Statistics of GPS scintillations over South America at three levels of solar activity

Authors


Abstract

[1] This study characterizes low-latitude scintillations at L-band frequency in South America on daily, monthly, and seasonal time scales at three levels of solar activity: high, moderate and low levels. Three years (November 2001–October 2002, 2004 and 2008) of amplitude scintillation data from GPS receivers at three locations: Cuzco (14.0°S, 73.0°W, dip 1.0°S), Iquitos (3.8°S, 73.2°W, dip 7.0°N), and Bogota (4.4°N, 74.1°W, dip 16.0°N) were used for the investigation. These data were grouped into daily, monthly, and seasonal sets. We introduced tests on the data to reject signal fluctuations from non-ionospheric sources, such as multipath from terrestrial objects. The scintillation events from the data were further classified into three levels: weak (0.3 ≤ S4 < 0.4), moderate (0.4 ≤ S4 < 0.7), and intense (S4 ≥ 0.7) scintillations and their monthly percentage of occurrences and durations were determined. We also used the seasonal averages of daily (1900 LT–2000 LT) TEC values to observe the variability of the equatorial anomaly using a chain of ten GPS receivers that are located at a latitudinal span of 10°N–40°S along the west coast of South America. This study concludes that GPS scintillations in this longitude sector are post-sunset events and decay before or around local midnight, with intense activity and longer durations in the months of March and January. On seasonal time scales, the highest frequencies and longest durations of events were recorded during March Equinox and December Solstice. The months of May–July (June Solstice) had the least frequencies of occurrence and durations of events. Scintillation activity also increases with solar activity. Finally, scintillations increase in frequencies of occurrence and durations from Cuzco (near the magnetic equator) toward Bogota (near the crest of the equatorial anomaly) during solar maximum. However, a gradual collapse of the anomaly crest, away from Bogota toward Iquitos was observed as solar activity decreases, and as a result, the occurrence frequencies of scintillations at Iquitos increase relative to those at Bogota.

1. Introduction

[2] Ionospheric effects have raised some technical concerns that must be resolved before the implementation of space-based navigation for aviation applications can become a global reality. In order to resolve these issues, additional research efforts that are focused on characterizing the equatorial ionosphere are required. The impacts of the ionosphere on space-based communication and navigation manifest in two main categories: group delays and scintillations. While the group delay issue is being addressed by the dual-frequency technique, that of scintillations still remains a concern. This is because scintillations exhibit extreme variability in space and time [Aarons, 1982; Basu et al., 2002]. Consequently, for life critical applications, such as aviation navigation, real-time predictability is required.

[3] During post-sunset hours, the F region of the ionosphere often becomes turbulent and develops electron density irregularities [Woodman and LaHoz, 1976]. These irregularities scatter radio waves to cause amplitude and phase scintillations [Fremouw et al., 1978; Rino and Owen, 1980], and consequently degrade the performance of satellite-based communication and navigation systems [Groves et al., 1997; Basu et al., 1999, 2002; Doherty et al., 2002]. Intense scintillations can cause GPS signals to fade below the threshold margin of the receiver, in which case the signal becomes buried in noise, leading to signal loss and cycle slips. Scintillations are strong at high latitudes, weak at midlatitudes and intense at the equatorial region [Aarons, 1982; Basu et al., 1988].

[4] The equatorial zonal electric field plays a dominant role in shaping the development of the equatorial anomaly. This field is eastward during the day and reverses to the west after sunset. At low latitudes during sunset, an enhanced eastward electric field, known as the pre-reversal enhancement in zonal electric field, develops at F region heights [Woodman, 1970; Fejer, 1991; Heelis, 2004]. The increased electric field causes a redistribution of ionization by a fountain-like effect, thereby increasing the ionization density near the equatorial anomaly crests at the expense of that at the trough over the magnetic equator [Wright, 1962; Anderson, 1973a, 1973b; Balan and Bailey, 1995; Balan et al., 2009]. The vertical upwelling of the ionosphere leads to steep density gradients in the bottomside F region, and becomes unstable to the Rayleigh-Taylor instability [Kelley, 1989; Basu et al., 2002]. Large scale plasma depletions (bubbles) develop in the bottomside of the F region, which later transform into a plethora of small scale irregularities as they ascend upwards [Woodman and LaHoz, 1976; Valladares et al., 2001, 2004; Gwal et al., 2004]. The steep gradients on the edges of the depletions develop into a cascade of small scale irregularities as the bubbles gain altitude, sometimes exceeding 1000 km above the magnetic equator [McClure et al., 1977]. The bubbles extend in altitude, and then map down along magnetic field lines to anomaly locations of about 15°N and 15°S magnetic latitudes. The above effect is most pronounced during the equinoctial months of solar maximum years [Abdu et al., 1998, 2000]. Many authors have reviewed the physics of the dynamics of the equatorial ionospheric irregularities [e.g., Keskinen and Ossakow, 1983; Fejer, 1991; Titheridge, 1995; Heelis, 2004]. The generation of these irregularities and the persistence of high ambient ionization around the equatorial latitudes often lead to severe scintillation effects near the anomaly crest during post-sunset hours.

[5] The goal of the present study is to characterize low-latitude scintillations at L-band in western sector of South America on daily, monthly, seasonal and solar activity time scales.

2. Data and Method of Analysis

[6] The amplitude scintillation data that were used for this research were acquired at three low-latitude stations that are located in the South American western longitude sector, namely, Cuzco (14.0°S, 73.0°W, dip 1.0°S), Iquitos (3.8°S, 73.2°W, dip 7.0°N), and Bogota (4.4°N, 74.1°W, dip 16.0°N) using CRS1000 Leica GPS receivers with a sampling rate of 10 Hz. These receivers are being managed by Boston College. The Leica GPS receiver at Cuzco stopped operation in April, 2004, and has since been replaced with a Novatel GSV4004B GPS receiver. Furthermore, ten GPS receivers (including those used for the scintillation measurements) that are located in a latitudinal span between 10°N and 40°S were used for the TEC measurements. Valladares et al. [2004] gave detailed description of the TEC data resources. We considered seasonal averages of the verticalized TEC distributions during the post-sunset hours (1900 LT–2000 LT) in conformity with the daily onset time of ionospheric irregularities formation and evolution. Figure 1 shows the locations of these sites, the stations that provided both TEC and scintillation data are represented by red-filled circles, while those that provided only TEC data are represented by blue-filled circles. The data sets cover three years at three different levels of solar activity (November 2001–October 2002 (maximum), 2004 (moderate), and 2008 (minimum)). 2001 and 2002 were years of high-solar-activity (HSA) with average annual sunspot numbers (Rz): 111, and 104 respectively; 2004 was a year of moderate-solar-activity (MSA), Rz: 41; and 2008 was a year of low-solar-activity (LSA), Rz: 03.

Figure 1.

A map showing the locations of the GPS stations. The data for both TEC and scintillation analysis were acquired from the three stations with red-filled circles, while those stations that provided only TEC data for the analysis are represented by the blue-filled circles.

[7] To ensure a reliable statistical inference, we introduced three data cut-off criteria: (1) data for only equatorial scintillation hours (1800 LT–0600 LT) were used [Aarons, 1982; Basu et al., 1988, 2002; Ezquer et al., 2003], (2) only satellites with elevation angles greater or equal to 30° were used, so as to reject signal fluctuations from non-ionospheric origins such as multipath [Matsunaga et al., 2002; Carrano and Groves, 2010], and (3) each one-minute event was characterized by the data from the satellite which provided the highest scintillation index S4 (standard deviation of the factor I/〈I〉, where I is the intensity of the received signal and 〈I〉 is its average value).

[8] Following Bilitza et al. [2004] and Akala et al. [2010a, 2010b] the data were grouped into daily and monthly sets and further into different seasons by using the associated 3 months of data for each season: December Solstice (November–January), March Equinox (February–April), June Solstice (May–July), and September Equinox (August–October). Scintillation events from the data sets were further classified into three levels: weak (0.3 ≤ S4 < 0.4), moderate (0.4 ≤ S4 < 0.7), and intense (S4 ≥ 0.7) scintillations.

[9] The determination of the frequency of occurrence of scintillation for any particular level is based on whether or not scintillation at that level was detected on a given night. A given night is considered to experience scintillation if the observed scintillations persisted for at least four minutes during that night. The monthly percentage of occurrence at any given scintillation level was then calculated based on the number of nights of the month on which events at that level were recorded and the number of days of the observation during that month [Sobral et al., 2002]. This definition was subsequently adapted for the seasonal percentage of occurrence. Finally, we defined the duration of a scintillation patch on a given night as the time interval between the first sample (at a given scintillation level) and the last sample for that night, although, on a continuous scale. However, for rare cases whereby gaps were observed between patches on the same night, the continuous time intervals between the first and last samples of each patch were added together to obtain the patch duration. Following Rama Rao et al. [2006], the patch durations at moderate and intense levels were grouped into 30-min bins with their corresponding seasonal frequencies of occurrence for the HSA and MSA data, while 5-min bins were used for the LSA data. The choices of the bin intervals of 30 min and 5 min were borne out of the need to avoid empty bins, considering the size of the available samples.

3. Results and Discussions

[10] Figures 2, 3, and 4 show the amplitude scintillations that were observed by GPS receivers during the HSA year (2001/2002) for December Solstice, March Equinox, June Solstice, and September Equinox at Cuzco, Iquitos, and Bogota respectively. The contours in these plots are colored according to the S4 index. The same format is used in the two groups of Figures 5, 6, and 7 and 8, 9, and 10 to represent the corresponding results for MSA (2004) and LSA (2008) years, respectively. From these figures, scintillations were observed to have a daily trend of occurrences during the hours of 1900 LT–0200 LT, with marked highest occurrences during the HSA year. These observations are consistent with earlier observations by Aarons [1982], Basu et al. [1988], and Dubey et al. [2006]. Close inspection of the data on daily basis further revealed that the moderate and intense scintillations were localized within the hours of 2000 LT–2300 LT at all the stations. During these hours of the day, the ionospheric density is largely dependent on the electric fields, recombination rate, and the neutral winds [Titheridge, 1995]. During the period of local sunset in the equatorial region, the zonal neutral wind and the rapid decay of the E region density interact to develop an enhanced eastward electric field on the day-side of the terminator and a westward electric field on the nightside [Rishbeth et al., 1963; Woodman, 1970; Anderson and Haerendel, 1979; Kelley, 1989; Basu et al., 2002; Heelis, 2004]. The enhanced eastward electric field, otherwise known as the pre-reversal enhancement in the zonal electric field, causes vertical up-welling of the F region, and steepens the bottomside density gradient to trigger the Rayleigh–Taylor instability [Kelley, 1989; Rishbeth, 1998; Heelis, 2004]. Consequently, the low-density plasma from the bottomside percolates into the topside ionosphere to develop a plethora of plasma bubbles that transform to a cascade of irregularities of different scale sizes, and cause scintillation of radio signals, even at L-band frequencies [Basu et al., 2002].

Figure 2.

Amplitude scintillations at L-band frequency observed at Cuzco during the HSA year (2001/2002) (a) December Solstice, (b) March Equinox, (c) June Solstice, and (d) September Equinox.

Figure 3.

Amplitude scintillations at L-band frequency observed at Iquitos during the HSA year (2001/2002) (a) December Solstice, (b) March Equinox, (c) June Solstice, and (d) September Equinox.

Figure 4.

Amplitude scintillations at L-band frequency observed at Bogota during the HSA year (2001/2002) (a) December Solstice, (b) March Equinox, (c) June Solstice, and (d) September Equinox.

Figure 5.

Amplitude scintillations at L-band frequency observed at Cuzco during the MSA year (2004) (a) December Solstice and (b) March Equinox (no data for June Solstice and September Equinox).

Figure 6.

Amplitude scintillations at L-band frequency observed at Iquitos during the MSA year (2004) (a) December Solstice, (b) March Equinox, (c) June Solstice, and (d) September Equinox.

Figure 7.

Amplitude scintillations at L-band frequency observed at Bogota during the MSA year (2004) (a) December Solstice, (b) March Equinox, (c) June Solstice, and (d) September Equinox.

Figure 8.

Amplitude scintillations at L-band frequency observed at Cuzco during the LSA year (2008) (a) December Solstice, (b) March Equinox, (c) June Solstice, and (d) September Equinox.

Figure 9.

Amplitude scintillations at L-band frequency observed at Iquitos during the LSA year (2008) (a) December Solstice, (b) March Equinox, (c) June Solstice, and (d) September Equinox.

Figure 10.

Amplitude scintillations at L-band frequency observed at Bogota during the LSA year (2008) (a) December Solstice, (b) March Equinox, (c) June Solstice, and (d) September Equinox.

[11] Figure 11 shows the seasonal distributions of TEC values over the west coast of South America during the HSA, MSA and LSA years respectively. A typical feature of the TEC distribution, most especially for the December Solstice, March and September Equinoxes are the existence of troughs that reside at the magnetic equator or at latitudes close to it. Both crests of the equatorial (Appleton) anomaly at 1900 LT–2000 LT are located at magnetic latitudes varying between ±10° and ±18° depending on solar flux, and they are not always placed symmetrically with respect to the magnetic equator. Generally, at the three levels of solar activity, the March Equinox recorded the highest anomaly amplitudes, while the June Solstice showed weak anomaly amplitudes (within the range 10–25 TEC units) with characteristic troughs that almost level-off with the peaks. For instance, during the HSA year, the March Equinox showed a northern anomaly crest with a peak value of 116 TEC units at +14° magnetic latitude, while the southern crest attained a peak value of 106 TEC units at −15°. For the MSA year, the March Equinox showed a northern crest with a peak value of 60 TEC units at +12° magnetic latitude, while the southern crest had a peak value of 63 TEC units at −15°. The March Equinox of the LSA year showed a northern anomaly crest with a peak value of 17 TEC units and a southern crest with a peak value of 15 TEC units at ±11° magnetic latitude. The observed asymmetry in the amplitude of the crests may be attributed to the effect of pronounced transequatorial meridional winds [Valladares et al., 2001, 2004].

Figure 11.

Latitudinal TEC distributions over the west coast of South America during different seasons at (a) HSA, (b) MSA, and (c) LSA.

[12] Although, the present study did not investigate electric fields, we may presume that an enhanced eastward electric field have caused a redistribution of ionization by a fountain-like effect to increase the ionization density near the crest at the expense of the density of ionization at the magnetic equator [Balan and Bailey, 1995; Balan et al., 2009]. Furthermore, ambipolar diffusion [Rishbeth et al., 1963; Titheridge, 1995; Rishbeth, 1998] supported the enhanced eastward electric fields and neutral winds in transporting fountain plasma from dip equator toward the anomaly crest. At the crest, the continuous and fresh influx of ionization from the equator combines with the neutral wind to counteract the conventional decay process of ionization, thereby producing a secondary peak or a ledge in the ionization distribution [Anderson and Klobuchar, 1983], which extends the lifespan of the associated irregularities.

[13] Figure 12 shows the monthly frequencies of occurrence of different scintillation levels (weak, moderate, and intense) during the HSA year at Cuzco, Iquitos, and Bogota respectively. On a monthly basis, at Cuzco, the percentage of scintillation occurrence for intense, moderate and weak scintillations had maximum occurrences in the months of January and March, and had minimum occurrences in the months of April–September, May–July, and June–July respectively. At Iquitos, the intense, moderate and weak scintillations showed maximum occurrences in the months of January and March, and minimum occurrences in the months of May–July, June (moderate and weak levels). At Bogota, the percentage of scintillation occurrence for intense scintillation had maximum occurrences in the months of January and December, and minimum occurrences in the months of June and July. At the moderate scintillation level, the maximum occurrences were recorded in the months of January and February, and the minimum occurrences in the months of June and July. The weak scintillation showed maximum occurrence during the month of January, and the minimum occurrence during the month of June.

Figure 12.

Monthly frequencies of occurrences of different levels of scintillations during the HSA year (2001/2002) (a) Cuzco, (b) Iquitos, and (c) Bogota.

[14] Figures 13 and 14 show the same plots as Figure 12 during the MSA and LSA years at Cuzco, Iquitos, and Bogota respectively. At Cuzco during the MSA year, intense scintillations were only observed during the months of January–March, although data were only available for the months of January–April. At the moderate scintillation level and with the available data, the maximum occurrence was recorded in the months of February and the minimum occurrence in the month of January. The weak scintillation showed the maximum occurrences during the months of February and March, and the minimum occurrence during the month of April. At Iquitos, the percentage of scintillation occurrence for intense scintillation on monthly basis had maximum occurrences in the months of November and October, and the minimum occurrences in the months of May–September. For the moderate scintillation, the maximum occurrences were recorded in the months of October, November and March, and the minimum occurrences in the month of May–July. The weak level of scintillation during the MSA year showed maximum occurrences in the months of March and October, and the minimum occurrence in the month of June. At Bogota, the monthly percentage of occurrence of intense scintillation showed maximum occurrences in the months of November and March, and the minimum occurrences in the months of May and June. For the moderate scintillation, the maximum occurrences were recorded in the months of September and November, and the minimum occurrences in the months of May and June. As for the weak scintillation, maximum occurrences were recorded in the months of August–December and January–March, and the minimum occurrence in the month of June.

Figure 13.

Monthly frequencies of occurrences of different levels of scintillations during the MSA year (2004) (a) Cuzco, (b) Iquitos, and (c) Bogota.

Figure 14.

Monthly frequencies of occurrences of different levels of scintillations during the LSA year (2008) (a) Cuzco, (b) Iquitos, and (c) Bogota.

[15] During the LSA year, there were no intense scintillations during any of the months of the year at Cuzco. However, at the moderate scintillation level, events were only recorded during the months of March, December and October. For the weak scintillation, the maximum occurrences were recorded in the months of January and March, and there were no occurrences during the months of May–September. At Iquitos, no intense scintillation event was recorded during any of the months of the LSA year. For the moderate scintillation, the maximum monthly occurrences were observed in the months of October and March, and the minimum occurrences in the months of May and June. The weak level of scintillation showed maximum occurrences during the month of March and the minimum occurrence during the month of June. At Bogota, with the exception of the months of March, September and October that recorded events at intense level, there were no events for all other months of the year at the intense scintillation level. For the moderate scintillation level, the maximum occurrence was recorded in the month of February, and there were no events during the months of May–July. At the weak scintillation level, maximum occurrences were recorded in the month of January, and the minimum occurrence in the month of June.

[16] Typically, the highest frequencies of scintillation occurrence were observed during the months of March and January, and the least occurrences were observed during the months of June and July. It is important to mention that significant scintillation occurrence frequencies were also recorded during the months of November, December, February, September, and October. As for the months of April and August, there were relatively fewer events. Seasonally, the highest occurrences of scintillations at the three stations were recorded during the March Equinox and December Solstice with a marked increase in scintillation intensity with solar activity. Scintillation events were also observed during the September Equinox, whereas, the June Solstice recorded the least events, especially at moderate and intense levels.

[17] Figures 15, 16, and 17 show the patch durations for moderate/intense scintillations for the four seasons with their corresponding frequencies of occurrence during the HSA, MSA and LSA years at Cuzco, Iquitos, and Bogota respectively. With the exception of the June Solstice that recorded little or no scintillation events at moderate and intense levels (S4 > 0.4), the bin for the 30 min duration has the highest frequencies of occurrences for all the seasons at the three stations for both the HSA and MSA years. At Cuzco during the HSA year, the event with the longest lifespan (2 h) occurred on the 29th of March 2002, a representative of March Equinox and 1% of the nights of the season. At Iquitos, the events with the longest lifespan (2.5 h) occurred on the 17th of January 2002 during the December Solstice, which represents 1% of the nights of the season, and 10th of February, 9th, and 19th of March 2002 for the March Equinox, 4% of the nights of the season. At Bogota, the event with the longest lifespan (about 3.5 h) occurred on the 14th of February 2002. Although, data were not available for March and April at Bogota for a collective seasonal statistical inference, we have conservatively employed the February data as a representative of the season [Akala et al., 2010a]. Consequently, considering the number of nights that data were available for the season, the event of 14th of February represents 4% of the observed nights of the season.

Figure 15.

The Patch durations for moderate/intense scintillations (S4 ≥ 0.4) for the four seasons and their corresponding frequencies of occurrences during the HSA year (a) Cuzco, (b) Iquitos, and (c) Bogota.

Figure 16.

The Patch durations for moderate/intense scintillations (S4 ≥ 0.4) for the four seasons and their corresponding frequencies of occurrences during the MSA year (a) Cuzco, (b) Iquitos, and (c) Bogota.

Figure 17.

The Patch durations for moderate/intense scintillations (S4 ≥ 0.4) for the four seasons and their corresponding frequencies of occurrences during the LSA year (a) Cuzco, (b) Iquitos, and (c) Bogota.

[18] At Cuzco during the MSA year, the event with the longest lifespan (1 h) occurred on the 20th of January 2002, a representative of December Solstice and 5% of the observed nights in the season. However, only January–April data were available at Cuzco during the MSA year. At Iquitos, the event with the longest lifespan (2 h) occurred on the 27th of January 2002 during the December Solstice, which represents 1% of the nights of the season. At Bogota, the events with the longest durations (about 3 h) occurred on the 25th of January 2002 (December Solstice), and 1% of the nights of the season, 10th of March 2002 (March Equinox), and 1% of the nights of the season, and on the 30th October 2002, representing 1% of the nights of the September Equinox. For the LSA year at Cuzco, all the events clouded within the 5 min bin, 1%, 2% and 1% of the December Solstice, March Equinox, and September Equinox respectively. The only events for the June Solstice, although very brief, were recorded at Iquitos, and they all fall within the 5 min bin. At this station, the event with the longest lifespan (30 min) occurred on the 9th of February 2002 during the March Equinox, which represents 1% of the nights of the season. At Bogota, the event with the longest duration (30 min) occurred on the 16th of March (March Equinox), and 1% of the nights of the March Equinox season.

[19] Generally, patches with the longest durations were observed during the March Equinox and the occurrence frequencies increase from the magnetic equator toward the crest of the anomaly, in this case, from Cuzco toward Bogota (although, depending on the level of the solar flux). Comparatively, the longevity experienced by Bogota scintillation patches could be attributed to the large ambient plasma density of the equatorial ionization anomaly (EIA) crest location due to the post sunset plasma fountain [Wright, 1962; Anderson, 1973a, 1973b; Balan and Bailey, 1995]. However, it is important to point out that the TEC measurements showed a gradual collapse of the anomaly crest from Bogota toward Iquitos as solar activity decreased, and consequently locating the crest at magnetic latitudes in-between the two stations. By implication, during MSA and LSA years, comparatively high occurrences and longer durations of scintillation events like those observed at Bogota were also observed in parallel at Iquitos. On a station-to-station comparison, Bogota recorded the highest occurrences of moderate and intense scintillations, especially during the HSA year, followed by Iquitos, while, Cuzco had the least occurrences.

4. Conclusions

[20] We have characterized the equatorial and low latitude scintillations at L-band frequency over the western longitude sector of South America on daily, monthly, and seasonal time scales at high, moderate, and low levels of solar activity. Three years of amplitude scintillation data from three GPS stations, namely, Cuzco (14.0°S, 73.0°W, dip 1.0°S), Iquitos (3.8°S, 73.2°W, dip 7.0°N), and Bogota (4.4°N, 74.1°W, dip 16.0°N) were used for the investigation.

[21] Scintillations were observed to have a daily trend of occurrence during the hours of 1900 LT–0200 LT. Close inspection of the data further revealed that the moderate and intense levels of scintillations were localized within the hours of 2000 LT–2300 LT. The monthly characteristics of the observed scintillations showed that the months of March and January had highest frequencies of scintillation occurrence and longest patch durations, while the least occurrences and shortest patch durations were recorded in the months of June and July. Seasonally, the March Equinox and December Solstice showed maximum occurrences of scintillations and patches durations, while the June Solstice showed the minimum occurrence. These results are in agreement with the results of the seasonal percentage of occurrence of Spread-F and equatorial plasma bubbles (EPB) over the South American sector obtained by Valladares et al. [1996, 2001] using multiinstrument observations, Sobral et al. [2002] using optical observations, and Burke et al. [2004] using in situ data from satellites observations (DMSP and ROCSAT-1).

[22] From a station-to-station point of view, Bogota recorded the highest occurrences of scintillation and the longest patch durations, especially during the HSA year, followed by Iquitos, while Cuzco had the least occurrences and shortest patch durations. Bogota, generally being located on the northern crest of the equatorial anomaly, has a comparatively higher ambient ionization than the other two locations considered in this investigation. For instance, Iquitos is generally located within the inner flank of the anomaly, and Cuzco directly beneath the magnetic equator. However, a gradual collapse of the anomaly crest (away from Bogota) toward Iquitos was observed as solar activity decreased. The enhanced eastward electric field at Cuzco (magnetic equator) during local sunset may have caused a redistribution of ionization by fountain effect, leading to increase in ionization density near the crest of the anomaly at the expense of that at the trough [Anderson, 1973a, 1973b; Balan and Bailey, 1995; Balan et al., 2009].

[23] Overall, GPS scintillations are post sunset events, and they decay before or around local midnight. In South America, the most intense and longest duration of scintillations occur in the months of March and January, and during March Equinox and December Solstice in terms of the seasons. These events increase in occurrences and lifetimes as one move away from the magnetic equator toward the crest of the equatorial anomaly, and also with solar activity. In future efforts, we hope to expand the scope of the present study by analyzing similar data from other longitudinal sectors, such as the African and Southeast Asian/Australian sectors for a wider characterization of low-latitude scintillations at GPS frequency on a global level. This may be necessary for the development of future models that will support real-time predictability of low-latitude scintillations, and the subsequent implementation of space-based navigation for aviation applications on a global level.

Acknowledgments

[24] The first author thanks the U. S. Government for the Fulbright Scholarship grant, and the Institute for Scientific Research, Boston College for hosting him. The authors also gratefully acknowledge the National Science Foundation (NSF) for support under grant ATM-05-0521487.