Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
 The Conjugate Point Equatorial Experiment (COPEX) campaign was carried out in Brazil, between October and December 2002, to study the conjugate nature of plasma bubble irregularities and to investigate their generation mechanisms, development characteristics, spatial-temporal distribution, and dynamics. In this work we will focus mainly on the zonal spaced GPS (1.575 GHz) and VHF (250 MHz) receivers' data collected simultaneously at two magnetic conjugate sites of the COPEX geometry: Boa Vista and Campo Grande. These GPS/VHF receivers were set up to detect the equatorial scintillations and to measure ionospheric scintillation pattern velocities. Then, the zonal irregularity drift velocities were estimated by applying a methodology that corrects the effects caused by vertical drifts and geometrical factors. The results reveal the coexistence of kilometer- (VHF) and hundred-meter-scale (GPS L-band) irregularities into the underlying depletion structure. Over the conjugate site of Campo Grande, the average zonal velocity at VHF seems to be consistently larger than the estimated GPS velocities until ∼0200 UT, whereas over Boa Vista the irregularities detected from both techniques are drifting with comparable velocities. The hundred-meter-scale structures causing L-band scintillations appear to be drifting with comparable velocities over both the conjugate sites, whereas the kilometer-scale structures are drifting over Campo Grande with larger average velocities (before 0300 UT). Complementary data of ionospheric parameters scaled from collocated digital ionosondes are used in the analysis to explain differences/similarities on the scintillation/zonal drift results.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 Equatorial ionospheric spread F/plasma bubble (ESF/EPB) irregularities and their zonal drift velocities have been extensively studied in the past several decades by means of different experimental techniques. Depending on the observational technique and the radiofrequency used to monitor the ionosphere, the variability, structure, and dynamics of the irregularities with different scale sizes can be probed by different instruments. As for the most recent technologies of the Global Navigation Satellite System (GNSS) applications, such as the Global Positioning System (GPS), single and spaced antenna systems have been used to monitor, respectively, the activity and dynamics of the irregularities that cause fluctuations/scintillations in the L-band signals [Pi et al., 1997; Beach and Kintner, 1999; Kil et al., 2000; de Paula et al., 2002, 2007; Ledvina et al., 2004; Kintner et al., 2004; Muella et al., 2009].
 It is well known that the rapid increase of the eastward electric fields during the postsunset hours (and the bottomside instabilities generated by Rayleigh-Taylor plasma mechanism) is the main prerequisite (and source) responsible for the generation (triggering) of ESF/EPB irregularities. According to Haerendel , the primary instabilities can be followed by other secondary plasma instabilities that can lead to a cascading process and originate a wide spectrum of irregularity structures with different scale sizes and characteristics. Basu et al.  have shown, on the basis of multifrequency observations that irregularities with scale sizes varying from several meters to ∼1 km may coexist during the generation/growth phase of EPB irregularities. Depending on the apex height reached above the magnetic equator, the large depleted density bubbles can extend/expand along the magnetic field lines to F-region heights over latitudes off the equator. Then, the extremities “feet” of the bubbles intersect the crests of the equatorial ionization anomaly (EIA) at off-equatorial latitudes, a region of increased background electron density and sharp density gradients [Muella et al., 2008b, 2010]. The presence of these gradients give rise, inside the bubbles, to conditions favorable for the generation of irregularities with different scale sizes [Rao et al., 2006]. According to Muralikrishna , small-scale irregularities associated with large-scale depletions can be generated in a region of downward electron density gradients since the ambient electric field is also downward. Therefore, the cross-field instability is possibly a causative secondary mechanism for the generation of small-scale plasma irregularities.
 It is known that EPBs are large-scale plasma-depleted regions aligned along the magnetic flux tubes that may extend beyond the poleward edge of the equatorial anomaly to dip latitudes ranging up 20° from the magnetic equator [Sahai et al., 1981]. This feature allows the investigation from coordinated observations at conjugate sites of important aspects of the equatorial plasma bubbles, such as their development characteristics and connections with sporadic E-layers, symmetry/asymmetry conditions, the role of the evening vertical drift (transequatorial/meridional wind) in their development (suppression), day-to-day variability, latitudinal/temporal dependence of their zonal drift velocity, etc. For a better understanding of many of these aspects, the Conjugate Point Equatorial Experiment (COPEX) campaign of ∼70 days duration was conceived and carried out at four observation sites in Brazil from October to December 2002. The COPEX campaign combined the efforts of many scientific groups (e.g., National Institute for Space Research/INPE, Brazil; Air Force Research Laboratory, USA; University of Massachusetts Lowel; Brazilian Air Force; and others), and the data collected during the observation period were used to validate the Communication/Navigation Outage Forecasting System (C/NOFS) satellite [de La Beaujardière et al., 2004] prior to its launching in April 2008. The scientific objectives of the COPEX campaign and other relevant results published elsewhere using different instruments and techniques can be found in the work of Abdu et al. , McDougall et al. , Sobral et al. , Batista et al. , McNamara et al. , Muella et al. [2008b], and Reinisch et al. .
 In this paper, we present for the first time COPEX observations of simultaneous VHF (250 MHz) and GPS L-band (1.575 GHz) scintillations, which are sensitive to irregularities with different scale sizes (VHF is sensitive to ∼1 km and GPS L1 frequency is sensitive to ∼360–400 m for an irregularity layer at 350 km). The scintillations are known to be caused by irregularities embedded in the large-scale plasma bubble structures, and their eastward drift can be derived from ground-based GPS/VHF systems. Figure 1 shows the COPEX location and geometry on the central Brazilian territory. The configuration of the COPEX sites provides a unique opportunity to study the conjugate nature of the irregularities and the temporal/latitudinal dependence of their zonal drift velocities in a region of large magnetic declination. Two of the four COPEX sites were located at conjugate points, Boa Vista (2.8°N; 60.7°W; dip angle 22.05°N) and Campo Grande (20.5°S; 54.7°W; dip angle 22.32°S), and two near the magnetic equator, Cachimbo (9.5°S; 54.8°W; dip angle 4.25°S) and Alta Floresta (9.7°S; 56.0°W; dip angle 3.38°S) (not shown in Figure 1). The observations that we present here are intended to contribute to a better understanding of certain aspects of low-latitude ionospheric dynamics at conjugate points, such as, evolution and characteristics of different irregularity scale sizes, their variability, and temporal and latitudinal dependence. The main objectives of this paper are to present a comparative study of scintillations and zonal drifts observed at GPS L1 and VHF frequencies, and to explain differences/similarities on the basis of ionospheric and irregularity characteristics. Complementary ionospheric data from collocated digital ionosondes are used in the present study.
 The remaining sections of this work are as follows: section 2 presents the instrumentation and methodology to measure the scintillations and to estimate the irregularity drifts. Section 3 presents the results and discussions on the scintillation and zonal drift velocities monitored by ground-based GPS and VHF receivers. Section 4 presents the remarks and conclusions.
2. Experimental Techniques
2.1. VHF Spaced Antenna System
 The VHF system used during the COPEX campaign was developed to monitor scintillations in the frequency range of 240–260 MHz from radio beacon signals received from geostationary satellites FLEETSAT-7 (longitude 100°W) and FLEETSAT-8 (longitude 23oW). Each VHF system consists of RF and IF converters, one dedicated computer and a multiplex module for data processing, and a set of four Yagi antennas. The antennas are installed in a magnetically spaced configuration: two in the east direction and two in the west direction. A pair of antennas (CH 1–2) is separated by a distance of 100 m in the magnetic east-west direction and set up to detect signals from satellite FLEETSAT-7 (westerly link). The other pair (CH 3–4) of antennas with the same separation detects signals from satellite FLEETSAT-8 (easterly link). The characteristics of the signal paths from the two geostationary satellites are summarized in Table 1. It is noted that the signal raypath of the westerly link has a lower elevation angle than that of the easterly link. For this reason we will analyze in this study only the scintillations measured from easterly link FLEETSAT-8 satellite whose elevation angle is above 46° in all COPEX sites, which reduces the effects of multipath and tropospheric scattering on the scintillation measurements. The equatorial site of the VHF receiver was located at Alta Floresta, which is separated by a distance of only ∼130 km from Cachimbo.
Table 1. Azimuth and Elevation of the Signal Paths as Observed at Three of the COPEX Sites
Channel 1–2 (West)
Channel 3–4 (East)
 The signal power data samples (100 Hz) recorded by the VHF data acquisition system are digitalized and used to compute accurate values of S4 scintillation index for every 81.92 s. The S4 index is the most used parameter to study amplitude scintillations due to radio wave diffractions caused by Fresnel scale irregularities embedded in background plasma of large density, and is defined as the normalized standard deviation of the received signal power intensity. From the S4 records we show in this report the scintillation activity over all the COPEX sites throughout the campaign period. Then, for the nights with simultaneous VHF and GPS scintillation data, the morphology of the irregularities is discussed in terms of their intensity, onset time, and time duration.
 Zonal drift velocities for the nights with simultaneous scintillation records over the three conjugate sites are also estimated every 81.92 s using the scintillation-based spaced-antenna technique. Assuming frozen-in irregularity structures, the zonal scintillation pattern velocity is calculated from the time lag of the maximum cross-correlation of the scintillation pattern detected on the ground by the two spaced antenna (easterly link). However, it is known that, depending on the orientation of the signal path and the localization of the ground receiver, this scintillation pattern velocity (Vscint) may have contributions of the vertical drift of the irregularities and, consequently, may be biased in relation to the zonal irregularity velocity at the ionospheric puncturing point (IPP) [Ledvina et al., 2004], where
Here Vion is the zonal irregularity drift (positive eastward), VZ is the contribution due to the vertical drift of the irregularities (positive upward), ψdip is the magnetic dip angle at the origin of the receiver coordinate system, ϕ is the satellite zenith angle, and θ is the azimuth angle of the signal path obtained in the receiver coordinate system and measured clockwise with respect to the magnetic north. The contribution of the drift in the meridional direction can be assumed to be negligible because of the extended nature of the irregularities along the magnetic field line [Kintner et al., 2004]. Thus, for each COPEX site we estimate the zonal irregularity drift velocities (Vion) from the scintillation pattern velocity (Vscint) by applying the cross-correlation technique and equation (1). In order to improve the validity of the assumptions in the drift estimations we employ a cross-correlation threshold (ci) ≥ 0.9 and a scintillation level threshold of (S4) ≥ 0.25. The 0.25 threshold for the S4 index calculated from the VHF measurements can be considered to be above the level of noise and multipath effects. Thus, by considering high threshold limit of 0.9 for peak cross-correlation and well-developed 250 MHz scintillations (i.e., S4 > 0.25), we can ensure that the corrections to the estimated scintillation pattern velocity (Vscint) required to yield the “true velocity” can be considered negligible in our study (R. Livingston, personal communication, 2010). Moreover, because the focus of the present study does not concentrate on small differences in velocity at plasma bubble onset or other short periods of non-stationary scintillation time series, the time lag separation method in estimating Vscint can be considered appropriate.
 However, to correct the contribution of the vertical irregularity drift, an approximation is made by computing the time derivative of the F-region base height [VZ = d(h′F)/dt] at 4 MHz return signal (close to the true height) scaled from the ionograms recorded at 10 min intervals. The contribution of VZ is expected to be significant during the growth phase of the irregularities in the early evening owing to the electrodynamical E × B drift; however, the present method may underestimate the vertical plasma drift because of the effects of plasma recombination (for low F-layer heights (< 300 km)) and meridional winds (depending on their orientation). The signal path from the geostationary satellite FLEETSAT-8 to the VHF receivers punctures the ionosphere, at 350 km, in a sector of the sky that spans horizontally from ∼303 to 312 km east of the overhead ionosphere seen by the digital ionosondes at the conjugate stations. We assume that the ambient ionospheric conditions in estimating VZ do not change substantially over this distance. There should be ∼10–11 min delay in the onset of ESF/EPB over Boa Vista and Campo Grande (as seen by the ionosondes) compared with that along the signal path to the VHF receivers. The subionospheric coordinates of the signal paths from the geostationary satellite FLEETSAT-8 intersecting the ionosphere at 350 km altitude, to the VHF receivers at Boa Vista and Campo Grande, are shown in Table 2. The 350 km ionospheric points of FLEETSAT-8 map up along the field line to a height of ∼620 km above the magnetic equator. Thus, scintillations observed at the conjugate sites of Boa Vista and Campo Grande correspond to equatorial bubbles at or above this altitude over the equatorial stations of Cachimbo/Alta Floresta. For this reason the subionospheric coordinates of the signal received at Alta Floresta is calculated for an ionospheric intersection point at 620 km. This corresponds to an horizontal distance of ∼332 km and an approximately 12 min delay in the onset of ESF/EPB overhead Cachimbo compared with that along the signal path to Alta Floresta.
Table 2. Coordinates of the Subionospheric Points From Signals of Satellite FLEETSAT-8 to the Receivers Located at Three of the COPEX Sites
Satellite-to-Receiver Signal Path
ϕ350 = 2.62°N
λ350 = 57.9°W
ϕ620 = 8.65°S
λ620 = 51.9°W
ϕ350 = 19.13°S
λ350 = 52.2°W
2.2. GPS Scintillation Monitoring System
 During the COPEX campaign, two ground-based GPS receivers specially modified (SCINTillation MONitor-SCINTMON) to detect scintillations at L1 frequency (1.575 GHz) were installed (at each COPEX site) in a magnetically spaced configuration, and separated by a distance of 100 m along an east-west baseline. The SCINTMON receivers are based on GEC Plessey development system and consist of a 16-bit PC card making signal strength data recordings at every 20 ms. For each 1 min interval (3000 data samples) the signal power amplitude data are averaged and the S4 scintillation index is computed using postprocessing software (for more details, see Beach and Kintner ).
 As mentioned before, for the nights with simultaneous GPS and VHF scintillation data over the COPEX sites, the morphology of the irregularities is studied in terms of their intensity, onset time, and time duration. For comparison with the VHF scintillations we used data from GPS satellites with elevation angle higher than 40°.
where Vscint is the zonal velocity in the receiver coordinated frame calculated by measuring the time lag of the maximum cross-correlation of the amplitude scintillation patterns between the two-spaced receivers; hsat and hion denote, respectively, the satellite height above the receivers' horizontal plane and the mean scattering height of the diffraction-causing irregularities; VZis the vertical irregularity velocity; vsatx, vsaty, and vsatzare the zonal, meridional, and vertical components of the satellite velocity, respectively; and (qy/qx) and (qz/qx) are the mapping factors of the scintillation pattern (and its velocity) that rotate all vectors into the local magnetic frame. The mapping factors are defined as [Ledvina et al., 2004]
where Bz/By and Bx/By are ratios of the components of the magnetic field vectors at the ionospheric puncture point, ϕ is the satellite zenith angle, and θ is the satellite azimuth angle. In equation (2)Vscint and VZare obtained as in equation (1), hsat and sat are determined from satellite-transmitted ephemerides, hion is considered as 350 km over the conjugate stations of Boa Vista and Campo Grande (as obtained from the mean values of hmF2), and the components of the magnetic field vector in the mapping factors qy/qxand qz/qxare determined at altitude hion from the International Geomagnetic Reference Field (IGRF) model. It is noted from equation (2) that, differently from equation (1) for VHF, the zonal drift estimates are height-sensitive.
 Equation (2) is an extension of the method used by Kil et al. , de Paula et al. , and Muella et al. [2008b] for a more precise estimation of the zonal irregularity drift velocities by including geometrical effects. Muella et al.  used equation (2) to estimate de zonal drifts at the equatorial station of São Luís (2.33°S, 44.21°W, dip latitude 1.3°S). However, in their estimations the vertical irregularity velocity could be neglected because of the numerous scintillating satellites signals used to average the scintillation pattern velocity over a period of several days and times [e.g., Kintner and Ledvina, 2005]. Averaging tends to reduce for quite small values the term tan(ϕ) in the mapping factors (qy/qx) and (qz/qx), which in turn minimizes the effects of the vertical irregularity velocity in equation (2) [Ledvina et al., 2004]. This approximation could also be used in the present study, but only those satellites whose signals are puncturing the ionosphere in the east sector of the sky (azimuth angle varying between 0° and 180°) above the stations were used in the zonal drift estimations. Since we limited our measurements to the easterly scintillating GPS signals, the number of satellites available declines, thus it seems appropriate to account in the estimations the contributions due to VZ. From previous observations [Kil et al., 2000; Kintner et al., 2004] it has been noted for the GPS L1 signal that the contributions in Vscint due to the random changes in the scintillation pattern is small (4% or less) and, thus, we can approximate the mean apparent velocities (Vscint) to the true velocities. In the case of GPS-derived zonal drifts we employ a cross-correlation threshold (ci) ≥ 0.9 and a scintillation level threshold of (S4) ≥ 0.2. In addition, with the threshold limit of 0.9 for the peak cross-correlation, apparent and true velocities do not differ significantly and can be considered approximately the same. The zonal irregularity drift velocities estimated from the GPS scintillation measurements are then compared with those inferred from the VHF scintillation data at both conjugate sites.
3. Results and Discussions
3.1. Simultaneous VHF and GPS L1 Scintillation Activity
 The contour plots in Figures 2 and 3 depict the amplitude scintillation activity as detected, respectively, by the GPS and VHF receivers installed at the COPEX observation stations. The results for each station are presented in three panels, each panel representing the daily observations available during the months throughout the campaign period. The time scale in universal time (UT) refers to the nocturnal period, where UT = LT + 4 hours. The color scale bars indicate the S4 index levels used in the contour plots and the contours within the dashed lines indicate the absence of data. For the GPS the contours in Figure 2 represent the maximum scintillation index S4 calculated for a GPS satellite with elevation angle >40°. It can be observed in Figure 2 that the scintillation intensity is lower at the equatorial station of Cachimbo and tends to increase toward the conjugate sites of Boa Vista and Campo Grande. The lower scintillation levels over the dip equatorial station do not necessarily mean that the plasma bubble depletions are absent or weaker at the dip equator than at off-equatorial latitudes, but lie in the fact that for the GPS L-band frequencies for the strong scintillations to occur requires a significant background electron density [Makela, 2006]. In other words, the strength of the scintillations depends essentially on the integrated electron density deviation (ΔN), which is obtained from the product of the irregularity amplitude (ΔN/N) and the background plasma density (N). The amplitude of the irregularity may be the same along the magnetic field lines but the background electron density varies with latitude. Thus, high scintillation levels could be expected to occur more frequently at off-equatorial latitudes in the region of the crests of the equatorial ionization anomaly (EIA). However, results of Muella et al.  suggest that at latitudes of the EIA the most intense scintillations do not occur exactly at the crests, where the electron density is larger, but are more prone to be observed where sharp electron density gradients exist, that is, at the inner/outer edges of the enhanced anomaly crests.
Figure 2 also shows that except in the last campaign days during December, the GPS scintillation activity at the conjugate station of Campo Grande is somewhat more intense than over Boa Vista. Considering the elongated nature of the plasma bubble structures along the geomagnetic field lines, one should expect a strong correspondence among the scintillations observed at both conjugate sites. However, Figure 2 reveals asymmetrical features in the intensity and latitudinal distribution of the scintillations. It is known that large north-south asymmetry of the EIA, during the generation stage of the irregularities, reduces the instability growth rate through its control on the F-layer bottomside upward density gradients and the magnetic flux tube integrated conductivities [Maruyama, 1988; Abdu, 1997]. Otherwise, when the asymmetries in the ambient ionization at latitudes adjacent to the regions of the EIA arise during the developed phase of the irregularities, this could be a possible cause for a corresponding asymmetry in the scintillation activity at magnetic conjugate points. Asymmetries in the electron density distribution in the regions of the EIA can be attributed to the action of transequatorial meridional neutral winds. Thus, the role of horizontal neutral winds is to modify the plasma surrounding the crests of the anomaly and to affect the location, width, and magnitude of the plasma density gradients. Because the occurrence of equatorial scintillations tends to be mostly confined within the boundaries of the anomaly, asymmetries in the ambient ionization and in the electron density gradients at latitudes adjacent to the regions of the EIA could be a possible cause for the corresponding scintillation asymmetries at the conjugate sites of Boa Vista and Campo Grande. More recently, Abdu et al.  designated such asymmetries of ionization and gradients over conjugate sites as possible causes for the corresponding asymmetries observed in the echo intensities of digital ionosonde data. The magnetic conjugate stations of Boa Vista and Campo Grande are not expected to be located exactly under the crests of the EIA, but under the equatorward edge of the anomaly peaks.
 Despite the lack of data during some days, it may be noted from the simultaneous VHF scintillation data (S4 index) in Figure 3 that, over Campo Grande, the VHF scintillation activity is also somewhat more intense than over Boa Vista. Figure 3 shows that the VHF scintillations in the conjugate stations do not differ significantly from each other in terms of intensity until ∼0300 UT. However, it seems that over Campo Grande the scintillations tend to cease later and remain until ∼0600 UT on some days, whereas over Boa Vista the scintillations are more concentrated in the premidnight hours (until 0400 UT). Figure 3 also shows that until ∼0100 UT the magnitude of the S4 index observed at the equatorial station of Alta Floresta does not differ significantly from that observed at the magnetic conjugate stations.
 The parameter foF2 scaled from ionograms recorded during the COPEX campaign can be used to evaluate the presence of asymmetries in the electron density distribution at the conjugate sites (foF2 is proportional to the square root of NmF2) and how it could be related with the scintillation asymmetries. McNamara et al.  analyzed 30 days of foF2 obtained during the COPEX campaign and reported fairly good agreement between Boa Vista and Campo Grande during the evening hours. However, after 0400 UT (postmidnight hours) they observed from the mean variations of foF2 that the collapse in the peak electron density occurs later in Campo Grande than in Boa Vista. As evidenced mainly from VHF data, this fact may explain, in part, the reason for the scintillations over Campo Grande to cease later than over Boa Vista. Another aspect that may help to understand the reason for the scintillations at GPS L-band and VHF frequencies to be more intense in Campo Grande than over the northern conjugate site of Boa Vista is presented in Figure 4. In Figure 4, mass plots of vertical drifts (VZ) calculated from the time rate of change of the F-layer true heights [d(h′F)/dt] for all days of the COPEX campaign is depicted for the three stations (see Abdu et al.  for more details). The validity of the present method to determine the vertical drift of the evening equatorial ionosphere, in comparison with the measurement by other techniques, has been well documented in the literature [see, e.g., Bertoni et al., 2006; Abdu et al., 2006]. The vertical drift velocities calculated at each station are plotted versus UT and LT, and the time interval shown in Figure 4 emphasizes the evening prereversal enhancement (PRE) in VZ. The solid curves in the plots denote mean VZ. We call attention in Figure 4 to the asymmetry in the PRE amplitude between the two conjugate sites, where the amplitude of the PRE over Campo Grande is larger (in ∼6 ms−1) than over Boa Vista. According to Abdu et al.  such asymmetry in the PRE amplitude over the conjugate sites is due to the presence of a transequatorial neutral wind blowing northward. A large vertical drift over Campo Grande is displacing the ionosphere at F-layer heights to comparatively high altitudes and, probably, it is modifying the local ionosphere in such a way that creates more conditions favorable to the triggering of secondary instability process. In addition, if plasma is moved upward to heights of decreased chemical losses, it contributes to an enhancement of electron density deviation of the irregularity and, consequently, may increase the strength of scintillations.
 With the availability of simultaneous data of amplitude scintillation index (S4), an attempt is made to compare the characteristics of the scintillations registered from the collocated GPS and VHF receivers at the COPEX sites. The results shown in Figures 2 and 3 present a high degree of correspondence in terms of morphological features of the scintillation pattern detected by the two instruments. A careful observation of the scintillation activity among the two conjugate stations and from the two different techniques shows nearly symmetrical features in the latitudinal extension of the scintillations during the first evening hours. Thus, it reveals the coexistence of kilometer- and hundred-meter-scale irregularities in the underlying depletion structure. Another noticeable aspect seems to be related to the intensity of the scintillations. Figures 2 and 3 clearly show that scintillations at VHF frequency are more intense than at GPS L1 frequency. While between 2300 and 0200 UT the scintillation index S4 at VHF is mostly within 0.7–1.1, the magnitude of the S4 index at GPS L-band frequency ranges from 0.3 to 0.7. It can also be seen from Figures 2 and 3 that the hundred-meter-scale structures responsible for the scintillations at GPS L1 frequency appear to decay faster than the kilometer-scale structures that provoke scintillations at VHF. Thus, while the kilometer-scale structures are still visible after 0400 UT (postmidnight hours), the smaller scale structures causing the L-band amplitude scintillations are practically absent. Again, we call attention to the fact that, at the VHF frequencies, the importance of the background electron density seems to be comparatively lower than at the high L-band frequencies. This is clearly noted by comparing the scintillation activity detected close to the dip equator by the GPS receiver at Cachimbo, and that from the VHF receiver at Alta Floresta. Although the magnitude of the S4 index from the GPS L-band scintillations at Cachimbo rarely exceeds 0.4, the VHF scintillations at Alta Floresta are consistently higher and may assume saturated values of S4 index. Moreover, the difference in the scintillation intensity between the magnetic conjugate points and the dip equator are more drastic for the GPS L1 frequency than for the VHF frequency, which also indicates the relative importance of the background electron density in causing the amplitude fluctuations at L-band.
 The S4 level that is above the noise and multipath effects is the other important point to be noted from the comparison of the scintillations registered by the GPS and VHF receivers. The noise level of a receiver is critical to how it calculates S4 and this includes several factors: The receiver's thermal noise, the antenna front to back gain ratios, the preamplification in the antenna, the length of the cable connecting the antenna to the receiver, the loss in the cables, and other concerns related to the receiver design such as the bandwidth of the tracking loop, the type of the control system (whether it is phase-locked-loop or frequency-locked-loop), and in the case of analog to digital (A/D) conversion, the receiver needs the use of an automatic gain control (AGC) circuit to compensate for variations in RF signals strength (P. M. Kintner, personal communication, 2008). The multipath effect is probably the most critical and difficult to quantify because it varies over a wide range depending on satellite position, obstructions such as buildings, trees, hills, and the surrounding terrain. Multipath occurs when both direct and reflected signals reach the antenna causing signal variations on short time scales. Noise and multipath effects in general produce very weak scintillations. For the GPS receivers installed at the COPEX sites a S4 level above 0.2 can be considered to overcome the noise and multipath effects. However, for the VHF receivers the scintillation intensity shown in Figure 3 revealed, for all stations, a background blue contour during the time of no scintillation occurrence, which suggests that the S4 level of noise and multipath effect is between 0.18 and 0.36. We analyzed the VHF scintillation data and we found that a S4 threshold of 0.25 can be considered, for all practical purposes, to be above the noise of level and multipath. These threshold levels of 0.2 for the GPS scintillation data and 0.25 for the VHF scintillation data have been used in the zonal drift estimations presented in section 3.2.
 At VHF frequencies the occurrence of scintillations may be due to two types of irregularities, the EPB irregularities and the bottomside sinusoidal (BSS) irregularities. While the former maximizes during postsunset hours and can extend up to the topside F-region, the BSS irregularities according to Valladares et al.  are observed mainly during postmidnight hours in a narrow altitude range about the bottomside F region. Another difference between the two types of irregularities is related to their horizontal expansion. The BSS irregularities are in general confined to the equatorial belt of ±12° magnetic latitude and may extend up to 7500 km in the east-west direction [Valladares et al., 1983], whereas the EPB irregularities extend to approximately ±20° from the dip equator and from tens to hundreds of kilometers in the east-west direction [Schunk and Demars, 2003]. Fortunately, from the amplitude scintillation pattern signatures, recorded by the VHF receivers, it is usually possible to differentiate the fluctuations caused by each type of irregularity. The scintillations due to EPBs in general present an abrupt onset and descent with high magnitude of S4 during postsunset and premidnight hours. On the other hand, the scintillations due to BSS irregularities show a continuous patch that exists for several hours (more than 6 hours) around postmidnight with moderate to high intensity, and with quite a slow conclusion of the scintillation event [Basu et al., 1986]. The L-band frequencies have the advantage to be “invisible” to the BSS irregularities and, thereby, the GPS scintillation records can be used to identify the similar plasma bubble structures sampled simultaneously by the VHF receiver. Different signatures of these irregularities can also be found from simultaneous observations with ionospheric sounders. For example, EPB irregularities manifest in the form of range-type spread-F on ionograms, whereas BSS irregularities cause frequency-type spread-F [Rao et al., 2006]. The association of VHF scintillations with spread-F on ionograms and the symmetry/asymmetry characteristics of BSS irregularities during the COPEX campaign are not part of the scope of the present study, but deserve further investigation in a future work.
 In terms of the time onset of the scintillations we have noticed from the VHF data that it may occur ∼5–10 min earlier in Alta Floresta than at the conjugate stations of Boa Vista and Campo Grande. This time delay in the onset of the scintillations for the conjugate stations is due to the time required for the large-scale bubbles, generated at the dip equator, to reach the apex height that maps along the magnetic field line to F-region heights over off-equatorial latitudes. Thus, a day-to-day variability in the time onset of the scintillations can be expected to occur as result of the daily variations in the polarization electric fields causing the vertical movement of the equatorial irregularities. A difference in the time onset of the scintillations compared from the GPS and VHF data has been found to vary mostly between 10 and 20 min at the conjugate sites. It is observed that the onset of the scintillations appears first at VHF. The frequency dependence on the onset of scintillations at VHF and L-band frequencies is known to be very small [Franke et al., 1984]; thus, such differences can probably be due to geometrical factors, for which scintillations depend upon the position of the receiver relative to the irregularities in the ionosphere.
3.2. VHF and GPS Zonal Irregularity Drift Estimations
 The zonal irregularity drift velocities discussed in this report are based on 11 geomagnetically quiet nights (Kp < 3) of simultaneous VHF and GPS scintillation-based spaced-antenna measurements. Only those amplitude scintillation envelopes having a similarly distinguishing structure at both VHF and GPS L-band data are considered in the zonal drift estimations. The top panels in Figure 5 show the mean vertical drift velocity (VZ) calculated from 4 MHz ionospheric true height data (for the 11 selected nights) and used to correct the contribution caused by the vertical movement in the zonal drift estimations over Boa Vista (Figure 5, top left) and Campo Grande (Figure 5, top right). Figure 5 (middle and bottom) shows the mean zonal drift velocity of the scintillation pattern (Vscint) on the receiver plane and the mean zonal irregularity drift (Vion) at IPP calculated, respectively, for the GPS L-band and VHF frequencies. It is seen from Figure 5 that during the hours of the prereversal enhancement no drifts were calculated from the scintillation data. During this time period the scintillations are weak and the cross-correlation index (ci) tends to be small because of the difference of the signal to noise ratio detected at the two antennas. After 2300 UT the first zonal drifts are estimated and until ∼0100 UT the largest contributions of VZin the scintillation pattern velocities (Vscint) are found. In the station of Boa Vista during this time interval the largest difference between Vscint and Vion observed from GPS calculations is 21 ms−1 and from VHF data is 39 ms−1. For the northern conjugate station of Campo Grande they are 26.5 ms−1 and 38.9 ms−1 for the GPS and VHF calculations, respectively. The overall magnitude of the zonal irregularity drift velocities (Vion) at Boa Vista for the GPS and VHF frequencies are 4.3 ms−1 and 6.3 ms−1 larger than the zonal scintillation pattern velocity (Vscint), respectively. Whereas at the conjugate site of Campo Grande they are 4.8 ms−1 and 3.2 ms−1, respectively, for the GPS and VHF zonal drift estimations.
Figure 6 presents a comparison of the averaged (30 min bins) zonal drift velocities estimated between the magnetic conjugate sites of Boa Vista and Campo Grande. It is observed that for both GPS (Figure 6, top) and VHF (Figure 6, bottom) measurements, the differences in the zonal drift velocities estimated at the magnetic conjugate sites are mostly within the standard deviation bars. In the case of GPS measurements Figure 6 (top) shows that the Fresnel-scale (hundred meters) irregularity structures at Boa Vista and Campo Grande are drifting eastward with comparable velocities. The average GPS velocity estimated over Campo Grande is ∼6.5% larger than over Boa Vista. In the case of VHF measurements (Figure 6, bottom), even though the σ bars overlap most of the time, it is possible to see that the averaged velocities are somewhat larger over the southern conjugate station of Campo Grande than in Boa Vista. A nighttime average velocity ∼12% larger is found in Campo Grande than over Boa Vista, whereas before 0300 UT such differences may assume values of ∼18%. One plausible explanation for this finding could be attributed to the differences in the magnitude of the total magnetic field intensity (B). It is known from published results that the zonal drifts of the well-developed irregularities can be considered to be equivalent to the ambient zonal plasma drift velocities [Terra et al., 2004; Arruda et al., 2006; Sobral et al., 2009]. Thus, the irregularity drifts estimated in this report can be used as tracers of the ambient plasma. According to Haerendel et al. , the resultant zonal velocity (Vs) of the plasma at F-region heights equals −Eq/B, where Eq denotes the vertical electric field at any given geomagnetic (dipole) latitude and B is the intensity of the total magnetic field, in which both are not invariant along the magnetic flux tubes. Recently, Sobral et al.  showed from theoretical model that the zonal velocities of the ambient plasma using IGRF magnetic field values (at 250 km height) are ∼24% larger over Campo Grande than those calculated over Boa Vista. The conjugate site of Campo Grande is close to the South Atlantic magnetic anomaly region (SAMA) where the intensity of B is lower than that at the northern conjugate point over Boa Vista. For example, according to the IGRF model the total magnetic field intensity at 350 km over Campo Grande (20,000 nT) is ∼20% lower than over Boa Vista (25,150 nT). Thus, as a result of the low magnetic field intensity, a large zonal plasma drift may exist over the South Atlantic hemisphere and, consequently, a comparative larger irregularity zonal velocity could be expected at the southern conjugate station of Campo Grande.
 A comparison of the zonal irregularity drift velocities derived from GPS and VHF observations at each conjugate site is shown in Figure 7. The estimated zonal irregularity drift velocities (Vion) during the 11 nights are averaged for each 30 min time bins and plotted as function of universal and local time. The vertical bars indicate the standard deviations (σ) of the velocities. It is seen from Figure 7 that the differences in the zonal drift velocities at Boa Vista estimated from GPS and VHF data (Figure 7, top) are mostly within the deviation bars and the magnitude of the velocities vary between ∼100 and 185 ms−1. The average zonal velocity estimated from VHF technique is ∼4.75% larger than the average GPS velocity. Thereby, over Boa Vista the differences in the zonal velocities from the two techniques are not significant. Otherwise, in the southern conjugate station of Campo Grande (Figure 7, bottom), even though the σ bars from both techniques overlap most of the time, it is clear from Figure 7 that before 0200 UT the average velocities estimated from VHF data are larger than the GPS zonal drift estimations. While at Campo Grande a maximum magnitude of ∼185 ms−1 is computed from GPS measurements at ∼0200 UT, in the case of VHF a maximum of ∼225 ms−1 is observed around the same time. The average nighttime zonal drift velocity estimated from VHF technique is ∼10% larger than the average GPS drift velocity, however, before 0300 UT the average VHF velocities may attain values of ∼19% larger than the GPS-estimated velocities. From the comparison of the velocities estimated at the two frequencies it is also found that the zonal drifts show a tendency to increase slowly until their peak value between ∼0100 and 0200 UT (before local midnight). After that the velocities start to abate until their minimum magnitudes at ∼0400–0500 UT when the large variations due to the electric fields (that cause large vertical drifts) become relatively small, which is consistent with the results reported previously by Muella et al. [2008b] and Sobral et al. . The discussions on the differences of the zonal drift velocities attributed to the possible effects of low magnetic field intensity at latitudes of Campo Grande are based only on the analysis of the average magnitude of the estimated zonal velocities. However, as reported from previous theoretical and experimental studies [e.g., Haerendel et al., 1992; Arruda et al., 2006; Sobral et al., 2009] the lower local total magnetic field intensity will result in larger plasma and irregularity zonal drift velocities.
 By comparing the results from L-band and VHF frequencies in terms of the scale size of the scintillation-producing irregularity density structures, the results for the magnetic conjugate stations suggest that the hundred-meter and the kilometer-scale irregularities can be considered drifting eastward at comparable velocities. Whether at Boa Vista the VHF and GPS drift velocities are fitting very well, in the conjugate station of Campo Grande one might suggest that the kilometer-scale irregularities inside the bubbles (causing VHF scintillations) seem to be drifting slightly faster than the hundred-meter structures. Any difference in the zonal drift velocities estimate from GPS and VHF measurements could be associated with the fact that, in the case of GPS, the orbiting satellites are puncturing different irregularity structures in the sky above the station, whereas the geostationary satellite is always puncturing the ionosphere in the same point. Consequently, this geometrical factor might be contributing to discrepancies in the magnitude of the velocities when comparing both techniques. However, could any difference in the height distribution of the irregularities (with different scale sizes) probed by each technique explain the differences observed at both conjugate sites from the simultaneous GPS/VHF-estimated zonal drift velocities?
 In Figures 4 and 5 larger vertical drifts are noticed over Campo Grande than over Boa Vista, which probably contributes to the development of irregularities at comparative higher altitudes over the northern conjugate station. Thereby, if during the upwelling plasma the kilometer-scale irregularities inside the bubbles are located in a wider altitude range above the F peak, it is possible that an altitude gradient (shear) could be contributing for their velocities to be higher than the velocities of the Fresnel-scale irregularities causing GPS scintillations. Therefore, this effect could explain the reason for the magnitude of the averaged VHF-estimated zonal velocities in Campo Grande (Figure 7, bottom) to be larger than the GPS-estimated velocities. For example, de Paula et al.  compared GPS and airglow measurements and showed that the smaller-scale irregularities can drift with zonal velocities larger than the large-scale plasma bubble structures. However, we point out that such differences are expected once the airglow-inferred velocities by de Paula et al.  are restricted to ∼250 km, whereas the small-scale irregularities tend to be located in a more narrow altitude range about the F-peak where the electron density is high and steep density gradients exist. Therefore, GPS-estimated drift velocities being larger than the airglow-inferred drift velocities indicates the presence of a positive velocity shear with height [Sobral et al., 2009], rather than a difference associated solely with the scale-size structures.
 In terms of errors in the zonal drift calculations one must consider the experimental errors. Assuming a drift velocity (Vscint) of 200 ms−1 the error will be 4% and 2%, respectively, for the GPS (50 HZ) and VHF (100 Hz) technique. The errors due to the difference between the apparent and true velocities (mainly before 2100 UT) can be considered to be <10 ms−1 for GPS [Kil et al., 2000] and <20 ms−1 for VHF (according to estimations not shown here). Other sources of error may come from inaccuracies in the scaled ionospheric true height used to calculate vertical drift velocity contribution on equations (1) and (2). According to Muella et al. [2008a] and Fagundes et al. , digital ionosonde data ensure accuracies of ∼1 km in scaling h′F, which result in an estimated error of about ± 7 ms−1 in the effective vertical drift (VZ). Our results also rely on data obtained from the IGRF model, and it is crucial to have in mind that the outputs of the model are assumed initially to be correct. Thus, uncertainties arising from inaccuracies in the IGRF model are not provided. These factors are found in both equations (1) and (2) for VHF and GPS zonal drift estimations, respectively. However, for GPS drifts uncertainties may also be arising from GPS ephemeris (satellite velocity and position) error and on the assumed mean scattering height (hion). Satellite ephemeris bias can be considered a minor effect, since we used postprocessed precise ephemerides and data between the spaced receivers, which minimizes error because of its high correlation over baselines. Regarding the mean scattering height the well-accepted level of 350 km relates the location of the scintillation-causing irregularities to the height of maximum electron density. Although assumed, this mean scattering height is a good approximation since we used averaging over the scintillating GPS satellites [Kil et al., 2000; de Paula et al., 2002; Ledvina et al., 2004] and observed nighttime mean values of the F2 peak height (hmF2) parameter scaled from ionograms. The use of hmF2 is valid because the maximum contribution to the observed amplitude scintillation is from irregularities near the peak of the F region. In an average sense, the estimate of the zonal velocity with the mean scattering height of 350 km and the estimate with either 300 and 400 km produced a maximum error of 4 ms−1 (not shown here), which is also in agreement with the residual error calculations reported previously by Ledvina et al. . Hence, the factors mentioned above can be considered the main causes for discrepancies between the conjugate stations in the zonal drift velocities estimations for VHF and GPS techniques.
4. Remarks and Conclusions
 This report concerns the results of magnetic conjugate point observations of kilometer and hundred-meter-scale irregularity structures monitored by VHF and GPS ground-based instruments, respectively. The measurements of amplitude scintillations at VHF and GPS L-band frequencies were obtained from examining data acquired during the COPEX campaign and carried out from October to December 2002. The zonal drift velocities of the different irregularity scale sizes during the campaign period were estimated over 11 geomagnetically quiet nights of simultaneous VHF and GPS spaced antenna system observations. The main results of the present work are summarized as follows.
 VHF and L-band scintillations measurements at the COPEX sites reveal the coexistence of kilometer- and hundred-meter-scale irregularity structures into the underlying plasma depletion. However, the kilometer-scale structures causing VHF scintillations tend to decay slower than the hundred-meter-scale structures, which demonstrates the relative importance of background electron density in causing L-band scintillations. It is noticed from the results that scintillation activity is more intense at the conjugate sites of Boa Vista and Campo Grande than over the equatorial stations, owing to the increased background electron density and sharp electron density gradients that exist at latitudes adjacent to the crests of the equatorial ionization anomaly. Moreover, the scintillations seem to be more intense over Campo Grande than over Boa Vista, which can be attributed to the combination effects of large postsunset vertical plasma drift and delayed collapse in peak electron density (over Campo Grande). For future work, the present observations suggest that it would be interesting to analyze some important aspects of the thermosphere-ionosphere system, such as the role in equatorial scintillation activity played by the E × B drift velocities, the F2-peak density gradients, and the thermospheric neutral winds.
 The daily observations of GPS and VHF measurements reveal asymmetrical features in the intensity and latitudinal distribution of the scintillations. Thus, the day-to-day variability in the ambient ionization and in the electron density gradients could be a possible cause for the corresponding asymmetries observed over the magnetic conjugate sites. This implies that whether intense horizontal F2-peak density gradients exist between the dip equator and the EIA crests, a secondary instability process that creates conditions favorable for the generation of smaller scale plasma irregularities may be set up at the walls of a well-developed large-scale bubble that extends/expands along the magnetic field lines.
 The VHF-estimated zonal irregularity drift velocities are on average ∼4.75% and ∼10% larger than the GPS velocities for Boa Vista and Campo Grande, respectively. Over Campo Grande and before 0300 UT the average VHF velocities may attain values of ∼19% larger than the GPS-estimated velocities. However, the differences in the zonal drift velocities are mostly within the standard deviation bars. Whether during the upwelling plasma the kilometer-scale irregularities (inside the bubbles) causing VHF scintillations are located in a wider altitude range above the F-peak, and the smaller scale structures are in a more narrow altitude range about the F-peak (where the electron density is high and steep density gradient is present), it is possible that an altitude gradient (shear) could be contributing for the VHF velocities to be higher than the GPS-estimated velocities. There should also be a contribution due to geometrical factors on the differences seen from both techniques. In turn, the comparative large vertical plasma drift velocity over Campo Grande is probably contributing to a higher vertical shear of the plasma flow over this station.
 On average the GPS zonal drift velocities over Campo Grande are ∼6.5% greater than over the magnetic conjugation site of Boa Vista, whereas for the VHF estimations the irregularity drift velocities over the southern conjugate station of Campo Grande are ∼12% greater than over Boa Vista. Before 0300 UT the VHF velocities over Campo Grande may assume values of ∼8% higher than over Boa Vista. Based on previous modeling results [e.g., Sobral et al., 2009] we may conclude that the larger eastward drift velocities observed over Campo Grande can be attributed to the comparatively weaker total magnetic field intensity (considering a dipole configuration of the magnetic field) that exists over the South Atlantic hemisphere. In future work, we hope that the present report stimulates further experimental and theoretical studies about the effects of the geomagnetic field strength that drive the overall electrodynamics and its relation to the variation of irregularity occurrence and zonal drift velocities.
 The authors thank the Brazilian Air Force for their logistical support and facilities provided for the carrying out of the COPEX campaign. The authors also thank Robert Livingston (SCION Associates), Robert Sheehan, and Santimay Basu for the valuable comments on the methodologies used in the present work. Marcio Muella acknowledges the postdoctoral fellowship provided by the Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) under project 2008/04892-5. This research was also supported in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) under process BEX3367/05-3.