 13 days (day 346–358) in November of 1989. Three, the averaged monthly value of solar flux indices in 1989 is always higher in magnitude than that in 1988 except for the month of December in 1989. The solar activity is in the range 105-154, 169-222, respectively, for moderate and high solar activity period. Fourth, Figure 9b reveals strong anti-correlation between the averaged monthly solar flux intensity in 1988 and 1989.
 In this work we have used the hourly values of the ionosonde hmF2 time variations to derive F region vertical plasma drifts in order to investigate the seasonal and solar cycle effects on the electrodynamics of the equatorial ionosphere in the African region, assumed that the ΔhmF2/Δt is a realistic vertical drift velocity during the daytime sector; since hmF2 is mostly far above the threshold value of 300 km. However, there are several areas of comparison between the present study and previous seasonal and solar activity behavior of F region vertical drifts that employ different techniques [e.g., Batista et al., 1986; Namboothiri et al., 1989; Scherliess and Fejer, 1999; Fesen et al., 2000]. Our results are similar to Figure 2 of Batista et al.  who presented a comparison of the vertical F layer velocities measured by the Jicamarca radar with those simultaneously deduced as Δh′F/Δt from Huancayo (12°S, 75.3°W. dip 2°N, magnetic declination 1°W) ionograms, during solar maximum conditions for equinox, winter, and summer seasons. Namboothiri et al.  used a database from HF Doppler observations within 0700–0100 LT sector during the period of 1984 to1986, representative of low, medium, and high solar flux conditions for equinox, summer, and winter seasons; reported similar results. The authors noted that the characteristics pre-reversal enhancements that exhibit the most striking seasonal, solar, and magnetic activity effects dominate the diurnal pattern. The observations presented in Figures 2a–2c and 3a in their work are found to be in broad agreement with our present results. Furthermore solar cycle dependence of the empirical model drifts in four longitudinal sectors, namely: African-Indian (0°–150°E), Pacific (150°–210°E), Western American (210°–300°E), and Brazilian (300°–360°E) equatorial regions presented in Figure 7 in the work of Scherliess and Fejer  are perfectly in accord with the current study. At low flux conditions F10.7 = 90 in Figure 8 of Scherliess and Fejer  and F10.7 = 85 in Figure 3 of Fesen et al. ; the pre-reversal enhancement velocities are completely absent during the June solstice season, also in excellent agreement with our data (when F10.7 = 85). This result is in sharp contrast with that of Namboothiri et al.  who reported that the pre-reversal enhancement is absent during the summers of low solar years at Trivandrum (8.3°N, 76.9°E, dip; 0.9°S, magnetic declination, 3°W), India. The magnitudes of the observed daytime and nighttime are found to be consistently lower in all the three levels of solar activity periods. Fesen et al.  reported that the simulation model daytime drifts remain too small in both the zonal and vertical components.
 An interesting and unique feature of the low-latitude ionosphere is the pre-reversal enhancement [e.g., Woodman, 1970], a sharp upward spike in the vertical plasma velocities that occurs shortly after local sunset, superimposed on the typical diurnal variation of daytime upward and nighttime downward drifts. The value of PRE peak E × B drift determines the onset (or inhibition) conditions of spread-F irregularities. At low solar epoch the values of PRE are 5.1 m/s and 5.7 m/s, respectively for December solstice and equinoxes. The magnitudes of PRE during medium solar activity period are 8.7 m/s, 13.1 m/s, and 18.7 m/s for June solstice, December solstice and equinoxes, in that order. At high solar activity period, the values of PRE are found to be 15.6 m/s, 14.1 m/s, and 17.3 m/s for June solstice, December solstice, and equinoxes, respectively. Typical range found in the present study ∼5–20 m/s is in agreement with the results of Goel et al., . Oyekola et al.  estimated vertical plasma drifts using Δh′F/Δ for Ibadan (7.4°N, 3.9°E, 6.0°S dip) and found the seasonal quiet averages solar maximum PRE velocities as 36.4, 29.5 and 26 m/s during equinoxes, summer, and winter periods, respectively. A close look at Ouagadougou solar maximum PRE values shows that Ibadan PRE magnitude is a little more than about a factor of two higher than that of Ouagadougou values during similar solar and geomagnetic quiet conditions. Recently, Oyekola  reported that the equinoctial peak velocities are about 36 m/s and 35 m/s for Ibadan ionosonde and satellite drifts, respectively; but Jicamarca equinoctial peak velocity is less than 30 m/s.
 The values of F region vertical E × B drift velocities estimated from the time variation of ionospheric peak height hmF2 are observed to be smaller in values particularly in the dayside than those derived for magnetic equator of Trivandrum and Fortaleza and compared with that measured under similar conditions at Jicamarca as reported previously by Balan et al. . They noted that the vertical velocities derived from the ionospheric peak height are in good agreement with the velocities measured by the incoherent scatter radar at Jicamarca as long as the ionospheric peak is within the altitude range (300–400 km) of the radar at Jicamarca; the two velocities differ significantly (in magnitude during the evening hours) when the ionospheric peak moves outside the altitude of the radar. Blanc and Houngninon  using radar measurements performed in West Africa at Korhogo-Ivory Coast (9° 24′N, 5° 37′W, dip 4°S) in the period of March–July 1993 during International Equatorial Electrojet Year (IEEY) reported weak variations in F region vertical velocity. The authors noted that the velocity does not exceed 15 m/s either downward or upward. Therefore, it appears that equatorial ionosphere in Africa continent is characterized by smaller vertical E × B drift velocities. Further we expect that the observed results are not connected with the difference in the height range of observation. For instance, the estimated velocities (shown in the upper panel of Figures 1a–1c) during 1987 low, 1988 moderate, and 1989 high solar activity periods give the average velocity of the ionosphere between ∼266–368 km, ∼288–474 km, and ∼308–538 km, respectively. Thus on the average, the altitude to which they correspond can be taken to be 320 km, 380 km, and 420 km, for solar minimum, moderate, and high activity, respectively.
 When looking for a cause of low velocity values observed especially in the dayside sector, it must be remembered that lower thermospheric tidal forcing could influence the daytime vertical E × B drift velocity. Millward et al.  using coupled thermosphere-ionosphere-plasmasphere (CTIP) model showed that semidiurnal tide plays a large role in determining upward ion motion and thus the daytime equatorial ionospheric anomaly. In particular, variability within the tidal forcing such as that observed and discussed on a seasonal and year to year basis by Goncharenko and Salah  and on a daily basis by Zhou et al. , will undoubtedly play a role in determining the observed variability within the daytime vertical E × B drift, and the distribution of equatorial F region plasma.
 Ionospheric east-west electric fields can be affected by the dynamic conditions at the base of the thermosphere. Hence, the quiet-time diurnal pattern and its variations with season and solar activity are now fairly well understood, at least from the point of view of F region dynamo effects.
 The observed anti-solar activity correlations between the threshold parameters essential to initiate equatorial spread F irregularities could be due to the suppression of the instability growth rates by various factors. It is reported that transequatorial wind plays a pivotal role in the suppression of the generalized Rayleigh-Taylor (R-T) growth rate and that the wind effect is more pronounced if the E × B drift reversal time is late Maruyama , during the post-sunset periods. Abdu et al.  reported that small angles between sunset terminator and magnetic meridian favor the post-sunset ESF development, but perfect alignment of the terminator with magnetic meridian tends to decrease the efficiency of ESF development. In addition, vertical winds also control the R-T growth rates [Raghava Rao et al., 1992]. A downward wind pushing the F region into lower altitudes of higher recombination rates does favor the R-T growth rate, while upward wind acts otherwise. Of recent, Su et al.  using ROCSAT-1 observations during high solar activity year of 2000 (average F10.7 = 179) against moderate solar activity year of 2003 (average F10.7 = 128), reported an anti-solar activity correlation between the solar activity effect and the topside ionospheric density irregularity occurrence rate at longitude of bad alignment in a solstice season. They concluded that transequatorial meridional wind induced instability suppression process that is more effective to offset the growth of R-T instability during a solstice season in a high solar activity than in a low solar activity year. Then we attribute the cause of anti-solar activity correlation results obtained in the current work to the greatly larger suppression effects in the R-T instability growth rate from transequatorial component of the thermospheric wind over Ouagadougou longitude sector during high solar activity period than in a low solar activity year.
 We have examined the seasonal and solar cycle dependence of the Ouagadougou F region plasma drifts. By and large, the daytime upward drifts remain too small and are practically independent of solar activity. The variations are found to be dominated by the characteristics morning peak and evening enhancement velocities, which exhibit the most striking seasonal and solar activity effects, particularly, pre-reversal enhancements. The average equinoctial peak velocity drops almost by a factor of 3 as the solar activity index (F10.7) falls from 214 to 85 units.
 The reversal times near sunrise do not reveal any dependence on season and solar cycle, and are less variable. On the other hand, reversal times near sunset are in excellent agreements during the three levels of solar activity except winter months that exhibits considerable variations. The evening reversal times occur latest during June solstice for all levels of solar activity. The most probable time of occurrence of morning crossover for Ouagadougou is around 0530 LT, whereas the most likely time of occurrence of evening crossover is near 2000 LT. The onset (or suppression) parameters required to cause spread F irregularities at Ouagadougou longitudinal sector are found to be ∼5.5 m/s and ∼372 km, ∼13.6 m/s and ∼474 km, ∼17.1 m/s and ∼535 km for pre-reversal peak E × B drift and corresponding hmF2p, in that order, for low, medium, and high solar activity periods.
 The present observations indicate that the apparent vertical F layer drifts, inferred from ionosonde measurements as ΔhmF2/Δt, is, as theoretically predicted Bittencourt and Abdu , representative of the vertical E × B plasma drift velocity only around the times of pre-reversal vertical drift enhancements when the layer is high, as can be verified from the hmF2 values (see Figures 1a–1c). Our data are found to be completely consistent with some previous results [e.g., Namboothiri et al., 1989; Fejer et al., 1991; Sastri, 1996; Scherliess and Fejer, 1999; Fesen et al., 2000] that estimated equatorial ionospheric vertical ion drift from time variation of ionosonde virtual height, h′F and other techniques.
 There is a clear and unique link between peak hmF2 and prereversal enhancement at different levels of solar activity. In this regard, from the analysis of ground-based ionosonde measurements, we find evidence, although is for the bottomside ionosphere that supports the recent report by Su et al.  that space-based (ROCSAT-1) observations indicate anti-correlation between the solar activity effect and topside density irregularity occurrence rate at longitude of bad alignment in a solstice season.
 Finally, as for applications of our results: descriptions of F layer low-latitude electrodynamics variability are valuable for assessing the reliability of global thermospheric and ionospheric models. Also, as such, the results obtained here should provide input parameters for regional and global quiet ionospheric models, whenever the results over the sub-Sahara West African region are of interest.
 We would like to acknowledge the reviewers for their assistance in evaluating this paper. The useful comments and suggestions made by the reviewers particularly on re-processing of Figures 4 and 5 and inclusion of Figures 6, 8, and 9 improved the paper and make it clearer and easier to read. The authors thank the Abdus Salam ICTP for literature used in this work. One of the authors (O. S. Oyekola) is sincerely grateful to the following colleagues for assistance in the literature during the revised version of the work: V. H. Rios, Head of Physics Department, Tucuman University, Tucuman, Argentina, D. V. Pancheva, Center for Space, Atmospheric, and Oceanic Sciences, Department of Electronic and Electrical Engineering, University of Bath, Bath, U. K., and J. Nair, NC A & T State University, College of Arts and Sciences, Department of Physics, NC, USA.
 Amitava Bhattacharjee thanks R. S. Dabas and another reviewer for their assistance in evaluating this paper.