A comparison of ground and satellite observations of F region vertical velocity near the dip equator



[1] Nighttime F region vertical electrodynamic drifts were made at the magnetic equatorial stations in Africa, Ibadan (7.4°N, 3.9°E; 6°S dip), and Ouagadougou (12°N, 1.5°W; 5.9°N dip) using ionosondes hmF2 data during 1957/1958 International Geophysical Year (IGY) and 1990 periods, respectively; for high solar flux and geomagnetic quiet time conditions. We compare the seasonal averages of vertical drifts with observations made by Incoherent Scatter Radar (ISR) at Jicamarca (11.95°S, 76.87°W; 2°N dip) and Atmospheric Explorer E (AE-E) satellite for equatorial F layer vertical drifts reported by Fejer et al. (1995). The results indicate good accord between the three techniques at periods when convection dominates other factors (e.g., around prereversal enhancement) except for June solstice drifts. However, when the drifts are completely downward (negative) between 2000 and 0500 LT sector, the mean discrepancies between ionosondes and ISR drifts are of 8–11 m/s (December solstice), 14–17 m/s (equinoxes) but comparable during June solstice with a value of about 14 m/s. Conversely, the typical values of the differences among ionosondes and AE-E downward velocities are 7–10 m/s (December solstice), 9–12 m/s (equinoxes), yet again similar in winter with a significantly smaller value of about 3 m/s. The evening reversal times are in excellent agreement, apart from June solstice drifts, which exhibits large fluctuations. The morning reversal times show small variations. Equinoctial prereversal enhancement velocities have amplitudes of approximately 17–35 m/s between the methods. Our data are useful for global ionospheric modeling and for the predictions of development of nighttime equatorial F region irregularities at the African region where there is paucity of data.

1. Introduction

[2] The unique arrangement of the geomagnetic field and gravitational field at the equator provides for a number of fascinating physical phenomena, such as, equatorial ionization anomaly, equatorial electroject, evening prereversal enhancements, and equatorial spread F [e.g., Eccles, 1998]. These phenomena result from the electric field structure created by the complex interaction of thermospheric neutral wind, the geomagnetic field, gravity, and the ionosphere. However, several studies of F region plasma drifts have been generally more concentrated in the Peruvian, Brazilian, and Indian equatorial sectors, using various techniques of measurements. In contrast, there is an acute shortage of these measurements over the equatorial Africa ionosphere. The need of more spatial coverage and also continuous observations help improve the understanding of the climatology and weather of the low latitude electrodynamics [e.g., Fejer and Scherliess, 1997].

[3] Since the late 1960's the Jicamarca Incoherent Scatter Radar (ISR) has measured with very good accuracy both vertical and zonal E × B drifts [e.g., Woodman and Hagfors, 1969; Woodman, 1970; Woodman and Hoz, 1976; Fejer et al., 1979, 1985, 1989, 1991; Gonzales et al., 1979; Kudeki et al., 1999a]. However, these observations are limited to (a) few days (on average less than 20) a year, and (b) the Jicamarca longitude sector. Given the poor time and spatial coverage of ISR drift measurements, alternative means have been investigated using coherent scatter techniques with low power systems [e.g., Balsley, 1969; Hysell and Burcham, 2000; Chau and Woodman, 2004].

[4] Digisonde appears to be another means of getting ionospheric drifts, during day and night [Belehaki et al., 2006; Bertoni et al., 2006]. At equatorial latitudes, comparison and validation studies have not been yet conducted on the digisonde drift measurement [Woodman et al., 2006].

[5] The use of satellite measurements to study equatorial F region electric fields and plasma drifts [Maynard et al., 1988; Coley and Heelis, 1989; Coley et al., 1990; Heelis and Coley, 1992] and the solar cycle, seasonal, and longitudinal effects in them [Fejer et al., 1995] has contributed to a global view of the longitudinal behavior of these drifts and associated electric fields. Although the satellite data are, in principle, the most appropriate to give the broadest longitudinal coverage, the averaging procedure used to group the drifts in sector can sometimes mask important features that occur in some specific regions.

[6] Nevertheless, ground-based ionosonde measurement in the Peruvian, Brazilian, Indian, and African equatorial regions were used to infer the vertical plasma drifts during some specific local times, more precisely, during the prereversal enhancement that occurs around sunset, and during nighttime sector [Abdu et al., 1981; Batista et al., 1986, 1996; Hari and Krishna Murthy, 1995; Sastri, 1996; Oyekola et al., 2006; Oyekola, 2006]. Although the ionosonde drifts have well-known limitations, they have long shown clear indication of large longitudinal variations of the vertical plasma drift prereversal enhancement over the South American region.

[7] A comprehensive understanding of the variability of the vertical plasma drifts, particularly, in the evening and night hours are of fundamental importance, as they appear to have a direct link with equatorial spread F/plasma bubbles [e.g., Dabas et al., 2003; Fejer et al., 1999; Jyoti et al., 2004].

[8] The morphology of equatorial F region vertical plasma drifts determined by various techniques are generally in good agreement, but the drift amplitude indicates noticeably substantial unexplained disparities.

[9] In this study, we use vertical drifts derived from ionosonde measurements for the equatorial stations within African longitude sectors, Ibadan, Nigeria, and Ouagadougou, Burkina Faso during the nighttime from 1800–0600 LT; to examine the discrepancies and relationship between ionosonde, ISR, and AE-E drifts through a comparison of our result with those reported by Fejer et al. [1995] for Jicamarca and from AE-E satellite data.

2. Measurement Technique and Data

[10] The data used in this work consists of hourly median values of ionosonde hmF2 (the peak height of F2 layer) obtained from equatorial stations Ibadan (7.4°N, 3.9°E, 6°S dip). Ibadan data were taken from January to December 1958 (era of International Geophysical Year-IGY). That year was classified as a period of high solar flux with a mean sunspot (R12) of 185. The F10.7 radio flux was in the range 180–230 × 10−22 W/m2/Hz, with a mean value of 208 (in flux units). The calculated hmF2 values for Ibadan around the time of IGY were used. However, at Ouagadougou, hmF2 values were not given, only M(3000)F2 values were given, so, hmF2 values were then obtained from M(3000)F2 using Shimazaki's [1955] formula. The values of hmF2 span from January to December 1990, the period of high solar activity conditions with an average sunspot (R12) of 145. The F10.7 radio flux was in the range 156–205, with an average value of 190. For both data sets, geomagnetic quiet conditions (Kp ≤ 3.0) were considered.

[11] Since the number of scientific equipment measuring ionospheric drifts at equatorial African stations is very limited, it is always desirable to have other ways of inferring those parameters [Abdu et al., 1981]. Bittencourt and Abdu [1981] showed that, at special time periods during sunset and evening hours, when the F layer height is above a threshold of 300 km, the apparent F layer vertical displacement velocity estimated from ionosonde measurements, is the same as the vertical E × B plasma drift velocity, such as that determined from Incoherent Scatter Radar measurements of the F region. For heights smaller than 300 km, the apparent vertical velocity starts to depart significantly from the vertical E × B drift velocity, owning to the increasing dominance of the recombination process at these lower heights. For the periods 1800–0600 LT the F2 layer peak height remained far above 300 km for most of the time and hence no correction has been applied to account for the effects of the layer decay.

[12] The 12 months were grouped under equinoxes (March–April, September–October), December solstice (November, December, January, and February), and June solstice (May through August). Vertical F layer ion velocities at both stations were deduced from 4-month seasonal averages of hmF2 by taking Δ(hmF2)/Δt, where Δ(hmF2) is the difference between the hmF2 values at two consecutive 1-hour intervals. The uncertainties in the determination of hmF2 at Ibadan are around ±23 km, ±28 km, and ±22 km during June solstice, December solstice, and equinoxes, in that order; while those of Ouagadougou are roughly ±26 km, ±21 km, and ±22 km for June solstice, December solstice, and equinoxes, respectively. The corresponding ionosonde estimated vertical ion drift uncertainties are about ±2.43 m/s (June solstice), ±1.68 m/s (December solstice), and ±1.43 m/s (equinoxes) for Ibadan, but that of Ouagadougou are only near ±1.82 m/s (June solstice), ±1.23 m/s (December solstice), and ±1.56 m/s (equinoxes). The uncertainty associated with the vertical drift measurement with signal integration time of about 5 min for most of the observations reported from Jicamarca is stated to be of the order of 1–2 m/s [Fejer et al., 1979]. Clearly, ionosonde and radar data uncertainties are consistent.

[13] The method described above has been used for several years to derive the vertical motion of the equatorial ionosphere in the American and Indian sectors, as Δ(h′F)/Δt, where h′F is the virtual height of the bottomside of F region [e.g., Abdu et al., 1981; Batista et al., 1986; Hari and Krishna Murthy, 1995]. It is of note that the ionosonde measurements used in this work are reliable and seem sufficient to study short-term variability of the ionospheric evening and nightside vertical ion E × B drifts at two African stations that are symmetrically placed about the same dip equators and their comparisons with radar velocities data at Jicamarca and satellite Atmospheric Explorer E results reported by Fejer et al. [1995].

3. Observations and Discussion

[14] Figure 1 shows the temporal patterns of F region vertical plasma drifts for Ibadan and Ouagadougou ionosonde-derived, Jicamarca VHF radar, and AE-E drifts during June solstice season. The radar and satellite data were reproduced from the work of Fejer et al. [1995]. That investigation used 2 year of data from January 1978 and December 1979 to examine the longitudinal (+5 to –5 dip latitude) dependence of the satellite-observed drift at an epoch of moderate to high solar activity (160 ≤ F10.7 ≤180 in units of 10−22 W/m2/Hz). The average sunspot number, R12 for 1978 = 86.9; and 1979, with mean R12 = 146. Notice that the observational results presented here correspond essentially to quiet geomagnetic conditions. The vertical bars shown on the ionosonde drifts in Figures 123are the standard deviations from the mean values.

Figure 1.

Examples of average nighttime Ibadan (solid up triangle) and Ouagadougou (open circle) ionosonde-derived F region vertical velocities and the measured mean F region vertical plasma drifts at Jicamarca with ISR (solid diamond) and longitudinal averaged AE-E satellite (cross) during quiet time June solstice solar maximum conditions. Ibadan drift is during the period of International Geophysical Year (1957/1958), Ouagadougou is during 1990 year, and ISR and AE-E data are during 1978–1979 periods and are after Fejer et al. [1995]. Vertical bars superposed on Ibadan and Ouagadougou drifts are the standard deviations.

Figure 2.

Same as Figure 1, but for quiet December solstice high solar flux conditions.

Figure 3.

Same as Figure 1, but for quiet equinox solar maximum conditions.

[15] Figure 1 indicates the characteristics of nighttime June solstice patterns of variability among the probing techniques. Ibadan drift rises sharply from downward direction with a value of about 20 m/s at 1800 LT to upward direction with a value of about 23 m/s during the prereversal enhancement period (1900 LT). Thereafter, it decreases gradually until it reverses direction gently downward at 2200 LT, followed by a systematic hourly variation until the drift is upward at 0600 LT. On the other hand, Ouagadougou drift rises gently from downward direction with significantly smaller value of just 3 m/s at 1800 LT to upward direction with a value of about 10 m/s at 2000 LT. Thereafter decreases steadily until local midnight, followed by systematic increase until it is about to reverse to upward direction at 0600 LT. Jicamarca drift is upward at 1800 LT and suddenly become negative from 1900 LT, and decreases gradually till 2300 LT, followed by continuous increase until the drift is positive at 0600 LT. Also, satellite drift is upward between 1800 and 1900 LT followed by a sudden reversal from upward to downward direction at about 2000 LT. The variation then continues in the downward direction until about 0600 LT.

[16] As can be seen in Figure 1, Jicamarca drift show greater values of vertical drifts between 1800 and 0200 LT in comparison to other techniques. There is fair agreement between drift measurement methods during 0400 and 0600 LT. Of course, the magnitude of Ibadan and satellite average downward drifts between 2000 and 0500 LT are found to be similar, whereas the seasonally averaged magnitude of Ouagadougou drift is ∼ 2 m/s lower in value than the satellite drift June solstice season.

[17] In Figure 2 we present the observational results among ionosondes, radar, and AE-E vertical drifts during December solstice. From the plot one may note that there are differences between the solstices when compared to Figure 1. The variability pattern around the prereversal enhancement period for December solstice is slightly opposite that of June solstice. Nonetheless, the principal feature remains the same. Notice also that the variations in Figure 2 have extensive similarities for the three methods. The drifts are positive between 1800 and 1900 LT, followed by a reversal to a downward direction at 2000 LT; thereafter continue to exhibit significant hour-to-hour variation till 0500 LT, until the drifts crossover to upward direction at 0600 LT for all techniques. Evening upward peaks are not noticeable here except for Ibadan ionosonde drift with a peak velocity of about 27 m/s, and Ouagadougou ionosonde drift with a maximum velocity of about 11 m/s. It is as well noted from Figure 2 that radar and AE-E techniques exhibit greater values of drifts between local midnight and 0500 LT in comparison with ionosonde drifts. In magnitude, Ibadan ionosonde drifts between 2000 and 0500 local times are about a factor of two smaller than both radar and AE-E drifts, while Ouagadougou ionosonde drifts at the same local times interval is approximately a factor of three lower than Jicamarca and satellite drift values.

[18] Figure 3 illustrates typical characteristics of geomagnetic quiet nighttime and high solar activity conditions of Ibadan and Ouagadougou ionosondes, Jicamarca radar, and AE-E satellite vertical plasma drifts during equinoctial period. Comparing Figure 3 with results for other season (Figures 1 and 2), we find that these patterns reveal a unique feature of the low latitude ionosphere known as prereversal enhancements for all methods. The amplitudes of prereversal enhancement is lowest in Ouagadougou ionosonde drift with a value of almost 16.6 m/s; comparable for Ibadan ionosonde and Jicamarca radar drifts with a value near 27 m/s; and highest for AE-E drift (∼35 m/s). These peak vertical velocities occur shortly after local sunset (1900 LT) for all techniques. After 1900 LT, the velocities are downward, followed by significant hourly variations and eventually to a reversal in the drift direction at 0600 LT, except for Jicamarca drifts. Clearly, the patterns of variations are in good accord between the three methods around prereversal enhancement period. Again, between 2100 and 0600 LT, radar drift are larger in values than the ionosondes and AE-E vertical plasma drifts. Ibadan ionosonde drifts (when the drifts are completely downward) are smaller than satellite drift data and Jicamarca drifts by a factor of 2 and 3, in that order; whereas Ouagadougou ionosonde vertical drifts at similar local times are lower in magnitude than AE-E and radar drifts by a factor of 3 and 4, respectively.

[19] The vertical velocity differences between Incoherent Scatter Radar and Ibadan ionosonde (ISR-IBI), incoherent radar and Ouagadougou ionosonde (ISR-OUI), Atmospheric Explorer E (AEE) and Ibadan ionosonde (AEE-IBI), and Atmospheric Explorer E and Ouagadougou (AEE-OUI), respectively, are brought out in detail in Figures 4a–4c for June solstice, December solstice, and equinoxes, respectively. The main results noted from Figures 4a–4c are summarized below.

Figure 4a.

Variation of equatorial average F region vertical drift velocities measurement techniques differences. ISR, IBI, OUI, and AEE denote Incoherent Scatter Radar, Ibadan Ionosonde, Ouagadougou Ionosonde, and Atmospheric Explorer E, respectively, for June solstice season.

Figure 4b.

Same as Figure 4a, but for December solstice season.

Figure 4c.

Same as Figure 4a, but for equinox season.

[20] During June solstice the vertical drift difference between ISR and ionosondes (both IBI and OUI) are consistent with each other with a large value of ∼14 m/s, while the differences between satellite and ionosondes (IBI and OUI) are similarly in agreement with a value of approximately 3 m/s. On the other hand, the discrepancies during December solstice between ISR and ionosondes (IBI and OUI) are of 8–11 m/s, whereas the differences between satellites and ionosondes are 7–10 m/s. However, at equinoxes, the observed discrepancies are of order 14–17 m/s and 9–12 m/s, respectively, for radar and ionosondes, and for satellite and ionosondes.

[21] Conventional ionosonde measurements do not provide accurate plasma drifts measurements most of the day; they determine precisely the evening reversal times of the vertical plasma drifts [Fejer et al., 1991]. Figure 5 shows the inferred reversal times from ionosonde data from Ibadan, Ouagadougou, Fortaleza, and Kodaikanal and the reversal times determined from the Jicamarca drift observations all for solar maximum. Kodaikanal and Fortaleza reversal times data were inferred from the time variation of h′F (virtual height of the bottomside of the F layer), while reversal times data were derived from the time variation of hmF2 for Ibadan and Ouagadougou. Jicamarca F region time of drift evening reversal, Fortaleza, and Kodaikanal drift reversal times were deduced from the work of Fejer [1981] and Fejer et al. [1991]. Figure 5 also shows that the reversal times in the Peruvian, Indian and African equatorial regions have similar annual variations but are shifted by 6 months (Kodaikanal) and 7 months (Ouagadougou, Ibadan, and Fortaleza). The reason for this shift is probably the location of Fortaleza (dip = 7.5°S), Ibadan (dip = 6°S), Ouagadougou (dip = 5.9°N), and Kodaikanal (dip = 3°N) stations in the southern and Northern hemisphere, respectively. In addition, Indian station is closer to dip equator than the African and Peruvian stations.

Figure 5.

Comparison between the Jicamarca F region vertical drift reversal times with the reversal times inferred from the time of h'F maximum (Kodaikanal and Fortaleza) and the time of hmF2 maximum (Ibadan and Ouagadougou) all during solar maximum. Jicamarca and Kodaikanal are after Fejer [1981], while Fortaleza data are after Fejer et al. [1991].

[22] Woodman [1970] reported that in the nighttime, F region drift velocities are downwards between 1900 and 0600 LT and that there is a sudden intense upward drift prior to the evening field reversal, notably in high solar activity periods. He further reported that Jicamarca backscatter radar shows that daytime drift in the F region is directed vertically upward with a value of about 20 m/s. At the Indian equatorial sector, Rajaram [1977] found that the nighttime vertical velocities are directed downwards, and are greater after midnight than before midnight, both during magnetically quiet and perturbed times; the vertical velocity does not vary with height above 280 km. Fejer and Scherliess [1997] reported that the equatorial (≤∣7.5°∣) satellite daytime upward drifts are about 20 m/s, and are in good agreement with the comparable Jicamarca radar data, particularly for equinoxes and December solstice, but the nighttime downward drifts are usually smaller than the radar results, particularly during June solstice. In the work of Fejer et al. [1989] they found that Jicamarca evening vertical drifts during the December solstice and equinoxes near solar maximum are noticeably larger than the ionosonde drifts inferred from observations over the nearly Huancayo Radio Observatory (HRO) (12.0°S, 75.3°W; magnetic dip 2°N). Rastogi et al. (1991) suggested that the Jicamarca evening drifts are about 30% larger than the Huancayo ionosonde drifts. Fejer et al. [1995] showed that the F region vertical drifts from AE-E satellite are in agreement with the Jicamarca data. Sastri et al. (1995) found fair agreement between these AE-E drifts for the Indian sector and the Kodaikanal (10.3°N, 77.5°E; dip 4°N) ionosonde drifts. Batista et al. [1996] reported significant differences between the AE-E evening drifts for the Brazilian sector and the Fortaleza, Brazil (4°S, 38°W; dip 7.5°S) ionosone drifts. Scherliess and Fejer [1999] reported that the global equatorial vertical drift model results are in good agreement with the Fortaleza ionosonde drifts observations, except for June solstice solar maximum conditions. Also, empirical model results are found to be in good agreement with the ionosonde results over Trivandrum, India; but underestimate the HF radar evening prereversal velocities by about 10 m/s. In an investigation carried out by Woodman et al. [2006] comparing ionosonde and incoherent scatter drift measurements at the magnetic equator, they concluded that although nighttime vertical and horizontal drifts deduced from Doppler and diffraction pattern velocities of the F region HF reflected waves have shown to be a good proxy of the vertical and horizontal drift velocities at high latitudes, this is not the case at equatorial latitudes. They showed that those measurements generally do not agree with incoherent scatter (ISR) measurements at latitudes close to the magnetic equator, except during very special conditions. As an exception, ionosonde vertical drifts are in fair agreement with ISR drifts around the prereversal peak times. On the other hand, ionosonde zonal drifts can be used when the F region height is structured and the E region density is low or stable. Going by this conclusion the results obtained in this study agree well with part of their findings. Balan et al. [1996] using the Sheffield University Plasmasphere Ionosphere Model (SUPIM) in an attempt to investigate the modulating effects of neutral winds demonstrated that the vertical velocities derived for the magnetic equator of Trivandrum (India) and Fortaleza (Brazil) from the ionospheric peak height (hmF2) are in good agreement with the velocities measured by the Incoherent Scatter Radar at Jicamarca as long a the ionospheric peak is within 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. They reported that the derived velocity for Trivandrum and Fortaleza corresponds to the velocity of the ionospheric peak, which varies from about 300 to 600 km altitude (since the two stations have almost identical hmF2 values. Obrou et al. [2003] found a positive correlation between hmF2 and the equatorial F region vertical drifts given by the model of Scherliess and Fejer [1999].

[23] Most recently, using the rate of change of h′F data at Ibadan, Nigeria, to estimate the nighttime F region vertical velocities; Oyekola et al. [2006] noted that Scherliess and Fejer [1999] global equatorial drift model results undervalue Ibadan ionosonde-derived (as Δ(h′F)/Δt) average evening prereversal velocities by about 9 m/s in all seasons. They also observed that the lowest value (35 m/s) of longitudinal variation of the evening peak velocity obtained using this model is in excellent agreement with ionosonde, HF Doppler radar, and Jicamarca radar observational data only during the equinoxes conditions. In addition, Oyekola [2006] presented seasonal vertical drift velocity values from the ionosonde observations at Ibadan as Δ(h′F)/Δt and compared them with Jicamarca Incoherent Scatter Radar and AE-E satellite F region vertical drift results during the nighttime sector. The author found a comparable variability patterns during periods of high F layer heights during equinoxes and the December solstice, and the contrary behavior occurs during June solstice. Nevertheless, it is particularly important to emphasize the differences between Oyekola [2006] and current analysis.

[24] In the present comparative studies, seasonal averages of vertical drift values estimated from ionosonde data as Δ(hmF2)/Δt are compared with the same data sets used by Oyekola [2006], specifically Jicamarca incoherent radar (ISR) and AE-E satellite F region vertical drift measurements. Hence, five main points are apparent. First, the ionosonde drift derived from Δ(h′F)/Δt reflects the vertical movement of the base of the F layer, while the ionosonde drift calculated as Δ(hmF2)/Δt reflects F layer peak motions in the vertical direction. Second, the observed results in the work of Oyekola [2006] are in the altitude range between ∼230–460 km, so that the average height to which they correspond can be taken to be ∼330 km. In contrast, the present analyses indicate a height range of ∼280–560 km (Ibadan) and ∼300–500 km (Ouagadougou). The average altitude to which they both correspond to can be taken to be ∼400 km. According to Fejer et al. [1991], Jicamarca vertical drift is studies in the height range 250 to 600 km, but the drift average is between 300 to 400 km altitude range, and so the average height to which they correspond can be taken to 350 km. On the other hand, AE-E measurements show an altitude range of ∼230–470 km. Here, the mean height is also ∼350 km. A careful look of the average height where the drift results correspond, one may note that drift estimated from the time variations of ionosonde h′F should gives a fairly better results than that inferred from time variations of ionosonde hmF2 when compared with Jicamarca radar and AE-E satellite model measurements. Third, the current work provide valuable information on the hemispherical disparity in the calculation of vertical ion drifts from ionosondes and thus offer a better understanding of the morphology and behavior of the drifts in the African longitudinal sector under similar solar and geophysical conditions. Fourth, the role of vertical drifts continue to emerge, required, and dominant agent for occurrence of equatorial and low-latitude electrodynamical phenomena; therefore, it is interesting to employ other observed ground-based parameter in deriving ionospheric E × B drifts. Before now, hmF2 parameter has been proposed for studying the dynamical feature of the F2 region [e.g., Buonsanto, 1991]. Finally, the results presented in this paper should stress the relationship between ionosonde, Incoherent Scatter Radar, and satellite F region ion drifts, particularly at night and at equatorial latitudes as better knowledge of ion drafts are extremely useful in the electrodynamics investigations.

[25] The data presented in this analysis have clearly shown that ionosonde drifts compare fairly well with comparable Jicamarca Incoherent Scatter Radar, Ion Drift Meter (IDM) on the AE-E satellite observations of F region vertical plasma drifts at other equatorial stations for the same quiet magnetic periods but under different solar activity period (F10.7 solar indices of about 160–180, 190, and 208).

[26] The quiet time vertical drifts of equatorial F region plasma are due to electric fields generated by the dynamo action of tidal winds in E region and thermospheric winds in F region. During the day the electric fields are determine primarily by the tidal E region dynamo because the fields due to F region dynamo are largely shorted out by the highly conducting E region. Beginning at sunset, the control of the electrodynamics passes from the E region to the F region due to the significant decrease in E region to the F region conductivity that allows the development of polarization electric fields of the F region dynamo. The numerical modeling studies have, in fact, highlighted the importance of the F region dynamo to the F region plasma motions at night, in particular to the postsunset enhancement of the upward vertical drift near the dip equator [e.g., Crain et al., 1993a; Fesen et al., 2000]. It should be noted that the relative efficiencies of the E and F region dynamos in generating electric fields change significantly throughout the day, and also with season, solar activity conditions, and with altitude [Fejer and Scherliess, 2001] and may well contribute to variations observed in the present study.

[27] A possible cause of serious substantial disparities in the observational techniques of F region vertical ion drifts may perhaps be due to altitudinal effects. Spatial variation, such as, dip latitude could also have some influence on the measured values of vertical E × B drift velocities. Moreover, a number of theoretical and model calculations have shown that transequatorial magnetic meridional wind intensity could play a critical role in the variability of ion drifts, especially around the prereversal hours [e.g., Maruyama, 1988; Mendillo et al., 1992; Abdu et al., 2006] since Ibadan and Ouagadougou are not located exactly at the magnetic dip equator.

[28] Over the past years, temporal, spatial, seasonal, solar flux, as well as longitudinal dependence of lower atmospheric forcing on the upper ionosphere parameters, particularly, vertical plasma drifts have produced plentiful literature of results [e.g., Chen, 1992; Forbes and Leveroni, 1992; Parish et al., 1994; Forbes et al., 2000; Kazimirouvsky, 2002; Immel et al., 2006; Lastovicka, 2006; Pancheva et al., 2006]. Forbes et al. [2000] using hourly foF2 data from over 100 ionosonde stations during 1967–1989 to quantify F region ionospheric variability, and to asses to what degree the observed variability may be attributed to various sources, i.e., solar ionizing flux, meteorological influences, and changing solar wind conditions and noted that the variability associated with day-to-day solar photon flux changes is small compared with that ascribed mainly to meteorological effects. Chen [1992] from the investigation of 2-day oscillation of the Equatorial Ionization Anomaly (EIA), concluded that the atmospheric planetary wave can modulate the tidal wind and cause the 2-day oscillation (or day-to-day variation) of the EIA and some related parameters such as ionospheric electric field, the Sq current, and the equatorial electrojet via the dynamo effect and fountain effect. The author reported that this is important coupling between the low-latitude ionosphere and the middle and lower atmosphere. Forbes and Leveroni [1992] presented evidence for the thermospheric penetration of a quasi 16-day oscillation excited in the winter stratosphere, and the generation of electric field, current and plasma density in the ionosphere. They found effects of the same order as other solar-geophysical variations of significance. Parish et al. [1994] suggested that planetary waves could make a significant contribution to the dynamics and electrodynamics of the lower ionosphere and thermosphere and that they could be capable of causing significant day-to-day variability. They also noted that solar flux variations may possibly have a strong influence on periods of the waves in the range of 10 to 20 days during years of high solar activity.

[29] Of recent, Pancheva et al. [2006] reported that the ionospheric F region response at low latitudes is very sensitive to the dynamo generated electric fields because of the special geometry of the magnetic fields. In this case the 2-day variability of the electric fields generated by the modulated tides is easily transferred to the upper ionosphere by producing 2-day variability of the vertical plasma drift. This variable plasma drift then generates 2-day variability in the maximum electron density of the F region. Immel et al. [2006] using the IMAGE satellite data found a clear connection between the tides and the longitudinal distribution of equatorial F region E × B upward drifts. Lastovicka [2006] highlighted that waves such as tides and planetary waves coming from below the ionosphere could effect the behavior of the ionosphere, and so the knowledge of the influence of the waves is very desirable for understanding the vertical coupling in the atmosphere/ionosphere system, for the energy budget of the ionosphere for ionosphere dynamics, and for predictions of the state of the ionosphere for communication and other purposes.

4. Conclusion

[30] The F region vertical ion drifts near the magnetic dip equatorial stations in Africa were made using the hourly values of the ionosonde hmF2 time variations, assumed that Δ(hmF2)/Δt is a realistic vertical ion drift during the nighttime conditions. There is a comparable variability patterns and trends in comparison with behavior of the seasonal averages seen by the Jicamarca, Peru, ground-based radar, and those seen by the AE-E satellite only around the prereversal peak maximum except for June solstice. The results also indicate some serious significant discrepancies between 2100 to 0500 LT sector. Examination of the magnitude of the seasonal disparities between ionosondes and Incoherent Scatter Radar drifts show a range of ∼8–11 m/s (December solstice) and 14–17 m/s (equinoxes) in contrast to ∼7–10 m/s (December solstice) and ∼9–12 m/s (equinoxes) determined by ionosondes and ion drift meter onboard the AE-E satellite. The drift values are roughly similar during the June solstice for all techniques. In order words, the magnitudes of Ibadan ionosonde and AE-E downward drifts are similar during the June solstice, but lower in values by either a factor of two or three than other methods in December solstice and equinoxes. While Ouagadougou ionosonde drifts are about 2 m/s lower in magnitude than AE-E June solstice data, but smaller in value by either a factor of three or four than other techniques for other seasons. The observations here indicate the unique position of ionospheric station in view of the fact that Ibadan and Ouagadougou are located almost at the same but opposite distance from the dip equator and have the same magnetic declination of about 3°W. In spite of these considerable disparities among these methods of F region plasma drift measurements, the apparent ion drifts derived as Δ(hmF2)/Δt can be considered representative of the F region vertical drift at local times around prereversal peak excluding June solstice. In this regard, the present data give direct confirmation of some past theoretical and experimental results [e.g., Batista et al., 1986; Woodman et al., 2006].

[31] Additionally, using hmF2 data at Ibadan and Ouagadougou, the threshold values essential to trigger equatorial nighttime F region irregularities at the African sector are in the order of 23.3 m/s and 370 km, 27.7 m/s and 389 km, 30.8 m/s and 433 km for prereversal E × B drift amplitude and the corresponding peak hmF2, respectively, for June solstice, December solstice, and equinoctial period at Ibadan. Whereas, at Ouagadougou, the values are around 14.0 m/s and 406 km, 15.9 m/s and 383 km, and 16.6 m/s and 401 km for prereversal peak vertical drift velocity and the peak hmF2, respectively, during June solstice, December solstice, and equinoxes. Also, the evening reversal times from upward to downward direction are in close agreement, apart from June solstice again, which exhibits large fluctuations. On the other hand, morning inversion times from downward to upward direction show small variations. The current database appear adequate to examine the relationship between ionsonde, radar, and satellite F region vertical ion drifts so that future study should produce a more complete link among the three techniques since ion drifts play a fundamental role in the electrodynamics of the equatorial ionosphere. Finally, our data are useful for global ionospheric modeling and for the predictions of development of nighttime equatorial F region irregularities at the African region where there is paucity of data.


[32] The authors wish to thank the referees for their assistance in evaluating this paper. Their comments and suggestions have drastically improved the paper. This work was supported by ONRG grant N00014-07-1-4019.